** rnA United States
Environmental Protection
^#^1 Agency
EPA/600/R-13/241 | June 2015 | www.epa.gov/research
Assessment of Mitigation
Systems on Vapor Intrusion:
Temporal Trends, Attenuation
Factors, and Contaminant
Migration Routes under Mitigated
And Non-mitigated Conditions
RESEARCH AND DEVELOPMENT
-------
-------
Assessment of Mitigation
Systems on Vapor Intrusion:
Temporal Trends, Attenuation
Factors, and Contaminant
Migration Routes under Mitigated
and Non-mitigated Conditions
Prepared for
U.S. Environmental Protection Agency
National Exposure Research Laboratory
Las Vegas, Nevada
Prepared by
RTI International
3040 E. Cornwallis Road
Research Triangle Park, NC 27709
and
ARCADIS U.S., Inc.
4915 Prospectus Drive, Suite F
Durham, NC 27713
EPA Contract EP-C-11-036, Task Order 009
Although this work was reviewed by EPA and approved for publication, it may not necessarily
reflect official Agency policy. Mention of trade names and commercial products does not
constitute endorsement or recommendation for use.
U.S. Environmental Protection Agency
Office of Research and Development
Washington, DC 20460
-------
-------
Table of Contents
1.0 Executive Summary 1-1
1.1 Background 1-1
1.2 Purpose and Objectives 1-1
1.3 Methods 1-2
1.4 Conclusions 1-3
1.4.1 Conceptual Site Model: VOC Data 1-3
1.4.2 Mitigation System Performance—Radon 1-3
1.4.3 Mitigation System Performance—VOCs 1-4
1.4.4 Meteorological Effects on Vapor Intrusion 1-5
1.4.5 Preferential Pathways and Conceptual Site Model: Helium Tracer and
Geophysical Tests 1-6
1.4.6 Temporal Variability and Trends 1-8
1.4.7 Summary 1-8
2.0 Introduction 2-1
2.1 Background 2-2
2.1.1 Variability in Vapor Intrusion Studies 2-4
2.1.1.1 Spatial V ariability 2-5
2.1.1.2 Temporal Variability 2-7
2.1.1.3 Measurement Variability 2-8
2.1.2 Vapor Attenuation Factors 2-9
2.1.3 Potential for Use of Radon as a Surrogate for VOC Vapor Intrusion 2-9
2.1.4 Passive VOC Sampling 2-12
2.2 Objectives 2-16
2.2.1 Time Scale and Measurement of Independent and Dependent Variables 2-18
2.2.2 Data Quality Objectives and Criteria 2-18
3.0 Methods 3-1
3.1 Site Description 3-1
3.1.1 Area Geology/Hydrogeology 3-1
3.1.2 Area Potential Sources 3-1
3.1.3 Building Description 3-7
3.1.3.1 Building Age, Condition, and HVAC 3-7
3.1.3.2 Building Utilities/Potential Entry Points 3-8
3.1.4 Building Occupancy During Sampling 3-9
3.1.5 Investigation History 3-9
3.2 Evolution of Conceptual Site Model 3-10
3.2.1 Priorto 2011-2012 Investigations 3-10
3.2.2 After 2011-2012 Investigations (U.S. EPA, 2012a) 3-10
3.2.3 Refinements in Conceptual Site Model Sought in this 2012-2013 Study 3-11
l
-------
3.3 Building Renovation and Mitigation 3-12
3.3.1 Subslab Depressurization Mitigation System Installation 3-12
3.4 Monitoring, Sampling, and Analysis 3-19
3.4.1 Indoor and Outdoor Air VOC Monitoring 3-19
3.4.2 Subslab and Soil Gas (TO-17) 3-21
3.4.3 Online Gas Chromatograph 3-22
3.4.4 Groundwater 3-23
3.4.5 Subslab Depressurization System Stack Gas Sampling 3-24
3.5 Radon Sampling and Analysis 3-24
3.5.1 Indoor Air Radon Sampling and Analysis 3-24
3.5.2 Subslab and Soil Gas Radon Sampling and Analysis 3-25
3.5.3 Continuous (Real-Time) Indoor Air Radon Sampling and Analysis 3-26
3.6 Physical Parameters Monitoring 3-26
3.6.1 On-Site Weather Station 3-26
3.6.2 Indoor Temperature 3-28
3.6.3 Soil Temperature 3-28
3.6.4 Soil Moisture 3-28
3.6.5 Potentiometric Surface/Water Levels 3-29
3.6.6 Differential Pressure 3-29
3.6.7 Air Exchange Rate 3-30
3.6.8 Crack Monitoring 3-30
3.7 Data Aggregation Methods 3-31
4.0 Results and Discussion: Quality Assurance Checks of Individual Data Sets 4-1
4.1 VOC Sampling—Indoor Air-Passive—Air Toxics Ltd. (ATL) 4-1
4.1.1 Blanks 4-1
4.1.2 Surrogate Recoveries 4-3
4.1.3 Laboratory Control Sample Recoveries 4-3
4.1.4 Duplicates 4-4
4.2 VOC Sampling—Subslab and Soil Gas (TO-17)—U.S. EPA 4-5
4.2.1 Blanks 4-5
4.2.2 Calibration Verification 4-7
4.2.3 Internal Standard Recoveries 4-8
4.2.4 Surrogate Recoveries 4-8
4.2.5 Laboratory Control Sample Recoveries 4-9
4.2.6 Field Duplicates 4-9
4.3 Online Gas Chromatograph (Soil Gas and Indoor Air) 4-10
4.3.1 Blanks 4-10
4.3.2 Initial Calibration 4-10
4.3.3 Continuing Calibration 4-11
4.3.4 Calibration Check via Comparison to Fixed Laboratory (TO-15 vs. Online GC) 4-12
4.3.5 Agreement of Online GC Results with TO-17 Verification Samples 4-14
ii
-------
4.3.6 Agreement of Integrated Online GC Results with Passive Samplers 4-16
4.3.7 Overall Assessment of Online GC Data 4-37
4.4 Radon 4-43
4.4.1 Indoor Air: Comparison of Electrets Field, ARCADIS to Charcoal Analyzed by
U.S. EPA R&IE National Laboratory 4-43
4.4.2 Comparison of Average of Real-Time AlphaGUARD to Electrets and Charcoal
Canisters 4-45
4.4.3 Quality Assurance Checks of Electrets 4-48
4.5 On-Site Weather Station vs. National Weather Service (NWS) 4-48
4.6 Groundwater Analysis—EPA NERL 4-51
4.6.1 Blanks 4-51
4.6.2 Surrogate Recoveries 4-53
4.7 Groundwater Analysis—Pace Laboratories 4-54
4.8 Database 4-54
4.8.1 Checks on Laboratory Reports 4-54
4.8.2 Database Checks 4-54
4.9 Air Exchange Rate Measurements 4-55
5.0 Subslab Depressurization Mitigation System Monitoring Results 5-1
5.1 Differential Pressure and Mitigation System Flow 5-1
5.1.1 Radon System Design Standards for Differential Pressure 5-1
5.1.2 Differential Pressure Monitoring of this SSD System 5-3
5.1.3 Mitigation System Flow 5-21
5.2 Radon Monitoring: Hourly and Weekly Time Scales 5-21
5.3 VOC Monitoring During Mitigation Testing 5-30
5.3.1 Descriptive Statistics 5-33
5.3.2 Effect of Mitigation System Status on Indoor Air VOC Levels 5-45
5.3.3 Discussion 5-46
5.4 Stack Gas Monitoring 5-47
5.4.1 Is Stack Gas an Indicator of System Performance in Protecting Indoor Air? 5-47
5.4.2 Air Exchange Rate Measurements 5-50
5.4.3 Stack Gas Measurements to Define Flux to Structure 5-53
6.0 Results and Discussion: VOC Concentration Temporal Trends and Relationship to HVAC
and Mitigation 6-1
6.1 VOC Seasonal Trends Based on Weekly, Biweekly, and Monthly Measurements for
52+Weeks 6-1
6.1.1 Indoor Air 6-1
6.1.2 Subslab Soil Gas 6-4
6.1.3 Shallow and Deep Soil Gas 6-15
6.2 Radon Seasonal Trends (based on Weekly Measurements) 6-45
6.3 VOC Short-Term Variability (Based on Daily and Hourly VOC Sampling) 6-45
6.3.1 Indoor Air 6-45
iii
-------
6.3.1.1 Chloroform 6-45
6.3.2 Subsurface Soil Gas Data 6-51
6.3.2.1 Chloroform 6-51
6.3.2.2 Tetrachloroethylene (PCE) 6-53
6.4 Radon Short-Term Variability (Based on Daily and More Frequent Measurements) 6-58
6.5 Outdoor Climate/Weather Data 6-59
6.5.1 Indianapolis Weather Compared with VOCs and Radon 6-65
6.5.1.1 Wall Port VOC Concentrations as a Function of Barometric
Pressure and Wind Speed 6-51
7.0 Results and Discussion: Establishing the Relationship between VOCs and Radon in
Subslab/Subsurface Soil Gas and Indoor Air 7-1
7.1 Previously Reported Tests of Radon as a Semiquantitative VOC Tracer 7-1
7.2 Understanding the Performance of the Radon Tracer During Mitigation Testing 7-1
7.3 Attenuation Factors Derived Using the Radon Tracer 7-4
7.4 Radon Tracer in Statistical Time Series Analysis 7-4
8.0 Results and Discussion: Attenuation of Soil Gas VOCs and Radon 8-1
8.1 Calculation of Daily Attenuation Factors without Mitigation 8-1
8.1.1 Daily Radon Attenuation Factors without Mitigation 8-2
8.1.2 Daily VOC Attenuation Factors without Mitigation 8-2
8.2 Subslab and Soil Gas to Indoor Air Weekly Attenuation Factors 8-8
8.2.1 Radon Weekly Attenuation Factors 8-8
8.2.2 VOC Weekly Attenuation Factors 8-11
8.3 Effect of Mitigation 8-19
8.3.1 Subslab to Indoor Air Daily Attenuation 8-20
8.3.2 Subslab and Soil Gas to Indoor Air Weekly Attenuation 8-23
9.0 Results and Discussion: Determine If Observed Changes in Indoor Air Concentration of
Volatile Organics of Interest are Mechanistically Attributable to Changes in Vapor
Intrusion 9-1
9.1 Large Differential Pressures, Pressure Changes and Meteorological Factors Analysis
with Mitigation Off 9-1
9.1.1 Temperature Effects on Differential Pressure 9-1
9.1.2 Barometric Pressure Effects on Differential Pressure 9-4
9.1.3 Precipitation Effects on Differential Pressure 9-8
9.1.4 Wind Effects on Differential Pressure 9-10
9.1.5 Assessment of Whether High Observed Differential Pressures Could be
Artifactual 9-16
9.1.6 Examination of High-Resolution Time Series Data for Individual Extreme
Differential Pressure Events 9-16
9.2 Influence of Meteorological Conditions on Indoor VOC Concentration; Mitigation
Off 9-44
9.2.1 Temperature Effect on VOCs 9-44
IV
-------
9.2.2 Barometric Pressure Effect on VOCs 9-47
9.2.3 Precipitation Effects on VOCs 9-47
9.2.4 Effect of Wind on VOC Concentrations 9-52
9.3 Summary of Meteorological Effects on Vapor Intrusion—Evidence Presented in
Sections 6 and 9 9-59
10.0 Time Series Analysis 10-1
10.1 Time Series Analysis of Indoor Radon Data (AlphaGUARD) Aggregated with 1-Day
Time Resolution 10-2
10.2 Correlation between Radon Concentration and Categorical Predictors 10-18
10.3 Correlation between Radon Concentration Time Series for 422 Basement South in
2011-2012 (X422BN-1) and Continuous Predictor Variables 10-19
10.4 Correlation between Radon Concentration Time Series for 422 2nd Floor Office
(2011-2012) and Predictor Variables 10-25
10.5 Correlation between Radon Concentration Time Series for 422 Basement South
(2012-2013) and Predictor Variables 10-32
10.6 Correlation between Radon Concentration Time Series for 422 Office on 2nd Floor
and Predictor Variables 10-37
10.7 Correlation between VOC (Radiello) Time Series and Predictor Variables in 422
Basement South 10-41
10.7.1 Stationarity and Serial Correlation Analysis 10-41
10.7.2 Predictor Variables Modeled and Their Potential for Autocorrelation 10-50
10.7.3 Time Series Analysis Results for 2011-2012 Chloroform Data Set 10-53
10.7.4 Time Series Analysis Results for 2011-2012 PCE Data Set 10-57
10.7.5 Time Series Analysis of 422 Basement South Chloroform Data Set from the
period Sept 2012-Apr 2013 10-65
10.7.6 Time Series Analysis Results of 422 Basement South PCE Data from Sept 2012
through April 2013 10-71
Addendum 10-79
11.0 Results and Discussion: Do Groundwater Concentrations Control Soil Gas Concentrations
at this Site? And Thus Indoor Air Concentrations? 11-1
11.1 Groundwater Level Changes 11-1
11.2 Groundwater Concentration Trends 11-2
11.2.1
Is the Groundwater Concentration Trend Correlated to Well or Water Table Depth? 11-5
11.3 Revision to the Conceptual Model—Is the Groundwater Concentration Related to
Soil Gas and Indoor Air Concentrations? 11-6
12.0 Results and Discussion: Special Studies 12-2
12.1 Summary of Geophysics Study 12-2
12.2 Summary of Tracer Testing 12-4
12.2.1 Introduction to Tracer Testing 12-4
12.2.2 Tracer Testing Objective 12-4
12.2.3 Tracer Test Experimental Methods 12-4
v
-------
12.2.4 Tracer Test Results and Discussion 12-5
12.2.5 Summary of Tracer Test Conclusions 12-16
12.3 Testing Utility of Consumer Grade Radon Device (Safety Siren Pro) 12-16
12.3.1 Introduction to the Use of Consumer Grade Radon Monitoring Equipment
in Vapor Intrusion 12-16
12.3.2 Objective of Consumer-Grade Radon Device Testing 12-16
12.3.3 Methods of Consumer-Grade Radon Device Testing 12-16
12.3.4 Consumer Grade Radon Detector Results and Discussion 12-17
13.0 Conclusions and Recommendations 13-1
13.1 Conclusions 13-1
13.1.1 Mitigation System Performance 13-2
13.1.2 Meteorological Effects on Vapor Intrusion 13-4
13.1.3 Preferential Pathways and Revisions to Conceptual Site Model: Helium Tracer
and Geophysical Tests 13-5
13.1.4 Revisions to Conceptual Site Model: VOC Data 13-7
13.1.5 Temporal Trends 13-8
13.2 Practical Implications for Practitioners 13-9
13.2.1 Mitigation Design Implications 13-9
13.2.2 Sampling Planning Implications 13-9
13.2.3 Delineating Preferential Pathways 13-10
13.3 Recommendations 13-10
14.0 References 14-1
Appendix A. Radon Mitigation System Photos and Field Diagnostic Report
Appendix B. Geophysics Report
Appendix C. Additional Statistical Analysis of Effect of Mitigation on Indoor VOC Concentrations
VI
-------
List of Figures
1-1. Temporal coverage of data sets collected (red line indicates the cutoff date for this
report) 1-5
1-2. Indoor air concentrations of PCE (top panel) and radon (middle and bottom panels) with
mitigation on (black bars), passive mitigation (no fan, gray bars), and mitigation off (no
bar). Snow and frozen ground events are indicated by blue bars and red dots 1-9
2-1. An overview of important vapor intrusion pathways (U.S. EPA graphic) 2-3
2-2. Soil gas and groundwater concentrations below a slab (Schumacher et al., 2010) 2-6
3-1. Lithological fence diagram showing the major soil types beneath the 422/420 house 3-2
3-2. Aerial view of duplex, 420/422 East 28th Street, showing nearby sanitary and storm
sewers 3-3
3-3. East side of house (on right) and adjoining commercial quadraplex visible (left) 3-4
3 -4. Roof of adj acent commercial quadraplex 3-4
3 -5. Looking toward southeast corner of adj acent commercial quadraplex 3-5
3-6. Visual evidence of historic dry cleaners in area 3-6
3-7. Front view of house during summer 2011 sampling, with fan testing and weather station 3-7
3-8. Front view of duplex under winter conditions showing designation of sides and HVAC
setup 3-8
3-9. 422 (left) and 420 East 28th Street in January 2011 3-9
3-10. Map view of the 422-side basement showing SSD mitigation system legs, subslab soil
gas extraction pits (red circles), and the position of the passive "sampling racks."
Horizontal divisions are walls between "north" (top in figure), "central," and "south"
(bottom in figure) sections of basement with open walkways between (cistern is in the
central basement) 3-14
3-11. Map view of the 420-side basement showing the SSD mitigation system legs, subslab soil
gas extraction pits (red circles), and the passive "sampling racks." Horizontal divisions
are walls between "north" (top in figure), "central," and "south" (bottom in figure)
sections of basement with open walkways between (cistern is in the central basement) 3-15
3-12. Photos of mitigation system: (left) SSD blower and stack on northeast corner of duplex;
right) SSD extraction point, showing valve and U-tube manometer 3-16
3-13. Cross-section showing the general layout of the 422/420 north and central basements
with the positioning of the extraction legs, exterior blower, and exhaust stack 3-17
3-14. Subsurface soil gas monitoring probes (SGP), subslab sampling ports (SSP), and
groundwater monitoring wells (MW). Horizontal divisions are walls between "north,"
"central," and "south" sections of basement with open walkways between (cisterns are in
the central basements). Probes/ports in red were sampled by the on-site GC. Soil
temperature and moisture probes were installed in the 422 basement between SGP 8 and
MW 3 and in the backyard to the north of MW 2 3-18
3-15. Passive indoor air sampling rack: 422 first floor 3-20
3-16. Ambient sampler shelters on telephone pole near duplex 3-21
3-17. Monitoring well MW-3, installed in the basement and completed on the first floor 3-26
Vll
-------
3-18. Front view of 420/422 duplex with location of weather station sensors indicated with red
arrow 3-27
3-19. Calibrated crack monitor 3-30
4-1. TCE Continuing Calibration Standard Analyses, Hartman 3 Period 4-12
4-2. XY Comparison plot of Radiello and GC indoor air concentration measurements
(|ig/m3), Hartman 1 sampling period 4-30
4-3. XY Comparison plot of Radiello and GC indoor air concentration measurements
(|ig/m3), Hartman 2 sampling period 4-31
4-4. Time series comparison of field GC and passive sampling data: 422 basement,
Hartman Period 2, chloroform. Horizontal gray line is calculated GC reporting limit.
Red hash marks on y-axis indicate missing values 4-33
4-5. Time series comparison of field GC and passive sampling data: 422 first floor,
Hartman Period 2, chloroform. Horizontal gray line is calculated GC reporting limit.
Red hash marks on y-axis indicate missing values 4-34
4-6. Time series comparison of field GC and passive sampling data: 422 basement,
Hartman Period 2, PCE. Horizontal gray line is calculated GC reporting limit. Red
hash marks on y-axis indicate missing values 4-35
4-7. Time series comparison of field GC and passive sampling data: 422 first floor, Hartman
Period 2, PCE. Horizontal gray line is calculated GC reporting limit. Red hash marks on
y-axis indicate missing values 4-36
4-8. XY Plot of field GC vs. passive sampler data, Hartman Period 3 4-38
4-9. Time series comparison of field GC and passive sampling data: 422 basement, Hartman
Period 3, chloroform. Horizontal gray line is calculated GC reporting limit. Red hash
marks on y-axis indicate missing values 4-39
4-10. Time series comparison of field GC and passive sampling data: 422 first floor, Hartman
Period 3, chloroform. Horizontal gray line is calculated GC reporting limit. Red hash
marks on y-axis indicate missing values 4-40
4-11. Time series comparison of field GC and passive sampling data: 422 Basement, Hartman
Period 3, PCE. Horizontal gray line is calculated GC reporting limit. Red hash marks on
y-axis indicate missing values 4-41
4-12. Time series comparison of field GC and passive sampling data: 422 first floor, Hartman
Period 3, PCE. Horizontal gray line is calculated GC reporting limit. Red hash marks on
y-axis indicate missing values (none in this case) 4-42
4-13. Correlation between radon measured using the electret and charcoal methods 4-45
4-14. Aerial view of study house, showing potential influences on wind velocity, red arrow
indicates study house 4-49
4-15. Comparison of National Weather Service Indianapolis temperature data to weather
station at 422 East 28th Street 4-50
4-16. Comparison of National Weather Service Indianapolis relative humidity to weather
station at 422 East 28th Street 4-50
4-17. Comparison of National Weather Service wind speed data to weather station at 422 East
28th Street 4-51
Vlll
-------
5-1. Subslab vs. basement differential pressure: 422 side during mitigation testing 5-19
5-2. Subslab vs. basement differential pressure: 420 side during mitigation testing 5-20
5-3. Deep soil gas vs. shallow soil gas differential pressure during mitigation testing 5-21
5-4. Basement vs. upstairs differential pressure: 422 side during mitigation testing 5-22
5-5. Basement vs. exterior differential pressure: 422 side during mitigation testing 5-22
5-6. Stack gas flow velocity from SSD system 5-23
5-7. Real-time radon monitoring: 422 basement 5-23
5-8. Real-time radon monitoring: 422 second floor 5-24
5-9. Weekly integrated radon (electret) during mitigation testing 5-24
5-10. Passive sampler monitoring of PCE during mitigation testing 5-31
5-11. Passive sampler monitoring of chloroform during mitigation period 5-31
5-12. Indoor air PCE, real-time monitoring during mitigation testing 5-32
5-13. Boxplots of mitigation effect on indoor air concentrations 5-46
5-14. Stack gas monitoring during mitigation te sting: chloroform 5-48
5-15. 422 first floor versus stack gas chloroform concentrations: mitigation on 5-48
5-16. Stack gas monitoring during mitigation testing: PCE 5-49
5-17. Stack gas versus 422 first floor PCE concentrations: mitigation on 5-49
6-1. PCE concentrations in indoor and ambient air vs. time (7-day Radiello samples) 6-2
6-2. Chloroform concentrations in indoor and ambient air vs. time (7-day Radiello samples) 6-2
6-3. Benzene concentrations in indoor air 6-3
6-4. Toluene concentrations in indoor air 6-4
6-5. Interior and exterior sampling port locations. Sampling ports sampled by the on-site GC
are shown in red, with parenthetical notes indicating which SGP depths were sampled by
the GC 6-5
6-6a. Plot of subslab chloroform concentrations vs. time (TO-17 data) 6-6
6-6b. Plot of subslab chloroform concentrations vs. time, first intensive sampling period (TO-
17 data) 6-6
6-6c. Plot of subslab chloroform concentrations vs. time mitigation testing period (TO-17
data) 6-7
6-7a. Plot of subslab PCE concentrations vs. time. (TO-17 data) 6-7
6-7b. Plot of subslab PCE concentrations vs. time, first intensive sampling period. (TO-17
data) 6-8
6-7c. Plot of subslab PCE concentrations vs. time, mitigation testing period (TO-17 data) 6-8
6-7d. Plot of subslab PCE concentrations vs. time, mitigation testing period; real time GC 6-9
6-8a. Plot of wall port chloroform concentrations vs. time (method TO-17) 6-11
IX
-------
6-8b. Plot of WP-3 chloroform concentrations vs. time (online GC) 6-12
6-9a. Plot of wall port PCE concentrations vs. time (method TO-17) 6-13
6-9b. Plot of WP-3; PCE concentrations vs. time (online GC) 6-14
6-10. Chloroform concentrations at subslab and 6-ft soil gas ports directly under the 420 side of
duplex 6-18
6-11. PCE concentrations at 6-ft soil gas ports and subslab immediately below the 420 side of
the duplex 6-19
6-12. Chloroform concentrations at 9-ft soil gas ports below 420 side of the duplex 6-20
6-13. PCE concentrations at soil gas points 9 ft below the 420 side of duplex 6-21
6-14. Chloroform concentrations in soil gas at 13 ft below the 420 side of the duplex 6-22
6-15. PCE concentrations in soil gas at 13 ft below the 420 side of duplex 6-23
6-16. Chloroform concentrations in soil gas at 16.5 ft below the 420 side of duplex 6-24
6-17. PCE concentrations in soil gas at 16.5 ft below the 420 side of the duplex 6-25
6-18. Chloroform concentrations in 6-ft soil gas and subslab ports immediately below the 422
side of the duplex 6-26
6-19. PCE concentrations in 6-ft soil gas ports and subslab ports directly below the 422 side of
the duplex 6-27
6-20. Chloroform concentrations in soil gas port at 9-ft depth below 422 side of duplex 6-28
6-21. PCE concentrations in soil gas at 9 ft below the 422 side of the duplex 6-29
6-22. Chloroform concentrations in soil gas at 13 ft below the 422 side of the duplex 6-30
6-23. PCE concentrations in soil gas at 13 ft below the 422 side of the duplex 6-31
6-24. Chloroform concentrations at 16.5 ft below the 422 side of the duplex 6-32
6-25. PCE concentrations at 16.5 ft below the 422 side of the duplex 6-33
6-26. Chloroform concentrations in exterior soil gas at 3.5 ft bis 6-34
6-27. PCE concentrations in exterior soil gas at 3.5 ft bis 6-35
6-28. Chloroform concentrations in exterior soil gas at 6 ft. bis 6-36
6-29. PCE concentrations in exterior soil gas at 6 ft bis 6-37
6-30. Chloroform concentrations in exterior soil gas at 9 ft bis 6-38
6-31. PCE concentrations in exterior soil gas at 9 ft bis 6-39
6-32. Chloroform concentrations in exterior soil gas at 13 ft bis 6-40
6-33. PCE concentrations in exterior soil gas at 13 ft bis 6-41
6-34. Chloroform concentrations in exterior soil gas at 16.5 ft bis 6-42
6-35. PCE concentrations in exterior soil gas at 16.5 ft bis 6-43
6-36. Subslab PCE concentrations over a 1-week period during the first intensive round 6-44
6-37. Subslab PCE concentrations over a 1-week period during the second intensive round 6-44
6-38. Radon: Weekly time integrated samples (electret) 6-45
x
-------
6-39. Online GC chloroform indoor air data for 422 first floor 6-46
6-40. Online GC chloroform indoor air data for 422 basement 6-47
6-41. Online GC chloroform indoor air data for 420 first floor 6-47
6-42. Online GC chloroform indoor air data for 420 basement 6-48
6-43. Online GC PCE indoor air data for 422 first floor 6-49
6-44. Online GC PCE indoor air data for 422 basement 6-49
6-45. Online GC PCE indoor air data for 420 first floor 6-50
6-46. Online GC PCE indoor air data for 420 basement 6-50
6-47. Online GC subsurface chloroform soil gas data—Phase 1 and Phase 2 6-52
6-48. Online GC subsurface chloroform soil gas data—Phase 1 6-52
6-49. Online GC subsurface chloroform soil gas data—Phase 2 6-53
6-50. Online GC subsurface PCE soil gas data—Phase 1 and Phase 2 6-54
6-51. Online GC subsurface PCE soil gas data—Phase 1 6-54
6-52. Online GC subsurface PCE soil gas data—Phase 2 6-55
6-53. Method 10-17 data for SSP-4 6-55
6-54. Online GC PCE measurements in SSP-4 6-57
6-55. Comparison of online GC measurements of PCE and chloroform in SGP9 at 6 ft 6-57
6-56. Real-time radon levels (422 basement) 2011-2013 6-58
6-57. Real-time radon levels (422, 2nd floor office), 2011-2013 6-59
6-58. Temperature records from the external temperature monitor and the HOBO devices at
seven indoor locations on the 422 and 420 sides of the house 6-61
6-59. Stacked hydrological graph with rainfall in inches (top—green line), depth to water in
feet (middle—red circles), and discharge at Fall Creek in ft3/s (bottom—blue line) 6-62
6-60. Plot of high wind speed for measurement period, wind run and wind speed (average over
measurement period) at 422/420 house over time 6-63
6-61. Weather variables measured inside 422 office (2nd floor) and on roof: a. barometric
pressure (in Hg); b. indoor air density, c. indoor air equilibrium moisture content, d,
indoor percent humidity, f. outdoor percent humidity, g. rain (inches total in measurement
period), h. rain rate—the most intense rainfall during the measurement period in
inches/hour 6-64
6-62. Snow depth vs. time (data are from NCDC records for the Indianapolis International
Airport) 6-65
6-63. GC Phase 2 VOCs at WP-3 compared with 422/420 house external pressure 6-66
6-64. GC Phase 3 VOCs at WP-3 compared with 422/420 house external pressure 6-67
6-65. GC Phase 2 VOCs at WP-3 compared with 422/420 house external wind speed 6-68
6-66. GC Phase 3 VOCs at WP-3 compared with 422/420 house external wind speed 6-69
XI
-------
6-67. GC chloroform at subslab and soil gas ports versus radon from stationary
AlphaGUARDs 6-72
6-68. GC PCE at subslab ports versus radon from stationary AlphaGUARDs 6-73
6-69. GC chloroform in indoor air versus radon from stationary AlphaGUARDs 6-74
6-70. GC PCE in indoor air versus radon from stationary AlphaGUARDs 6-75
6-71. GC chloroform concentrations in indoor air, 420 first floor 6-76
6-72. GC chloroform concentrations in indoor air, 420 basement south 6-76
6-73. GC PCE concentrations in indoor air, 420 first floor 6-77
6-74. GC PCE concentrations in indoor air, 420 basement south 6-77
7-1. Comparison of mean concentrations among wall and subslab ports under different
mitigation conditions (heat on data only): radon (top), PCE (middle), and chloroform
(bottom) 7-2
7-2. Long-term trends in radon and VOCs with shading showing mitigation status during
Phase 2 7-5
8-1. AlphaGUARD concentrations and daily radon attenuation factors (number of samples
indicated below each bar graph pair). Sample pairs represented indicated along x-axis.
SSP-7 is on the 420 side of the building and so may not be representative of subslab
attenuation as the other sample pairs. Sampling period = January 2011-May 2013,
mitigation system off. 8-3
8-2. Daily indoor air and soil gas PCE concentrations used in attenuation factor calculations
(number of samples with heat on and off are indicated below each bar graph pair) 8-4
8-3. Daily PCE attenuation factors (number of samples indicated below each bar graph pair) 8-5
8-4. Daily chloroform concentrations used in attenuation factor calculations (number of
samples indicated below each bar graph pair) 8-6
8-5. Daily chloroform attenuation factors (number of samples indicated below each bar graph
pair) 8-7
8-6. Electret radon concentrations used in weekly attenuation factor calculations. Number of
measurements indicated below each box and whiskers pair 8-9
8-7. AlphaGUARD soil gas radon concentrations (pCi/L) used in weekly attenuation factor
calculations. Number of measurements noted below each box and whiskers pair. Note
that these concentrations represent averages across all samples taken at the same depth
and time (e.g., subslab samples are averages across all subslab points) 8-10
8-8. Weekly radon attenuation factors (number of samples indicated below each bar graph
pair). Note that these attenuation factors were calculated from indoor air and soil gas
concentrations averaged at the same depth and time (as described in Figure 8-7) 8-11
8-9. Radiello PCE concentrations used in attenuation factor calculations. Numbers at the
bottom of each column indicate the number of available readings for unheated (left) and
heated (right) conditions 8-12
8-10. TO-17 PCE concentrations used in attenuation factor calculations 8-13
8-11. Weekly PCE attenuation factors (numbers of samples indicated by numbers beneath each
bar graph pair) 8-14
Xll
-------
8-12. Radiello chloroform concentrations used as numerators for weekly attenuation factor
calculations 8-15
8-13. TO-17 chloroform concentrations used as denominators for weekly attenuation factor
calculations 8-16
8-14. Weekly chloroform attenuation factors (number of samples indicated below each bar
graph pair) 8-17
8-15. Attenuation of PCE, chloroform, and radon juxtaposed 8-18
8-16. Mitigation status and schedule 8-19
8-17. Radon indoor air and soil gas concentrations used in daily attenuation factor calculations 8-20
8-18. Daily radon attenuation factors 8-21
8-19. Daily PCE attenuation factors 8-22
8-20. Daily chloroform attenuation factors 8-23
8-21. Electret indoor radon concentrations 8-24
8-22. AlphaGUARD subsurface radon concentrations (pCi/L) 8-25
8-23. Weekly radon attenuation 8-26
8-24. Radiello indoor air PCE concentrations 8-27
8-25. TO-17 soil gas PCE concentrations 8-28
8-26. Weekly PCE attenuation 8-29
8-27. Radiello weekly indoor air chloroform concentrations 8-30
8-28. TO-17 soil gas chloroform concentrations 8-31
8-29. Weekly chloroform attenuation factors 8-32
8-30. Attenuation of PCE, chloroform, and radon juxtaposed 8-33
9-1. Long-term trend in subslab vs. basement differential pressure (Pa) compared with
exterior temperature and the first derivative of exterior temperature (°F) 9-2
9-2. XY plot of subslab vs. basement differential pressure vs. daily low exterior temperature 9-3
9-3. XY plot of daily low exterior temperature vs. (subslab vs. basement differential pressure) 9-4
9-4. Long term pressure trends in subslab vs. basement differential pressure (Pa) compared
with external barometric pressure (inches) with derivative plots 9-5
9-5. XY plot of external barometric pressure vs. (422 subslab vs. basement differential
pressure) 9-6
9-6. XY plot of barometric pressure drop (per hour) vs. (422 subslab vs. basement differential
pressure) 9-7
9-7. Long term trends in subslab vs. basement pressure (Pa) compared to rainfall and snow
depth (inches) 9-8
9-8. XY graph of total daily rainfall vs. (subslab vs. basement differential pressure) 9-9
9-9. XY graph of total snow depth (inches) vs. differential pressures (Pa) 9-10
Xlll
-------
9-10. Long-term trends in subslab vs. basement differential pressure (Pa) compared to
maximum wind speed and average wind speed data (mph) 9-11
9-11. Long-term trend in subslab vs. basement differential pressure (Pa) vs. wind direction
parameters (change in direction, maximum direction, average direction) (degrees) 9-12
9-12. XY plot of daily high wind speed vs. (subslab vs. basement differential pressure) 9-13
9-13. XY plot of daily average wind speed vs. (subslab vs. basement differential pressure) 9-14
9-14. XY plot of wind direction effects on subslab vs. basement differential pressure (upper
plot) and radon concentrations in the 422 basement (lower plot) 9-15
9-15. Extreme Event 1: subslab vs. basement differential pressure (positive difference indicates
flow into the structure) 9-17
9-16. Extreme Event 2: subslab vs. basement differential pressure (positive difference indicates
flow into the structure) 9-18
9-17. Extreme Event 3: subslab vs. basement differential pressure (positive difference indicates
flow into the structure) 9-19
9-18. Extreme Event 4: subslab vs. basement differential pressure (positive difference indicates
flow into the structure) 9-20
9-19. Extreme Event 5: subslab vs. basement differential pressure (positive difference indicates
flow into the structure) 9-21
9-20. Extreme Events 6/7: subslab vs. basement differential pressure (positive difference
indicates flow into the structure) 9-22
9-21. Detailed time series of unusual pressure Event 1 showing barometric pressure changes
(barometric pressure in inches of Hg, differential pressure in Pa) 9-23
9-22. Detailed time series of unusual pressure Event 1 showing precipitation events (rainfall in
inches, snow depth in inches, differential pressure in Pa) 9-24
9-23. Detailed time series of unusual pressure Event 2 showing barometric pressure changes
(barometric pressure in inches of Hg, differential pressure in Pa) 9-25
9-24. Detailed time series of unusual pressure Event 2 showing temperature changes
(differential pressure in Pa, temperature in °F) 9-26
9-25. Detailed time series of unusual pressure Event 3 showing barometric pressure changes
(barometric pressure in inches of Hg, differential pressure in Pa) 9-27
9-26. Detailed time series of unusual pressure Event 3 showing temperature changes
(differential pressure in Pa, temperature in °F) 9-28
9-27. Detailed time series of unusual pressure Event 3 showing wind direction variables (wind
direction-related variables in degrees, differential pressure in Pa) 9-29
9-28. Detailed time series of unusual pressure Event 3 showing PCE and radon 9-30
9-29. Detailed time series of Event 4 showing wind direction variables (wind direction-related
variables in degrees, differential pressure in Pa) 9-32
9-30. Detailed time series of Event 4 showing wind speed variables (wind speed variables in
MPH, differential pressure in Pa) 9-33
9-31. Detailed time series of Event 4 showing temperature variables (differential pressure in
Pa, temperature in °F) 9-34
xiv
-------
9-32. Detailed time series of Event 5 showing precipitation event (rainfall in inches, snow
depth in inches, differential pressure in Pa) 9-35
9-33. Detailed time series of Event 5 showing wind speed variables (wind direction-related
variables in degrees, differential pressure in Pa) 9-36
9-34. Detailed time series of Event 5 showing wind direction variables 9-37
9-35. Detailed time series of Event 5 showing temperature variables (differential pressure in
Pa, temperature in °F) 9-38
9-36. Detailed time series of Event 5 showing PCE and radon 9-39
9-37. Detailed time series of Event 6/7 showing barometric pressure (barometric pressure in
inches of Hg, differential pressure in Pa) 9-40
9-38. Detailed time series of Event 6/7 showing precipitation (rainfall in inches, snow depth in
inches, differential pressure in Pa) 9-41
9-39. Detailed time series of Event 6/7 showing wind speed variables (wind direction-related
variables in degrees, differential pressure in Pa) 9-42
9-40. Detailed time series of Event 6/7 showing wind direction variables (wind direction-
related variables in degrees, differential pressure in Pa) 9-43
9-41. Detailed time series of Event 6/7 showing radon and PCE 9-44
9-42. XY graph of total heating degree days per week vs. weekly PCE concentration (Radiello
data) 9-45
9-43. XY graph of heating degree days vs. chloroform concentration (weekly Radiello data) 9-46
9-44. XY graph of average weekly barometric pressure vs. PCE concentration 9-47
9-45. XY graph of weekly average snow depth vs. PCE concentration 9-48
9-46. XY graph of weekly average snow depth vs. chloroform concentration 9-49
9-47. XY graph of weekly average snow depth vs. radon concentration 9-50
9-48. XY graph of total weekly rainfall vs. PCE concentration in indoor air 9-51
9-49. XY graph of total weekly rainfall vs. chloroform concentration in indoor air 9-52
9-50. XY graph of wind direction vs. PCE concentration in indoor air in 422 basement 9-53
9-51. XY graph of wind direction vs. chloroform concentration in indoor air in 422 basement 9-54
9-52. XY graph of wind direction vs. radon concentration in indoor air in 422 basement 9-55
9-53. Modeled effect of building wind loads on ground surface and subslab gauge pressure
distribution (adapted from U.S. EPA, 2012d) 9-56
9-54. Modeled effect of building wind load on subslab soil vapor distribution for recalcitrant
and aerobically biodegradable VOCs (adapted from U.S. EPA, 2012d) 9-57
9-55. XY graph of wind speed vs. PCE concentration in indoor air in 422 basement 9-58
9-56. XY graph of wind speed vs. chloroform concentration in indoor air in 422 basement 9-59
xv
-------
10-1. Daily time series of radon concentrations (pCi/L) 422 basement north, 2011-2012 with
rolling-averages, and autocorrelation (ACF) and partial autocorrelation function (PACF)
plots 10-4
10-2. Time series of first difference of daily radon concentrations (pCi/L) 422 basement north,
2011-2012 with rolling-averages, and autocorrelation (ACF) and partial autocorrelation
function (PACF) plots 10-5
10-3. Daily time series of radon concentrations (pCi/L) 422 basement north, 2012-2013 with
rolling-averages, and autocorrelation (ACF) and partial autocorrelation function (PACF)
plots 10-6
10-4. Time series of first difference of daily radon concentrations (pCi/L) 422 basement north,
2012-2013 with rolling-averages, and autocorrelation (ACF) and partial autocorrelation
function (PACF) plots 10-7
10-5. Time series of daily radon concentrations: 422 office (2nd floor), 2011-2012 with
rolling-averages, and autocorrelation (ACF) and partial autocorrelation function (PACF)
plots 10-8
10-6. Time series of first difference of daily radon concentrations: 422 office (2nd floor),
2011-2012 with rolling-averages, and autocorrelation (ACF) and partial autocorrelation
function (PACF) plots 10-9
10-7. Time series of radon concentrations: 422 office (2nd floor), 2012-2013 with rolling-
averages, and autocorrelation (ACF) and partial autocorrelation function (PACF) plots 10-10
10-8. Time series of differences of daily radon concentrations: 422 office (2nd floor), 2012-
2013 with rolling-averages, and autocorrelation (ACF) and partial autocorrelation
function (PACF) plots 10-11
10-9. Correlation between radon concentration and predictors for both sites (basement north
and office 2nd floor). Time period 2011-2012. NOTE: See Table 10-2 for "Plain
Language" key of abbreviations used in this figure 10-16
10-10. Correlation between radon concentration and predictors for both sites (basement north
and office 2nd floor). Time period 2012-2013. NOTE: See Table 10-2 for "Plain
Language" key of abbreviations used in this figure 10-17
10-11. Time series plot, ACF and PACF for weekly Radiello chloroform. Location X422
basement south. Time period: Jan 5, 2011-Feb 15, 2012 10-42
10-12. Time series plot, ACF and PACF for first difference of weekly Radiello. Chloroform.
Location X422 basement south. Time period: Jan 5, 2011-Feb 15, 2012 10-43
10-13. Time series plot, ACF and PACF for weekly Radiello PCE. Location X422 basement
south. Time period: Jan 5, 2011-Feb 15, 2012 10-44
10-14. Time series plot, ACF and PACF for weekly chloroform. Location X422 basement south.
Time period: Sept 26, 2012-April 10, 2013 10-46
10-15. Time series plot, ACF and PACF for first difference weekly Radiello CHCI3-2
(X422BaseS_Radiello_Weekly_CHC13). Location X422 basement south. Time period:
Sept 26, 2012-April 10, 2013 10-47
10-16. Time series plot, ACF and PACF for weekly Radiello PCE-2
(X422BaseS_Radiello_Weekly_PCE). Location X422 basement south. Time period: Sept
26, 2012-April 10, 2013 10-48
xvi
-------
10-17. Time series plot, ACF and PACF for first difference weekly Radiello PCE-2
(X422BaseS_Radiello_Weekly_PCE). Location X422 basement south. Time period: Sept
26, 2012-April 10, 2013 10-49
10-18. XY plot of weekly average snow depth vs. PCE concentration 10-63
10-19. XY Plot of change in weekly average snow depth vs. PCE concentration 10-64
10-20. XY Plot of weekly average snow depth vs. change in PCE 10-65
11-1. Stacked hydrological graph with rainfall in inches (top—green line), depth to water in
feet (middle—red circles), and discharge at Fall Creek in ft3/s (bottom—blue line) 11-1
11-2. Actual groundwater levels along with the daily time series predicted from Fall Creek
gage height data 11-2
11-3. Groundwater concentrations over time for Indianapolis duplex 11-4
11-4. PCE groundwater concentrations over time, showing concentrations by individual well
and soil gas ports 11-4
11-5. Plot of groundwater PCE and chloroform concentrations against well screen (or soil gas
port) depth (well depth measured to top of screen) 11-5
11-6. Plot of groundwater PCE and chloroform concentrations against groundwater depth 11-6
12-1. Response at all locations to first helium injection in front yard 12-5
12-2. Response at all locations to second helium injection in front yard 12-6
12-3. Response at all locations to third helium injection in backyard 12-6
12-4. Response at all locations to fourth helium injection in backyard 12-7
12-5. Helium response at SGP2 cluster (south of duplex) directly above injection point after
injections 1 (top graph, mitigation off) and 2 (bottom graph, mitigation on) 12-8
12-6. Helium response at SGP6 cluster at SGP6 cluster (north of duplex) directly above
injection point after injections 3 (top graph, mitigation off) and 4 (bottom graph,
mitigation on) 12-9
12-7. Helium response at SGP1 cluster (south of duplex, approximately 6 ft closer to duplex
than injection point) after first Injection (top graph, mitigation off) and second injection
(bottom graph, mitigation on) 12-10
12-8. Helium response at SGP5 cluster (north of duplex, approximately 6 ft closer to duplex
than injection point) after third injection (top graph, mitigation off) and fourth injection
(bottom graph, mitigation on) 12-11
12-9. Helium response to first and second helium injections at SGP9 (interior) 12-12
12-10. Helium response to third and fourth helium injections at SGP10 (interior) 12-12
12-11. Cross-section view of helium tracer arrival at 6-ft depth intervals 12-14
12-12. Comparison of electret and Safety Siren results for first phase of the project (top graph)
and second project phase (bottom graph) 12-17
12-13. Comparison between the 422 office Safety Siren and electret 12-22
12-14. Comparison between the 422 basement N Safety Siren and electret 12-22
XVll
-------
xviii
-------
List of Tables
1 -1. Summary of Lines of Evidence for Meteorological Factors Influencing Vapor Intrusion in
This Study (Blank cells reflect types of analysis not completed for a given parameter) 1-7
2-1. VOC Indoor Air Sampling Method Options 2-15
2-2. Continuing Proj ect Obj ectives Addressed in this Document 2-17
2-2. Continuing Project Objectives Addressed in this Document (continued) 2-18
2-3. Factors Causing Temporal Change in Vapor Intrusion and How They are Observed and
Measured 2-19
2-4. Data Quality Objectives and Criteria 2-20
3-1. Pressure Readings Taken During Extraction Point Testing 3-13
3-2. Data Aggregation Applied to Predictor Variables 3-31
4-1. Indoor Air Passive Field Blank Summary—Radiello 130 4-2
4-2. Indoor Air Passive Trip Blank Summary—Radiello 130 4-2
4-3. Indoor Air Passive Laboratory Blank Summary—Radiello 130 4-2
4-4. Indoor Air Passive Surrogate Summary—Radiello 130 4-3
4-5. Indoor Air Passive LCS Summary—Radiello 130 4-4
4-6. Indoor Air Passive Laboratory Precision (LCS/LCSD) Summary—Radiello 130 4-4
4-7. Subslab and Soil Gas—EPA Field Blank Summary—TO-17 4-5
4-8. Subslab and Soil Gas—EPA Trip Blank Summary—TO-17 4-6
4-9. Subslab and Soil Gas—EPA Laboratory Blank Summary—TO-17 4-6
4-10. Subslab and Soil Gas—EPA Fridge Blank Summary—TO-17 4-7
4-11. EPA TO-17 Calibration Verification (CV) Summary 4-8
4-12. EPA TO-17 Internal Standard (IS) Summary 4-8
4-13. EPA TO-17 Surrogate Recovery Summary 4-9
4-14. EPA TO-17 Laboratory Control Sample (LCS) Summary 4-9
4-15. EPA TO-17 Field Duplicate Summary 4-10
4-16. Field GC Estimated Minimum Detection Limits and Practical Quantitation Limits 4-11
4-17. Result of Repeated TCE Calibration Standard Analyses on On-line GC in March 2013
(Hartman Period 3) 4-14
4-18. Results of Repeated PCE Calibration Standard Analyses on Online GC in March 2013
(Hartman Period 3) 4-14
4-19. Interlaboratory Results: Spiked Verification Samples 4-15
4-20. Interlaboratory Statistics: Spiked Verification Samples 4-16
4-21. Comparison of Online GC to Radiello Results by Week 4-17
xix
-------
4-22. Comparison between Electrets and Charcoal Canisters at the 422/420 EPA House from
January 19-26, 2011 4-43
4-23. Comparison of Electret and Charcoal Canister Data from April 27 to May 4, 2011 4-44
4-24. Comparison of Charcoal and Electret Radon December 28, 2011, to January 4, 2012 4-44
4-25. Comparison between 422 Basement N AlphaGUARDs and Electrets from March 30,
2011, and May 18,2011 4-45
4-26. Comparison of Real-Time AlphaGUARD to Integrated Electret August through October 4-46
4-27. Comparison of Real-Time AlphaGUARD to Integrated Electret Measurements
December 28, 2011, to January 4, 2012 4-46
4-28. Comparison of Real-Time AlphaGUARD to Integrated Electret Measurements January
through March 2012 4-47
4-29. Comparison of Real-Time AlphaGUARD to Integrated Electret Measurements January
through March 2013 4-47
4-29. Comparison of Real-Time AlphaGUARD to Integrated Electret Measurements January
through March 2013 4-48
4-30. Groundwater (5 mL)—EPA Field Blank Summary 4-52
4-31. Groundwater (5 mL)—EPA Laboratory Blank Summary—TO-17 4-52
4-32. Groundwater (25 mL)—EPA Field Blank Summary—TO-17 4-52
4-33. Groundwater (25 mL)—EPA Laboratory Blank Summary—TO-17 4-53
4-34. EPA Groundwater (5 mL) Surrogate Recovery Summary 4-53
4-35. EPA Groundwater (25 mL) Surrogate Recovery Summary 4-53
5-1. Subslab vs. Basement Differential Pressures Measured with Handheld Micromanometer
at Permanent SSPs (negative pressure indicates flow out of building; yellow indicates
mitigation off) 5-4
5-2. Wall Port vs. Basement Differential Pressure Measured with Handheld Micromanometer
(negative pressure indicates flow out of building) 5-8
5-3. Shallow Interior SGP (6 ft bis) vs. Basement Differential Pressure Measured with
Handheld Micromanometer (negative pressure indicates flow out of building) 5-9
5-4. Shallow Exterior SGP (3.5 ft and 6 ft bis) vs. Basement Differential Pressure Measured
with a Handheld Micromanometer (negative pressure indicates flow out of building) 5-10
5-5. Deep Interior SGP (9 ft and 13 ft bis) vs. Basement Differential Pressure Measured by
ARCADIS with Handheld Micromanometer (negative pressure indicates flow out of
building) 5-12
5-6. Deep Exterior SGP (9 ft and 13 ft bis) vs. Basement Differential Pressure Measured by
ARCADIS with Handheld Micromanometer (negative pressure indicates flow out of
building) 5-15
5-7. Comparison of Setra Continuous Sensor Differential Pressure vs. Airdata Multimeter
ADM-870 with SSD System Operating: December 29, 2012 (yellow shaded data reflects
an "off scale" response on the Setra) 5-20
5-8. Electret Radon Descriptive Statistics by Mitigation and Heating Status (pCi/L) 5-25
xx
-------
5-9. Indoor Air Radon Descriptive Statistics by Mitigation and Heating Status: From
Stationary Real Time AlphaGUARD (pCi/L) 5-26
5-10. Indoor Radon Descriptive Statistics—Individual Locations by Mitigation and Heating
Status: Electret Data (pCi/L) 5-26
5-11. Descriptive Statistics: Radon in Subslab and Wall Ports by Individual Location and
Mitigation and Heating Status (pCi/L) 5-28
5-12. Radon Descriptive Statistics by Location Type and Mitigation and Heating Status (pCi/L) 5-30
5-13. Descriptive Statistics of Weekly Passive VOC Measurements ((ig/m3) in Indoor Air by
Mitigation Status and Heating Use (yellow indicates statistics during active mitigation) 5-33
5-14. Distribution of Concentrations ((.ig/ni3) by VOC and Mitigation and Heating Status:
Indoor Air, Week-Long Passive Samples (yellow indicates statistics during active
mitigation) 5-34
5-15. Descriptive Statistics of Indoor VOC Concentrations (|ig/m3) During Mitigation Testing
by Location and Mitigation and Heating Status (yellow indicates statistics during active
mitigation) 5-35
5-16. Descriptive Statistics: Average Subslab and Wall Port VOC Concentrations (|ig/m3) by
Mitigation and Heating Status (yellow indicates statistics during active mitigation) 5-39
5-17. Distribution of Subslab and Wall Port VOC Concentrations ((ig/m3) by Mitigation and
Heating Status (yellow indicates statistics during active mitigation) 5-40
5-18. Descriptive Statistics of Subslab and Wall Port VOC Concentrations ((.ig/ni3) by Location
and Mitigation and Heating Status (yellow indicates statistics during active mitigation) 5-41
5-19. April/May 2011 Air Exchange Rate Measurement Results 5-51
5-20. September 2011 Air Exchange Rate Measurement Results 5-51
5-21. October 2011 Air Exchange Measurement Results (during and after fan testing) 5-52
5-22. April 2013 Air Exchange Measurement Results (During Mitigation) 5-52
5-23. National Survey of Air Exchange Rates, Reprinted from the EPA Exposure Factor
Handbook (U.S. EPA, 2011) 5-52
5-23. National Survey of Air Exchange Rates, Reprinted from the EPA Exposure Factor
Handbook (U.S. EPA, 2011) (continued) 5-53
5-24. Stack Gas Discharge Measurements During Mitigation 5-54
6-1. Frequency of Nondetectable Samples (%) by Soil Gas Point or Cluster 6-16
6-2. Frequency of Nondetects in TO-17 VOC Data by Soil Gas Sampling Depth 6-16
6-3. Summary Meteorological Data for Central Indiana Note that the symbols "A" and "v"
mean "above" and "below" normal, respectively, and that the weekly values show how
the weekly averages differ from normal (from Scheeringa and Hudson, 2012, 2013) 6-60
6-4. Summary of Meteorological Data During the 2012/2013 Snow and Ice Events 6-78
9-1. Summary of Qualitative Lines of Evidence for Meteorological Factors Influencing Vapor
Intrusion in This Study 9-60
10-1. Significant Lags for ACF and PACF for Each Time Series and Regression Model 10-12
10-2. Significant Lag and AR Model by Predictor and Site Location 10-13
xxi
-------
10-3. Model Parameters, Standard Errors by Model, Predictor and Time Series: All Radon
Time Periods, Categorical Variables 10-18
10-4. Model Parameters, Standard Errors by Model and Predictor for X422baseN_AG_radon
(X422BN-1): 2011-2012, Lag 1 Models 10-20
10-5. Model Parameters, Standard Errors by Predictor for X422baseN_AG_radon (X422BN-
1): 2011-1012, Lag 2 Models 10-25
10-6. Model Parameters, Standard Errors by Predictor for X422baseN_AG_radon (X422BN-1),
Lag 4 Models 10-25
10-7. Model Parameters, Standard Errors by Predictor for X422office_2nd_AG_radon
Concentration (X4220F2-1): 2011-2012, No Lag Terms in Model 10-26
10-8. Model Parameters, Standard Errors by Model and Predictor for Time Series Analysis of
Radon in 422 Office: 2011-2012, Lag 1 Models 10-26
10-9. Model Parameters, Standard Errors by Predictor for X422office_2nd_AG_radon
Concentration (X4220F2-1): 2011-2012. Lag 2 Models 10-32
10-10. Model Parameters, Standard Errors by Predictor for X422baseN_AG_radon
(X4220F2-1): 2011-2012. Lag4 Models 10-32
10-11. Model Parameters, Standard Errors by Predictor for 422 Basement Radon: 2012-2013.
No Lag Terms in Model 10-33
10-12. Model Parameters, Standard Errors by Model and Predictor for 422 Basement Radon:
2012-2013 Lag 1 Models 10-33
10-13. Model Parameters, Standard Errors by Predictor for X422office_2nd_AG_radon
Concentration (X4220F2-2): 2012-2013. No Lag Terms in Model 10-37
10-14. Model Parameters, Standard Errors by Model and Predictor for
X422office_2nd_AG_radon Concentration (X4220F2-2): 2012-2013 Lag 1 Models 10-38
10-15. Name, Periodicity, Time Period, and Location of Time Series (Outcome) Considered 10-41
10-16. Transformation and Terms Required by Time Series 10-45
10-17. Continuous Covariates by Time Period 10-50
10-18. Analysis for Outcome First Difference of X422BaseS_Radiello_Weekly_CHCl3.
Variables That Did Not Need Lag Terms. Period Jan 5, 2011-Feb 15, 2012 10-53
10-19. Analysis for Outcome X422BaseS_Radiello_Weekly_CHCl3. Variables that Needed a
lag-1 Term. Period Jan 5. 201 1 Feb 15,2012 10-54
10-20. Analysis for Outcome X422BaseS_Radiello_Weekly_CHCl3. Variables that Needed Lag-
1 and Lag-2 Terms. Period Jan 5, 2011-Feb 15, 2012 10-58
10-21. Time Series Analysis for Outcome First Difference of 422 Basement South PCE
Concentration Variables that Did Not Need Lag Terms. Period Jan 2011 to Feb 2012 10-59
10-22. Analysis for PCE Concentration at 42 Base South, Variables that Needed a Lag-1 Term.
Period Jan 2011 to Feb 2012 10-60
10-23. Analysis for PCE Concentration at 422 Base South Variables that Needed Both Lag-1
Week And Lag-2 Week Terms. Period Jan 2011 to Feb 2012 10-64
10-24. Analysis for First Difference of Chloroform Concentration at 422 Basement South.
Variables that Did Not Need Lag Terms. Period Sept 2012 to April 2013 10-66
XXll
-------
10-25. Analysis Chloroform Concentration at 422 Base South. Variables that Needed A Lag-1
One Week Term. Period Sept 2012 to April 2013 10-67
10-26. Analysis for Chloroform Concentration at 422 Base South. Variables Needing Lag-1 and
Lag-2 Week Terms. Period Sept 2012 to April 2013 10-71
10-27. Analysis for First Difference of422 Base South PCE Concentration. Variables that Did
Not Need Lag Terms. Period Sept 2012 to April 2013 10-72
10-28. Analysis for422 Base South PCE Concentration. Variables that Needed A Lag-1 Week
Term. Period Sept 2012 to April 2013 10-73
10-29. Analysis for PCE Concentration at 422 Base South. Variables Needing Both Lag-1 And
Lag-2 Week Terms. Period Sept 2012 to April 2013 10-77
11-1. Groundwater Monitoring Locations 11-3
12-1. Data Quality Objectives and Performance/Acceptance Criteria for Special Studies 12-2
12-2. Comparison of Safety Siren, AlphaGUARD, and Electret Radon Data 12-18
12-3. Comparison of Safety Siren and Electret Radon Data 12-20
13-1. Summary of Lines of Evidence for Meteorological Factors Influencing Vapor Intrusion in
This Study 13-6
XXlll
-------
xxiv
-------
Notice
The information in this document has been funded wholly by the United States Environmental Protection
Agency under contract number EP-C-11-036 to the Research Triangle Institute. It has been subjected to
external peer review as well as the Agency's peer and administrative review and has been approved for
publication as an EPA document. Mention of trade names or commercial products does not constitute
endorsement or recommendation for use.
Preface
This report entitled, "Assessment of Mitigation Systems on Vapor Intrusion: Temporal Trends,
Attenuation Factors, and Contaminant Migration Routes under Mitigated and Non-mitigated Conditions"
(EPA/600/R-13/241) is the second in a series of reports based on research performed to look at vapor
intrusion into a historical duplex in Indianapolis, Indiana. The research is being conducted to look at the
general principles of how vapors enter into this single residence.
The study was initiated in 2011 with the primary initial goal to investigate distributional changes in VOC
and radon concentrations in the indoor air, subslab, and subsurface soil gas from an underground source
(groundwater source and/or vadose zone source) proximal to a residence. Currently, the study has
extended more than 3.5 years in order to evaluate the effects due to seasonal variations on radon and VOC
vapor intrusion. As a result, a significant dataset has been generated that can be used to advance and
inform the understanding of vapor intrusion.
A series of at least four (4) reports are anticipated from the research at the Indianapolis duplex.
• The initial report entitled, "Fluctuation of Indoor Radon and VOC Concentrations Due to
Seasonal Variations" (EPA/600/R-12/673) examined the distributional changes in VOC and
radon concentrations in the indoor air, subslab, and subsurface from the ground water source into
a residence.
• This second report examines: (a) subsurface conditions that influence the movement of VOCs
and radon into the home; (b) effects of an installed mitigation system on VOC and radon
concentration into the residence; and (c) the influence of a winter capping event on vapor
movement into the home.
• The third report entitled, "Simple, Efficient, and Rapid Methods to Determine the Potential for
Vapor Intrusion into the Home: Temporal Trends, Vapor Intrusion Forecasting, Sampling
Strategies, and Contaminant Migration Routes" (EPA/600/R-15/XXX) will examine the use of
radon and other variables; such as weather data changes in temperature and differential pressure
between indoors and outdoors, as potential low-cost, easily monitored indicators of when to
sample for vapor intrusion events and when to turn on the mitigation system to reduce vapor
intrusion exposure to the residents. Select data trends through the years of study at this site are
also presented.
• The fourth report will provide information regarding the effectiveness of a soil vapor extraction
system in preventing vapor intrusion into the residence.
In general, because this work was conducted at a single residential duplex, it cannot be representative of
all sites and site conditions subject to vapor intrusion. However, it should be useful to compare the results
of this study of an older building in a temperate Midwest climate with other ongoing detailed studies,
such as the one conducted in a newer home in Layton, Utah for common threads that can be applied
across all vapor intrusion sites.
xxv
-------
A separate research report will be looking at the performance of passive sorbers for the monitoring of
vapor intrusion at multiples sites, including the Indianapolis duplex. It is anticipated that this report will
be released in late 2015.
It is anticipated that research will continue as new areas of scientific concern are identified and build on
the research that has been conducted to date. The publication of peer-reviewed journal articles on select
topics is also anticipated.
Acknowledgments
This project was conceived, directed, and managed by Brian Schumacher and John Zimmerman of U.S.
EPA NERL. Robert Truesdale (RTI International) and Chris Lutes (ARCADIS) led the project, with
report input by Brian Cosky (ARCADIS), Breda Munoz and Robert Norberg (RTI), Heidi Hayes (Air
Toxics Ltd.), and Blayne Hartman (Hartman Environmental Geoscience). The authors would also like to
thank the following for their valuable input to the project:
¦ Leigh Riley Evens, Executive Director of Mapleton-Fall Creek Development Corporation, the not
for profit that provided the house used for this work.
¦ Dale Greenwell, U.S. EPA NRMRL, and Ron Mosley (retired), radon and equipment support
¦ Gregory Budd, U.S. EPA Radiation and Indoor Environments National Laboratory, radon QC
sample analysis and instrument support
¦ Ausha Scott, analytical support, Air Toxics Ltd.
¦ Alan Williams and Jade Morgan, U.S. EPA NERL, TO-17 analytical support
¦ Robert Uppencamp ARCADIS - field, data interpretation, and site selection support; Sara Jonker,
ARCADIS, field support
¦ Rebecca Forbort and Valerie Kull, ARCADIS, Shu-yi Lin, RTI, data management and analysis
¦ Scott Forsberg, Harvard School of Public Health, air exchange rate measurement support
¦ Susan Beck and Sharon Barrell, RTI, document preparation and editing support.
xxvi
-------
Section 1—Executive Summary
Table of Contents
1.0 Executive Summary 1-1
1.1 Background 1-1
1.2 Purpose and Objectives 1-1
1.3 Methods 1-2
1.4 Conclusions 1-3
1.4.1 Conceptual Site Model: VOC Data 1-3
1.4.2 Mitigation System Performance-Radon 1-3
1.4.3 Mitigation System Performance-VOCs 1-4
1.4.4 Meteorological Effects on Vapor Intrusion 1-5
1.4.5 Preferential Pathways and Conceptual Site Model: Helium Tracer and
Geophysical Tests 1-5
1.4.6 Temporal Variability and Trends 1-8
1.4.7 Discussion 1-8
List of Figures
1-1. Temporal coverage of data sets collected (red line indicates the cutoff date for this
report) 1-4
1-2. Indoor air concentrations of PCE (top panel) and radon (middle and bottom panels) with
mitigation on (black bars), passive mitigation (no fan, gray bars), and mitigation off (no
bar). Snow and frozen ground events are indicated by blue bars and red dots 1-8
List of Tables
1 -1. Summary of Lines of Evidence for Meteorological Factors Influencing Vapor Intrusion in
This Study (Blank cells reflect types of analysis not completed for a given parameter) 1-6
1-i
-------
Section 1—Executive Summary
-------
Section 1—Executive Summary
1.0 Executive Summary
1.1 Context in Overall Research Program
Vapor intrusion is the migration of subsurface vapors, including radon and volatile organic compounds
(VOCs),1 in soil gas from the subsurface to indoor air. Vapor intrusion happens because there are pressure
and concentration differentials between indoor air and soil gas. Indoor environments are often negatively
pressurized with respect to outdoor air and soil gas, for example, from exhaust fans or the stack effect,2
and this pressure difference allows soil gas containing subsurface contaminant vapors to flow into indoor
air through advection. In addition, concentration differentials cause VOCs and radon to migrate from
areas of higher to lower concentrations through diffusion, which is can lead to vapor intrusion.
While vapor intrusion investigations have been ongoing for many years, several issues still remain. Vapor
intrusion site investigation costs are driven higher by the need for multiple samples per structure to
characterize the commonly observed spatial and temporal variability in indoor, subslab, and deep soil gas
concentrations. However, relatively few vapor intrusion assessment data sets have been published that
include both long-term monitoring and high-frequency sample collection for VOCs. Temporal variability
in VOC concentrations in indoor air is expected to be driven by variation in barometric pressure, house
operations, temperature, water table, and soil moisture. These phenomena have known but irregular
cycles on multiple time scales.
Subslab depressurization (SSD) is the predominant technology used for mitigating vapor intrusion.
Design practices for SSD systems have been adapted essentially verbatim from radon mitigation
experience. Few highly detailed long-term data sets have been published from tests of the effectiveness of
SSD mitigation systems on indoor air VOC concentrations from vapor intrusion.
1.2 Purpose and Objectives
The main initial goal of this project was to investigate distributional changes in VOC and radon
concentrations in the indoor air, subslab, and subsurface soil gas from an underground source
(groundwater source and/or vadose zone source) proximal to a residence. The time frame of this study
was more than 2.5 years in order to evaluate the effects due to seasonal variations on radon and VOC
vapor intrusion. Because this work was conducted at a single residential duplex, it cannot be
representative of all sites and site conditions subject to vapor intrusion. However, it should be useful to
compare the results of this study of an older building in a temperate mid-West climate with other ongoing
detailed studies, such as the one conducted in a newer home in Logan, Utah (Johnson et al., 2012; Holton
et al., 2013) for common threads that can be applied across all vapor intrusion sites.
We reported previously on our results from studies conducted in 2011-2012 prior to mitigation testing
(U.S. Environmental Protection Agency [EPA], 2012a). Here we report primarily new results of studies
conducted in 2012-2013, additional analyses of temporal variability encompassing the entire data set
collected, and the results of continuing research on six objectives initially established for the previous
(U.S. EPA, 2012a) research effort:
'Mercury in certain forms is sufficiently volatile and toxic to pose a vapor risk, but mercury vapor investigations are very site
specific and much rarer than those addressing VOCs and radon. As a result, mercury vapor intrusion is not discussed further in
this document.
2The stack effect is the movement of air into, upwards, and out of buildings, chimneys, flue gas stacks, or other containers
resulting from indoor/outdoor air density differences due to temperature and moisture gradients between indoor and outdoor air.
1-1
-------
Section 1—Executive Summary
1. Identify seasonal fluxes in radon and VOC concentrations as they relate to a typical use of
heating, ventilation, and air conditioning (HVAC) in the building.
2. Establish relationship between subslab/subsurface soil gas and indoor air concentrations of VOCs
and radon.
3. Determine the relationship of radon to VOC concentrations in, around, and underneath the
building.
4. Characterize the near-building environment sufficiently to explain the observed variation of
VOCs and radon in indoor air.
5. Determine whether the observed changes in indoor air concentration of volatile organics of
interest can be mechanistically attributed to changes in vapor intrusion.
6. Evaluate the extent to which groundwater concentrations and/or vadose zone sources control soil
gas and indoor air VOC concentrations at this site.
New objectives established for the 2012-2013 studies include the following:
¦ Better define the particular subsurface conditions that influence the movement of VOCs and
radon into this home. These conditions were expected to include differences in air permeability
on a spatial scale of 1 to 20 ft in the vadose zone beneath and immediately adjacent to the
structure, along with information on potential preferential pathways and conditions beneath the
foundation slab.
¦ Design, install, and monitor a mitigation system based on the predominant vapor intrusion
mitigation technology—SSD. We wish to determine how well the mitigation system worked in
reducing indoor radon and VOC concentrations for this particular well-studied duplex.
¦ Capture a winter snow/ice capping event to monitor its influence on radon and VOC vapor
movement into the home.
Because this report is the second report in a series of four, it should be regarded as an interim field report
that provides results through the spring of 2013 along with initial information on the performance of the
mitigation system at the duplex. Additional studies and reports are in progress that will test some of the
associations developed from this study and provide longer term tests of mitigation system performance.
As a result, interpretations and conclusions drawn in this report are subject to change as additional
information and insights are gained on vapor intrusion processes at this duplex.
1.3 Methods
This study was conducted at a highly instrumented pre-1920 residential duplex. The house was devoid of
potential indoor VOC sources, but one half of the structure was operated as if occupied (provision of
heating and cooling). To characterize the basement of this residential duplex, serving as a vapor intrusion
research house, several sampling devices have been installed: seven conventional subslab ports, four ports
similar to conventional subslab ports, seven external nested soil gas points (5 depths per point), and five
nested soil gas points below the basement (4 depths per point). This provides for collection of an
unusually comprehensive data set to formulate three-dimensional visualizations of seasonal VOC
concentrations.
In our overall study design, we used weekly measurements to observe our dependent variable—indoor air
concentration. We expected the indoor air concentration to depend on the flux from vapor intrusion from
soil gas. Our dependent variable is thus controlled by a series of independent variables with different time
cycles that affect the vapor intrusion process, including air temperature, barometric pressure, wind, soil
moisture, soil temperature, groundwater level, and HVAC operation. In the course of this study, we
1-2
-------
Section 1—Executive Summary
monitored or measured most of these independent variables or their surrogates and different frequencies
balancing on the general desire for continuous measurements against logistic considerations.
The strategy for the SSD mitigation system installation was to select an experienced radon and VOC
mitigation contractor and ask them to perform a "typical" active SSD system installation but with greater
documentation and reporting for the research purpose. We also added some additional valves and
sampling ports to the "typical" system to facilitate much more intensive monitoring than is usually
conducted for an SSD mitigation system designed for radon.
Figure 1-1 shows the various types of samples and sampling frequency employed for each across this
study. The more continuous variables (shown with black lines) are used in time series analysis in Sections
9 and 10. With respect to radon measurements, continuous measurements (weeklong electrets or
continuous AlphaGUARD data) were taken for indoor air, while short-term grab samples were used to
characterize soil gas. Similarly, the primary VOC measurements were weeklong Radiello samples (for the
entire project) and continuous measurements with an on-site gas chromatograph (GC) (during critical
project phases) for indoor air, with TO-17 grab samples to characterize soil gas on a weekly basis.
Meteorological, observational, and pressure differential (Setra) data were collected essentially
continuously during the entire project.
1.4 Conclusions
As noted above, this document is a field report updating results of a study in progress. Because
measurements and analyses are ongoing and future work is planned to test the validity of some of the
correlations and conclusions drawn at this point in the project, the results should be regarded as
preliminary and subject to change in the third and fourth reports of this series.
1.4.1 Conceptual Site Model: VOC Data
Although chloroform was detected in groundwater, the currently measured concentrations were too low to
account for the peak chloroform concentrations observed in soil gas. This suggests that there may be
(1) other sources of chloroform such as combined sewers3 or drinking water mains4 that leak below grade,
(2) higher groundwater concentrations at some locations near the site, or (3) chloroform mass stored in
the vadose zone from a historic release. For PCE, the results indicate a groundwater source, but the
narrow range of variability in PCE concentrations over time make it unlikely that variability in
groundwater concentrations is the only source of the observed changes in soil gas or indoor air
concentrations observed in this study. The variability in indoor air PCE concentrations is also influenced
by subsurface, building-related, and meteorological variables. The potential that other sources of PCE
may exist in the vadose zone or combined sewer lines cannot be ruled out at this point.
1.4.2 Mitigation System Performance—Radon
The mitigation system installed in the duplex met or exceeded all conventional performance tests as well
as more comprehensive tests involving pressure differentials and continuous indoor radon monitoring.
Radon reductions greater than 90% were observed, and all measured radon levels were below 4 picocuries
per liter (pCi/L) with the mitigation system on.
3Chloroform can form in sewers that receive bleach-containing products.
4Groundwater chloroform concentrations at this duplex are lower than the mean and peak drinking water concentrations for
Indianapolis (19 ppb and 82 ppb).
1-3
-------
Section 1—Executive Summary
1.4.3 Mitigation System Performance—VOCs
The mitigation system did not perform as well with VOCs as it did with radon. During 7 months of
mitigation system operation, immediate VOC reductions in indoor air were observed but the system only
achieved a reduction of just over 60% of VOC indoor air concentrations before mitigation. However,
additional decreases in indoor VOC levels were observed near the end of the monitoring period reported
in this document (May 2013). During these periods of mitigation system operation, the system was also
observed to increase soil gas levels below the slab and at depth below the duplex, suggesting that VOCs
are being redistributed by the mitigation system and that soil gas concentrations close to the building may
be enhanced by drawing higher concentrations of VOCs from greater depths. In addition, several snow
events corresponded to increases in indoor air VOC levels during mitigation that were not observed for
radon.
1-4
-------
Section 1—Executive Summary
¦O
O
.c
a>
Well Monitoring-
Water Level Monitoring
TO-17 Soil Gas-
TO-15 (Soil Gas)-
TO-15 (Indoor Air)
Stationary Alphaguard Indoor
Soil Temperature
SKC575i
Setra
Rain Level Monitoring
Radiello Indoor Air-
Portable Alphaguard Soilgas
Passive Ultra III
On-site GC-
Observation
Moisture-
Met Data Station -
Indoor Temp-
Fall Creek Monitoring
Electret-
Cts. Water Level Monitoring
* • •• * «
T3
&
&
Figure 1-1. Temporal coverage of data sets collected (red line indicates the cutoff date for this
report).
Dots represent discrete sampling events. Bars represent continuous sampling methods. The red line indicates the cutoff date for
data used in this study (May 2013). TO-15 is a summa cannister sampler; TO-17 is an active (pumped) sorbent tube sampler;
SKC 575 and Ultra III are badge-style passive sorbent samplers; Setra is a differential pressure measurement device; GC = gas
chromatograph; Electret and AlphaGUARD are radon measurement devices; Cts = continuous. See Section 3 for additional
information on measurements and methods.
1.4.4 Meteorological Effects on Vapor Intrusion
To assess the relationship between meteorological parameters and vapor intrusion, we used visual
examination of temporal trends in stacked plots of indoor air, soil gas, and subslab concentration data
1-5
-------
Section 1—Executive Summary
along with quantitative time series methods. Results from these lines of evidence are summarized in
Table 1-1. In summary, the data suggest that multiple meteorological variables likely interact in complex
ways to affect VOC vapor intrusion at this duplex.
As expected, based on stack effects, cold temperatures contributed to greater vapor intrusion. This was
expected from knowledge of the stack effect mechanism. The evidence also indicates that both snowfall
and snow/ice accumulation can increase VOC vapor intrusion, although this effect may be absent for
radon and is complex for VOCs. Snow varies in moisture content and; thus, air permeability from one
snow event to another and as a snow accumulation ages over time. There is relatively little evidence of
rain effects on VOCs, but there is evidence suggestive of a rain effect on radon. Barometric pressure
change appears to have effects on radon and probably VOCs, although the interactions are complex and
additional work on the time series data is needed to determine how best to analyze the effects of
barometric pumping on vapor intrusion in the duplex. There also is evidence of an association between
winds from westerly directions with vapor intrusion in the 422 portion of the duplex, but the evidence for
an effect of wind velocity is equivocal. Additional study is needed to assess how to best model the
complex interactions between meteorological variables and vapor intrusion at this site, as well as to see
how different meteorological and building conditions can lead to different results at other sites.
1.4.5 Preferential Pathways and Conceptual Site Model: Helium Tracer and
Geophysical Tests
Four helium tracer tests performed pairwise with common subsurface injection locations yielded similar
overall patterns of tracer distribution in soil gas outside the building with and without mitigation system
operation. The variability between paired tests (mitigation on and mitigation off) was more pronounced
beneath the building where the mitigation system would have been expected to have the most significant
influence on airflow. The similar patterns between tests performed in different subsurface areas (i.e.,
different injection points) suggest control by common features of soil stratigraphy or the building
envelope. Helium tracer concentrations suggest easy horizontal migration toward the building over
distances of up to 20 ft and rapid vertical migration from 13 ft to 6 ft bis at the injection cluster. However,
lower helium concentrations at certain ports suggest subsurface heterogeneity and preferential flow paths
that could not be fully mapped with only the four tests conducted.
Geophysical tests confirmed the location of many known features in and around the duplex, including the
shallow, moist silty clay layer overlying the deeper sand/gravel outwash layer and the shallower (7-7.5 ft
bis) silt/clay layer. Ground penetrating radar (GPR) results suggest that the concrete slab varies from 0.5
to 0.7 ft in thickness with an irregular undulating contact with the underlying fill material and resulting
gaps where soil gas may pool or move preferentially.
1-6
-------
Table 1-1. Summary of Lines of Evidence for Meteorological Factors Influencing Vapor Intrusion in This Study (Blank cells reflect types
of analysis not completed for a given parameter)
Snow or Ice
Accumulation
on Ground
Cold Exterior
Temperatures (or
Substantial Change
in temperatures)
Rain
Events/
Rainfall
Amount
Barometric
Pressure
Changes
Snowfall
West to
NW Winds
High Wind
Velocity
Apparent Temporal Association with
VOC Concentrations in Indoor Air
(Section 6, also EPA 2012a)
Yes
Yes
Yes
Possibly for
chloroform
Apparent Temporal Association with
VOC Concentrations in Wall Ports
or Subslab Ports (Section 6)
Yes
Yes
Weak
Some
Apparent Temporal Association with
Large Subslab to Indoor Differential
Pressure Events (Section 9.1)
Yes in
some
cases
Yes in some cases
Yes in some
cases
Yes in a few
cases
Yes in a few
cases
Apparent Trend in XY Graph of.
Meteorological Parameter vs.
Subslab/lndoor Differential Pressure
(Section 9.1 and EPA 2012a)
Yes
No
Yes
No
Apparent Trend in XY Graph of
Meteorological Parameter vs. VOC
Concentration (Section 9.2)
Yes for PCE, not definitive
for chloroform
Yes
No clear
relationship
Not definitive
Yes for
PCE, No for
chloroform
No for PCE,
Yes for
chloroform
Correlation with Radon in
Quantitative Time Series Analysis
(sections 10.1 to 10.4); 422
Basement and Office
NO
No
Yes in most analyses
Yes in some
analyses
Yes in most
analyses
Yes in some
analyses
Correlation with Chloroform in
Quantitative Time Series Analysis
(Sections 10.5 and 10.7); 422
Basement
Yes in one of two cases with
opposite signs for the
coefficients ofthe current
and past weeks.
Yes
No
Yes in some
analyses
Correlation with PCE in Quantitative
Time Series Analysis (Sections 10.6
and10.8), 422 Basement
Yes in one of two cases but
with an unexpectedly
negative coefficient for the
current week.
Yes, although
coefficients are both
positive and negative
No
Yes
No
No
O"
3
&
si
£
-------
Section 1—Executive Summary
1.4.6 Temporal Variability and Trends
PCE levels in indoor air follow the general trend of starting higher at the beginning of the project (January
2011), dropping to a low in early summer, and rising slightly and leveling out through the end of the
intensive premitigation study period (February 2012). This general trend was attributed primarily to
temperature, because the winter of 2010-2011 was much more severe than the winter of 2011-2012.
During mitigation testing, which began in October 2012, radon dropped quickly with mitigation and
remained low, while indoor air PCE concentrations first dropped then rose to levels above those observed
in the March 2011 to September 2012 time period (see February and March period in Figure 1-1). Given
that soil gas levels also tended to rise at times during mitigation, we postulate that VOCs can be moved
close to the structure either by a cumulative stack effect during a severe winter or by operation of an SSD
mitigation system. It is unknown whether this VOC migration effect toward the slab will be common at
sites with other geological formations or contaminant distributions or whether it would continue to occur
at this site with longer operation of the mitigation system. Because our mitigation testing included several
on-off cycles over one winter, we also do not know whether more substantial reductions in indoor air
VOC concentrations would be achieved with continuous operation of the SSD mitigation system.
However, spatial patterns changed dramatically when mitigation was operating, indicating that at least
during initial operation the mitigation system was influencing both soil gas and indoor air concentrations.
1.4.7 Summary
These results suggest that current chemical vapor intrusion mitigation system designs, based on radon
systems experience, may produce designs that are highly effective for radon but not as effective for
VOCs, at least during the initial months of system operation. This finding suggests a need for longer term
confirmation of post-mitigation VOC concentrations and replication of this study's findings in other
environments. Specifically, buildings of other ages/designs in conditions similar to this (15 to 20 ft to
groundwater, moderate strength source, and coarse deep geology) should be tested. This finding should
also prompt more intensive studies of long-term mitigation system performance in commercial buildings
and in other geographies. The current trend of TCE being managed based on short-term exposure
thresholds provides additional impetus for such studies, because radon and other VOCs are not usually
managed based on short-term health effects.
The results reported here provide little support for the common guidance that vapor intrusion sampling
must be timed around rain events greater than one half inch. While there may have been some effects on
vapor intrusion of major seasonal flooding events that changed the local water table by approximately 5
ft, there was not any apparent effect on indoor air concentrations from more moderate rain events. The
results reported here do suggest that snow events, snow cover, and/or frozen soils may temporarily
increase vapor intrusion.
1-8
-------
Section 1—Executive Summary
I ndoor Air Radon a nd Indoor Air GC PC£
•
«
*
•
• imrnm o •
1
f *
o • 1
>>
•
• •
« ¦
• »» • m
mm mm -
TT
9 DownsUiriHn
• ypMarsFhi
10
9 c O AlHrtoMltPCE
Btl
7 £ • 420firitPCE
6 5J'
S* O 422fir$tK£
* 3
3-a O 420 Ba»S PCE
2 n
1> • 422BaJ*SPCE
0—
¦1 Mlt^itionOr
— Jfaurrt Mitigation
Blizjird and minor
SIMWJ
—Ground Ff oien
* BriefSmw
Ov\0%^
Date
Figure 1-2. Indoor air concentrations of PCE (top panel) and radon (middle and bottom panels)
with mitigation on (black bars), passive mitigation (no fan, gray bars), and mitigation
off (no bar). Snow and frozen ground events are indicated by blue bars and red dots.
A large number of variables have been shown here to most likely have an interactive effect on VOC vapor
intrusion, including cold temperatures, snow/ice, barometric pressure, and wind direction. Practitioners
should thus expect to not be able to explain in detail temporal patterns drawn from small data sets (for
example, three or four rounds of VOC sampling). However, after results from this study are confirmed in
studies of other buildings, it may be possible to develop recommendations to guide selection of "near
worst case" indoor air sampling conditions for specific sites based on each site's known characteristics
such as climate, stratigraphy, and source characteristics.
1-9
-------
Section 1—Executive Summary
1-10
-------
Section 2—Introduction
Table of Contents
2.0 Introduction 2-1
2.1 Background 2-2
2.1.1 Variability in Vapor Intrusion Studies 2-4
2.1.1.1 Spatial Variability 2-5
2.1.1.2 Temporal Variability 2-7
2.1.1.3 Measurement Variability 2-8
2.1.2 Vapor Attenuation Factors 2-9
2.1.3 Potential for Use of Radon as a Surrogate for VOC Vapor Intrusion 2-9
2.1.4 Passive VOC Sampling 2-12
2.2 Objectives 2-16
2.2.1 Time Scale and Measurement of Independent and Dependent Variables 2-18
2.2.2 Data Quality Objectives and Criteria 2-18
List of Figures
2-1. An overview of important vapor intrusion pathways (U.S. EPA graphic) 2-3
2-2. Soil gas and groundwater concentrations below a slab (Schumacher et al., 2010) 2-6
List of Tables
2-1. VOC Indoor Air Sampling Method Options 2-15
2-2. Continuing Proj ect Obj ectives Addressed in this Document 2-17
2-2. Continuing Project Objectives Addressed in this Document (continued) 2-18
2-3. Factors Causing Temporal Change in Vapor Intrusion and How They are Observed and
Measured 2-19
2-4. Data Quality Objectives and Criteria 2-20
2-i
-------
Section 2—Introduction
-------
Section 2—Introduction
2.0 Introduction
Vapor intrusion is the migration of subsurface vapors, including radon and volatile organic compounds
(VOCs), in soil gas from the subsurface to indoor air. Vapor intrusion happens because there are pressure
and concentration differentials between indoor air and soil gas. Indoor environments are often negatively
pressurized with respect to outdoor air and soil gas, for example, from exhaust fans or the stack effect,5
and this pressure difference allows soil gas containing subsurface contaminant vapors to flow into indoor
air through advection. In addition, concentration differentials cause VOCs and radon to migrate from
areas of higher to lower concentrations through diffusion, which is another cause of vapor intrusion.
For VOCs, the vapor intrusion exposure pathway extends from the contaminant source, which can be free
product (nonaqueous phase liquids or NAPLs), VOCs sorbed to the geologic matrix, or contaminated
groundwater, to indoor air exposure points. Contaminated matrices may include groundwater, soil, soil
gas, and indoor air. VOC contaminants of concern typically include halogenated solvents such as
trichloroethene (TCE), tetrachloroethene (PCE), chloroform, and the degradation products of TCE and
PCE, including dichloroethenes and vinyl chloride. These halogenated VOCs were widely used and are
toxic and degrade very slowly in the subsurface, making them priority contaminants of concern through
the vapor intrusion exposure pathway at many hazardous waste sites nationwide. Petroleum
hydrocarbons, such as the aromatic VOCs of benzene, toluene, ethylbenzene, and xylenes (BTEX), are
also contaminants of concern for vapor intrusion, but because they degrade much more readily in the
subsurface, they are much less likely to lead to a vapor intrusion problem (U.S. Environmental Protection
Agency [U.S. EPA], 2012p).
Radon is a colorless radioactive gas that is released by radioactive decay of radionuclides in soil, where it
migrates into homes through vapor intrusion in a similar fashion to VOCs. Radon is high in areas where
the radioactive precursors to radon occur at relatively high concentrations in soil (as with the subject
house of this investigation) and affects many more homes across the United States than halogenated
VOCs. Low-cost testing and effective mitigation methods are available for radon, and the radon exposure
pathway has been studied extensively by EPA and other organizations and thus contributes to a
conceptual understanding of the vapor intrusion process.
VOC vapor intrusion is less well studied than radon and is the primary focus of this research project. In
particular, the study focuses on halogenated VOCs, which are relatively recalcitrant (resistant) to
biodegradation in aerobic soils and groundwater (with typical half-lives of a year or more; Howard et al.,
1991); in contrast, radon has a radioactive half-life of about 3.8 days (Cohen, 1971). Of the two primary
VOCs subject to investigation under this project, PCE is generally considered quite recalcitrant, with an
aerobic half-life in groundwater of 1 to 2 years (Howard et al., 1991). Studies of chloroform
biodegradation under aerobic conditions are mixed, with some showing recalcitrance (e.g., a 0.2- to 5-
year half-life in Howard et al., 1991) and others showing moderate cometabolic biodegradation with
methylene chloride and chloromethane as sequential degradation products (Air Force Center for
Environmental Excellence [AFCEE], 2004; Agency for Toxic Substances and Disease Registry
[ATSDR], 1997).
Current practice for evaluating the vapor intrusion pathway involves a multiple line of evidence approach
based on direct measurements in groundwater, external soil gas, subslab soil gas, and/or indoor air.
Modeling approaches ranging from simple constructs, such as attenuation factors to one-dimensional
models to three-dimensional models, are frequently used as an aid to data interpretation and predictive
5The stack effect is the overall upward movement of air inside a building that results from heated air rising and escaping through openings in the
building super structure, thus causing an indoor pressure level lower than that in the soil gas beneath or surrounding the building foundation
(http://www.epa. gov/iaq/glossarv.htmlV
2-1
-------
Section 2—Introduction
tool. No single line of evidence is considered definitive, and direct measurements can be costly, especially
where significant spatial and temporal variability require repeated measurements at multiple locations to
assess the chronic risks of long-term VOC exposure accurately.
The main focus of this report is to better characterize this variability by collecting a detailed long-term
data set of week-long measurements of subslab soil gas, external soil gas, and indoor air, on a single
building that is affected by vapor intrusion of radon and VOCs. By examining both short-term and long-
term (average annual) concentrations, the project provides valuable information on how to best take and
evaluate measurements to estimate long-term, chronic risk for VOCs. Special attention was paid to
snow/ice events and flooding events as potential causes of dramatic temporal variability. We then
implemented a common mitigation technology—subslab depressurization (SSD)—to evaluate the
effectiveness of this approach as a tool for reducing indoor concentrations and the temporal variability.
Radon concentration fluctuations were also measured because if radon can be shown to indicate when
there is a potential for chemical (i.e., VOC) vapor intrusion, radon, which is much cheaper to measure
than VOCs, could be an important tool in improving the investigation and mitigation of chemical vapor
intrusion using SSD. In addition, there is much research on radon intrusion into indoor air that could
provide valuable lessons for chemical vapor intrusion.
The study reported here is an extension of work conducted and published in a previous report (U.S. EPA,
2012a). The earlier study examined the:
¦ passive sorbent performance over various timescales,
¦ an evaluation of the usefulness of soil gas samples taken externally to the building,
¦ heating, ventilation, and air conditioning (HVAC) system cycles,
¦ the use of temporary vs. permanent subslab ports, and
¦ induced depressurization within a building as a vapor intrusion evaluation strategy (fan testing).
The previous study helped define additional research questions addressed in this study such as the:
¦ specific geologic and anthropogenic features that influence contaminant transport in this specific
case (and by implication may be important in other similar urban neighborhoods),
¦ relative role of groundwater and vadose zone sources, and
¦ control of radon and VOC vapor intrusion variability through SSD mitigation.
In addition to the effects of installing and operating an SSD mitigation system on VOC and radon levels
in soil gas and indoor air, the new research report provides additional research on the effect of snow and
frozen ground cover on vapor intrusion, as well as a longer term data set. For topics where significant new
data were obtained after September 2012, we interpret the new evidence in conjunction with that reported
previously.
2.1 Background
An overview of the VOC vapor intrusion pathway is shown in Figure 2-1; the building in which exposure
occurs is shown in the center. Three main routes of VOC migration have been defined:
¦ Movement of VOC vapors from shallow soil sources through the unsaturated (vadose) zone
¦ Transport of VOCs through groundwater, followed by partitioning of VOCs from the most
shallow layer of groundwater into vadose zone soil gas
¦ Vapor movement through preferential pathways such as utility corridors
2-2
-------
Section 2—Introduction
In portions of these three routes, advective forces predominate, and in others diffusive forces dominate
transport. The final step of vapor intrusion typically involves soil gas moving from immediately below
the building slab into the indoor air, which is normally envisioned as an advective process for most slabs,
although it may be diffusive with very well-sealed slabs. This subslab space is often significantly more
permeable than the bulk vadose zone soil, either because a gravel drainage layer was intentionally used or
the soils have shrunk back from the slab in places. In those cases, the subslab space is expected to serve as
a common plenum allowing the lateral mixing of VOCs that reach the building through multiple
pathways. In other cases, the subslab space may not be so interconnected, resulting in differing subslab
VOC concentrations at different locations across the slab.
S5r«n*mliri<
Advatifcjrt
Top Of
r rincjif
DltsoCveti Cont*nHr»fton
Figure 2-1. An overview of important vapor intrusion pathways (U.S. EPA graphic).
It has been argued that in addition to the average advective force, there is an important and even dominant
role in transport under some conditions (such as high permeability) for the fluctuating element of the
pressure field, which, like diffusion, contributes to the movement of mass from high to low concentration
zones (Robinson and Sextro, 1997; Robinson, Sextro and Fisk, 2007; DeVaull 2012, 2013).
Vapor and liquid transport processes and their interactions with various geologic and physical site settings
(including building construction and design), under given meteorological conditions, control migration
2-3
-------
Section 2—Introduction
through the vapor intrusion pathway. Variations in building design, construction, use, maintenance, and
subslab composition and temporal variation in meteorological factors (e.g., atmospheric pressure,
temperature, and precipitation and its infiltration) all influence vapor intrusion. Utility corridors, such as
the backfill around water lines or partially full sanitary or combined sewers, can provide routes of
preferential migration through the vadose zone. Advective flow into a building can occur through cracks
in the floor, below grade walls, or at incompletely sealed utility penetrations in the building envelope. NJ
DEP (2013) summarizes other important factors that can affect vapor intrusion at many sites:
¦ biodegradation of VOCs as they migrate in the vadose zone,
¦ site stratigraphy,
¦ soil moisture and groundwater recharge,
¦ fluctuations in water table elevation, and
¦ temporal and inter-building variations in the operation of ventilation systems in
commercial/industrial buildings.
These and other factors combine to create a complex and dynamic system controlling vapor intrusion at a
particular site.
This project explored and further developed several promising cost-effective techniques to evaluate the
vapor intrusion pathway and improve data quality. Two primary tools were investigated: (1) using
modified sorbent-based measurement techniques for time-integrated measurements of indoor air VOCs
and (2) using radon measurements for assessing VOC vapor intrusion. The project also investigated
measurements of pressure differentials (subslab vs. indoor), meteorological conditions, crack size, and air
exchange rates in the context of the chemical-specific measurements described above. These physical
measurements are not stand-alone tools nor are they the emphases of the current research program, but are
necessary supporting tools for developing a conceptual understanding of spatial variability, temporal
seasonal effects, and a mass balance around a building subject to vapor intrusion.
2.1.1 Variability in Vapor Intrusion Studies
This project focused on observing changes in vapor intrusion over a >2-year period both with and without
SSD mitigation. In order to express quantitatively our goals for this project, it is necessary to understand
the causes and typical ranges of spatial and temporal variation in various matrices studied for vapor
intrusion assessment.
Through measurements of radon and VOC vapor intrusion under various conditions, several studies have
provided insight into the complexity of temporal variability in indoor air concentrations attributable to
vapor intrusion—the primary focus of this work. Nazaroff et al. (1987) studied how induced-pressure
variations can influence radon transport from soil into buildings with roughly hourly resolution. In a more
recent study, Mosley (2007) presented the results of experiments, showing that induced building-pressure
variations influence both the temporal and spatial variability of both radon and chlorinated VOCs
(CVOCs) in subslab samples and in indoor air (hourly sampling for radon). Schuver and Mosley (2009)
have also reviewed numerous studies of radon indoor concentrations, in which multiple repeated indoor
air samples were collected with hourly, daily, weekly, monthly, 3-month, and annual sample durations for
study periods of up to 3 years; however, detailed soil gas radon data sets are much rarer.
Several radon studies have demonstrated that barometric pressure fluctuations can affect the transport of
soil gas into buildings (Robinson and Sextro, 1997; Robinson et al., 1997). The impact of barometric
pressure fluctuations on indoor air is influenced by the interaction of the building structures and
conditions, as well as other concurrent factors, such as wind (Luo et al., 2006, 2009). Changes in
2-4
-------
Section 2—Introduction
atmospheric conditions (e.g., pressure, wind) and building conditions (e.g., open doors and windows) may
temporarily over- or under-pressurize a building. Based on long-term pressure differential data sets
acquired by ARCADIS and EPA's National Risk Management Research Laboratory (NRMRL) at a
different Indianapolis study site (the Wheeler building) at which both radon and VOCs are being
measured in both subslab and indoor air, other factors that may cause temporal and spatial variability in
soil vapor and indoor air concentrations include:
¦ fluctuation in building air exchange rates due to resident behavior/HVAC operations,
¦ fluctuations in outdoor/indoor temperature difference, and
¦ rainfall events and resultant infiltration and fluctuations in the water table elevation.
The pressure difference between a house-sized building and the surrounding soil is usually most
significant within 1 to 2 m of the structure, but measurable effects have been reported up to 5 m from the
structure (Nazaroff et al., 1987). Temperature differences or unbalanced mechanical ventilation are likely
to induce a symmetrical pressure distribution in the subsurface, but the wind load on a building adds an
asymmetrical component to the pressure and distribution of contaminants in soil gas.
Folkes et al. (2009) summarized several large groundwater, subslab, and indoor air data sets collected
with sampling frequencies ranging from quarterly to annually during investigations of vapor intrusion
from chlorinated VOC plumes beneath hundreds of homes in Colorado and New York. They analyzed
these data sets to illustrate the temporal and spatial distributions in the concentration of VOCs. Their
analysis demonstrated that although the areal extent of structures affected by vapor intrusion mirrored the
plume of chlorinated VOCs in groundwater, not all structures above the plume were affected. It addition,
they found that measured concentrations of VOCs in indoor air and subslab soil gas can vary considerably
from month to month and season to season, and that sampling results from a single location or point in
time cannot be expected to represent the range of conditions that may exist spatially or at other times.
In a study of the vapor intrusion pathway at the Raymark Superfund site, DiGiulio et al. (2006) showed
that measured concentrations of CVOCs in subslab exhibited spatial and temporal variability between
neighboring houses and within individual houses. Similar variability in subslab CVOC concentrations
within and between houses has been observed during vapor intrusion evaluations of several sites in New
York State (Wertz and Festa, 2007).
In scenarios with coarser soils (e.g., sands, gravels), the soil gas permeability is high, and changes in
building pressurization may affect the airflow field and the resultant soil vapor concentration profiles near
buildings. In scenarios with fine-grained soils (e.g., silts, clays), the soil gas permeability is low and soil
gas flow rates (Qs) may be negligible and not affect the subsurface concentration. Nevertheless, in both
soil-type scenarios, over-pressurization of the building may still significantly reduce the indoor air
concentration because of the reversal of soil gas flow direction from the building into the soil (Abreu and
Johnson, 2005, 2006).
A wind-induced, non-uniform pressure distribution on the ground surface on either side of a house may
cause spatial and temporal variability in the subslab soil vapor concentration distribution if the wind is
strong and the soil gas permeability is high (Luo et al., 2006, 2009). In addition, during or after a rainfall
event, the subsurface beneath the building may have a lower moisture content than the adjacent areas
because of water infiltration.
2.1.1.1 Spatial Variability
Spatially, reports of several orders of magnitude variability without apparent patterns between indoor air
and subslab concentrations for adjacent structures in a neighborhood are very common (see, for example,
2-5
-------
Section 2—Introduction
U.S. EPA, 2012c). Six orders of magnitude in subslab concentration variability were reported by Eklund
and Burrows (2009) for a commercial building of 8,290 sq ft. As shown in Figure 2-2, Schumacher and
coworkers (2010) observed more than three orders of magnitude concentration variability in shallow soil
gas below a slab over a span of 50 lateral feet, suggesting a strong effect of impervious surfaces both in
limiting soil gas exchange with the atmosphere and in maintaining relatively high concentrations of
VOCs in shallow groundwater. They also observed two orders of magnitude concentration variability
with a depth change of 10 ft in the unsaturated zone within one borehole. Although these publications,
unlike this study, are for larger nonresidential properties, they do support the general conclusion that
spatial variability can be over several orders of magnitude at vapor intrusion sites.
Lee and coworkers (2010) observed two orders of magnitude variability in subslab concentration beneath
a small townhouse. Studies by McHugh and others (2007) have generally found markedly less variability
in indoor air concentrations than in subslab concentrations, probably due to the greater degree of mixing
in the indoor environment.
»EPA
United Slates
Environments* Protection
Agency
Measured Soil Gas
Profile for TCE - Phase 2
ST—2 ST-3
17777777777777777777777%
yy/yyy, U MP AVJED AREA#:
10
1 50,000
39*0Q0
2901 r
8^,000 | | 52,0001 | 2,70Q|
EE DEI
10 20 30 40 50 60 70 80 90
FEET
Figure 2-2. Soil gas and groundwater concentrations below a slab (Schumacher et al., 2010).
2-6
-------
Section 2—Introduction
2.1.1.2 Temporal Variability
Temporal variability has been summarized by ITRC (2007),
which states in Section D.4.10:
Variations in soil gas concentrations due to temporal effects are
principally due to temperature changes, precipitation, and activities
within any overlying structure. Variations are greater in samples taken
close to the surface and dampen with increasing depth. In 2006 there
were a number of studies on temporal variation in soil gas
concentrations, and more are under way or planned in 2007 by USEPA
and independent groups. To date these studies have shown that short-term variations in soil gas concentrations at
depths 4 feet or deeper are less than a factor of 2 and that seasonal variations in colder climates are less than a
factor of 5 (Hartman 2006). larger variations may be expected in areas of greater temperature variation and
during hecny periods of precipitation, as described below.
• Temperature. Effects on soil gas concentrations due to actual changes in the vadose zone
temperature are minimal. The bigger effect is due to changes in an overlying heating or
HVAC system and the ventilation of the structure due to open doors and windows. In colder
climates, worse-case scenarios are most likely in the winter season. The radon literature
suggests that temporal variations in soil gas are typically less than a factor of 2 and that
seasonal effects are less than a factor of 5. If soil gas values are more than a factor of 5
below acceptable levels, repeated sampling is likely not necessary regardless of the season.
If the measured values are within a factor of 5 of allowable risk levels, then repeated
sampling may be appropriate.
• Precipitation. Infiltration from rainfall can potentially impact soil gas concentrations by
displacing the soil gas, dissolving VOCs, and by creating a "cap " above the soil gas. In
many settings, infiltration from large storms penetrates into only the uppermost vadose zone.
In general, soil gas samples collected at depths greater than about 3-5 feet bgs or under
foundations or areas with surface cover are unlikely to be significantly affected. Soil gas
samples collected closer to the surface (<3 feet) with no surface cover may be affected. If the
moisture has penetrated to the sampling zone, it typically can be recognized by difficulty in
collecting soil gas samples. If high vacuum readings are encountered when collecting a
sample or drops of moisture are evident in the sampling system or sample, measured values
should be considered as minimum values.
• Barometric Pressure. Barometric pressure variations are unlikely to have a significant e ffect
on soil gas concentrations at depths exceeding 3-5 feet bgs unless a major storm front is
passing by. A recent study in Wyoming (Luo et cd. 2006) has shown little to no relationship
between barometric pressure and soil gas oxygen concentrations for a site with a water
table at -15 feet bgs.
In summary, temporal variations in soil gas concentrations, even for northern climates, are minor compared with
the consen'ative nature of the risk-based screening levels. If soil gas values are a factor of 5-10 times below the
risk-based screening levels, there likely is no need to do repeated sampling unless a major change in conditions
occurs at the site (e.g., elevated water table, significant seasonal change in rainfall)...
Temporal Variability Example:
IBM site, Endicott, New York
Recent data from a large site in
Endicott, New York, collected over a
15-month period showed soil gas
concentration variations of less than
a factor of 2 at depths greaterthan
5 feet bis.
2-7
-------
Section 2—Introduction
And in Section D.8 of the same document, ITRC notes:
Short-term temporal variability in subsurface vapor intrusion occurs in response to changes in weather conditions
(temperature, wind, barometric pressure, etc.), and the variability in indoor air samples generally decreases as the
duration of the sample increases because the influences tend to average out over longer intervals. Published
information on temporal variability in indoor air quality shows concentrations with a range of a factor of 2-5 for
24-hour samples (Kuehster, Folkes, and Wannamaker 2004; McAlary et al. 2002). If grab samples are used to
assess indoor air quality, a factor of safety (at least a factor of 5) should be used to adjust for short-term
fluctuations before comparing the results to risk-based target concentrations. long-term integrated average samples
(up to several days) are technically feasible, using a slower flow rate this is the USEPA recommended approach for
radon monitoring). Indoor air sampling during unusual weather conditions should generally be avoided.
In Section D.l 1.8, ITRC goes on to discuss the effect of meteorological changes on vapor intrusion:
A variety of weather conditions can influence soil gas or indoor air concentrations. The radon literature suggests
that temporal variations in the soil gas are typically less than a factor of 2 during a season and less than a factor of
5 from season to season). Recent soil gas data from Endicott, New York and Casper, Wyoming are in agreement
with the radon results. For soil gas, the importance of these variables will be greater the closer the samples are to
the surface and are unlikely to be important at depths greater than 3-5 feet below the surface or structure
foundation.
The most frequent time interval of observation in routine vapor intrusion practice has been 8- to 24-hour
integrated samples. In this project, multiple durations of observation of indoor air concentrations were
compared, including automated discrete samples collected on 3-hour intervals and passive samples with
varying integration times: 24-48 hours, 7 days, 14 days, 28 days, 91 days, 182 days, and 364 days.
A team led by Paul Johnson (Johnson et al., 2012; Holton et al., 2013) reported more than 2 years of high
frequency observation of a home overlying a chlorinated solvent groundwater plume in Layton, Utah. At
the time of this report, key preliminary observations at that site with regard to temporal variability
included the following:
¦ Indoor air variability in TCE of about three orders of magnitude was observed.
¦ The near-source data, such as deep soil gas, were more consistent in time than the near-surface
data sets such as subslab air or indoor air.
¦ The temporal trend was characterized by "long periods of relative VI inactivity with sporadic VI
activity" and "long periods of relative VI activity with sporadic VI inactivity" (Johnson, 2012).
¦ "24-h samples are not a very practicable option at the resolution required for robust VI pathway
assessment" (Holton, 2013).
2.1.1.3 Measurement Variability
Beyond spatial and temporal variability, the underlying uncertainty of the measurements used to assess
vapor intrusion must also be considered. Many measurements of vapor intrusion, both in indoor air and
subslab soil gas, have traditionally relied on Summa canister samples analyzed by methods TO-14/TO-15.
(U.S. EPA, 1999a, 1999b). Method TO-15 specifies an audit accuracy of 30% and a replicate precision of
25% as performance criteria. But even those figures do not fully convey the interlaboratory variability
observed for these methods when applied to the low concentrations typical of indoor air studies. As Lutes
and coworkers (2010) reported:
¦ "In two recent TO-15 or 8260 interlaboratory comparisons administered by the company ERA
for gas phase samples the acceptance range for tetrachloroethylene results were: 4.31-22.3 ppbv
(July-Sept 2009 study) and 31.6-74.1 (ig/L (October-November 2007 study)."
2-8
-------
Section 2—Introduction
¦ "For comparison in a 2007 TO-14/TO-15 study conducted by Scott Specialty Gasses the reported
values for toluene reported by 12 labs varied from 3.1 to 18.6 ppbv."
2.1.2 Vapor Attenuation Factors
One common way of evaluating the impact of subsurface vapors on indoor air quality is to compute the
ratio of indoor air concentration to subslab soil vapor concentration. EPA has defined the resulting
"attenuation factor" as follows: "The attenuation factor, a, is a proportionality constant relating indoor air
concentrations (Cmdoor air) to the concentrations of vapors in soil gas (Csoii gas) or groundwater (Cgr0undwater)
concentrations." For soil gas to indoor air, the equation is as follows:
Cindoor air OlSG x Csoil gas-
For groundwater, a similar equation is used except that the dimensionless Henry's Law Constant (H) is
used to convert the dissolved VOC concentration in groundwater to the corresponding equilibrium vapor
concentration:
Cindoor air = (XGW x Cgroundwater x H.
A larger a indicates less attenuation, and a smaller value indicates more attenuation. The greater the
attenuation factor, the greater the indoor air concentration.
Note that both of these equations assume that all of the indoor air VOC concentration (Cindoor air) is from
vapor intrusion. In many cases, this is not the case because of VOC-containing products in the indoor
environment. At this site, VOCs are not in use because the house is not occupied, so all VOCs over the
outdoor ambient air concentration can be attributed to vapor intrusion.
Within any one given site, the attenuation factors
¦ between groundwater and indoor air typically vary 2 to 3 orders of magnitude and
¦ between external soil gas and indoor air typically vary 2 to 4 orders of magnitude.
Subslab soil gas and indoor air typically vary 2 to 4 orders of magnitude (Dawson and Schuver, 2010).
EPA recently published a compilation of attenuation factor data (U.S. EPA, 2012c) that analyzes spatial
and temporal variability. Because the case with the most rounds of data discussed in the compilation as an
example of temporal variability has only six rounds, this report is expected to provide a valuable addition
to the literature regarding attenuation factors in a residential structure.
2.1.3 Potential for Use of Radon as a Surrogate for VOC Vapor Intrusion
Radon, a naturally occurring radioactive gas, is a potentially useful surrogate for assessing VOC vapor
intrusion because the physics of radon intrusion into indoor air is similar to VOC vapor intrusion. Radon
is ubiquitous in the soil and present at measurable quantities in soil gas throughout the United States.
Indeed, much of the research in VOC vapor intrusion is an expansion of earlier work on radon intrusion.
Applications of radon as a VOC surrogate have been proposed for the following reasons (Lutes, 2009;
Mosley, 2007; Mosley, 2008; Schuver, 2009):
¦ Estimating attenuation factors, with the measured radon attenuation factor serving as a surrogate
for the attenuation that may be occurring for VOCs
¦ Screening of large populations of housing units/buildings, with the presence of radon above
ambient levels in indoor air serving as evidence of soil gas influence
2-9
-------
Section 2—Introduction
¦ Use as a line of evidence to help distinguish indoor sources of VOCs where VOC indoor
concentrations are higher than would be inferred based on the radon attenuation factor. Also,
differing responses of radon and VOCs to building pressurization/depressurization tests could be
used to assess the potential for indoor sources
¦ Locating soil gas entry points, when higher radon readings are observed near entry points
¦ Verifying SSD mitigation system performance based on the reduction of indoor air radon
concentrations during SSD system operation.
Radon provides a nearly unique surrogate for VOC vapor intrusion because its presence in the indoor
environment is usually a result of radon in the soil gas immediately surrounding a building. In theory, the
entry mechanisms are believed to be the same for VOCs and radon in soil gas. Thus, measured radon
entry rates should be a good predictor of relative entry rates for VOCs. The advantages of using radon as
a surrogate measure for VOC vapor intrusion characterization include:
¦ Measurements of radon are easier, more accurate and precise, and much less expensive than
canister measurements of VOCs (typically less than 10% of the VOC analysis cost). Passive
indoor sampling for radon costs approximately $5 to $20 per sample. Active radon sampling
(indoor air and subslab) uses some of the same equipment and setup as for VOCs. This
minimizes sampling times and cost. Continuous measurement devices for radon are also
available ranging from consumer grade devices under $150 to professional grade instruments
under $10,000.
¦ High levels of indoor radon identify buildings that are vulnerable to soil gas entry.
¦ Because of the low sampling/analytical costs for radon, it is possible to conduct more field
measurements than with VOCs. This, in turn, can increase confidence in the field evaluation.
¦ Because the SSD mitigation systems can be expected to behave similarly for radon and VOCs in
the vicinity of the building, radon measurements before and after installation of vapor intrusion
mitigation systems may be useful for assessing SSD mitigation system performance for VOCs as
well.
In summary, the limited data gathered to date suggest that radon measurement may be an inexpensive,
semi-quantitative surrogate for VOC measurement when characterizing vapor intrusion and may
significantly enhance vapor intrusion characterization and decision making, particularly when used in
conjunction with subslab sampling. However, several key aspects and assumptions of this approach need
to be verified before it can be put into widespread use. For radon to be a valuable surrogate for VOCs:
¦ Radon detection in building interiors should be quantitatively possible across the wide range of
subslab concentrations encountered in the United States. Ideally these measurements can be
made with inexpensive passive methods (i.e., charcoal or electrets).
¦ The radon route and mechanism of entry should be similar to that of VOCs of interest, once both
species are present in the subslab soil gas. This would imply that the subslab attenuation factors
for radon and VOCs are similar.
¦ Variance in the natural vadose zone (unsaturated soil) radon concentration across a given
building footprint should be low enough to allow radon to be a useful indicator.
¦ Concentrations of radon and the VOCs of concern should be well correlated in subslab soil gas.
This would not necessarily be expected based on the fact that radon and VOCs have different
sources. However, they may indeed be approximately correlated if the VOC(s) of interest and
radon are both widely dispersed in deep soil gas. In this case, the concentrations of both radon
and VOCs at various locations in the subslab may be controlled primarily by the ratio of flow
from the deep soil gas to the flow from ambient air (in which both VOC and radon
concentrations would be expected to be low).
2-10
-------
Section 2—Introduction
¦ Interior sources of radon should be negligible.
The loss rates to sink effects in the indoor environment should be similar or negligible for radon and
VOCs so that the air exchange rate forms the primary control on indoor air concentration once vapor
intrusion has occurred.
To our knowledge, this concept was first applied in a relatively small study (Cody et al., 2003) at the
Raymark Superfund Site in Connecticut. The study compared the intrusion behavior of radon and
individual VOCs by determining attenuation factors between the subslab and indoor (basement) air in 11
houses. The results indicated that the use of radon measurements in the subslab and basement areas was
promising as a conservative predictor of indoor VOC concentrations when the subslab VOC
concentrations were known. Further work at the Raymark site (U.S. EPA, 2005b) statistically compared
basement and subslab concentration ratios for radon and VOCs associated with vapor intrusion. Of six
test locations, three showed that basement/subslab concentration ratios for radon and VOCs associated
with subsurface contamination were similar. Three had statistically different ratios, suggesting that further
research was needed to evaluate the usefulness of radon in evaluating vapor intrusion. Conservative
VOCs (those believed to be associated only with subsurface contamination) were a better predictor of
other individual volatile compounds associated with vapor intrusion than was radon.
A three-building complex, commercial case study of the radon tracer approach was published by Wisbeck
et al. (2006). Radon and indoor air attenuation factors were calculated for five sampling points and were
generally well correlated. Subslab radon concentrations varied by approximately a factor of 10 across the
five sampling points.
Results of an earlier test program at Orion Park Housing units at Moffett Field have been preliminarily
reported (Mosley, 2007). Results showed:
¦ Low levels of radon can be measured with sufficient accuracy to be used in analysis of vapor
intrusion problems.
¦ Radon is a promising, low-cost surrogate for soil gas contaminants; however, as with VOCs
themselves, the complete distribution under the slab must be known in order to properly interpret
its impact on indoor measurements.
¦ Unexpectedly, the subslab areas under each unit were segmented. The four subslab sampling
points installed in one unit were not in good communication with one another. An introduced
tracer, SF6, moved very slowly and not very uniformly under the slab.
¦ Results showed that for soils like these with poor communication, a subslab measurement at a
single point is not very reliable for estimating potential vapor intrusion problems. The average
value of subslab measurements at several locations also may not yield a reliable estimate of
indoor concentrations. When subslab communication is poor, one must identify a connection
between subslab contaminants and a viable entry path.
The potential usefulness of the radon tracer was studied in 2007 to 2010 by EPA NRMRL at Moffett
Field in California and in the Wheeler building in Indianapolis. These studies are summarized in three
draft peer-reviewed papers that have been submitted for EPA internal review:
¦ Vapor Intrusion Evaluation Using Radon as a Naturally Occurring Tracer. In this paper, we
compile data from five study sites where radon has been used in VOC vapor intrusion
investigations and attenuation factors were calculated. A total of 17 buildings are included in the
data set, a mix of commercial and residential, in a wide variety of geographical areas within the
United States. Attenuation factors were roughly correlated between radon and VOCs.
2-11
-------
Section 2—Introduction
¦ Randomized Experiment on Radon Tracer Screening for Vapor Intrusion in a Renovated
Historical Building Complex: This study focused on a renovated former industrial facility now
being reused as residential, public, and office space. Fifty locations within the complex were
originally screened for radon using passive sampling techniques. Then two subsets of these
sample locations were selected for passive VOC sampling, one randomly and the other based on
the radon information. The upstairs radon-guided samples were significantly higher in TCE than
the randomly selected locations. The portions of the building complex where the radon guidance
appeared to provide predictive power were understandable in terms of the building design and the
concept of the open basement serving as a common plenum.
¦ Case Study: Using Multiple Lines of Evidence to Distinguish Indoor and Vapor Intrusion Sources
in a Historic Building: This paper uses data sets developed at the Southeast Neighborhood
Development Corporation (SEND) Wheeler Arts Building site in Indianapolis, Indiana, to
demonstrate the use of multiple lines of evidence in distinguishing indoor from subsurface
sources in a complex multiuse, multiunit building. The use of radon as a quantitative tracer for
vapor intrusion source discrimination is shown as well as the use of differential pressure data as
an additional line of evidence. Box and whisker plots of the distribution of indoor air pollutants
on multiple floors are used to distinguish pollutants with predominant subslab sources from those
with predominant indoor sources. Those pollutants that the box and whisker analysis suggest have
indoor sources are also corroborated from the literature as having very common indoor sources
expected in this building, including arts and crafts activities, human exhalation, consumer
products, and tobacco smoking.
A recent review presentation by Schuver (2013) thus summarizes the usefulness of radon as a
"qualitative/semi-quantitative indicator of building specific susceptibility to near-surface soil-gas/vapor
intrusion" and "a signature of the building-specific responses to environmental changes" and a "key (3rd-
strike) evidence basis for demonstrated-potential for chem-VI." Steck (2012) summarizes an EPA
document currently under development that describes lessons learned from radon that can be applied in
vapor intrusion research and practices.
2.1.4 Passive VOC Sampling
Sorbent-based methods are an emerging technology for vapor intrusion assessment. Current standard
practices for indoor air VOC monitoring in the United States include the use of negatively pressurized,
ultra-clean, passivated, stainless steel canisters for sample collection. Practitioners frequently use 8- to 48-
hour integrated samples with Summa canisters in an attempt to average over an exposure period. This is
the U.S. "gold standard" for indoor air analysis, but it is expensive to implement. Professional experience,
shows that the flow controllers currently used in commercial practice are subject to substantial flow rate
and final pressure errors when set for integration times in excess of 24 hours (Hayes, 2008).
Active and passive sorbent sampling techniques are already in use in the United States for personal air
monitoring for industrial workers and are outlined in both the Occupational Safety and Health
Administration (OSHA) Sampling and Analytical Methods (http://www.osha.gov/dts/sltc/
methods/toc.html') and National Institute for Occupational Safety and Health (NIOSH) Manual of
Analytical Methods (http://www.cdc. gov/niosh/nmam/). Typical sampling scenarios involve the collection
of active or passive samples to monitor a single chemical used in the workplace over a period of up to 10
hours. These methods are designed to meet OSHA permissible exposure limits (PELs), which are
typically in the ppm range and consequently several orders of magnitude higher than risk-based indoor air
screening levels and not suitable for ambient air measurements without modification.
Active sorbent methods (i.e., TO-17) have also been published by EPA for VOC measurements in
ambient air (U.S. EPA, 1999c). However, in those methods, air samples are normally actively collected
2-12
-------
Section 2—Introduction
over 1 hour, using a sample pump with a sampling rate of 16.7 mL/min to 66.7 mL/min, yielding total
sample volumes between 1 and 4 liters. Sampling intervals can be extended beyond 1 hour; however, care
must be taken to ensure breakthrough volumes are not exceeded in order to quantitatively retain the
compounds of interest on the sorbent tube. Given the minimum pump flow rate cited in TO-17 of 10
mL/min, the practical upper limit for chlorinated VOCs using a multi-bed thermal desorption sorbent tube
is on the order of 10 liters up to 20 L for select VOCs yielding a corresponding maximum collection
period of 8 to 24 hours (Marotta et al., 2012).
One way to lower the detection limits and control day-to-day variability is to sample over a longer period
of time. Recent studies have shown that it may be feasible to use passive sorbent samplers to collect a
continuous indoor air sample over several weeks. This approach would provide a lower detection limit, be
cost-effective, and result in a time-integrated composite sample. Laboratory and field evaluations of
passive samplers for ambient and indoor air applications have been published and showed promising
results for sampling durations of up to 14 days. Exposure of badge-type charcoal passive samplers to
controlled atmospheres of 10 to 200 ppb benzene, toluene, and m-xylene showed good performance when
deployed for 14 days (Oury et al., 2006). A field study published by Begerow et al. (1999) showed
comparability between two charcoal-based passive sampler geometries, badge and tube-style for 4-week
indoor and outdoor air samples. Field evaluations were also conducted using radial charcoal and thermal
desorption Radiello® samplers to determine performance over a 14-day period. Ambient BTEX
measurements using the Radiello samplers compared well to active sorbent sampling results (Cocheo et
al., 2009).
Testing at Orion Park, Moffett Field in California by EPA NRMRL Air Pollution Prevention and Control
Division (APPCD), EPA Region IX, and ARCADIS compared measurements of VOCs by Method TO-15
to three different radial and axial tube-type sorbent systems6:
7. Radial: activated charcoal (with carbon disulfide [CS2] extraction: gas chromatography-mass
spectrometry [GC/MS])
8. Radial: carbograph 4 (TO-17: thermal desorption [TD] GC/MS)
9. Axial: chromosorb 106 thermal desorption tube (TO-17: TD GC/MS)
Performance for the two radial methods was superiorto the axial method (Lutes, 2010). Testing was also
performed at the Wheeler site in Indianapolis comparing Summa canisters to Radiello radial solvent-
extracted samplers. Across the two sites, the Radiello solvent extracted showed good agreement to TO-15
and precision at both sites for chlorinated compounds. Agreement was poor for polar compounds: ethanol,
methyl ethyl ketone (MEK), methyl isobutyl ketone (MIBK), and acetone. Radiello TD correlated well
with Summa TO-15 but gave noticeably lower concentrations, suggesting that 2 weeks is too long an
integration time for these samplers. The agreement of the axial (tube) method was inferior (Mosley et al.,
2008; Lutes et al., 2010).
Table 2-1 compares the characteristics of commercially available passive sampler geometries and
available sorbent configurations. The geometry of the sampler (radial, badge, or axial tube) largely
determines the sampling rate or uptake rate with the radial design resulting in the highest sampling rate
and the tube-style the lowest sampling rate. The permeation sampler relies on permeation of the vapor-
phase compound through the polydimethylsiloxane (PDMS) membrane and adsorption to the sorbent bed
behind the membrane. The greater the sampling rate, the greater the mass of VOCs adsorbed onto the
sorbent bed. In addition to the passive geometries available, sorbent pairings fall into two main
6Radial samplers are sorbent-containing tubes where diffusion from the surrounding air occurs radially along the entire length of the tube. Axial
samplers are tubes containing sorbent where diffusion occurs axially through one open end of the tube. Because of the higher surface area
exposed for diffusion, radial samplers have higher uptake rates than axial tube-type samplers.
2-13
-------
Section 2—Introduction
categories—charcoal based and thermally desorbable. Charcoal-based materials are characterized as very
strong sorbents with a large surface area and a corresponding high adsorption capacity. To efficiently
extract adsorbed compounds for measurement in the laboratory, an aggressive solvent extraction is
required. The thermally desorbable sorbents are generally much weaker than charcoal with a smaller
surface area, allowing for analysis of the adsorbed compounds through thermal extraction. As Table 2-1
shows, when comparing the same passive geometry, the thermally desorbed model provides the lowest
detection limits, while the charcoal-based solvent-extracted system allows for longer sampling times as
well as a greater dynamic range because the high capacity of the charcoal minimizes sorbent saturation
under conditions of high analyte or background matrix.
European agencies have developed standard methods for passive sampling for VOCs that are applicable
to the range of concentrations and durations to be tested in this project:
¦ Methods for the Determination of Hazardous Substances (MDHS) 88: Volatile Organic
Compounds in Air: Laboratory Method Using Diffusive Samplers, Solvent Desorption and Gas
Chromatography, December 1997. Published by the Health and Safety Executive of the United
Kingdom: http://www.hse.gov.uk/index.htm.
¦ MDHS 80: Volatile Organic Compounds in Air: Laboratory Method Using Diffusive Solid
Sorbent Tubes, Thermal Desorption and Gas Chromatography, August 1995. Published by the
Health and Safety Executive of the United Kingdom: http://www.hse.gov.uk/index.htm.
¦ Ambient air quality: Standard Method for Measurement of Benzene Concentrations—Part 4:
Diffusive Sampling Followed by Thermal Desorption and Gas Chromatography, EN 14662-
4:2005. Published by the European Committee of Standardization.
¦ Ambient air quality: Standard Method for Measurement of Benzene Concentrations—Part 5:
Diffusive Sampling Followed by Solvent Desorption and Gas Chromatography, EN 14662-
5:2005. Published by the European Committee of Standardization. (Also published as the British
Standard BS EN 14662-5:2005).
Indoor air quality: Diffusive Samplers for Determination of Concentrations of Gases and Vapors: Guide
for Selection, Use, and Maintenance, EN 14412:2004. Published by the European Committee of
Standardization.
2-14
-------
Section 2—Introduction
Table 2-1. VOC Indoor Air Sampling Method Options
Parameter
Whole Air
Sorb© n t_
Active
Sorbent-Diffusive
Collection media
Summa
canister (TO-
15)
Multi-bed
ATD sorbent
tubes (TO-
17)
Radial:
Charcoal
(Radiello
130)
Radial: TD
sorbent
(Radiello 145)
Badge: Charcoal
type
(SKC 575, 3M
OVM3500)
Badge: TD
sorbents selected
by deployment
time: (SKC Ultra I,
II, III)
Tube: TD
sorbents (e.g.,
Chromosorb 106)
Permeation:
Charcoal
type
(WMS™)
Permeation:
TD sorbent
(WMS™)
Ease of deployment
Good
Good
Excellent
Excellent
Excellent
Excellent
Excellent
Excellent
Excellent
Estimated media &
shipping costb
$$$
$$
$
$$
$
$$
$$
$
$
Method and analysis
TO-15 GC/MS
TO-17
GC/MS
Solvent
Extraction
GC/MS or
GC/FID
TO-17
GC/MS
Solvent extraction
GC/MS or
GC/FID
TO-17 GC/MS
TO-17 GC/MS
Solvent
extraction
GC/MS
TO-17
GC/MS
Estimated analytical
reporting limit
0.05-0.1
ng/m3
1-10 ng
100-200 ng
1-10 ng
75-200 ng
1-10 ng
1-10 ng
50-200 ng
1-10 ng
Expected sampling
rate
0.5-3.5
mL/min
10-200
mL/min
-60 mL/min
-25 mL/min
-10 mL/min SKC
-30 mL/min 3M
-10 mL/min
-0.5 mL/min
-0.5-5
mL/min
-0.5-5
mL/min
Recommended
sampling duration
Typically 24
hours
8-24 hours
Up to 30
days
Up to 7 days
for chlorinated
solvents
Up to 4 weeks
1 -7 days
In general, up to
4 weeks)
Up to 30
days
Up to 30
days
Estimated sample
reporting limits3
-0.05 (SIM)—
0.1 |jg/m3
-0.1-1 |jg/m3
-0.1-0.4
|jg/m3
-0.005-0.05
|jg/m3
-0.25-2 |jg/m3
-0.01-0.1 |jg/m3
-0.2-2 |jg/m3
-1-40 |jg/m3
-1-40 |jg/m3
Applicable range of
chlorinated solvents
(based on available
sampling rates)
TCE/PCE and
all breakdown
products
including vinyl
chloride (VC)
TCE/PCE
and all
breakdown
products
including VC
TCE, PCE,
111-TCA,
chloroform
TCE, PCE,
111-TCA
Validated for a
wide range of
chlorinated
solvents for 8
hours, several for
up to 30 days
TCE, PCE, DCE,
111-TCA,
chloroform, 12-
DCA, cis-12-DCE,
trans-12-DCE, 11-
DCA
TCE, PCE, 111-
TCA
TCE, PCE
and most
breakdown
products
TCE, PCE
and most
breakdown
products
Approximate costs: $ <$50, $$ = $50 to $100, $$$ >$100.
b Normalized to a 7-day period for diffusive samplers.
-------
Section 2—Introduction
Given the wide range of sampling durations required for this project, several diffusive sampler
configurations are recommended to meet anticipated project objectives for indoor air measurements. For
short-term samples (less than 7 days), the sampler must have sufficient sensitivity to measure the low
VOC concentrations that are expected in the indoor air. Thermally desorbable sorbents paired with a
badge or radial-style geometry can be used effectively for the 24-hour samples and yield low reporting
limits. The badge style is recommended over the radial style given the larger number of chlorinated
compounds for which sampling rates have been generated and validated. For durations of greater than 7
days, stronger sorbents with higher adsorptive capacity are recommended, which require solvent
extraction. Although the solvent extraction is less sensitive than thermal desorption, the high sampling
rate of the radial sampler geometry over durations of 7 to 30 days will result in sample reporting limits
essentially equivalent to or lower than those generated using the thermal desorption technique.
Very few studies have evaluated VOC measurements using diffusive samplers beyond 30 days, and
determining if this is possible is one objective of this study. The sorbent selection, the sampler geometry,
and the target chemical's volatility all may have a significant impact on the successful application of
diffusive samplers to extended deployment periods. The few published studies evaluating sampling
intervals greater than 30 days are largely focused on measuring BTEX (Bertoni et al., 2001; Brown and
Crump, 1993), and the stability of chlorinated compounds on sorbents in the presence of humidity and the
variability of the sampling rate past 30 days are not well understood for any of the diffusive samplers
under consideration for this study.
Given the previous studies discussed above and the existence of standard methods for this application in
Europe, the 1- and 2-week Radiello passive samplers for VOCs are considered sufficiently accurate and
precise to be the primary VOC measurement tool in this project and are used as a basis of comparison for
longer duration samples.
Results from our previous report on studies of this house (U.S. EPA, 2012a) led to these conclusions
regarding the performance of the solvent extracted radial style charcoal passive sampler:
¦ Excellent agreement was observed between numerical averages of successive 7-day exposure
samples with the results of single passive samplers exposed for 14 days (almost always within
+/- 30%) for all compounds despite dramatic temporal variability. This suggests uniform uptake
rates for these time periods.
¦ PCE, benzene, hexane, and toluene performed well for 28 days.
¦ PCE and toluene performed well for 91, 182, and 364 days.
¦ Temporal variability is substantial and for certain compounds passive samplers allow cost-
effective acquisition of long-term average concentration data.
¦ Compound vapor pressure correlates with the relative performance of different compounds with
the passive samplers. Method accuracy over different durations increases with increasing vapor
pressure because of better sorbent retention of the VOC.
2.2 Objectives
The overall goal of this project was to investigate distributional changes in VOC and radon concentrations
in the indoor air, subslab, and subsurface soil gas of a residential building from an underground
(groundwater or vadose zone) source adjacent to the residence. The time frame of this study was more
than 2 years in order to evaluate the effects due to seasonal variations on radon and VOC vapor intrusion.
This report describes the second phase of this project, with the first phase described in U.S. EPA (2012a).
Several objectives that were established for the initial, first phase research effort were continued in this
study:
2-16
-------
Section 2—Introduction
10. Identify seasonal fluxes in radon and VOC concentrations as they relate to atypical use of HVAC
in the building.
11. Establish relationship between subslab/subsurface soil gas and indoor air concentrations of VOCs
and radon.
12. Determine the relationship of radon to VOC concentrations in, around, and underneath the
building.
13. Characterize the near-building environment sufficiently to explain the observed variation of
VOCs and radon in indoor air.
14. Determine whether the observed changes in indoor air concentration of volatile organics of
interest can be mechanistically attributed to changes in vapor intrusion.
15. Evaluate the extent to which groundwater concentrations and/or vadose zone sources control soil
gas and indoor air VOC concentrations at this site.
In October 2012, we critically evaluated our progress against these objectives, assessed the additional
information that could be gained from further study, and used that assessment to define the objectives of
this study. Table 2-2 provides additional detail on the objective statements continued from the previous
project. Data quality objectives for these objectives can be found in U.S. EPA (2012a).
Table 2-2. Continuing Project Objectives Addressed in this Document
Original Statement of Objective
Current/Ongoing Status
Determine relationship of radon to
VOC concentrations.
This study continued to address how radon relates to VOC vapor intrusion,
including statistical analysis of correlations. The relative effect of SSD
mitigation on radon and VOC concentrations and attenuation factors was
also addressed.
Establish relationship between
subslab/subsurface soil gas and
indoor air concentrations of VOCs and
radon.
Soil gas and indoor air concentrations are compared graphically in a similar
fashion as the previous project. In general, soil gas concentrations continue
to increase with depth and appear to drive indoor air concentrations.
Identify any seasonal fluxes in radon
and VOC concentrations as they
relate to a typical use of HVAC in the
building.
Although the previous report adequately addressed this objective, it left
unexplained why the relationship between stack effect driving force and
indoor concentration appears to be nonlinear. This study examined the
interactions of multiple meteorological factors. In this study we also
investigated how the HVAC and the SSD mitigation system interact to
influence vapor intrusion processes
Determine if observed changes in
indoor air concentration of volatile
organics of interest are
mechanistically attributable to
changes in vapor intrusion.
The current study looked at both preferential flow pathways using helium
tracer and geophysical tests. We also looked at the effects of a frozen soil
capping event and conducted a more detailed time series analysis of how
meteorological and building factors (such as differential pressures) interact to
influence vapor intrusion.
Characterize the near-building
environment sufficiently to allow future
3D modeling of this site.
Helium tracer tests, groundwater level monitoring, and a geophysical
investigation were conducted to provide a better understanding of the near-
building subsurface conditions.
Evaluate the extent to which
groundwater concentrations control
soil gas concentrations at this site and
thus indoor air concentrations (as
distinguished from vadose zone
sources).
Previous work established that soil gas concentrations of both chloroform
and PCE peak just above the water table. PCE groundwater concentrations
measured continued to correlate well with deep soil gas, but although
analytical improvements (lower detection limits) enabled chloroform to be
detected in groundwater, current concentrations were too low to drive the
measured soil gas chloroform concentrations.
(continued)
2-17
-------
Section 2—Introduction
Table 2-2. Continuing Project Objectives Addressed in this Document (cont.)
Original Statement of Objective
Current/Ongoing Status
Collect additional data to evaluate the
possibility of a "capping" effect from
snow and ice cover.
A colder 2012/2013 winter allowed us to collect the additional data on the
capping effect. We showed that both snow/ice cover and snowfall events
without accumulation appear to affect vapor intrusion.
Evaluate the ability of a low-cost
($129) consumer-grade radon
detector provide a continuous
indication of soil gas entry into the
structure.
Continued testing during SSD system installation and manipulation to test the
commercial radon detector under varying radon levels, and confirm good
performance at and below 4 pCi/L and good utility overall in spite of a high
bias at higher radon concentrations.
The following additional studies were undertaken with new objectives developed for this report:
¦ Better define the particular subsurface conditions that influence the movement of VOCs and
radon into this home. Conditions investigated include differences in vadose zone air permeability
beneath and immediately adjacent to the structure; definition of entry routes for soil gas into the
building envelope; the degree to which utility corridors function as preferential transport
pathways, either through the vadose zone or through the building envelope; and how the
structure of the foundation may subdivide the subslab air space.
¦ Design, install, and monitor a mitigation system based on the predominant vapor intrusion
mitigation technology—SSD—with the objective of determining how efficiently SSD works for
mitigating radon and VOCs in this particular well-studied case. We monitored VOCs and radon
from the SSD system exhaust pipe along with flow with the objectives to calculate flux through
the system and to determine if this flux can be usefully correlated with indoor VOC and radon
concentrations.
¦ Capture an additional winter snow/ice capping event to monitor its influence on vapor movement
into the home with an SSD mitigation system installed.
Characteristics of the experimental design and data quality objectives developed to meet these objectives
are described below.
2.2.1 Time Scale and Measurement of Independent and Dependent Variables
In our overall study design, we used weekly measurements to observe our dependent variable—indoor air
concentration. We expected the indoor air concentration to be dependent on the flux from vapor intrusion
from soil gas. Our dependent variable is thus controlled by a series of independent variables with different
time cycles that affect the vapor intrusion process, including air temperature, barometric pressure, wind,
soil moisture, soil temperature, groundwater level, and HVAC operation.
In the course of this study we monitored or measured most of these independent variables or their
surrogates and different frequencies balancing on the general desire for continuous measurements against
logistic considerations. Table 2-3 was prepared to consider these time-scale issues and the implications
they may have for our test matrix. Figures in Nazaroff and Nero (1988) show examples of how such
independent variables controlled indoor radon concentrations in previous studies.
2.2.2 Data Quality Objectives and Criteria
Table 2-4 summarizes the data quality objectives and criteria for the new elements of this project. Each
objective is expressed first qualitatively and then each objective is expressed in quantitative and statistical
terms where possible. The measurements that were used to achieve each objective are also listed. More
details on the specific test methods used are provided in Section 3 of this report. Specific sampling and
2-18
-------
Section 2—Introduction
analytical data quality objectives can be found in the quality assurance/quality control section of this
report (Section 4).
Table 2-3. Factors Causing Temporal Change in Vapor Intrusion and How They are Observed and
Measured
Independent
Variables/
Causes of
Variability
Expected Time
Cycle
Indoor VOC & Soil Gas Measurement
Intervals Available to Observe at these
Time Scales
Measurements of
Independent Variables
Available
HVAC system
on/off
¦ 10 min-1 hour
The influence of the HVAC system in
general was observed on a scale of days
and weeks by comparing sides of the
duplex and periods of "on" and "off' for AC
units. Although the individual cycles of the
forced air heating system were not visible in
measurements taken over a 24 hours or
longer time scale or even in the 2-hour
online GC data, the cumulative impact of
heating system "on" and "off' over exposure
periods of weeks to was relevant.
Measurement with data
logger was planned every
5 minutes within heating
season.
Diurnal
temperature/
wind (night/day)
¦ 24 hour
Measurements with the online GC and
continuous radon instruments have
sufficient time resolution on the scale of
hours to observe this.
Weatherstation: at least
one data point per hour.
Barometric
pumping from
weather fronts
¦ 2-3 days typical
Weekly, except for daily samples and
continuous measurements during intensive
sampling events.
Weather station: ambient
pressure logging with at
least one data point per
hour.
Water table
fluctuations
¦ Barometric
pressure: 2-3
days
¦ Rain events:
irregular
¦ Seasonal climate:
monthly
¦ Surface water
level: hours
Weekly and monthly integrated indoor air
samples. Measurements with the
continuous radon and on-site GC
instruments have time resolution on the
scale of minutes to hours.
Monthly water-level
measurements;
supplemented beginning in
fall 2012 with real-time
data logger at one station
on site; strong correlation
of gauge height and
groundwater level enabled
hind casting of
groundwater levels for
entire project.
Soil and
groundwater
temperature
change
¦ Annual/seasonal
Weekly, biweekly, and quarterly samples of
indoor air and soil gas.
Soil temperature logging
with thermocouples: one
or more points per hour.
Groundwater temperature
monthly during sampling.
Vadose zone
moisture
change
¦ Seasonal major
rain events?
Weekly samples of indoor air and soil gas.
Measurements with the online GC and
continuous radon instruments have time
resolution on the scale of hours.
Once per hour at five
depths.
Stack-effect,
heating vs.
cooling season
¦ Daily and
seasonal
Weekly samples of indoor air and soil gas.
Measurements with the online GC and
continuous radon instruments have time
resolution on the scale of hours.
Differential pressures,
indoor temperatures: 15-
minute rolling average.
2-19
-------
Table 2-4. Data Quality Objectives and Criteria
Task Order
Objective
Study Purpose/
Objectives
Study Questions
Measurement Used To
Support Study Questions
Measurement Performance or Acceptance
Criteria
Determine
subsurface
effects on
building
envelop
Better define the
particular subsurface
conditions that influence
the movement of VOCs
and radon into this
home.
What is the nature of the
subsurface environment—fine
features of the stratigraphic
environment that influence
contaminant transport?
Geophysical techniques,
including ground penetrating
radar (GPR), electromagnetic
resistivity (EM), electrical
resistivity (ER), vacuum
testing, and subsurface tracer
testing (to be addressed in a
separate geophysical report)
Geophysical data to be acquired from multiple
transects within each third of the basement.
Vacuum testing from at least two extraction points.
At least two helium tracer injections with monitoring on
1-hour frequency at five or more locations.
Continue data analysis of soil gas information to
provide data for a three-dimensional picture of radon
and VOC concentration changes at different locations
overtime to discern seasonal patterns, whether
caused by temperature or water level. Observe effect
of on/off cycles of the SSD system on contaminant
movement.
(NOTE: analytical data quality objectives are defined in
Section 4)
Determine the effects of
utility corridors on
subsurface movements
of VOCs and radon.
How do the house foundation
features and utility corridors
affect subsurface movements of
VOCs and radon?
Active and passive sampling
methods.
Geophysical measurements
(GPR)
(to be addressed in future
reports).
Determine
effect of
SSD
mitigation
on VOC and
radon levels
Install and monitor an
SSD mitigation system
for VOCs and radon.
How efficiently does the SSD
system work at this particular
location?
Active and passive VOC and
radon sampling of both the
subslab environment and the
stack with the SSD system on
and off in 1-week intervals.
Indoor air VOC and radon
data, differential pressure data
(differential pressure to be
addressed in a future report).
At least three on/off cycles of the SSD system.
One 6-week period of on-site GC operation for soil
gas, indoor air, and ambient VOC levels.
Continuous indoor air radon measurements during
entire test period.
Weekly VOC and radon measurements during entire
test period.
Weekly radon and VOC grab samples from soil gas
ports (subslab and deeper).
Continuous Setra differential pressure measurements
(NOTE: analytical data quality objectives are defined in
Section 4)
Can a relationship be
established between the flux of
VOC/radon exhausted from the
stack and the effectiveness of
the system in reducing indoor
air concentrations?
What is the differential pressure
required to counter the
observed substantial seasonal
variability in VOC
concentrations?
Does the effect of the SSD
system reduce the variable
component of the vapor
intrusion driving force (in
addition to changing the
average differential pressure)?3
(continued)
-------
Table 2-4. Data Quality Objectives and Criteria (cont.)
Task
Order
Objective
Measurement Performance or Acceptance
Criteria (for this question/# of data points
anticipated)
Study Purpose/
Objectives
Study Questions
Measurement Used To
Support Study Questions
Determine
the effect
of a winter
ice cap on
vapor
intrusion
Capture a winter ice
capping event and
monitor its influence on
vapor movement into the
home.
Does a winter ice capping event
or heavy snowfall affect VOC
and radon concentrations?
If so, through what mechanism
does the winter event affect
concentrations (increase in
subslab concentration? effects
on stack effect induced
pressure field?)?
Can an effect of snow/ice cover
outside the residence be
discerned that is separate from
the effect of temperature?
Active and passive sampling
in an intensive mode. Reports
of snow cover made during
week days by ARCADIS on-
site personnel. We
supplemented these
observations with publically
available information from the
National Climatic Data
Center/National Weather
Service for the Indianapolis
International Airport
http://www.ncdc.noaa.qov/sno
w-and-ice/dlv-data.php.
Multiple snow and ice events (as observed in the
winter of 2012/2013);
Statistical data analyses (also performed for previous
winters).
8 weeks of online GC for VOCs
Continuous AlphaGUARD data for radon
Maintain index of snow cover for the long-term analysis
of the indoor air data set.
(NOTE: analytical data quality objectives are defined in
Section 4)
O'
S
Kj
O
o'
s
-------
Section 2—Introduction
2-22
-------
Section 3—Methods
Table of Contents
3.0 Methods 3-1
3.1 Site Description 3-1
3.1.1 Area Geology/Hydrogeology 3-1
3.1.2 Area Potential Sources 3-1
3.1.3 Building Description 3-7
3.1.3.1 Building Age, Condition, and HVAC 3-7
3.1.3.2 Building Utilities/Potential Entry Points 3-8
3.1.4 Building Occupancy During Sampling 3-9
3.1.5 Investigation History 3-9
3.2 Evolution of Conceptual Site Model 3-10
3.2.1 Priorto 2011-2012 Investigations 3-10
3.2.2 After 2011-2012 Investigations (U.S. EPA, 2012a) 3-10
3.2.3 Refinements in Conceptual Site Model Sought in this 2012-2013 Study 3-11
3.3 Building Renovation and Mitigation 3-12
3.3.1 Subslab Depressurization Mitigation System Installation 3-12
3.4 Monitoring, Sampling, and Analysis 3-17
3.4.1 Indoor and Outdoor Air VOC Monitoring 3-17
3.4.2 Subslab and Soil Gas (TO-17) 3-23
3.4.3 Online Gas Chromatograph 3-22
3.4.4 Groundwater 3-23
3.4.5 Subslab Depressurization System Stack Gas Sampling 3-24
3.5 Radon Sampling and Analysis 3-24
3.5.1 Indoor Air Radon Sampling and Analysis 3 -24
3.5.2 Subslab and Soil Gas Radon Sampling and Analysis 3-25
3.5.3 Continuous (Real-Time) Indoor Air Radon Sampling and Analysis 3-26
3.6 Physical Parameters Monitoring 3-26
3.6.1 On-Site Weather Station 3-26
3.6.2 Indoor Temperature 3-28
3.6.3 Soil Temperature 3-28
3.6.4 Soil Moisture 3-28
3.6.5 Potentiometric Surface/Water Levels 3-29
3.6.6 Differential Pressure 3-29
3.6.7 Air Exchange Rate 3-30
3.6.8 Crack Monitoring 3-30
3.7 Data Aggregation Methods 3-31
List of Figures
3-1. Lithological fence diagram showing the major soil types beneath the 422/420 house 3-4
l
-------
Section 3—Methods
3-2. Aerial view of duplex, 420/422 East 28th Street, showing nearby sanitary and storm
sewers 3-5
3-3. East side of house (on right) and adjoining commercial quadraplex visible (left) 3-6
3 -4. Roof of adj acent commercial quadraplex 3-6
3 -5. Looking toward southeast corner of adj acent commercial quadraplex 3-7
3-6. Visual evidence of historic dry cleaners in area 3-8
3-7. Front view of house during summer 2011 sampling, with fan testing and weather station 3-9
3-8. Front view of duplex under winter conditions showing designation of sides and HVAC
setup 3-10
3-9. 422 (left) and 420 East 28th Street in January 2011 3-11
3-10. Map view of the 422-side basement showing SSD mitigation system legs, subslab soil
gas extraction pits (red circles), and the position of the passive "sampling racks."
Horizontal divisions are walls between "north" (top in figure), "central," and "south"
(bottom in figure) sections of basement with open walkways between (cistern is in the
central basement) 3-14
3-11. Map view of the 420-side basement showing the SSD mitigation system legs, subslab soil
gas extraction pits (red circles), and the passive "sampling racks." Horizontal divisions
are walls between "north" (top in figure), "central," and "south" (bottom in figure)
sections of basement with open walkways between (cistern is in the central basement) 3-15
3-12. Photos of mitigation system: (left) SSD blower and stack on northeast corner of duplex;
right) SSD extraction point, showing valve and U-tube manometer 3-16
3-13. Cross-section showing the general layout of the 422/420 north and central basements
with the positioning of the extraction legs, exterior blower, and exhaust stack 3-17
3-14. Subsurface soil gas monitoring probes (SGP), subslab sampling ports (SSP), and
groundwater monitoring wells (MW). Horizontal divisions are walls between "north,"
"central," and "south" sections of basement with open walkways between (cisterns are in
the central basements). Probes/ports in red were sampled by the on-site GC. Soil
temperature and moisture probes were installed in the 422 basement between SGP 8 and
MW 3 and in the backyard to the north of MW 2 3-18
3-15. Passive indoor air sampling rack: 422 first floor 3-20
3-16. Ambient sampler shelters on telephone pole near duplex 3-21
3-17. Monitoring well MW-3, installed in the basement and completed on the first floor 3-24
3-18. Front view of 420/422 duplex with location of weather station sensors indicated with red
arrow 3-27
3-19. Calibrated crack monitor 3-30
List of Tables
3-1. Pressure Readings Taken During Extraction Point Testing 3-13
3-2. Data Aggregation Applied to Predictor Variables 3-31
11
-------
Section 3—Methods
3.0 Methods
3.1 Site Description
The test house is a vacant residential duplex at 420/422 East 28th Street in the Mapleton Fall Creek
neighborhood of Indianapolis. This area of Indianapolis was initially a farming settlement known as
Mapleton founded in the 1840s. The primary residential development in this area occurred in the late
1800s and early 1900s. Commercial development on the immediate cross street, Central Avenue, began in
the 1920s.
3.1.1 Area Geology/Hydrogeology
Several soil borings were advanced in the area immediately surrounding the house, during monitoring
well (MW) construction and soil gas port installation. SGP-1A, SGP-1B, and SGP-1C, as well as MW-1A
and MW-1B, were installed April 29, 2010. All additional soil gas ports and MWs on the exterior of the
house were installed between August 30 and September 1, 2010. Soil gas ports and MW-3 located below
the footprint of the house were installed in September 2010. 3-D visualizations of subsurface lithology are
presented in Figure 3-1. Boring logs are included in Appendix A.
In the southern portion of the property, topsoil extends down to about 0.5 to 1 ft. Beneath the topsoil is
found sand or silt mixed with cinders, coal fragments, or ash to about 1.5 ft. Then to between 5 and 6 ft is
silt or silty sand with varying amounts of clay. Some trace gravels start at about 7 ft, and underlying that
layer are sands and gravels to between 15 and 16 ft. beneath that sand and gravel is generally sand.
To the east side of the property, at the surface are soils with a visibly high organic content and gravel or a
concrete sidewalk. Underlying that from 1 to 3 ft is sand or clayey sand, with some gravel and coal
fragments in some borings. Beneath that layer down to 7 ft is predominantly clay with some sand or silt.
Underlying that is sand with some clay and gravel down to about 12 to 14 ft. Down to 16.25 ft is sand,
with gravel added down to 16.5 ft.
To the north side of the property, the first foot is fill, sand, and gravel. Beneath that to 3 ft is brick, with
sand and weathered brick to 3.5 ft. The brick constituent in this location is possibly a remnant of a former
exterior basement stairwell. To 6.25 ft is silty, sandy clay. Down to 8 ft is sand, with sand, gravel, and
some clay down to 12 ft. From 12 to 16 ft is sand.
On the west side of the property, the first half-foot beneath the surface is the concrete sidewalk.
Underlying that to 1.25 ft is fill, cinders, and gravel. Down to 6.75 ft is silty, sandy clay with trace gravel.
The layer beneath that to 15.5 ft is sand and gravel with some clay followed by sand to the end of the
boring at 16.5 ft. See Section 6.1 for additional information on site soils.
In the top figure, the view is toward the north from the street in front of the house. The bottom figure
shows a view toward the south from the backyard. The empty white area at the top of the soil figure
represents the house's basement. In the immediate vicinity of the house, silt and clay (brown) are present
until 7.5 to 8 ft below land surface. After that, sand and gravel (burnt orange) alternate with layers of sand
(orange).
3.1.2 Area Potential Sources
The site location, as illustrated in Figure 3-2, is bounded to the north by 29th St., to the west by N. New
Jersey St., and to the east by Central Avenue. There is a river, Fall Creek, approximately 300 ft to the
3-1
-------
Section 3—Methods
IwiWHumtiawHnt
i '>K:RETE_SPTK>_fB-L_Ct>L.
To/umwi'i
sow
Ltthology Fences
6WMTI» )H*C K.
tofwSI
WAOjUKI OP^tLj
Mj./je.cur)
Figure 3-1. Lithological fence diagram showing the major soil types beneath the 422/420 house.
|BOM|
, C'Dplh (H by.} —
o~o:p!STiH
| Littiology Fences |
3-2
-------
Section 3—Methods
south of the site toward which groundwater flow generally trends. Across the street south of the site, there
is a parking lot and to the east there is an open field. Across an alley to the west of the site, there is an
open lot with a grassy area and a paved parking lot. Adjacent to the north side of the site there are
backyards of the residential buildings along Central Avenue.
420 E. 28th St, Indianapolis, IN
4Stormsewers.shp
4 Sanitarysewers.shp
Gas Lines
• Stormsewerstructures.shp
• Sanitarysewerstructunes.shp
® CSO's
Figure 3-2. Aerial view of duplex, 420/422 East 28th Street, showing nearby sanitary and storm
sewers.
Immediately adjacent to the studied duplex (approximately 10 ft east) lies a small commercial/residential
quadraplex (Figures 3-3, 3-4, and 3-5)) with a diverse, primarily commercial history dating back to 1930.
The four portions of the building are numbered as 424 East 28th Street, 426 East 28th Street, 2802
Central Avenue, and 2804 Central Avenue.
0.08 Miles
3-3
-------
Section 3—Methods
Figure 3-3. East side of house (on right) arid adjoining commercial quadraplex visible (left),
Figure 3-4. Roof of adjacent commercial quadraplex.
3-4
-------
Section 3—Methods
Figure 3-5. Looking toward southeast comer of adjacent commercial quadraplex.
Among the historic uses of parts of that building were a pharmacy, beauty supply, radio shop, fur store,
and detector companies. Regarding most of the businesses that occupied that space, only their name is
currently known, and those names do not match any businesses with a current local or Internet presence.
Thus, chemical uses, though probable, are not documented. The back part of the adjacent building at 2804
Central Ave has historically been occupied by "Wolf Fur Co." Later in 1954, the same location was
occupied by the 'Avideo Detectors Telaveta." In 1930 it was occupied by Gould & Schildmoler ENEN
and Home Radio Co. The records for the adjacent buildings (424 to 428 East 28th Street and 2802 to
2804 Central Avenue) show a number of drug store and beauty shop uses. There are substantial gaps in
the records for these properties, and there seems to be little or nothing reported about what was occupying
these locations between 1970 and 2000.
There were 9 to 10 histonc laundry cleaners located less than a quarter of a mile to the north of the
422/420 house, and one was a quarter of a mile to the west (Figure 3-6). These were listed as hand and
steam laundries, pressers, and driers. Tire most recent laundry was present in 1970 (EDR Radius Map,
June 15, 2010). In the fall of 2010, we observed Mapleton/Fall Creek Development Corporation
(MFCDC) staff excavating an underground storage tank that appeared to contain product at a dry cleaners
several blocks upgradient from the house.
There were three historic gas stations or auto service and repair shops within a quarter of a mile to the
north as well. The most recent auto repair shop was present in 1990 (EDR Radius Map, June 15, 2010).
3-5
-------
Section 3—Methods
I \ . I
—1—i'
Figure 3-6. Visual evidence of historic dry cleaners in area.
The property southwest of the intersection of East 28th and Central Avenue was historically mildly
impacted with petroleum hydrocarbons and managed as a brownfield named "Mapleton-Fall Creek Site"
or "Fall Creek Central Project." This site was closed after tank and soil removal. One round of VOC
groundwater data was acquired at that location that showed detectable chloroform at 8.9 to 22,1 ug/L in a
June 2005 sampling event. These previous studies showed that the study area has sand and gravel geology
from approximately 7 to 25 ft below land surface (bis) and groundwater at approximately at 16 ft bis. The
upper 7 ft of the stratigraphy is heterogeneous, variously described as including fill materials, loam, and
silty and moist sandy clay.
Based on the general topography of the area and professional experience in this portion of Indianapolis,
groundwater is thought to flow from the north of the 422/420 house south of the house to Fall Creek.
Thus, many of the historic laundries or auto shops that are potential contaminant sources are generally
upgradient of the study house.
The 422/420 house is located between Central Avenue and its associated alleyway on 28th Street. The
immediate area receives a moderate amount of traffic, but the Central Avenue/Fall Creek Parkway
intersection is very busy throughout most of the day. This could be a contributing factor to petroleum-
based contaminants in surface soils and ambient air.
3-6
-------
Section 3—Methods
3.1.3 Building Description
3.1.3.1 Building Age, Condition, and HVAC
The tested house located at 422/420 East 28th Street, Indianapolis, IN (Figure 3-7) is an early twentieth
ccntur\ duplex, dating from before 1915 because it is present on the 1915 Sanborn map of the area. Based
on the mirrored floor plans of the two sides, it is likely that the house was always a duplex. Construction
is wood frame on a brick foundation with a poured concrete basement floor. Interior floor materials
include tile, carpet, and wood flooring.
Figure 3-7. Front view of house during summer 2011 sampling, with fan testing and weather
station.
The duplex at 422/420 was initially abandoned and is now owned by MFCDC in Indianapolis. Before our
involvement, the house had been vandalized and stripped of all valuable metals and fixtures (e.g., copper
wiring and tubing, most plumbing fixtures, many outlets) and destroyed the previous HVAC unit.
A staff member from the Indianapolis ARCADIS office acquired the use of the house for the duration of
the project through the generosity of the MFCDC. A small rent is now being paid by ARCADIS to
MFCDC for use of the house for this study.
Power was restored to the house in September 2010. A gas-fired forced air HVAC unit was installed on
the 422 side in October 2010 by Edwards Electric and Mechanical for use in this project (Figure 3-8).
The house had no air conditioning (AC) system, and we chose to install window-mounted units, which
would have been the likely type used by any tenants in this house.
3-7
-------
Section 3—Methods
422
Heated
7
/!
/ -
Figure 3-8. Front view of duplex under winter conditions showing designation of sides and
HVAC setup.
There are internal and external visual clues indicating (Figure 3-9) the house has been updated several
times. For example, visual clues suggest that a previous HV AC unit had been installed that was not native
to the house's original construction. In the basement, there is evidence of former coal chutes (possibly)
and cisterns on both the 420 and 422 sides. The probable coal chutes and old windows had been blocked
by cmder blocks before ARCADIS occupancy. The cisterns had also been cemented over. Comments
made by electricians in the basement suggest that at one time the house had been heated by an old style
furnace, indicated by cemented-over holes in the walls, but that it had been gone for some time.
3.1.3.2 Building Utilities/Potential Entry Points
The electric lines connect to the house at the northwest comer of the 420 side. Since all original wiring
native to the house had been removed by vandals before the project, we had to have the junction box
rewired to the city electrical line and run new lines within the house to new outlets at designated points.
The gas line connects only to the furnace from an access line in the south wall of the 422 side. Both the
electrical lines and the gas line were emplaced by Edwards Electrical and Mechanical during the furnace
installation and enter the house at the original entry points for each utility.
Sanitary sewer lines run immediately south of the house along East 28th Street. Sanitary and combined
sewer lines run less than one block east and west of the house along Central Avenue and New Jersey
Street (see previous Figure 3-2). There is a sewer lateral running beneath the basement floor along the
length of the 422 side from north to south that was buried and cemented over sometime after the floor's
original construction. PVC drain lines join this lateral from the plumbing on both sides of the duplex. The
HVAC unit drains condensation into a floor grill leading to the lateral. A nonfunctional water line enters
the house from the south. Large, cinder-blocked portions of the north interior basement walls of both
sides of the duplex along with brick strata in borings have been observed. We interpret these to be
vestigial entranceways to the basement from a time when the basement was accessed from the back yard,
rather than from an interior basement door.
3-8
-------
Section 3—Methods
Figure 3-9. 422 (left) and 420 East 28th Street in January 2011,
3.1.4 Building Occupancy During Sampling
The initial concept for the 422/420 house was to create an environment free from lifestyle-related indoor
air sources, but operated as though the space were occupied in order to simulate a residential environment
free from indoor VOC sources. The 422/420 house was borrowed (and is now rented) from MFCDC that
owns the property. The house was an ideal location for this study because even though it is an older
residence typical of this area of Indianapolis, it had no occupants, was not subject to any use beyond the
project, was in a good location and price range, and had vapor intrusion present. Because the house was
unoccupied and in poor condition, we could set up ports, wells, and sensors for observations and install a
mitigation system without having to consider the occupants" comfort or convenience.
To more closely approach a living environment, a field scientist worked on site for several months before
sampling began. During most normal work weeks during the periods of active VOC sampling, the field
scientist was at the house at least 4 days per week. During the down times between VOC sampling efforts
(such as April to September 2012), visits to the house were less frequent. The intent during VOC
sampling periods was to have an individual who would open doors and windows, move through the
environment, and make temperature adjustments similar to the way a homeowner would. The constant
close proximity of the worker to the work zone also allowed for quick responses to environmental
changes and any issues with the sampling devices. A second floor bedroom on the 422 side of the duplex
was minimally modified and used as an office for the sampling staff member.
3.1.5 Investigation History
The selection and screening of this duplex was conducted in April to June 2010 as described in the first
report on this series of projects (U.S. EPA, 2012a). That report describes the design and results of an
3-9
-------
Section 3—Methods
extensive 14-month soil gas, groundwater, and indoor air sampling program conducted from December
2010 to March 2012. This report covers VOC and radon samples taken after that date, priorto and after
the October 2012 installation of an SSD mitigation system, and until May 2013. In many cases, the results
and analysis in this report build on the results presented in U.S. EPA (2012a). This report also describes
the effects of the SSD mitigation system on radon and VOC levels in the duplex and investigates the
factors that influence VOC and radon levels in and under the duplex. Where sampling locations and
techniques are common to both stages of the investigation, the descriptions from U.S. EPA (2012a) are
repeated here.
3.2 Evolution of Conceptual Site Model
This report provides an opportunity to document how a conceptual site model can evolve through
intensive study of the vapor intrusion situation surrounding a single building as one of the goals of this
study. This discussion updates the conceptual site model from the previous report (U.S. EPA, 2012a).
Subsequent reports in this series will make any updates to the conceptual site model that are made
necessary as additional information is collected about the site.
3.2.1 Priorto 2011-2012 Investigations
During site selection, the initial conceptual site model for this structure was that a vapor intrusion source
was most likely present in shallow and subslab soil gas due to historical dry cleaning facilities and
adjacent commercial uses. Radon impacts were suspected because Marion County, Indiana, is in EPA's
Zone 1—highest risk for radon. Detectable concentrations of chlorinated hydrocarbons were detected
during initial site screening and responded to depressurization of the structure by fans (U.S. EPA, 2012a).
The source of the primary VOCs (PCE and chloroform) observed at this duplex was initially suspected to
be transport of contaminants either:
¦ through a groundwater pathway from upgradient dry cleaners or
¦ released into the shallow vadose zone during the operations of the adjacent commercial
quadraplex.
Later observations and discussions suggested that disinfection byproducts in city drinking water could be
an additional potential source for chloroform detected in soil gas.
3.2.2 After 2011-2012 Investigations (U.S. EPA, 2012a)
The detailed 2011-2012 site investigation and monitoring work described in U.S. EPA (2012a) added the
following details to our conceptual site model of this duplex and the vapor intrusion exposure pathway:
¦ The groundwater and nearby Fall Creek are intimately connected. The groundwater level
beneath the house is subject to rapid swings of up to 5 ft over the course of a few days during
seasonal flooding in the creek. There also could be connections to the combined sewers that
discharge into Fall Creek.
¦ The stack effect caused by indoor/outdoor temperature differentials operates not only during the
heating season, but also during the summer as well, due to the "solar stack effect" and the
storage of heat in the building during cool late summer/fall nights. Differential pressure
measurements indicate that changes in building differential pressure are reflected in a
measureable advective driving force between the 13-ft depth near the water table and the 6-ft
depth directly beneath the basement. Therefore, in this case, advection may be the primary cause
of VOC migration through the deeper portions of the vadose zone.
3-10
-------
Section 3—Methods
¦ The heterogeneity of the subslab concentrations and geophysical result suggests the absence of
an engineered gravel layer beneath the duplex. Thus, the subslab does not behave here as a well-
mixed plenum.
¦ PCE is apparently widely spatially distributed in site groundwater at concentrations well below
the current 5 (ig/L MCL (U.S. EPA, 2012a). These shallow groundwater concentrations
apparently control deep soil gas concentrations. Only a moderate degree of attenuation occurs in
those deep soil concentrations as they are drawn toward the basement of the structure.
Substantial attenuation occurs in the upper 6 ft of the site external soil gas, which is finer grained
materials than the sandy deeper materials. It is currently unclear whether this is due to gas
permeability contrasts, sorption processes, or most likely barometric pumping dilution.
Substantial attenuation also occurs across the building envelope between subslab and indoor air.
¦ Chloroform is present in highest concentration in deep soil gas. Substantial chloroform has been
historically been detected in groundwater on a site 200 ft to the southwest. Chloroform was also
detected in groundwater at this house in preliminary sampling and at low levels (<0.6 jj./L) in
the spring of 2013. Studies were conducted that determined that the lack of detections in recent
groundwater samples on site is not from losses in the sampling and analysis process. Chloroform
attenuation in soil gas is substantial between the area just above the water table and the 6-ft
depth below the structure. Chloroform is also substantially attenuated between subslab air and
indoor air.
¦ The relative importance of the potential sources of PCE and chloroform—historic dry cleaners,
historical activities in the adjacent commercial/industrial quadraplex, and leaking storm
sewers/drinking water lines—is unclear.
¦ Sewer lines and laterals likely play some role in contaminant fate and transport in this system.
Elevated concentrations of PCE and chloroform were present in the headspace of sewer gas
(U.S. EPA, 2012a). As described in U.S. EPA (2012a), their role as a direct entry pathway has
been minimized through plumbing trap and vent maintenance and blocking the drains in the
house. Their role in lateral transport through the vadose zone and into the subslab of the duplex
will be elucidated through future geotechnical studies.
¦ There is a strong seasonal component to the PCE and chloroform indoor concentrations (see
Section 11, Figure 11-12). The seasonal component is partially but not completely correlated to
the strength of the stack effect (Section 10, Figures 10-10 and 10-11).
¦ Concentrations of benzene, hexane, and toluene in indoor air are quite similar to ambient levels
and appear to move in lockstep with ambient air, although there are some traces of benzene in
soil gas (Section 11, Figure 11-12). TCE in indoor air also tracks ambient concentrations when
TCE is low, but are very similar to the PCE plots when concentrations were high at the
beginning of the study, suggesting a contribution of subsurface sources to TCE indoor air
concentrations.
3.2.3 Refinements in Conceptual Site Model Sought in this 2012-2013 Study
In the Quality Assurance Project Plan (QAPP) for the current work, we defined several goals related to
improving our conceptual site model based on the investigation results available at that time:
¦ Better define the particular subsurface conditions that influence the movement of VOCs and
radon into this home. These conditions are expected to include differences in air permeability on
a spatial scale of 1 to 20 ft in the vadose zone beneath and immediately adjacent to the structure.
¦ Better define the particular entry routes of soil gas into the building envelope. Define the degree
to which utility corridors function as preferential transport pathways—either through the vadose
zone or through the building envelope.
3-11
-------
Section 3—Methods
¦ Determine how the structure of the foundation may subdivide the subslab air space.
¦ Capture a winter capping event to monitor its influence on vapor movement into the home.
Uncertainty remains about the relative importance of a groundwater source. The detected concentrations
in groundwater in the installed wells are able to account for the highest current deep soil gas
concentrations through a Henry's law calculation for PCE within a factor of 3, but the chloroform soil gas
concentrations were 12 times higher than the vapor concentrations calculated using Henry's Law off of
the groundwater concentrations. The highest soil gas concentrations are generally on the downgradient
(SE) side of the house. However, generally the highest concentrations observed in soil gas are just above
the water table. There is a large and rapid response of the potentiometric surface to rainfall, perhaps
related to the presence of combined sewers and surface water bodies in the vicinity of the study duplex.
There is a visual correlation between chloroform trends and changes in hydrogeology. Also, at a site 200
ft to the southeast, substantial chloroform was previously detected in groundwater (U.S. EPA, 2012a,
Section 11 and Section 13.1.6).
Several hypotheses could explain these observations:
¦ The primary stored mass is in the deep vadose zone either sorbed to soil particles, present in soil
moisture, or present as soil gas in the least permeable portions of the soils. (Others have
hypothesized that vadose zone soils will retain mass for a substantial time period after an
associated groundwater plume naturally attenuates).
¦ The primary source is affected groundwater lateral to the duplex location not observed by our
monitoring wells, but perhaps suggested by prior off-site detections. The primary source is from
water that is periodically transported along deep combined sewers and leaked water from those
sewers percolating downward toward the water table. This might manifest in higher VOC
concentrations in soil moisture in the vadose zone or in the capillary fringe than in the sampled
shallowest portion of the saturated zone.
General support for the importance of these hypotheses at other sites can be found in the literature (Carr,
2011; Christ, 2010). Particularly relevant is this statement from Carr (2010): "The common perception
that VI potential is largely a function of contemporaneous groundwater quality is flawed."
3.3 Building Renovation and Mitigation
Details of the original building renovations were presented in U.S. EPA (2012a). Generally, the house
was rewired, a heating system was installed on the 422 side of the building, window air conditions were
added, and locks and a security system were installed. The primary renovation for this phase was the
installation of the SSD mitigation system.
3.3.1 Subslab Depressurization Mitigation System Installation
The strategy for the SSD mitigation system installation was to select an experienced radon and VOC
mitigation contractor and ask them to perform a "typical" active SSD system installation but with greater
documentation and reporting for the research purpose. We also added some additional valves and
sampling ports to the "typical" system to facilitate monitoring.
On October 16, 2012, Brian Schumacher and John Zimmerman of EPA were present on site with
ARCADIS and our radon mitigation subcontractor Radon Environmental to oversee the installation of the
SSD mitigation system under both sides of the 422/420 duplex. The initial plan for the installation
planned for two extraction pits to be installed at the northern basement sections on the 422 and 420 sides
3-12
-------
Section 3—Methods
of the house. After installation, if pressures with only the two initial legs proved insufficient, two more
could be installed to provide the necessary negative pressures beneath the subslab.
Prior to the drilling start time, all Radiello passive VOC sampling was stopped, the electrets were read,
and SKC Ultra III badges were begun for the duration of the drilling/installation. This was to ensure
potential increases in radon and VOCs could be monitored during installation or would not interfere with
normal VOC monitoring for the time period.
Drilling began in the north basements at 10:00 am and used the core drill method to install the subslab
suction system. A section of the concrete slab was drilled out with a core drill, and schedule 40 3 in. PVC
piping was fit into the slab hole. The piping was then arranged in the basement to intersect a main line,
which would lead out of the basement and connect to a
fan. The fan (or blower) was a Radon Away RP 265
high-flow fan unit. The fan has an on/off switch and each
extraction leg possesses a ball valve, so the whole system
can be shut off or only select legs of the system can be
shut off.
The initial two legs in each of the north basements were
initially turned on at 16:30 on October 16, 2012. After
initial pressure drop testing the project team decided that
pressures were insufficient (generally < -0.04 in.WC; see
Table 3-1) and that two additional legs would need to be
installed the following day by Radon Environmental in
each of the central basement areas. The full system, then
consisting of four extraction legs total (Figures 3-10 and
3-11), was turned on at 17:20 on October 17, 2012.
Additionally, sampling ports for the SSD mitigation
system were drilled into the positive side of the SSD
mitigation system stack (i.e., above the blower). The
ports were drilled for WMS Waterloo samplers,
AlphaGUARD sampling, and a port for insertion of an
airfoil velocity measurement attachment for the
micromanometer that was used to test the system after
installation but prior to monitoring. Figure 3-12 shows
external and internal photographs of the system, and
Figure 3-13 is a cross-section diagram showing the
general layout of the 422/420 north and central
basements with the positioning of the extraction legs,
exterior blower, and exhaust stack Additional system
photographs and details on system testing can be found in
Appendix A.
After installation and testing, the system was operated
and monitored for three on periods, two passive periods,
and one fully off period. The three on periods ran from
October 17, 2012, to November 14, 2012; December 12,
2012, to December 29, 2012; and February 6, 2013, to April 24, 2013. The two passive periods ran from
November 14, 2012, to December 12, 2012, and from December 29, 2012, to January 16, 2013. The fully
off period ran from January 16, 2013, to February 6, 2013.
Table 3-1. Pressure Readings Taken
During Extraction Point Testing
Date: 10/16/12 with two suction lines
activated, one in 422 basement north, and
one in 420 basement north.
Pressure Reading
(in WC)
Location
1
-0.155
2
-0.058
3
-0.020
4
-0.018
5
-0.006
6
-0.035
7
-0.038
8
-0.017
9
-0.011
10
-0.003
Date: 10/17/12 with the two in each of the
north basement sections and two in each of
the center basement sections.
Pressure Reading
(in.WC)
Location
1
-0.092
2
-0.089
3
-0.046
4
-0.046
5
-0.009
6
-0.065
7
-0.066
8
-0.040
9
-0.035
10
-0.006
in. WC = inches water column
3-13
-------
Section 3—Methods
To
420
side
Basement
422 East 28th St
Mitigation Leg
±1=
Sr
House
Septic
Line
S |
I
Sampling
Rack
Up to
1st Floor
O
Open
Walkway
h—I-
"I 1"
Open
Walkway
\
Mitigation
Leg
o O SGP 9
SSP 4
E
ampluvg
Rack
Port 3
\1
Blower
§>
K
v
Figure 3-10. Map view of the 422-side basement showing SSD mitigation system legs, subslab
soil gas extraction pits (red circles), and the position of the passive "sampling
racks." Horizontal divisions are walls between "north" (top in figure), "central," and
"south" (bottom in figure) sections of basement with open walkways between
(cistern is in the central basement).
3-14
-------
Section 3—Methods
Cinderblock
Section of
Brick Wall
Basement
420 East 28th St
In
Wall
Port 4
Mitigation Leg
SSP 6
Sampling
Rack
Unknown
Remnant
Open \
Walkway Old Drain?
-I 1
SGP 11
Cistern
y ssp 3
SSP 7 Up to
Mitigation Leg
1st Floor
H h
+
Open Open
Walkway Walkway
Ssnpiing
Rack
O SSP 5
o
SGP 12
r
Figure 3-11. Map view of the 420-side basement showing the SSD mitigation system legs,
subslah soil gas extraction pits (red circles), and the passive "sampling racks."
Horizontal divisions are walls between "north" (top in figure), "central," and "south"
(bottom in figure) sections of basement with open walkways between (cistern is in
the central basement).
3-15
-------
Section 3—Methods
Figure 3-12. Photos of mitigation system: (left) SSD blower and stack on northeast corner of
duplex; (right) SSD extraction point, showing valve and U-tube manometer.
3-16
-------
Section 3—Methods
Basic Layout of the Mitigation
System--Cross-section View
Connection
between 422/420
Leg Open/Close
Figure 3-13. Cross-section showing the general layout of the 422/420 north and central
basements with the positioning of the extraction legs, exterior blower, and exhaust
stack.
3.4 Monitoring, Sampling, and Analysis
Section 3 in the previous project report (U.S. EPA, 2012a) provided details on the design and installation
of the monitoring infrastructure used in this project including wells, soil gas monitoring ports, soil
temperature and moisture sensors, differential pressure sensors, weather data, and indoor and outdoor
monitoring for VOCs and radon. This report updates the previous report, with this section summarizing
and updating the previous report's text on the VOC sampling configurations used for soil gas, air, and
groundwater monitoring. Section 3.5 describes radon sampling, and Section 3.6 describes monitoring of
physical parameters like weather, indoor temperature, and differential pressures. Figure 3-14 maps the
subsurface monitoring points including soil gas sampling ports and groundwater monitoring wells.
3.4.1 Indoor and Outdoor Air VOC Monitoring
The overwhelming majority of the indoor passive sampling was done with Radiello 130s supplied by and
analyzed by Air Toxics Ltd. For comparison, two different types of SKC badges were also used that were
specifically adapted to use at very short or long sampling durations.
3-17
-------
Section 3—Methods
| Storage
Closet
CD
-i
O
7?
S
Covered
Porch
>
Q.
0>*
O
-------
Section 3—Methods
For passive sampling, several racks were set up to facilitate arranging groups of samplers in consistent
locations for different durations during the run of the project. These racks were ordinary laundry drying
racks that can be purchased inexpensively at most department stores (Figure 3-15). The racks were ideal
in that they allowed multiple samplers to be placed at the same, or similar, levels within the normal
breathing zone. One rack was placed in the first floor center room of 420 and 422 and in the northern and
southern areas of each basement.
At each rack, a specific location was assigned for one of several day durations each approximately 6 in.
apart to minimize the potential for starvation effects. Durations of 7, 14, 28, 91, 182, and 364 days were
used from January 2011 to March 2012 to study the performance of varying durations of passive samplers
as well as temporal variability. From October 2012 through May 2013, the primary emphasis of the
passive sampling was to study the effects of the SSD mitigation system and snow/ice effects on VOC and
radon levels using weekly and quarterly sampling durations. Enough spaces on the rack remained for
duplicates of those durations, plus special locations occupied during intensive rounds. SKC badges were
primarily hung on the back portion of the racks, in a similar manner to the Radiellos.
In addition to these indoor racks, a special outdoor (ambient) location had to be made to accommodate the
samplers. A hood was purchased to house the samplers and mounted on a telephone pole by the alley near
the house (Figure 3-16). This hood housed all of the Radiellos and badges for the different day durations.
Sampling of Radiellos consisted of removing the white diffusive body from its backing shield, opening
the glass vial that contained the new screened Radiello 130, and allowing it to slide into the white body;
then the white body was replaced in its backing plate with a new sample number. The old one was then
sealed in a glass vial for shipping. Each week, Radiellos of the appropriate durations were stopped and
replacements were started. For example, when the 7- and 14-day Radiellos were stopped, new ones were
put up in their places. The 7-day samples were taken down the following week, followed by the 14-day
samples the week after. This arrangement allowed us to compare the results of different time durations to
each other (ex. four weekly samples against the monthly for the same time period). Additionally, during
some of the intensive rounds, daily Radiellos were taken to compare them to the weekly time increments.
SKC 575 badges with the secondary diffusion cover were used for comparing the longest Radiello
durations (the 182- and 364-day time periods). These solvent-extracted charcoal badges have been used in
the literature for durations of 4 weeks and longer. SKC Ultra Badges (thermally desorbed) were used for
24-hour and 7-day sampling during an intensive round and short-term sampling during a fan test. Both
Radiellos and SKC badges were provided by and returned to Air Toxics Ltd. for analysis.
Summa canisters (6 liter) were only used for preliminary site screening and indoor air before and after the
fan testing (Method TO-15). These were acquired from and returned to Air Toxics Ltd. for analysis.
3-19
-------
Section 3—Methods
Figure 3-15. Passive indoor air sampling rack: 422 first floor.
3-20
-------
Section 3—Methods
Figure 3-16. Ambient sampler shelters on telephone pole near duplex.
3.4.2 Subslab and Soil Gas (TO-17)
The primary method of subslab and soil gas sampling for VOCs dunng the second phase of the project
was by TO-17. In this method, a thermo-desorption tube, with a female Swagelok end, was connected to
each sampling port in turn. Each port had its own male union connected to a valve. Before sampling, the
port was purged with an SKC Universal XR pump set to lL/min. Five well volumes were then purged via
an exhaust line that ran away from the operator for exterior ports or out of a basement window in the case
of the interior ports. The fittings were attached with wrenches, and an air tight syringe was mounted onto
the other end of the TO-17 tube. Once this was done, the port's valve was opened, and the syringe was
used to draw 200 mL of air through the TO-17 tube over a period of a mmute. After this, the port valve
was closed, and the TO-17 tube was removed and sealed for shipping.
Samples were taken from the operational ports positioned at three (interior probes) or four (exterior
probes) depths each week from January 2011 through February 2012. Sampling was performed at least
monthly from October 2012 through May 2013. During this time period, sampling was also performed
3-21
-------
Section 3—Methods
more frequently at time points selected to observe the effects of snow/ice events and flooding events.
Initially, the preferred depths to sample were 3.5, 9, and 16.5 ft bis exterior and 6, 9, and 16.5 ft bis
interior. However, a higher than expected water table prevented the sampling of the 16.5 ft depths for
most of the duration of the project. Unusually high water tables or perched/infiltrating water occasionally
made other soil gas ports inoperative. In addition, all wall ports were sampled during most sampling
rounds, as well as a subset of the subslab ports. Details of the subslab, wall port, and soil gas probe
locations and construction can be found in U.S. EPA (2012a).
The majority of the TO-17 tubes collected were prepared by and analyzed by the EPA National Exposure
Research Laboratory (NERL) in Las Vegas, NV. For the extensive sampling of the intensive rounds
conducted in 2011 to 2012, additional TO-17 tubes were prepared by and analyzed by Air Toxics. An
intercomparison study of the two TO-17 laboratories was conducted in the previous project and showed
acceptable agreement between the two laboratories (see Section 4.2.4 of U.S. EPA [2012a]). During the
intensive rounds, all functioning ports (not made inoperative by water) were sampled at least once each
day of the round. For a few days of each round, several locations were sampled multiple times of the day
with the intention of comparing hourly and daily variability to the normal weekly variability.
3.4.3 Online Gas Chromatograph
An automated sampling and analysis system was provided by Hartman Environmental Geoscience for
two periods during the previous project and from December until early March for this project, and system
design and deployment are described here for all three sampling periods. The system consisted of the
following elements:
¦ gas chromatograph (GC) with an electron capture detector (ECD),
¦ 16-port stream selection valve,
¦ sample injection valve with an adsorbent trap or 1 cc sample loop,
¦ computerized data acquisition system (Peaksimple by SRI Instruments), and
¦ remote connection via wireless.
The GC was connected by gas tight tubing from selected sample points for first floor indoor air, soil gas,
subslab soil gas, and ambient air samples. The tubing from each sample location was connected to the
stream selector valve. At any time, one of the entering tubes was connected to the adsorbent trap or
sample loop depending on the position of the stream selector valve. A low-flow vacuum pump drew the
vapor sample through the tubing at a rate of 25 cc/min to 40 cc/minute for 30 to 90 seconds to purge the
sample tubing and ensure the sample in the sample loop was from the selected sample location. When
purging was complete, the sample injection valve would rotate and inject the sample into the GC for
analysis. Cycle time from start of purging to the end of the analysis was approximately 10 to 15 minutes.
When the analysis was complete, the stream selector valve advanced to the next position (next sample
location) and the process repeated itself. This sequence continued uninterrupted until stopped by the
operator. Approximately seven (7) to nine (9) samples from each sample location were analyzed each
day.
The data acquisition software (Peaksimple) acquired the chromatographic data and also controlled the
stream-selector valve, sample injection, and GC analysis and stored the data to a summary file on a
laptop. Remote access to the laptop and the data was enabled by a WiFi connection installed at the house
for this purpose.
The 16 sampling ports were distributed as follows: one was initially connected to a nitrogen tank but later
was connected to a line to outdoor air (~4 ft from the house), one was connected to a TCE standard
3-22
-------
Section 3—Methods
periodically, and two were blanks used to clear the instrument after each run. There were 12 sample
locations: four indoor air, three subslab ports, one wall port, and three house-interior and one exterior soil
gas probes with soil gas ports at multiple depths (3 ft, 6 ft, 9 ft, and 16.5 ft bis) in the subsurface. The 6 ft
bis soil gas probes corresponded to the subslab probes in terms of depth.
All sampling lines were constructed of 1/16 in. OD stainless steel tubing (except the 420 first floor line
had approximately a 20 ft section of 1/8 in. OD stainless steel tubing at the sampling end). The tubing for
all lines ran from the stream selector valve at the GC along interior walls to the sampling points. At the
sample locations, the indoor air lines hung suspended over passive sampler racks, within the breathing
zone. For soil gas ports and subslab ports, each tube was connected to a sampling port by means of
Swage-Lok male/female fittings.
In the first phase of the automated program (August 2011 to October 2011), the vapor sample from each
location was concentrated onto an adsorbent trap. Volumes passed over the trap were adjusted depending
on the vapor concentration at each location and ranged from 20 cc to 80 cc. Higher volumes were
collected on the trap for lower concentration locations such as indoor air. Lower volumes were used for
soil gas. Cycle time from start of purging to the end of the analysis was approximately 10 minutes.
Approximately nine (9) samples from each sample location were analyzed each day.
In the second phase of the program (December 2011 to February 2012), the adsorbent trap was eliminated
and the sample was passed through a 1 cc sample lop for direct injection into the GC. This modification
was made to minimize carry-over between the high-concentration soil gas samples and the low-
concentration indoor air samples and to speed up the analysis. Cycle time from start of purging to the end
of the analysis was approximately 10 minutes. Approximately nine (9) samples from each sample location
were analyzed each day.
In the third phase of the program (December 2012 to March 2013), the adsorbent trap was eliminated and
the sample was passed through a 1 cc sample loop for direct injection into the GC. The analysis was also
slowed down to enable lower detection limits for chloroform. Cycle time from start of purging to the end
of the analysis was approximately 15 minutes. Approximately seven (7) samples from each sample
location were analyzed each day.
3.4.4 Groundwater
Groundwater samples were taken approximately monthly with permeable diffusion bags (PDBs) from
EON Products Inc. However, because of difficulties sampling the indoor 1-inch well (MW-3) by PDB,
samples were taken by bailers from February 6, 2013, onward. The 422/420 duplex has six exterior MWs
(two clusters of three) and one single-depth interior well installed in the basement and completed on the
first floor (Figure 3-17). The exterior wells are arranged in groups of three in the front and the back
yards. Each group of three is divided into depths of 16 to 21 ft, 21 to 24 ft, and 24 to 26 ft bis. The interior
well (MW-3) is about 18 ft bis, but the casing extends up to the first floor for ease of access, so it is about
24 ft deep at its access point. The exterior wells are 2 inches in diameter, and the internal well is 1 inch in
diameter. PDBs for the exterior wells are 12 by 1.75 inches, and the interior is 18 by 0.75 inches. PDBs
were deployed for at least 2 weeks, and a new set of PDBs was cycled through almost monthly. PDBs
were filled initially with deionized water provided by the EPA NERL laboratory. Most groundwater
samples were shipped to EPA NERL-Las Vegas for VOC analysis by Method 8260. A few samples were
analyzed by Pace laboratories in Indianapolis as a quality control check.
Groundwater samples were also collected from soil gas points when they were temporarily flooded using
a peristaltic pump. Peristaltic pump samples were also collected from the monitoring wells for
comparison purposes.
3-23
-------
Section 3—Methods
I
UUI
\
Figure 3-17. Monitoring well MW-3, installed in the basement and completed on the first floor.
3.4.5 Subslab Depressurization System Stack Gas Sampling
Passive sampling from the stack for VOCs was done with a passive sampler, the Waterloo Membrane
Sampler. The Waterloo Membrane Sampler is preferred for sampling from the SSD stack because its
design makes it resistant to changes in air velocity and because of its small size. The sampler and its
validation are described at http://www.sirem-lab.com/images/PDF/wms.pdf (last accessed 10/9/2012).
Radon readings in the stack were taken using the portable AlphaGUARD instrument used for soil gas.
Stack velocity readings were taken with a Shortridge AND-870C Multimeter with the Airfoil Velocity
measurement probe.
3.5 Radon Sampling and Analysis
3.5.1 Indoor Air Radon Sampling and Analysis
The primary radon sampling method was electrets ion chambers collecting radon samples passively in
indoor air for the same 7-day intervals as Radiellos-collected VOCs. The following secondary methods
were, however, also used for radon in indoor air:
¦ stationary AlphaGUARDs at two locations to provide greater time resolution,
¦ carbon absorbers for a QC comparison, and
¦ consumer grade ionization chamber-based detector (Safety Siren Pro Series 3 manufactured by
Family Safety Products Inc.) for comparison.
Each method is described in detail below.
We used Rad Elec, E-Perm, ST-type (short-term) electrets according to EPA 402-R-92-004 (U.S. EPA,
1992). These were primarily deployed in s-chambers, but h-chambers were used on a few occasions. To
3-24
-------
Section 3—Methods
sample, electrets were opened within their chambers at their assigned locations for a week. After a week,
the chambers were closed, all electrets were allowed to equilibrate for an hour to the room temperature
where they would be read, and then their voltages were read on a Rad Elec electret voltage reader. Start
and stop times, as well as voltages, were recorded and the electrets redeployed. The voltages,
configurations (e.g., ST electrets in s-chambers), dates, and times would then be incorporated into a
calculation used to convert voltage to picoCuries per liter (pCi/L) (pCi/L), with background gamma
correction.
The electrets reader was calibrated weekly with three standards. In addition, an electret blank test was run
weekly to test for effects of the chamber on the electrets. In this test, an electret not used during the
sampling was inserted into one of the used electret chambers (closed) and then read to determine whether
there had been any voltage drop from the previous week's reading.
Electrets were hung in mesh bags, one at each of the same locations used for sampling Radiellos, plus a
duplicate at one location (three locations on the 422 side of the house and three on the 420 side). The
ambient electret was kept in a permeable bag and hung from a wooden dowel about 2 ft from the house.
Since December 28, 2011, a new electret was added in the 422 second floor office to be used in
conjunction with the radon siren testing.
Charcoal canisters from the U.S. EPA Radiation and Indoor Environments (R&IE) National Laboratory in
Las Vegas, NV, were set out on the sampling racks on three separate occasions to check the accuracy of
the electret readings (U.S. EPA, 1990). They were simply opened for a week (matching an electret
sampling period), closed, and shipped back to EPA for testing. Section 3.5.3 discusses the stationary
AlphaGUARDs that were also used on the project for indoor air radon measurement.
Consumer-grade radon detector (Safety Siren) testing was a later addition to the project. Six Pro Series 3
Safety Siren radon gas detectors were deployed on December 23, 2011, and in use until March 1, 2012.
They were tested again from October 2012 to May 2013 during a period of mitigation on/off testing. Each
was installed at one of six locations: 422 second floor office, 422 first floor center room, 422 basement
south, 422 basement north, 420 first floor center room (stolen October 11, 2012, not replaced), and 420
basement south. The intention of the test was to determine the agreement among the radon Safety Sirens,
electrets, stationary AlphaGUARDs, and (for 1 week) charcoal canisters. The Safety Sirens can be read
once each week, so their readings were taken when the other data types were being acquired and their
readings compared.
3.5.2 Subslab and Soil Gas Radon Sampling and Analysis
Radon readings were collected weekly in 2011-2012 and approximately monthly or as meteorological
conditions required in 2012-2013 with a portable AlphaGUARD Professional Radon Monitor from
Genitron instruments. Operations were based on EPA guidelines for using continuous radon monitors
(U.S. EPA, 1992). More information on the AlphaGUARD can be found at
www.genitron.de/products/products.html. During routine sampling, this device was connected to subslab,
soil gas, and wall ports with an SKC Universal XR pump set to 1 L/min. Tubes connected the sample port
to the pump (with a moisture filter on the sampling end) and the pump to the AlphaGUARD. A purge line
led away from the operator for exterior sampling and out of basement windows for interior sampling
locations. The AlphaGUARD requires a 10-minute cycle of uninterrupted air flow from the sample
location for an accurate reading. Because a certain amount of time was needed for movement between,
one 10-minute cycle was spent relocating and then another to sample at the next location. Thus, each
sample port needed 20 minutes to sample.
Because radon has a short half-life (3.8 days) and the migration time from substantial depths for soil gas
is estimated to be months to years (Kurtz and Folkes, 2008; Carr et al., 2011), radon sampling focused on
3-25
-------
Section 3—Methods
the shallowest depths and, thus, differed from the VOC sampling strategy. Exterior sampling consisted of
the shallowest ports available of the wells closest to the house. Usually, these were the 3.5- and 6-ft deep
ports of SGPs 1, 7, 4, and 5. Periodically, these depths would not yield a sample, presumably because of
moisture infiltration. In such cases, the next shallowest depths were chosen. Routine interior sampling
included all wall ports, five of the subslab ports, and the shallowest intervals of the nested interior soil gas
ports.
For routine sampling, first an ambient reading was taken outdoors and -20 ft away from the 422/420
house. Then, lines to be sampled would be purged with the SKC pump (five soil gas point volumes,
calculated based on the depth). Finally, the pump would be connected to the AlphaGUARD to acquire a
full 10-minute sample.
The AlphaGUARD has a readout screen that details the results of the analysis at the end of each
10-minute cycle. The data provided are radon concentration (Bq/m3), relative humidity (%), pressure
(mbar), and temperature (° C). These data were recorded each week in a spreadsheet and the Bq/m3
converted to pCi/L.
3.5.3 Continuous (Real-Time) Indoor Air Radon Sampling and Analysis
The real-time AlphaGUARDs are essentially the same as the handheld AlphaGUARD instrument used to
sample from the soil gas ports, except they are not fitted with the same nozzle type, because they are not
connected to external pumps. Rather, in this application they are operated in a diffusion mode. These
AlphaGUARDs are intended to be placed to give readings in specific rooms. In the case of the 422/420
duplex, one unit was placed in the 422 second floor office, and the other was placed in the 422 north
basement area. These units stayed in their locations, except for brief, periodic data downloadings. These
units were first regularly deployed on March 31, 2011, and were in near-continuous operation until May
2013 except for a period of off-site recalibration in late July through late September 2012.
The data are produced by the instrument in the same units as the portable AlphaGUARD (requiring
conversion to pCi/L), and data points are collected every 10 minutes. However, because these devices
were not moved, all 10-minute cycles are usable. The real-time AlphaGUARDs are used in conjunction
with Data Expert software, also from Genitron Instruments. Once each week, the AlphaGUARDs were
connected to the computer (the one in the basement required briefly moving the instrument to download),
and the software downloaded the readings for the week. These were then saved as text files for later
conversion to Excel spreadsheet files.
3.6 Physical Parameters Monitoring
3.6.1 On-Site Weather Station
This project used a Davis Vantage Vue Weather Station on site with Weather Link data logger and
software (Figure 3-18). The components consist of the outdoor monitoring unit, the indoor receiver, and
the computer connection. The outdoor monitoring unit was mounted on an accessible portion of the
422/420 house roof. The unit was mounted on steel pipes, but 5 ft above the highest roof deck (that of the
attic dormer).
The outdoor unit contains all the exterior monitoring equipment (e.g., wind speed cups, rain gauge) and
has a solar panel/battery backup for power. The outdoor unit transmits a radio signal to the indoor
receiver, which also records the data every half hour. The indoor unit is human readable and can also be
used to set a variety of parameters. The indoor unit also records the house interior data at its location, in
this case the 422 second floor office. Once each week, the data were downloaded from the indoor unit
onto the computer containing the Weather Link software. These data were saved as a text file and later
3-26
-------
Section 3—Methods
compiled in an Excel spreadsheet file. Many parameters are recorded; the key ones required for this
project are temperature (degrees F, interior and exterior), relative humidity (%), wind speed (mph), and
wind direction (16 points [22.5°] on compass rose).
Initially, and at least every 6 months, the results from this on-site system were compared with other
nearby weather stations in Indianapolis using at least 1 day's observations. The National Weather Service
(NWS) Indianapolis International Airport (KIND) is approximately 15 miles southwest from the site. The
Indianapolis NWS station at Eagle Creek Airpark (KEYE) is approximately 9 miles west of the site.
There is also a private weather station available online closer to the site in Indianapolis, IN
(KININDIA33).
Figure 3-18. Front view of 420/422 duplex with location of weather station sensors indicated with
red arrow.
During the autumn months of 2012, it was discovered that the weather station stopped recording data for
bnef periods each day, usually for approximately 2 hours in the early morning. It was detennined in the
winter of 2012 that the house exterior station needed its battery changed. Because the station's height
made it inaccessible under ordinary conditions, a subcontractor was hired. As soon as the winter ice
3-27
-------
Section 3—Methods
conditions allowed (January 15, 2013), Ping's Tree Service was hired to access the external station's
battery with a bucket truck. After the battery was changed, no further interruptions in data occurred.
3.6.2 Indoor Temperature
Although the indoor weather station unit can record temperature, it only does this in the 422 second floor
office where it is located. Because temperature readings were required at all sample locations to allow
adjustment of the passive sampler data for uptake rate variation due to temperature, another form of data
collection was necessary. HOBOs data loggers, made by Onset (http://www.onsetcomp.com/), were
placed—one at each of the six passive sampler racks in the house. HOBOs record temperature (degrees F)
and relative humidity (%) every 30 minutes. Once a week, these data were recorded by taking them to the
computer with the Hoboware reading software and later importing those data to an Excel spreadsheet file.
Special spreadsheets were created to provide this information for the different Radiello time durations to
the passive sampler analytical laboratory.
3.6.3 Soil Temperature
Soil temperature was recorded by thermocouples from Omega (Type T, hermetically sealed tip insulated
thermocouples, HSTC-TT-T-24S-120). During the initial house set up, holes were drilled beneath the 422
basement slab and backyard soils of the duplex (see Figure 3-14) to accommodate thermocouple probes
with end points set at different depths. Wires were inserted in ~2-in diameter holes with weights loosely
attached near the ends. The holes were allowed to cave in and backfill naturally. The thermocouple wires
run from their holes to male/female connectors (sealed from the elements in rubber "boots") and from
there to a data acquisition system (PDAQ 56 by IOtech), where the data were recorded to the software on
the computer. A reading was taken approximately every 15 minutes. The thermocouples wired to the
PDAQ roughly corresponded to the depths of the soil gas ports: inside at 6, 9, 13, and 16.5 ft bis; outside
at 1, 3.5, 6, and 13 ft bis. However, there is one thermocouple (outside 16.5 ft) that is wired into an
Omega data logger (OM-EL-USB-TC). The thermocouple data were most typically collected at 15-
minute intervals.
3.6.4 Soil Moisture
Soil moisture was recorded by implanted Watermark moisture sensors. The units of measurement for the
soil moisture sensors are explained by Smaj stria and Harrison (2002):
Water potential is commonly measured in units of bars (and centibars in the English
system of measurement) or kilopascals (in metric units). One bar is approximately equal
to one atmosphere (14.7 lb/in 2) ofpressure. One centibar is equal to one kilopascal.
Because water is held by capillary forces within unsaturated soil pore spaces, its water
potential is negative, indicating that the water is under tension and that work must be
done to extract water from the soil. A water potential reading of 0 indicates that the soil
is saturated, and plant roots may suffer from lack of oxygen. As the soil dries, water
becomes less available and the water potential becomes more negative. The negative sign
is usually omitted for convenience when soil water potentials are measured.
The soil water matrix potential can be converted into volumetric water content using known equations.
Moisture content is often measured in fixed laboratories as gravimetric water content. To convert
gravimetric water content to volumetric water, multiply the gravimetric water content by the bulk specific
gravity of the material.
These sensors were also installed in the holes drilled during the house set up. Before insertion, the sensors
had to be presoaked in water to prepare them. The sensors are pill-shaped devices at the end of a wire.
The wire was run up through a PVC pipe of the appropriate length for the depth and the wire grasped
3-28
-------
Section 3—Methods
manually. The sensor could then be placed to the appropriate depth within the hole, the PVC pipe
withdrawn, and the soil backfill allowed to fill in naturally. Wires extend to the Watermark 900M
monitor, which reads and records the data every 30 minutes. Once each week these data were downloaded
to the Watergraph 3.1 software on the computer. Data were recorded in centibars. Soil moisture probes
were installed near the soil temperature probes in the 422 basement and backyard (see Figure 3-14). The
sensors were installed to approximately correspond to the soil gas port depths: inside at 6, 13, and 16.5 ft
bis and outside at 3.5, 6, 9, 13, and 16.5 ft bis.
3.6.5 Potentiometric Surface/Water Levels
Water levels in the seven wells (three clusters) on site were taken periodically with a Solinst water-level
meter. The water-level results were compared against U.S. Geological Survey (USGS) stream gauge data
for Fall Creek at Millersville, site 03352500 near the house.
For this new phase of the project, a Solinst water Levelogger Model 3001 was used to obtain higher time
resolution, starting November 9, 2012; data are taken each half hour by the device. The device was
installed in the deepest well (1A, -26 ft) of the south yard monitoring well cluster (MW1). Installation
made use of the existing tether system originally used with PDB sampling. Approximately each month,
the water logger is retrieved and connected to a computer via a USB port. Using the Levelogger Series 4
Software, the data are downloaded from the logger and entered into a spreadsheet. The spreadsheet
contains formulas for converting the recorded height of the water column to one corrected for outdoor
pressure (using outdoor pressures from the weather station). Each month, depth to water is also manually
measured using a Solinst water level indicator for comparison to the data logger. Readings differ by
approximately 0.3 ft. The logger is then restarted and redeployed on its tether back in MW1-A.
3.6.6 Differential Pressure
Differential pressure readings were monitored by Setra Model 264 low differential pressure transducer.
These units contain a pressure-sensitive diaphragm that measures pressure changes from the exterior
high/low poles. The poles had tubing connected that ran from the areas to be measured. Some Setra poles
were left open as an interior reference at a particular location. The configurations on the 422 side were as
follows: subslab versus basement, basement versus upstairs, deep soil gas versus shallow soil gas, and
basement versus exterior (out of the basement window). Only one unit was located on the 420 side, and it
was used for subslab versus basement. Three lines used soil gas ports as access points: 422 deep soil gas
versus shallow soil gas used SGP8-6 and SGP8-13, 422 subslab versus basement used SSP-1, and 420
subslab versus basement used SGP11-9. When these locations had to be sampled for VOCs, the ports
would be closed, disconnected from the Setras, purged, and sampled. Afterward, the ports would be
reconnected to the Setras and opened again.
The four Setras on the 422 side of the house are wired into the Personal Measurement Device, PMD-
1208LS from Measurement Computing. The PMD is connected to the computer and uses TracerDaq
software. Readings are taken every 15 minutes. The one Setra on the 420 side is connected to the PDAQ
device and also takes a reading every 15 minutes (but not necessarily the same 15-minute interval as the
PMD Setras).
In the beginning of the project, the Setras were laid flat on their supporting surfaces. In February 2011,
manufacturer's guidance was found indicating that they should be mounted vertically. The manufacturer
stated that correcting for the different mounting could be done by blocking the poles in the horizontal
position to determine their "zero readings" and then recording those same readings in the vertical position
to determine the offset. The offset could then be factored in to change the horizontal position data to
vertical. By March 31, 2011, all were hung in this manner, and the early data corrected.
3-29
-------
Section 3—Methods
3.6.7 Air Exchange Rate
To determine the air exchange rate, capillary adsorption tubes (CATs) were used in conjunction with
perfluorodimethylcyclohexane (PDCH) and perfluoromethylcyclohexane (PMCH) emitters, provided by
the Harvard School of Public Health (HSPH) (EPA Method IP-4). The emitters are small metal shells
containing a fluid (either PDCH or PMCH), and the shells are contained within a foam wrapping. The
fluid releases a tracer gas at a measured constant rate, which is picked up by the CATs when in place. One
stopper end of the CAT is removed when the samplers were deployed for periods of 1 week to allow
sampling of the tracer gas by the adsorbent medium.
On April 22, 2011, in the 422 side of the house, 10 of the PDCH emitters were placed in the basement, 10
PMCH emitters were placed on the first floor, and 9 PMCH emitters were placed on the second floor.
Care was taken that emitters be placed far enough from each other and from walls (about 3 to 4 ft). The
placement locations also allowed unrestricted air flow.
CATs were used for sampling for air exchange rate measurement on four occasions. The first was from
April 27, 2011, to May 4, 2011; the second was from September 23, 2011, to September 29, 2011; the
third was from October 13, 2011, to October 14, 2011 and from October 18, 2011, to October 19, 2011;
and the fourth was from April 2, 2013, to April 9, 2013. On the first occasion, CATs were deployed—one
on the 422 first floor (center room) and two in the 422 basement (one duplicate). One was also placed in
420 on the first floor (center room) and in the 420 basement (center room). On the second occasion, CATs
were only deployed on the 422 side of the house. One was in the 422 office on the second floor, one on
the first floor (center room), and two were placed in the basement center room (one duplicate). On the
fourth occasion, CATs were deployed on the 422 second floor office, the 422 first floor central room, the
422 basement central room (and a duplicate there), and the 420 first floor central room. When sampling,
CATs were placed on their sides with one cap removed and slightly tipped at one end so the open end
pointed toward the ground. After sampling, the CATs were sealed and sent to HSPH for analysis.
3.6.8 Crack Monitoring
The basement floors and walls were visually inspected for significant cracks (i.e., ones where vapors
could infiltrate from subsurface soils). For the three most significant cracks, we installed a calibrated
crack monitor as shown in the Figure 3-19. This device consists of two plates that move independently.
One plate is white with a black millimeter grid; the other is transparent with red crosshairs centered over
the grid. Once the monitor is secured with epoxy or screws across a crack, the crosshairs shift vertically or
horizontally on the grid, making crack movement easily visible and trackable. It was installed with a 5-
Minute® Epoxy, a rapid-curing, general-purpose adhesive that bonds rigid, durable substrates such as
metals, glass, ceramics, concrete, and wood in all combinations. The position of the monitor was recorded
monthly and indicated that the monitored cracks did not move during the course of the study.
BEN MEADOWS Calibrated Crack Monitor
Figure 3-19. Calibrated crack monitor.
3-30
-------
Section 3—Methods
3.7 Data Aggregation Methods
In order to conduct statistical time series analysis in Sections 9 and 10, data had to be arranged into files
that contained one value for each predictor (independent) variable for each value of an outcome
(dependent) variable. Because of methodological constraints, not all data sets were acquired with exactly
the same time intervals. Therefore data were aggregated at the level of individual days and weeks for data
analysis. Professional judgment was used to determine the most appropriate method of aggregation for a
given parameter (e.g., mean, sum, mode, maximum); in most cases, the mean or mode was used as a
central tendency estimate to avoid any bias in the aggregated variable. The methods of aggregation for
each variable are provided in Table 3-2.
Table 3-2. Data Aggregation Applied to Predictor Variables
Method of
Aggregation
Variable Name (Plain Language)
Variable Code
Building Variables
420 air conditioning status (on/on briefly/off)
AC_on-off_420_daily
Mode
422 air conditioning status (on/on briefly/off)
AC_on-off_422_daily
Mode
422 fan status (on/off) (Note: fan was never used on 420)
Fan_on-off_422_daily
Mode
420 side heating status (on/off)
Heat_on-off_420_daily
Mode
422 heating status (on/off)
Heat_on-off_422_daily
Mode
House Mitigation Status (not yet installed/on/passive/off)
Mitigation_Status_Daily
Mode
Building Environment Variables
Air density interior
AirDens_422
Mean
Dew point, interior, Fahrenheit
Dew_pt_422_F
Mean
Humidity interior
Hum_422_%.
Mean
Interior heating Index
lndoor_Heat_lndex
Mean
420, subslab vs. basement differential pressure
Setra_420ss.base_Pa
Mean
422 basement vs. exterior differential pressure, Pascals
Setra_422base.out_Pa
Mean
422, basement vs. upstairs differential pressure, Pascals
Setra_422base.upst_Pa
Mean
422, deep vs. shallow soil gas differential pressure, Pascals
Setra_422SGdp.ss_Pa
Mean
422, subslab vs. basement differential pressure, Pascals
Setra_422ss. base_Pa
Mean
Temperature at 420 basement north sampling location from
HOBO
T_420baseN_C
Mean
Temperature at 420 basement south sampling location from
HOBO
T_420baseS_C
Mean
Temperature at 420 first floor sampling location from HOBO
T_420first_C
Mean
Temperature, 422 first floor from weather station
T_422_F
Mean
Temperature 422 basement north from HOBO
T_422baseN_C
Mean
Temperature 422 first floor from HOBO
T_422baseS_C
Mean
Temperature on first floor of 422 of duplex from HOBO
T_422first_C
Mean
Subsurface and Stream Variables
Height Measured at Fall Creek Stream Gauge in feet
Fall_Crk_Gage_ht_ft
Mean
Soil moisture, 13 ft bis beneath structure, cbar
Soil_H20_ln13._cbar
Mean
Soil moisture 16.5 ft bis beneath structure, cbar
Soil_H20_ln16.5._cbar
Mean
Soil moisture 6 ft bis beneath structure, cbar
Soil_H20_ln6._cbar
Mean
Soil moisture 13 ft bis exterior, cbar
Soil_H20_Out13._cbar
Mean
(continued)
3-31
-------
Section 3—Methods
Table 3-2. Data Aggregation Applied to Predictor Variables (cont.)
Variable Name (Plain Language)
Variable Code
Method of
Aggregation
Soil moisture, 3.5 ft bis exterior, cbar
Soil_H20_Out3.5._cbar
Mean
Soil moisture 6 ft bis exterior, cbar
Soil_H20_Out6._cbar
Mean
Soil temperature 13 ft bis beneath structure
Soil_T_C_MW3.13
Mean
Soil temperature 16.4 ft bis beneath structure
So i l_T_C_M W3.16.5
Mean
Soil temperature 6 ft bis beneath structure
Soil_T_C_MW3.6
Mean
Soil temperature 9 ft bis beneath structure
Soil_T_C_MW3.9
Mean
Soil temperature 1 ft bis exterior
Soil_T_C_OTC.1
Mean
Soil temperature 13 ft bis exterior
S o i l_T_C_OT C. 13
Mean
Soil temperature 16.5 ft bis exterior
Soil_T_C_OTC.16.5
Mean
Soil temperature 6 ft bis exterior
Soil_T_C_OTC.6
Mean
Weather Variables
Barometric pressure rate of change in inches of mercury per
hour
Bar_drop_.Hg.hr
Mean
Barometric pressure in inches of mercury
Bar_in_Hg
Mean
Net barometric pressure change over measurement period in
inches of mercury
BP_Net_Change
First-Last, by
date/time
Standard deviation of barometric pressure change over
measurement period in inches of mercury
BP_Pump_Speed
Standard Deviation
Largest barometric pressure change over measurement period
("stroke length" of barometric pumping) in inches of mercury
B P_Stro ke_Le n gt h
Maximum-Minimum
Cooling degree days
Cool_Degree_Day
Sum
Dew point, exterior
Dew_pt_out_F
Mean
Heating degree days
Heat_Degree_Day
Sum
Exterior Heating Index - calculated based on temperature and
humidity
Heat_lndex_F
Mean
Humidity exterior
Hum_out_%.
Mean
Rain (inches) totaled during observation period
Rain_ln_met
Sum
Rain highest rate during observation period in inches/hour
RainJPH
Maximum
Depth of snow on the ground, inches
Snowdepth_daily
Mean
Temperature exterior from HOBO
T_out_C
Mean
Exterior temperature from weather station (°F)
T_out_F
Mean
Temperature exterior, high during data collection period
T_out_Hi_F
Maximum
Lowest exterior temperature in Fahrenheit
T_out_Lo_F
Minimum
Temperature, humidity, and wind index
THW_F
Mean
Wind chill
Wind_Chill_F
Mean
Average wind direction in degrees
Wind_Dir
Trigonometric Mean
Wind direction of high speed during measurement period in
Degrees
Wind_Dir_Hi
Direction paired to
high speed
Wind run is a function of wind speed and duration
Wind_Run_mi
Sum
High wind speed during measurement period
Wind_Speed_Hi_MPH
Maximum
Average wind speed during measurement period
Wind_Speed_MPH
Mean
(continued)
3-32
-------
Section 3—Methods
Table 3-2. Data Aggregation Applied to Predictor Variables (cont.)
Variable Name (Plain Language)
Variable Code
Method of
Aggregation
Chemical Concentratio
n Measurements
Chloroform concentration at 420 basement north sampling
location, in |jg/m3, as measured by Radiello sample
420BaseN Radiello Weekly
CHCI3
Randomly choose
(when more than one
per week)
Chloroform concentration at 420 basement south sampling
location, in |jg/m3, as measured by Radiello sample
420BaseS Radiello Weekly
CHCI3
Randomly choose
(when more than one
per week)
Chloroform concentration at 422 basement north sampling
location, in |jg/m3, as measured by Radiello sample
422BaseN Radiello Weekly
CHCI3
Randomly choose
(when more than one
per week)
Chloroform concentration at 422 basement south sampling
location, in |jg/m3, as measured by Radiello sample
422BaseS Radiello Weekly
CHCI3
Randomly choose
(when more than one
per week)
Chloroform concentration at 422 first floor sampling location,
in |jg/m3, as measured by Radiello sample
420First_Radiello_Weekly_
CHCI3
Randomly choose
(when more than one
per week)
Chloroform concentration at 422 first floor sampling location,
in |jg/m3, as measured by Radiello sample
422First_Radiello_Weekly_
CHCI3
Randomly choose
(when more than one
per week)
Chloroform concentration at outside sampling location, in
|jg/m3, as measured by Radiello sample
Out_Radiello_Weekly_CHCI3
Randomly choose
(when more than one
per week)
Tetrachloroethene concentration at 420 basement north
sampling location, in |jg/m3, as measured by Radiello
sample
420BaseN Radiello Weekly
PCE
Randomly choose
(when more than one
per week)
Tetrachloroethene concentration at 420 basement south
sampling location, in |jg/m3, as measured by Radiello
sample
420BaseS Radiello Weekly
PCE
Randomly choose
(when more than one
per week)
Tetrachloroethene concentration at 422 basement north
sampling location, in |jg/m3, as measured by Radiello
sample
422BaseN Radiello Weekly
PCE
Randomly choose
(when more than one
per week)
Tetrachloroethene concentration at 422 basement south
sampling location, in |jg/m3, as measured by Radiello
sample
422BaseS Radiello Weekly
PCE
Randomly choose
(when more than one
per week)
Tetrachloroethene concentration at 422 first floor sampling
location, in |jg/m3, as measured by Radiello sample
420First Radiello Weekly
PCE
Randomly choose
(when more than one
per week)
Tetrachloroethene concentration at 422 first floor sampling
location, in |jg/m3, as measured by Radiello sample
422First Radiello Weekly
PCE
Randomly choose
(when more than one
per week)
Tetrachloroethene concentration at outside sampling
location, in |jg/m3, as measured by Radiello sample
Out_Radiello_Weekly_ PCE
Randomly choose
(when more than one
per week)
Radon concentration at 422 basement north sampling
location, in pCi/L, as measured by AlphaGUARD sample
422baseN_AG_radon
Mean
Radon concentration at 422 office sampling location, in
pCi/L, as measured by AlphaGUARD sample
422office_2nd_AG_radon
Mean
(continued)
3-33
-------
Section 3—Methods
Table 3-2. Data Aggregation Applied to Predictor Variables (cont.)
Variable Name (Plain Language)
Variable Code
Method of
Aggregation
Tetrachloroethene concentration at 420 basement south
sampling location, in |jg/m3, as measured by GC sample during
the first period of GC sampling
420baseS_GC1_PCE
Mean
Tetrachloroethene concentration at 422 basement south
sampling location, in |jg/m3, as measured by GC sample during
the first period of GC sampling
422baseS_GC1_PCE
Mean
Tetrachloroethene concentration at 420 first sampling location,
in |jg/m3, as measured by GC sample during the first period of
GC sampling
420first_GC1_PCE
Mean
Tetrachloroethene concentration at 422 first sampling location,
in |jg/m3, as measured by GC sample during the first period of
GC sampling
422fi rst_GC 1 _PC E
Mean
Tetrachloroethene Concentration at Wall Port 3 sampling
location, in |jg/m3, as measured by GC sample during the first
period of GC sampling
WP3_GC1_PCE
Mean
Tetrachloroethene concentration at Subslab Port 2 sampling
location, in |jg/m3, as measured by GC sample during the first
period of GC sampling
SSP2_GC1_PCE
Mean
Tetrachloroethene concentration at Subslab Port 4 sampling
location, in |jg/m3, as measured by GC sample during the first
period of GC sampling
SSP4_GC1_PCE
Mean
Tetrachloroethene concentration at Subslab Port 7 sampling
location, in |jg/m3, as measured by GC sample during the first
period of GC sampling
SSP7_GC1_PCE
Mean
Tetrachloroethene concentration at Soil Gas Port 11 sampling
location at a depth of 13 feet, in |jg/m3, as measured by GC
sample during the first period of GC sampling
SGP11-13_GC1_PCE
Mean
Tetrachloroethene concentration at Soil Gas Port 2 sampling
location at a depth of 9 feet, in |jg/m3, as measured by GC
sample during the first period of GC sampling
SGP2-9_GC1_PCE
Mean
Tetrachloroethene concentration at Soil Gas Port 8 sampling
location at a depth of 9 feet, in |jg/m3, as measured by GC
sample during the first period of GC sampling
SGP8-9_GC1_PCE
Mean
Tetrachloroethene concentration at Soil Gas Port 9 sampling
location at a depth of 6 feet, in |jg/m3, as measured by GC
sample during the first period of GC sampling
SGP9-6_GC1_PCE
Mean
Tetrachloroethene concentration at 420 basement south
sampling location, in |jg/m3, as measured by GC sample during
the second period of GC sampling
420baseS_GC2_PCE
Mean
Tetrachloroethene concentration at 422 basement south
sampling location, in |jg/m3, as measured by GC sample during
the second period of GC sampling
422baseS_GC2_PCE
Mean
Tetrachloroethene concentration at 420 first sampling location,
in |jg/m3, as measured by GC sample during the second period
of GC sampling
420first_GC2_PCE
Mean
Tetrachloroethene concentration at 422 first sampling location,
in |jg/m3, as measured by GC sample during the second period
of GC sampling
422fi rst_GC2_PC E
Mean
(continued)
3-34
-------
Section 3—Methods
Table 3-2. Data Aggregation Applied to Predictor Variables (cont.)
Variable Name (Plain Language)
Variable Code
Method of
Aggregation
Tetrachloroethene concentration at Wall Port 3 sampling
location, in |jg/m3, as measured by GC sample during the
second period of GC sampling
WP3_GC2_PCE
Mean
Tetrachloroethene concentration at Subslab Port 2 sampling
location, in |jg/m3, as measured by GC sample during the
second period of GC sampling
SSP2_GC2_PCE
Mean
Tetrachloroethene concentration at Subslab Port 4 sampling
location, in |jg/m3, as measured by GC sample during the
second period of GC sampling
SSP4_GC2_PCE
Mean
Tetrachloroethene concentration at Subslab Port 7 sampling
location, in |jg/m3, as measured by GC sample during the
second period of GC sampling
SSP7_GC2_PCE
Mean
Tetrachloroethene concentration at Soil Gas Port 11 sampling
location at a depth of 13 feet, in |jg/m3, as measured by GC
sample during the second period of GC sampling
SGP11-13_GC2_PCE
Mean
Tetrachloroethene concentration at Soil Gas Port 2 sampling
location at a depth of 9 feet, in |jg/m3, as measured by GC
sample during the second period of GC sampling
SGP2-9_GC2_PCE
Mean
Tetrachloroethene concentration at Soil Gas Port 8 sampling
location at a depth of 9 feet, in |jg/m3, as measured by GC
sample during the second period of GC sampling
SGP8-9_GC2_PCE
Mean
Tetrachloroethene concentration at Soil Gas Port 9 sampling
location at a depth of 6 feet, in |jg/m3, as measured by GC
sample during the second period of GC sampling
SGP9-6_GC2_PCE
Mean
Tetrachloroethene concentration at 420 basement south
sampling location, in |jg/m3, as measured by GC sample during
the third period of GC sampling
420baseS_GC3_PCE
Mean
Tetrachloroethene concentration at 422 basement south
sampling location, in |jg/m3, as measured by GC sample during
the third period of GC sampling
422baseS_GC3_PCE
Mean
Tetrachloroethene concentration at 420 first sampling location,
in |jg/m3, as measured by GC sample during the third period of
GC sampling
420first_GC3_PCE
Mean
Tetrachloroethene concentration at 422 first sampling location,
in |jg/m3, as measured by GC sample during the third period of
GC sampling
422first_GC3_PCE
Mean
Tetrachloroethene concentration at Wall Port 3 sampling
location, in |jg/m3, as measured by GC sample during the third
period of GC sampling
WP3_GC3_PCE
Mean
Tetrachloroethene concentration at Subslab Port 2 sampling
location, in |jg/m3, as measured by GC sample during the third
period of GC sampling
SSP2_GC3_PCE
Mean
Tetrachloroethene concentration at Subslab Port 4 sampling
location, in |jg/m3, as measured by GC sample during the third
period of GC sampling
SSP4_GC3_PCE
Mean
Tetrachloroethene concentration at Subslab Port 7 sampling
location, in |jg/m3, as measured by GC sample during the third
period of GC sampling
SSP7_GC3_PCE
Mean
(continued)
3-35
-------
Section 3—Methods
Table 3-2. Data Aggregation Applied to Predictor Variables (cont.)
Variable Name (Plain Language)
Variable Code
Method of
Aggregation
Tetrachloroethene concentration at Soil Gas Port 11 sampling
location at a depth of 13 feet, in |jg/m3, as measured by GC
sample during the third period of GC sampling
SGP11-13_GC3_PCE
Mean
Tetrachloroethene concentration at Soil Gas Port 2 sampling
location at a depth of 9 feet, in |jg/m3, as measured by GC
sample during the third period of GC sampling
SGP2-9_GC3_PCE
Mean
Tetrachloroethene concentration at Soil Gas Port 8 sampling
location at a depth of 9 feet, in |jg/m3, as measured by GC
sample during the third period of GC sampling
SGP8-9_GC3_PCE
Mean
Tetrachloroethene concentration at Soil Gas Port 9 sampling
location at a depth of 6 feet, in |jg/m3, as measured by GC
sample during the third period of GC sampling
SGP9-6_GC3_PCE
Mean
3-36
-------
Section 4—Results and Discussion: Quality Assurance Checks of Individual Data Sets
Table of Contents
4.0 Results and Discussion: Quality Assurance Checks of Individual Data Sets 4-1
4.1 VOC Sampling—Indoor Air-Passive—Air Toxics Ltd. (ATL) 4-1
4.1.1 Blanks 4-1
4.1.2 Surrogate Recoveries 4-3
4.1.3 Laboratory Control Sample Recoveries 4-3
4.1.4 Duplicates 4-4
4.2 VOC Sampling—Subslab and Soil Gas (TO-17)—U.S. EPA 4-5
4.2.1 Blanks 4-5
4.2.2 Calibration Verification 4-7
4.2.3 Internal Standard Recoveries 4-8
4.2.4 Surrogate Recoveries 4-8
4.2.5 Laboratory Control Sample Recoveries 4-9
4.2.6 Field Duplicates 4-9
4.3 Online Gas Chromatograph (Soil Gas and Indoor Air) 4-10
4.3.1 Blanks 4-10
4.3.2 Initial Calibration 4-10
4.3.3 Continuing Calibration 4-11
4.3.4 Calibration Check via Comparison to Fixed Laboratory (TO-15 vs. Online GC) 4-12
4.3.5 Agreement of Online GC Results with TO-17 Verification Samples 4-14
4.3.6 Agreement of Integrated Online GC Results with Passive Samplers 4-16
4.3.7 Overall Assessment of Online GC Data 4-37
4.4 Radon 4-43
4.4.1 Indoor Air: Comparison of Electrets Field, ARCADIS to Charcoal Analyzed by
U.S. EPA R&IE National Laboratory 4-43
4.4.2 Comparision of Average of Real-Time AlphaGUARD to Electrets and Charcoal
Canisters 4-45
4.4.3 Quality Assurance Checks of Electrets 4-48
4.5 On-Site Weather Station vs. National Weather Service (NWS) 4-48
4.6 Groundwater Analysis—EPA NERL 4-51
4.6.1 Blanks 4-51
4.6.2 Surrogate Recoveries 4-53
4.7 Groundwater Analysis—Pace Laboratories 4-54
4.8 Database 4-54
4.8.1 Checks on Laboratory Reports 4-54
4.8.2 Database Checks 4-54
4.9 Air Exchange Rate Measurements 4-55
4-i
-------
Section 4—Results and Discussion: Quality Assurance Checks of Individual Data Sets
List of Figures
4-1. TCE Continuing Calibration Standard Analyses, Hartman 3 Period 4-12
4-2. XY Comparison plot of Radiello and GC indoor air concentration measurements (|_ig/m3).
Hartman 1 sampling period 4-30
4-3. XY Comparison plot of Radiello and GC indoor air concentration measurements (|ig/m3),
Hartman 2 sampling period 4-31
4-4. Time series comparison of field GC and passive sampling data: 422 basement,
Hartman Period 2, chloroform. Horizontal gray line is calculated GC reporting limit. Red
hash marks on y-axis indicate missing values 4-33
4-5. Time series comparison of field GC and passive sampling data: 422 first floor,
Hartman Period 2, chloroform. Horizontal gray line is calculated GC reporting limit. Red
hash marks on y-axis indicate missing values 4-34
4-6. Time series comparison of field GC and passive sampling data: 422 basement,
Hartman Period 2, PCE. Horizontal gray line is calculated GC reporting limit. Red hash
marks on y-axis indicate missing values 4-35
4-7. Time series comparison of field GC and passive sampling data: 422 first floor, Hartman
Period 2, PCE. Horizontal gray line is calculated GC reporting limit. Red hash marks on
y-axis indicate missing values 4-36
4-8. XY Plot of field GC vs. passive sampler data, Hartman Period 3 4-38
4-9. Time series comparison of field GC and passive sampling data: 422 basement, Hartman
Period 3, chloroform. Horizontal gray line is calculated GC reporting limit. Red hash
marks on y-axis indicate missing values 4-39
4-10. Time series comparison of field GC and passive sampling data: 422 first floor, Hartman
Period 3, chloroform. Horizontal gray line is calculated GC reporting limit. Red hash
marks on y-axis indicate missing values 4-40
4-11. Time series comparison of field GC and passive sampling data: 422 Basement, Hartman
Period 3, PCE. Horizontal gray line is calculated GC reporting limit. Red hash marks on
y-axis indicate missing values 4-41
4-12. Time series comparison of field GC and passive sampling data: 422 first floor, Hartman
Period 3, PCE. Horizontal gray line is calculated GC reporting limit. Red hash marks on
y-axis indicate missing values (none in this case) 4-42
4-13. Correlation between radon measured using the electret and charcoal methods 4-45
4-14. Aerial view of study house, showing potential influences on wind velocity, red arrow
indicates study house 4-49
4-15. Comparison of National Weather Service Indianapolis temperature data to weather
station at 422 East 28th Street 4-50
4-16. Comparison of National Weather Service Indianapolis relative humidity to weather
station at 422 East 28th Street 4-50
4-17. Comparison ofNational Weather Service wind speed data to weather station at 422 East
28th Street 4-51
4-ii
-------
Section 4—Results and Discussion: Quality Assurance Checks of Individual Data Sets
List of Tables
4-1. Indoor Air Passive Field Blank Summary—Radiello 130 4-2
4-2. Indoor Air Passive Trip Blank Summary—Radiello 130 4-2
4-3. Indoor Air Passive Laboratory Blank Summary—Radiello 130 4-2
4-4. Indoor Air Passive Surrogate Summary—Radiello 130 4-3
4-5. Indoor Air Passive LCS Summary—Radiello 130 4-4
4-6. Indoor Air Passive Laboratory Precision (LCS/LCSD) Summary—Radiello 130 4-4
4-7. Subslab and Soil Gas—EPA Field Blank Summary—TO-17 4-5
4-8. Subslab and Soil Gas—EPA Trip Blank Summary—TO-17 4-6
4-9. Subslab and Soil Gas—EPA Laboratory Blank Summary—TO-17 4-6
4-10. Subslab and Soil Gas—EPA Fridge Blank Summary—TO-17 4-7
4-11. EPA TO-17 Calibration Verification (CV) Summary 4-8
4-12. EPA TO-17 Internal Standard (IS) Summary 4-8
4-13. EPA TO-17 Surrogate Recovery Summary 4-9
4-14. EPA TO-17 Laboratory Control Sample (LCS) Summary 4-9
4-15. EPA TO-17 Field Duplicate Summary 4-10
4-16. Field GC Estimated Minimum Detection Limits and Practical Quantitation Limits 4-11
4-17. Result of Repeated TCE Calibration Standard Analyses on On-line GC in March 2013
(Hartman Period 3) 4-14
4-18. Results of Repeated PCE Calibration Standard Analyses on Online GC in March 2013
(Hartman Period 3) 4-14
4-19. Interlaboratory Results: Spiked Verification Samples 4-15
4-20. Interlaboratory Statistics: Spiked Verification Samples 4-16
4-21. Comparison of Online GC to Radiello Results by Week 4-17
4-22. Comparison between Electrets and Charcoal Canisters at the 422/420 EPA House from
January 19-26, 2011 4-43
4-23. Comparison of Electret and Charcoal Canister Data from April 27 to May 4, 2011 4-44
4-24. Comparison of Charcoal and Electret Radon December 28, 2011, to January 4, 2012 4-44
4-25. Comparison between 422 Basement N AlphaGUARDs and Electrets from March 30,
2011, and May 18, 2011 4-45
4-26. Comparison of Real-Time AlphaGUARD to Integrated Electret August through October 4-46
4-27. Comparison of Real-Time AlphaGUARD to Integrated Electret Measurements
December 28, 2011, to January 4, 2012 4-46
4-28. Comparison of Real-Time AlphaGUARD to Integrated Electret Measurements January
through March 2012 4-47
4-29. Comparison of Real-Time AlphaGUARD to Integrated Electret Measurements January
through March 2013 4-47
4-iii
-------
Section 4—Results and Discussion: Quality Assurance Checks of Individual Data Sets
4-29. Comparison of Real-Time AlphaGUARD to Integrated Electret Measurements January
through March 2013 4-48
4-30. Groundwater (5 mL)—EPA Field Blank Summary 4-52
4-31. Groundwater (5 mL)—EPA Laboratory Blank Summary—TO-17 4-52
4-32. Groundwater (25 mL)—EPA Field Blank Summary—TO-17 4-52
4-33. Groundwater (25 mL)—EPA Laboratory Blank Summary—TO-17 4-53
4-34. EPA Groundwater (5 mL) Surrogate Recovery Summary 4-53
4-35. EPA Groundwater (25 mL) Surrogate Recovery Summary 4-53
4-iv
-------
Section 4—Results and Discussion: Quality Assurance Checks of Individual Data Sets
4.0 Results and Discussion: Quality Assurance Checks of Individual
Data Sets
This section describes the sampling and analytical quality assurance/quality control (QA/QC) checks
conducted for passive VOC sampling using Radiello samplers (4.1),*7 active sorbenttube sampling per
method TO-17 for soil gas samples (4.2)*, the on-site gas chromatograph used for continuous monitoring
of indoor air and soil gas (4.3), radon measurements by AlphaGUARD and electret instruments (4.4)*,
weather station measurements (4.5), groundwater sampling and analysis (4.6 and 4.7)*, and entry of the
compiled data into the project databases (4-8). Additional details on each of the sampling methods can be
found in Section 3.
4.1 VOC Sampling—Indoor Air-Passive—Air Toxics Ltd. (ATL)
QA/QC checks for the passive Radiello 130 samplers used for indoor and outdoor air sampling are
described in the following sections for blanks (4.1.1), surrogate recoveries (4.1.2), and laboratory control
surrogate (LCS) recoveries (4.1.3). For blanks, chloroform showed no detections, while PCE showed an
acceptably small percentage (3 to 9%) of detectable concentrations between the detection and reporting
limits. All surrogate recoveries met the laboratory control acceptance criteria. Chloroform failed to meet
the LCS recovery limits five times, while all PCE LCS recoveries met the control limits. These results
being above the control limits suggest that a minority of the time the laboratory may be overestimating
the concentration of chloroform and hexane by a factor of two times or less, which is acceptable quality
for the study's data quality objectives.
4.1.1 Blanks
Field blanks, trip blanks, and laboratory blanks were used to evaluate false positives and/or high bias due
to transport, storage, sample handling, and sorbent contamination.
¦ Field blanks were collected using a blank Radiello 130 cartridge from the media sample batch
sent to the field from the laboratory. The cartridge was removed from the sealed storage vial and
transferred to the diffusive housing in a similar manner to sample deployment. The cartridge was
then immediately removed from the housing, returned to the storage vial, and sealed for
shipment back to the laboratory with the field samples. In general, a field blank was collected
with each shipment to the laboratory. A total of 67 field blanks were submitted over the duration
of the project.
¦ Trip blanks were also assigned as blank Radiello cartridges from the media batches. The
cartridge was not opened or removed from the storage vial but was sent back to the laboratory
along with the field samples. There were 23 trip blanks submitted for analysis.
¦ For the laboratory blanks, a Radiello 130 cartridge was extracted with each analytical batch to
measure background from the sorbent and the extraction process. A total of 120 unique lab
blanks were analyzed and reported over the duration of the project.
To assist in data interpretation, all blank samples and all field sample results were evaluated down to the
method detection limit (MDL). The results of the field, trip, and laboratory blanks are summarized in
Tables 4-1, 4-2, and 4-3. The number of blanks with detections above the reporting limit (RL) and MDL
are tabulated. Summary statistics were then calculated on this subset of positive detections.
Measurements marked with an asterisk are designated as critical in the project QAPP.
4-1
-------
Section 4—Results and Discussion: Quality Assurance Checks of Individual Data Sets
Table 4-1. Indoor Air Passive Field Blank Summary—Radiello 130
Mean
Blank
Cone.
(H9)
FB
Cone. >
RL
RL> FB
Cone. >
MDL
% of Field
Blanks with
Detections
Std.
Dev.
(H9)
# FB
Analyzed
Max
(H9)
RL (Mg)
Min (|jg)
Benzene
0.4
67
0
58
87
0.122
0.04
0.04
0.21
Chloroform
0.1
67
0
0
0
NA
NA
NA
NA
cis-1, 2-DCE
0.1
67
0
0
0
NA
NA
NA
NA
Hexane
0.1
67
4
13
19
0.099
0.091
0.033
0.35
PCE
0.1
67
0
4
6
0.032
0.02
0.007
0.05
Toluene
0.1
67
1
25
37
0.044
0.037
0.014
0.17
TCE
0.1
67
0
5
7
0.015
0.009
0.006
0.03
NA= Not Applicable
Table 4-2. Indoor Air Passive Trip Blank Summary—Radiello 130
Mean
Blank
Cone.
(H9)
FB
Cone. >
RL
RL> FB
Cone. >
MDL
% of Trip
Blanks with
Detections
Std.
Dev.
(M9)
# FB
Analyzed
Max
(M9)
RL (Mg)
Min (|jg)
Benzene
0.4
23
0
21
91
0.102
0.039
0.042
0.16
Chloroform
0.1
23
0
0
0
NA
NA
NA
NA
cis-1, 2-DCE
0.1
23
0
0
0
NA
NA
NA
NA
Hexane
0.1
23
0
10
43
0.049
0.012
0.036
0.07
PCE
0.1
23
0
2
9
0.015
0.009
0.009
0.02
Toluene
0.1
23
0
18
78
0.02
0.008
0.012
0.041
TCE
0.1
23
0
4
17
0.024
0.016
0.009
0.043
NA= Not Applicable
Table 4-3. Indoor Air Passive Laboratory Blank Summary—Radiello 130
Mean
Blank
Cone.
(H9)
LB
Cone. >
RL
RL> LB
Cone. >
MDL
% of Lab
Blanks with
Detections
Std.
Dev.
(M9)
# LB
Analyzed
Max
(M9)
RL (Mg)
Min (|jg)
Benzene
0.4
120
9
113
94
0.1
0.056
0.038
0.34
Chloroform
0.1
120
10
0
0
NA
NA
NA
NA
cis-1, 2-DCE
0.1
120
0
0
0
NA
NA
NA
NA
Hexane
0.1
120
1
36
30
0.303
0.022
0.034
0.14
PCE
0.1
120
0
3
3
5.5
0.000
0.008
0.01
Toluene
0.1
120
3
72
60
0.454
0.026
0.005
0.13
TCE
0.1
120
0
5
4
0.372
0.006
0.013
0.03
NA= Not Applicable
4-2
-------
Section 4—Results and Discussion: Quality Assurance Checks of Individual Data Sets
Benzene was detected above the MDL but below the RL in a majority of the field, trip, and lab blanks at
similar background levels. The average of the positive detections was 0.121, 0.102, and 0.1 |Jg for the
field, trip, and lab blanks, respectively. The benzene blank levels are largely due to benzene
contamination present in the carbon disulfide extraction solvent. Although the laboratory used high purity
(99.99%) carbon disulfide reagent, benzene is present as a common contaminant in this solvent (White,
1964).
Although the benzene background levels were below the RL, a positive bias is expected for the daily
Radiello and a large subset of the weekly indoor air samples. Longer duration samples would normally
collect more mass and thus would not be significantly affected.
Hexane and toluene were also commonly detected in the field, trip, and lab blanks above the MDL. In the
case of the field and lab blanks, some had concentrations above the RL for hexane and toluene. All
detections in the trip blanks were below the RL but above the MDL. Similar to benzene, a positive bias
for hexane and toluene is anticipated for the daily Radiello samples due to the blank levels.
Because benzene, hexane, and toluene have a relatively constant low level blank contribution from the
media, the blank problems are more significant for the shortest duration samples (i.e., daily and to a lesser
extent weekly). See Section4.1.1 ofU.S.EPA (2012a) for a full discussion of these issues.
No detections of chloroform or cis-l,2-dichloroethene (cis-l,2-DCE) were measured in any of the blanks.
For a small percentage of the blanks, low concentration detections above the MDL were measured for
tetrachloroethene (PCE) and trichloroethene (TCE).
In summary, the contaminants of most concern in this study showed either no blank detections (for
chloroform) or an acceptably small percentage (3 to 9%) of low concentrations between the detection and
reporting limits (for PCE). The contaminants with highest blank detections (benzene, toluene, and
hexane) were not a primary focus for this study in that they were attributed to ambient outdoor air sources
and did not come from vapor intrusion.
4.1.2 Surrogate Recoveries
To monitor extraction efficiency, 5.0 (.ig oftoluene-d8 was spiked into each field sample and QC sample
Radiello 130 cartridge immediately prior to extraction. The recoveries were evaluated against laboratory
limits of 70 to 130%. All surrogate recoveries met the laboratory criterion, and summary statistics are
presented in Table 4-4.
Table 4-4. Indoor Air Passive Surrogate Summary—Radiello 130
Parameter
Result
Number of surrogate recoveries measures
1,681
Average recovery (%R)
103
Standard deviation (%R)
5.4
Minimum recovery (%R)
86
Maximum recovery (%R)
122
4.1.3 Laboratory Control Sample Recoveries
Accuracy of the extraction and analysis step for the target compounds was evaluated by analyzing an
LCS. An unused Radiello cartridge was spiked with a standard containing 5.0 |Jg of each compound of
4-3
-------
Section 4—Results and Discussion: Quality Assurance Checks of Individual Data Sets
interest. The laboratory acceptance criterion for LCS recovery was 70 to 130%. Chloroform and hexane
failed to meet the control limits five times each. These results being above the control limits suggest that a
minority of the time the laboratory may be overestimating the concentration of chloroform and hexane by
a factor of two times or less. Benzene, cis-1, 2-DCE, toluene, and TCE failed to meet the control limits
once each. PCE LCS recoveries met the control limits. Summary statistics are presented in Table 4-5.
Table 4-5. Indoor Air Passive LCS Summary—Radiello 130
Number of LCS
Analyzed
Mean LCS %
Recovery
LCS Std
Dev (%R)
Min
(%R)
Max (%R)
Benzene
113
96
14
70
147
Chloroform
113
100
17
70
206
cis-1,2-DCE
113
99
14
72
192
Hexane
113
103
20
71
219
PCE
113
100
11
80
130
Toluene
113
97
11
76
131
TCE
113
101
11
78
148
4.1.4 Duplicates
Sample precision was evaluated by collecting field duplicates and by analyzing LCSDs. Field duplicates
were collected for approximately every 10 field samples, and an LCSD was prepared and analyzed with
each sample preparation batch. Because the LCSD was a second cartridge prepared and extracted in the
same manner as the LCS, the relative percentage difference (%RPD) represents the precision of the
analytical method from extraction through analysis. The method precision is summarized in Table 4-6.
The laboratory acceptance criterion of %RPD < 25% was met by PCE, toluene, and TCE but exceeded in
2 batches by benzene, 5 by chloroform, 1 by cis-1, 2-DCE, and 11 by hexane.
Table 4-6. Indoor Air Passive Laboratory Precision (LCS/LCSD) Summary—Radiello 130
Number of
LCSD
Analyzed
Mean
%RPD
Std Dev.
(%RPD)
Min (%RPD)
Max (%RPD)
Number of
Exceedances
Benzene
113
9
8
0
42
2
Chloroform
113
10
8
0
35
0
cis-1,2-DCE
113
5
5
0
31
0
Hexane
113
13
11
0
47
5
PCE
113
4
4
0
19
0
Toluene
113
5
5
0
19
0
TCE
113
5
4
0
20
0
4-4
-------
Section 4—Results and Discussion: Quality Assurance Checks of Individual Data Sets
4.2 VOC Sampling— Subslab and Soil Gas (T0-17)—U.S. EPA
4.2.1 Blanks
Field, trip, refrigerator, and laboratory blanks were used to evaluate false positives and/or high bias due to
transport, storage, sample handling, and sorbent contamination. Field blanks were collected using a blank
Tenax TA TO-17 sorbent tube from the media sample batch sent to the field from the laboratory. The
Swagelok end caps were removed as if to prepare for sample collection; however, no soil vapor was
pulled through the tube. The end caps were immediately replaced, and the tube was sent back to the
laboratory with the field samples. Typically, a field blank was collected with each shipment to the
laboratory. A total of 121 field blanks were submitted over the duration of the project.
Blank Tenax TA TO-17 sorbent tubes from the media batches were also assigned as trip blanks. The tube
remained capped and wrapped in aluminum foil and was sent from the laboratory to the field and back to
the laboratory along with the field samples. There were 111 trip blanks submitted for analysis.
In the case of the laboratory blank, a Tenax TA TO-17 tube was analyzed with each analytical batch to
measure background from the sorbent tubes and instrumentation. A total of 387 lab blanks were analyzed
and reported over the duration of the project.
For a refrigerator (fridge) blank, a Tenax TA TO-17 tube was stored and analyzed with each sample batch
to measure background from the sample storage refrigerator. The tubes were stored in the refrigerator
capped and sealed in a zip lock bag on top of the jars containing the samples that were received as a
batch. The fridge blanks were placed in the refrigerator with a sample batch and remained in the
refrigerator with the batch until all the samples from that batch had been analyzed. So, the fridge blanks
were in the refrigerator longer than some of the samples within a batch. A total of 61 fridge blanks were
analyzed and reported over the duration of the project.
To assist in data interpretation, all blank samples and all field sample results were evaluated down to the
MDL. The results of the field, trip, laboratory, and fridge blanks are summarized in Tables 4-7, 4-8, 4-9,
and 4-10. The number of blanks with detections above the RL and MDL are tabulated. Summary statistics
were then calculated on this subset of positive detections.
Table 4-7. Subslab and Soil Gas—EPA Field Blank Summary—TO-17
Number of Field Blanks
Mean
Blank
Cone,
(ng)
% of Field
Blanks with
Detections
Std.
Dev.
(ng)
RL
(ng)
Min
(ng)
Max
(ng)
Analyzed
Cone. >
RL
RL>Conc.
> MDL
Benzene
5.0
121
0
53
44
1.4
0.5
0.81
3.0
Carbon disulfide
5.0
121
0
9
7
3.4
1.4
1.7
6.4
Chloroform
2.0
121
5
0
4
72
110
3.0
260
cis-1,2-DCE
2.0
121
0
1
1
1.5
N/A
1.5
1.5
Hexane
10
121
0
2
2
1.6
1.5
2.2
4.4
Methylene chloride
50
121
0
9
9
8.7
5.2
2.5
19
PCE
2.0
121
9
0
7
9.6
4.3
2.1
10
Toluene
5.0
121
0
18
15
2.2
2.0
1.1
7.7
TCE
2.0
121
1
0
1
2.8
N/A
2.8
2.8
N/A = Not Applicable
4-5
-------
Section 4—Results and Discussion: Quality Assurance Checks of Individual Data Sets
Table 4-8. Subslab and Soil Gas—EPA Trip Blank Summary—TO-17
Mean
Blank
Cone,
(ng)
RL
(ng)
% of Trip
Blanks with
Detections
Std.
Dev.
(ng)
Min
(ng)
Max
(ng)
Analyzed
Cone. >
RL
RL>Conc.
> MDL
Benzene
5.0
111
0
38
34
1.3
0.5
0.81
2.6
Carbon disulfide
5.0
111
0
9
8
2.6
0.8
1.6
4.0
Chloroform
2.0
111
6
1
6
32
41
2.0
120
cis-1,2-DCE
2.0
111
0
0
0
0
0
0
0
Hexane
10
111
0
2
2
2.0
1.5
1.0
3.0
Methylene chloride
50
111
0
4
4
2.8
0.8
2.2
4.0
PCE
2.0
111
4
0
4
18
11
2.3
27
Toluene
5.0
111
3
20
21
3.1
4.1
1.0
19
TCE
2.0
111
2
0
2
3.7
2.0
2.3
5.2
Table 4-9. Subslab and Soil Gas—EPA Laboratory Blank Summary—TO-17
Number of Lab Blanks
Mean
Blank
Cone,
(ng)
% of Lab
Blanks with
Detections
Std.
Dev.
(ng)
RL
(ng)
Min
(ng)
Max
(ng)
Analyzed
Cone. >
RL
RL>Conc.
> MDL
Benzene
5.0
387
7
99
27
1.8
1.9
0.80
12
Carbon disulfide
5.0
387
4
42
12
9.6
9.2
0.87
52
Chloroform
2.0
387
16
2
5
3.4
1.7
1.3
5.8
cis-1,2-DCE
2.0
387
0
4
1
4.1
0.7
4.9
3.4
Hexane
10
387
1
10
3
4.9
5.6
1.5
21
Methylene chloride
50
387
0
8
2
3.2
1.2
2.4
5.6
PCE
2.0
387
4
8
3
1.8
1.2
0.7
4.1
Toluene
5.0
387
5
47
13
2.7
3.4
1.0
16
TCE
2.0
387
4
3
2
5.6
5.3
1.4
16
N/A = Not Applicable
4-6
-------
Section 4—Results and Discussion: Quality Assurance Checks of Individual Data Sets
Table 4-10. Subslab and Soil Gas—EPA Fridge Blank Summary—TO-17
Mean
Blank
Cone,
(ng)
RL
(ng)
iMumuei
uirnuyt
DIctMftS
% of Fridge
Blanks with
Detections
Std.
Dev.
(ng)
Min
(ng)
Max
(ng)
Analyzed
Cone. >
RL
RL>Conc.
> MDL
Benzene
5.0
61
0
22
36
1.2
0.40
0.81
1.8
Carbon disulfide
5.0
61
0
2
3
2.3
0.69
1.8
2.8
Chloroform
2.0
61
2
0
3
2.3
0.29
2.1
2.5
cis-1,2-DCE
2.0
61
0
0
0
0
0
0
0
Hexane
10
61
0
3
5
1.0
0.08
0.88
1.0
Methylene chloride
50
61
0
4
7
6.1
7.5
1.8
17
PCE
2.0
61
6
4
16
3.2
1.7
0.8
3.5
Toluene
5.0
61
5
11
26
8.5
20
0.96
82
TCE
2.0
61
8
1
15
7.4
4.6
1.5
17
Benzene was detected above the MDL in 44%, 34%, 27%, and 36% of the field (Figure 4-7), trip (Figure
4-8), laboratory (Figure 4-9), and fridge (Figure 4-10) blanks, respectively. The average of the positive
detections was 1.4, 1.3, 1.8, and 1.2 nanogram (ng) for the field, trip, lab, and fridge blanks, respectively.
Seven laboratory blanks had benzene concentrations above the RL of 5.0 ng. The benzene blank levels are
largely due to background contribution from the Tenax TA polymer, which can break down during the
heating step to generate low levels of benzene (Middleditch, 1989).
The concentrations of benzene in the TO-17 soil vapor samples were similar in magnitude to those
measured in the field blanks. Of the 2844 TO-17 soil vapor samples analyzed by EPA, 59% of the
samples had a positive detection of benzene. Of the samples that had a positive detection for benzene,
only 2% had a detected concentration above the RL of 5.0 ng. The second most common contaminant in
these blank samples was toluene, which has also been reported as a Tenax breakdown product (MacLeod
and Ames, 1986; Cao and Hewitt, 1994).
Detections of the key compounds that form the focus of this work—PCE, chloroform, and TCE—
occurred in 3% or less of the hundreds of samples and field, trip, and lab blanks analyzed. However, the
percentage of refrigerator blanks with PCE and TCE contamination was considerably higher—16%.
4.2.2 Calibration Verification
The calibration relationship established during the initial calibration was verified at the beginning of each
24-hour analytical shift using a calibration verification standard concentration equal to the mid-point of
the initial calibration range. If the analyte concentration was within ±30% (40% for Carbon Disulfide and
Methylene Chloride) of the expected concentration of the calibration verification standard, then the initial
calibration was considered valid, and the analysis of samples was continued. Most analyte calibration
verification standard recoveries met the QAPP established criterion, and summary statistics are presented
in Table 4-11.
4-7
-------
Section 4—Results and Discussion: Quality Assurance Checks of Individual Data Sets
Table 4-11. EPA TO-17 Calibration Verification (CV) Summary
Number
of CV
Analyzed
Mean CV
% Recovery
CV Std
Dev (%R)
Min
(%R)
Max (%R)
CV Recovery
Limits
Number of
Exceedances
Benzene
665
97
18
2
276
70-130%
22
Carbon disulfide
665
84
52
0
664
60-140%
251
Chloroform
665
91
19
0
298
70-130%
53
cis-1,2-DCE
665
95
20
0
268
70-130%
45
Hexane
665
93
20
0
262
70-130%
51
Methylene chloride
665
94
67
0
818
60-140%
266
PCE
665
87
17
0
262
70-130%
75
Toluene
665
98
18
0
286
70-130%
23
TCE
665
95
17
0
276
70-130%
19
4.2.3 Internal Standard Recoveries
Two internal standards were utilized in the calibration of the TO-17 analytical instrumentation, 1,4-
difluorobenzene and chlorobenzene-d5. 4.7 ng of 1,4-difluorobenzene and 4.8 ng of chlorobenzene-d5 in
a gas phase standard were automatically introduced into the sample flow path by the instrumentation
during the initial tube desorption of all samples. The internal standard calibration was used to account for
routine variation in the response of the chromatographic system as well as variations in the exact volume
of sample introduced into the chromatographic system. The recoveries were evaluated against the QAPP
established criteria of 60 to 140% recovery. Most internal standard recoveries met the QAPP established
criterion, and summary statistics are presented in Table 4-12.
Table 4-12. EPA TO-17 Internal Standard (IS) Summary
Number
of IS
Analyzed
Mean IS
%
Recovery
IS Std
Dev (%R)
Min (%R)
Max (%R)
IS Recovery
Limits
Number of
Exceedances
1,4-Difluorobenzene
4620
99
34
15
373
60-140%
907
Chlorobenzene-d5
4620
99
30
18
358
60-140%
776
4.2.4 Surrogate Recoveries
To monitor analytical efficiency, 5.3 ng of bromochloromethane were loaded onto each QC and field
sample sorbent tube along with the vapor phase internal standard mix during sample analysis. Field
surrogates were not included in the scope of this project. The recoveries were evaluated against laboratory
limits of 70 to 130%. Most surrogate recoveries met the QAPP established criterion, and summary
statistics are presented in Table 4-13.
4-8
-------
Section 4—Results and Discussion: Quality Assurance Checks of Individual Data Sets
Table 4-13. EPA TO-17 Surrogate Recovery Summary
Parameter
Result
Number of surrogate recoveries measured
4,620
Average recovery (%R)
105
Standard deviation (%R)
14
Minimum recovery (%R)
22
Maximum recovery (%R)
360
4.2.5 Laboratory Control Sample Recoveries
Analytical accuracy was evaluated by analyzing an LCS. Two clean Tenax TA TO-17 sorbent tubes were
spiked with a calibration standard from a source independent from the primary calibration standard and
analyzed after each initial calibration. The spike contained approximately 100 nanograms of each target
compound. The performance of the EPA TO-17 LCS spikes is summarized in Table 4-14. A total of 10
LCS samples were evaluated, and all met the laboratory RLs with the exceptions of five outliers for
carbon disulfide, four outliers for methylene chloride, and one outlier for cis-l,2-DCE.
Table 4-14. EPA TO-17 Laboratory Control Sample (LCS) Summary
Number
of LCS
Analyzed
Mean LCS
%
Recovery
Min (%R)
Max (%R)
LCS
Recovery
Limits
LCS Std
Dev (%R)
Number of
Exceedances
Benzene
10
101
11
86
118
70-130%
0
Carbon disulfide
10
117
64
24
272
70-130%
5
Chloroform
10
96
11
82
122
70-130%
0
cis-1,2-DCE
10
105
10
96
133
70-130%
1
Hexane
10
98
gii
72
120
70-130%
0
Methylene chloride
10
111
71
29
291
70-130%
4
PCE
10
85
8.1
71
97
70-130%
0
Toluene
10
102
13
80
128
70-130%
0
TCE
10
100
12
80
120
70-130%
0
4.2.6 Field Duplicates
Sample precision was evaluated by collecting field duplicates. Field duplicates were collected for
approximately every 10 field samples. The sample precision is summarized in Table 4-15. The laboratory
acceptance criterion of %RPD < 50% was met by PCE, toluene, and TCE but exceeded in 2 batches by
benzene, 5 by chloroform, 1 by cis-1, 2-DCE, and 11 by hexane.
4-9
-------
Section 4—Results and Discussion: Quality Assurance Checks of Individual Data Sets
Table 4-15. EPA TO-17 Field Duplicate Summary
Number
Analyzed
Mean
%RPD
Std Dev.
(%RPD)
Min (%RPD)
Max (%RPD)
Number of
Sample
Exceedances
Benzene
173
40
36
0
163
35
Chloroform
173
30
49
0
197
26
cis-1,2-DCE
173
21
32
1
106
2
Hexane
173
45
34
13
119
5
PCE
173
23
40
0
197
20
Toluene
173
27
23
0
91
8
TCE
173
38
50
0
173
8
4.3 Online Gas Chromatograph (Soil Gas and Indoor Air)
The online GC was used in three distinct mobilization periods each of which had some differences in
instrument setup. These analyses were provided by Hartman Environmental Geosciences with logistical
support from ARCADIS. Therefore, we refer to these data sets in this discussion of QA checks as
¦ Hartman 1 = August 11, 2011, to October 17, 2011;
¦ Hartman 2 = December 1, 2011, to February 26, 2012; and
¦ Hartman 3 = December 14, 2012, to March 8, 2013.
4.3.1 Blanks
Instrument blanks were analyzed at least once per analysis cycle of the 12 sampling locations. Nitrogen or
outdoor air was analyzed at the beginning of the analysis cycle (stream selector valve port #1). System
blanks (no vapor sample injected) were analyzed twice per analysis cycle at the end of the analysis cycle
(stream selector valve ports #15 and #16) from August 26, 2011, through February 26, 2012.
Environmental analytical data are normally compared with blank data using approaches suitable for very
small numbers of samples (often just one blank vs. one sample). For example, data are often qualified if
the sample does not exceed a certain multiple of the blank concentration.8 However, this approach is no
longer a mandatory requirement of the EPA functional guidelines for low concentration VOCs (U.S.
EPA, 2008b). Those guidelines now call for professional judgment in cases where the sample result
exceeds the blank result itself. In this case, given that we have hundreds of measurements of indoor air
and either a blank or atmospheric air taken with the same instrument, it is appropriate to use other
statistical tests to judge whether the samples are significantly different from the blank (or atmospheric air
control).
4.3.2 Initial Calibration
For Hartman Period 1 (August 11, 2011, to October 17, 2011), initial calibration curves for PCE and
chloroform were performed at the start of the monitoring program as follows:
¦ PCE: Two points at concentrations of 14 ng/m3 and 70 ng/m3
8http://140.194.76.129/publications/eng-manuals/EM 200-l-10/c-10.pdf
4-10
-------
Section 4—Results and Discussion: Quality Assurance Checks of Individual Data Sets
¦ CHCh: A single point at a concentration 10 ng/m3, with a separate linearity study after the initial
deployment
Additional calibration points were not possible because of uncertainties with the calibration standards
brought to the site during instrument set-up.
For Hartman Period 2 (December 1, 2011, to February 16, 2012), initial calibrations were as follows:
¦ PCE low range: six points at concentrations from 0.7 ng/m3 to 23 ng/m3
¦ PCE high range: three points at concentrations from 3.5 ng/m3 to 69 ng/m3
¦ CHCI3 low range: four points at concentrations from 3.3 ng/m3 to 55 ng/m3
¦ CHCh high range: three points at concentrations from 55 ng/m3 to 270 ng/m3
For Hartman Period 3, the chloroform and PCE calibration ranges were:
¦ CHC13: 1.0 to 250 (ig/m3, 6 calibration points
¦ TCE: 5.5 to 220 (ig/m3, 6 calibration points
¦ PCE: 0.69 to 280 (ig/m3, 8 calibration points
Although a formal MDL determination was not conducted for the Hartman 1 and Hartman 2 field GC
periods, a formal MDL determination based on seven repetitive injections of a standard was performed
for Hartman 3. For the other periods, the MDL was estimated based on three times the concentration
observed in repetitive injections of nitrogen blanks or background air. This field MDL and PQL
information and its basis are summarized in Table 4-16. As discussed in Section 4.5.1, this three-times
blank (or ambient air) definition of a detection is more stringent than required by current EPA functional
guidelines and probably does not adequately capture the sensitivity of a data set with hundreds of
repetitive analyses of both the target atmosphere and the blank (or ambient air).
Table 4-16. Field GC Estimated Minimum Detection Limits and Practical Quantitation Limits
CHCh
PCE
PQL Low Cal
|jg/m3
Period
Dates
MDL
|jg/m3
PQL Low Cal
|jg/m3
MDL
|jg/m3
Hartman 1
8/2011-10/2011
1
10
0.84
14
Hartman 2
12/2011-2/2012
0.7
0.6
0.9-1.2
0.69
Hartman 3
12/2012-3/2013
0.7
1.0
0.7
0.69
Notes:
MDLs computed as three times air blanks
CHCI3 MDL from actual analyses; PCE from three times air blank
MDLs for PCE calculated from 14 runs of a low concentration standard
4.3.3 Continuing Calibration
Continuing calibration was not performed using the compounds of primary interest because of the
concern that the calibration standard could contaminate the indoor air values (since it necessarily would
be stored within the study duplex). Instead a surrogate compound, TCE, was used for continuing
calibration. The TCE was plumbed to stream selector port #14 with the intent it would be analyzed in
every analytical cycle of the 16 ports. However, during both the Hartman 1 and 2 periods of the program,
the TCE calibration standard quickly ran out because of a leak at port 14 in the stream selector valve. As
4-11
-------
Section 4—Results and Discussion: Quality Assurance Checks of Individual Data Sets
an alternative, a calibration check comparing the performance of the field instrument to a laboratory-
based instrument with site sample was performed as discussed in the next section for those periods. The
valve was replaced before Hartman 3.
The Hartman 3 data set had a continuing calibration standard that was repeatedly analyzed for 332
successive analyses over the first 48 days of the operational period until it was exhausted (Figure 4-1).
For those analyses, the mean was 74.20 (ig/m3 with a standard deviation of 0.99, indicating a very stable
measurement with a variability range well within +/- 2 (ig/m3. The TCE concentration in this standard
was measured at 32 (ig/m3 by H&P Laboratories.
After the continuing calibration standard was exhausted on January 31, attempts were made to provide
additional calibration standards. These culminated in a final calibration run between March 7 and
March 11 reported in Section 4.5.4. Taken together, these approaches to continuing calibration suggest
that the instrument maintained precision well within the quality assurance project plan (QAPP)
established precision goal of +/- 25%.
~
~
~
~
>
^ * ~~
~
~
12/7/2012 12/17/2012 12/27/2012 1/6/2013 1/16/2013 1/26/2013 2/5/2013
Date
Figure 4-1. TCE Continuing Calibration Standard Analyses, Hartman 3 Period.
4.3.4 Calibration Check via Comparison to Fixed Laboratory (TO-15 vs. Online GC)
Verification samples were collected and analyzed by H&P Mobile Geochemistry during each sampling
period as follows. The H&P fixed base lab is certified for a variety of tests such as EPA 8260B, EPA
TO-15, and CA LUFT/8015m. Key certifying bodies include:
4-12
-------
Section 4—Results and Discussion: Quality Assurance Checks of Individual Data Sets
¦ U.S. Department of Defense, Environmental Laboratory Accreditation Program (DoD-ELAP) -
PJLA Accreditation No. 69070 - Certificate No. LI 1-175 (Methods: H&P-SOP 8260SV; EPA
8260B; EPA T015; EPA T014A; H&P-S0P-T015M)
¦ California Department of Public Health, Environmental Laboratory Accreditation Program,
(ELAP) Certificate No. 2741; and
¦ New York State Department of Health, National Environmental Laboratory Accreditation
Conference (NELAC) Standards Certificate Lab ID No. 11845
Hartman Period 1: An indoor air sample was collected from the 422 first floor on October 11, 2011, and
compared with the on-site instrument to check on the reported concentration values. The results were as
follows (ng/m3):
On-site GC H&P TO-15
CHCls 1.7 0.8
PCE 3 1.3
In addition, a 24-hour time composite indoor air sample was collected from the 422 first floor and the
basement on September 22, 2011, and compared with the on-site instruments values over the same time
period to check on the reported low concentration values. The results were as follows (ng/m3):
On-site GC ATL TO-15
422first floor:
CHCls 1.0 0.24
PCE 1.75 0.40
422 basement.
CHCls 1.7 0.41
PCE 3.5 0.94
Based on these data and the data summarized in Section 4.5.6, we decided that the online GC chloroform
low values (<5 ng/m3) could be adjusted down by a factor of 2 and the online GC PCE low values
(<5 ng/m3) could be adjusted down by a factor of 3. Alternatively, the generally low bias exhibited by the
Hartman Period 1 samples could be adequate justification to not use these data in subsequent analysis or
at least to regard any conclusions drawn as less reliable than those drawn from the Periods 2 and 3 data.
Hartman Period 2: A sample was collected from probe SP8-9 on December 11, 2011, and compared with
the on-site instrument. The results were as follows (ng/m3):
On-site GC H&P TO-15
CHC13 118 100
PCE 140 160
Based on these results, no adjustments in the online GC data were made.
Hartman Period 3: A final calibration check was performed between March 7 and March 11, 2013.
During that period, 26 successive analyses of the same Hartman Environmental-prepared standard9 were
performed and the results compared with an analysis of the standard performed at a fixed based
9The standard was prepared in a Tedlarbag by diluting a liquid standard into 1,000 cc and was analyzed using TO-15 and the
auto GC. The target concentration range was 10 to 100 (ig/m3.
4-13
-------
Section 4—Results and Discussion: Quality Assurance Checks of Individual Data Sets
laboratory. The results indicated excellent precision and good accuracy for TCE (Table 4-17) and PCE
(Table 4-18).
Table 4-17. Result of Repeated TCE Calibration Standard Analyses on On-line GC in March 2013
(Hartman Period 3)
TCE
Hartman GC
Average
9.68
|jg/m3
Stdevp
0.57
|jg/m3
%RPD
Precision
5.9%
H&P value
8.12
|jg/m3
Accuracy
% RPD
17.6%
Table 4-18. Results of Repeated PCE Calibration Standard Analyses on Online GC in March 2013
(Hartman Period 3)
PCE
Hartman GC
Average
6.99
|jg/m3
Stdevp
0.51
|jg/m3
Precision
7.3%
H&P Value
8.25
|jg/m3
Accuracy
% RPD
-16.5%
4.3.5 Agreement of Online GC Results with TO-17 Verification Samples
Early in the Hartman 1 period, ATL prepared four 3 L Tedlar bags each containing approximately 2 L of
vapor labeled A, B, C, and D and sent them to the Indianapolis field site. Bags A and B were duplicate
nitrogen blanks. Bags C and D were duplicate spikes with chloroform, TCE, and PCE drawn from a
common Summa canister. Analyses were performed of these bags using the online GC and by ARCADIS
staff collecting TO-17 samples directly from the bags and submitting them to NERL for analysis. ATL
also performed analyses before sending the bags to Indianapolis and after their return from the field.
Results of these interlaboratory comparisons are provided in Table 4-19, and statistical comparison is
provided in Table 4-20. The agreement between the two fixed based laboratories where the RPD is <25%
is excellent; this is a considerably narrower range then is often seen in method VOC method
intercomparison studies (Lutes 2010B). The agreement between the field instrument and the fixed based
laboratories with all RPDs <50% is somewhat lesser, but still reasonable given that RPDs that large are
sometimes seen between fixed based laboratories running the same method (Lutes, 2010b) and that this
comparison is between methods - between an automated GC-ECD and an attended GC-MS.
4-14
-------
Table 4-19. Interlaboratory Results: Spiked Verification Samples
Subsample
Date
PCE
flag
PCE
ug/m3
PCE
ppbv
TCE
flag
TCE
ug/m3
TCE
ppbv
Chloroform
Flag
Chloroform
ug/m3
Chloroform
ppbv
Bag
Laboratory
Analysis Date
D
Air Toxics
8/9/2011
8/9/2011
21
34
42
A
Hartmann
8/11/2011
8/11/2011
<
2
<
2
<
2
B
Hartmann
8/11/2011
8/11/2011
<
2
<
2
<
2
C
Hartmann
8/11/2011
8/11/2011
20
28
40
D
Hartmann
8/11/2011
8/11/2011
20
23
40
C
Air Toxics
8/12/2011
8/12/2011
13
16
20
D
Air Toxics
8/12/2011
8/12/2011
12
16
21
B
EPA NERL
8/10/2011
8/14/2011
U
8.5
1.2
U
6.7
1.2
B
12
2.4
A
EPA NERL
8/10/2011
8/14/2011
U
8.5
1.2
U
6.7
1.2
U
6.2
1.3
D
EPA NERL
8/10/2011
8/14/2011
85
12.3
110
20.1
140
28.2
B
EPA NERL
8/10/2011
8/14/2011
u
8.5
1.2
u
6.7
1.2
B
12
2.4
A
EPA NERL
8/10/2011
8/14/2011
u
8.5
1.2
u
6.7
1.2
B
11
2.2
D
EPA NERL
8/10/2011
8/14/2011
80
11.6
110
20.1
130
26.2
C
EPA NERL
8/10/2011
8/14/2011
89
12.9
110
20.1
140
28.2
C
EPA NERL
8/10/2011
8/14/2011
84
12.2
110
20.1
130
26.2
O'
>3
Co
£
3
b
>3
SJ
S'
5
>3
a.
<"
I
a.
b
q
3"
-------
Section 4—Results and Discussion: OA Checks of Individual Data Sets
Table 4-20. Interlaboratory Statistics: Spiked Verification Samples
Interlab Comparison
Data Summary for Interlab
Data
Standard Samples after Pooling c and d: Interlab Comparison Using
Standard Results
Mean (ppbv)
% Difference (% error)**
EPA
Air Toxics
Air Toxics
Air Toxics
Actual
(TO-15)
Air Toxics
(N=3)
Hartmann
(N=2)
Chemical
NERL
vs. EPA
vs.
vs.
(N=4)
NERL
Hartman
Hartman
Chloroform
42
27.7
27.2
40.0
1.64
38.03
36.45
Tetrachloroethene
21
15.3
12.3
20.0
22.24
47.95
26.42
Trichloroethene
34
22.0
20.1
25.5
8.80
23.46
14.74
4.3.6 Agreement of Integrated Online GC Results with Passive Samplers
4.3.6.1 Hartman Periods 1 and 2
Table 4-21 and Figures 4-2 and 4-3 compare the concentrations measured by the 1-week Radiello
samples to the concentrations calculated by averaging the online GC results. In Figures 4-2 and 4-3, the
Radiello chloroform (red) and PCE (aqua) concentrations are plotted against their corresponding weekly
average GC values for Hartman 1 and Hartman 2, respectively. The grey line in each figure has a slope of
1 and an intercept of 0 and represents the ideal case where GC and Radiello measurements match exactly.
Most of the time, the weekly GC sample averages are higher than the corresponding Radiello weekly
samples, suggesting a consistent positive bias for the GC or a negative bias for the Radiellos. However,
this difference is small (mostly less than a factor of 2; Table 4-21), and the data are still usable, given the
purpose of the GC and weekly Radiello data, to measure short-term and long-term variability in indoor air
VOC concentrations.
For chloroform, agreement is generally remarkably good for the first 4 weeks of instrument operation.
The results for this period are generally within 50 relative percent difference, which we considered good
for this comparison between two different methods, given that variability in interlaboratory comparisons
for split samples of VOCs using one method can be larger. Expressed as a ratio during this period the
online GC result is always between 0.6 and 1.9 times the Radiello result.
However, for chloroform, agreement is noticeably worse in succeeding weeks (after September 14, 2011).
Generally the chloroform values reported from the online GC are one to three times higher than the values
from the corresponding Radiello sample, although higher ratios up to six times higher were occasionally
observed, associated with the lowest concentration Radiello results. During the period when ambient
samples were also collected with the online GC, those results tended to be a more significant fraction of
the measured indoor air values than was seen in the Radiello samples. This suggests the possible
existence of an elevated baseline in the online GC data.
Results were considerably improved in Hartman Period 2 over the results in the later portions of Period 1
(see Figure 4-2). This may be due to the instrument setup changes that were made (as described in
Section 3.4.3). This may also be due to the higher concentrations available for analysis in the field
samples. In general, the agreement across all periods is better at higher concentration levels (>0.5 |_ig/m3).
Hartman Period 1 was the only period conducted under summer conditions.
For PCE, the relationship between the online GC and the Radiello samples appears more stable with the
vast majority of the results showing online GC results one to three times higher than the corresponding
Radiello data.
4-16
-------
Table 4-21. Comparison of Online GC to Radiello Results by Week
= "D S
S " II
-2-1 «
;g 3 S
8. o >
o >
re U)
o .=
re U)
o .=
re U)
o .E
a u >
8/17/2011 22:55
8/24/2011 21:21
Hartmanl
420BaseS
Tetrachloroethene
0.38
0.695027
0.701027
0.699013
59%
59%
59%
1.83
1.84
1.84
8/24/2011 21:22
8/31/2011 20:51
Hartmanl
420BaseS
Tetrachloroethene
0.4
0.624743
0.624743
0.624743
44%
44%
44%
1.56
1.56
1.56
8/31/2011 20:52
9/7/2011 20:34
Hartmanl
420BaseS
Tetrachloroethene
0.36
0.539148
0.546148
0.541167
40%
41%
40%
1.50
1.52
1.50
9/7/2011 20:36
9/14/2011 23:09
Hartmanl
420BaseS
Tetrachloroethene
0.36
0.588986
0.588986
0.588986
48%
48%
48%
1.64
1.64
1.64
9/14/2011 23:11
9/21/2011 22:23
Hartmanl
420BaseS
Tetrachloroethene
0.63
0.633948
0.689419
0.666505
1%
9%
6%
1.01
1.09
1.06
9/21/2011 22:25
9/28/2011 21:09
Hartmanl
420BaseS
Tetrachloroethene
0.27
0.516929
0.706606
0.596752
63%
89%
75%
1.91
2.62
2.21
9/28/2011 21:12
10/6/2011 21:41
Hartmanl
420BaseS
Tetrachloroethene
0.52
0.755773
0.80244
0.797745
37%
43%
42%
1.45
1.54
1.53
8/17/2011 22:36
8/24/2011 21:14
Hartmanl
420First
Tetrachloroethene
0.3
0.655622
0.655622
0.655622
74%
74%
74%
2.19
2.19
2.19
8/24/2011 21:16
8/31/2011 20:44
Hartmanl
420First
Tetrachloroethene
0.33
0.578883
0.578883
0.578883
55%
55%
55%
1.75
1.75
1.75
8/31/2011 20:46
9/7/2011 20:27
Hartmanl
420First
Tetrachloroethene
0.23
0.65351
0.66751
0.661562
96%
97%
97%
2.84
2.90
9/7/2011 20:29
9/14/2011 22:48
Hartmanl
420First
Tetrachloroethene
0.22
0.777222
0.777222
0.777222
112%
112%
112%
3.53
3.53
3.53
9/14/2011 22:49
9/21/2011 22:1
Hartmanl
420First
Tetrachloroethene
0.35
0.870784
0.93418
0.950923
85%
91%
92%
2.49
2.67
2.72
9/21/2011 22:20
9/28/2011 20:58
Hartmanl
420First
Tetrachloroethene
0.696707
0.872836
0.896551
118%
132%
133%
3.87
4.85
4.98
9/28/2011 21:00
10/6/2011 21:32
Hartmanl
420First
Tetrachloroethene
0.34
0.830614
0.894781
0.904659
84%
90%
91%
2.44
2.63
2.66
8/17/2011 22:17
8/24/2011 20:58
Hartmanl
422BaseS
Tetrachloroethene
0.53
0.83145
0.83145
0.83145
44%
44%
44%
1.57
1.57
1.57
8/24/2011 21:00
8/31/2011 20:24
Hartmanl
422BaseS
Tetrachloroethene
0.2
0.538404
0.538404
0.538404
92%
92%
92%
2.69
2.69
2.69
8/31/2011 20:26
9/7/2011 20:20
Hartmanl
422BaseS
Tetrachloroethene
0.52
0.745522
0.752522
0.751039
36%
37%
36%
1.43
1.45
1.44
9/7/2011 20:22
9/14/2011 22:27
Hartmanl
422BaseS
Tetrachloroethene
0.89
1.160058
1.160058
1.160058
26%
26%
26%
1.30
1.30
1.30
9/14/2011 22:29
9/21/2011 22:02
Hartmanl
422BaseS
Tetrachloroethene
0.94
1.082404
1.114102
1.136477
14%
17%
19%
1.21
9/21/2011 22:05
9/28/2011 20:39
Hartmanl
422BaseS
Tetrachloroethene
0.6
0.877887
1.033693
1.147923
38%
53%
63%
1.46
1.72
1.91
9/28/2011 20:42
10/6/2011 21:1
Hartmanl
422BaseS
Tetrachloroethene
0.73
1.136962
1.184286
1.228005
44%
47%
51%
1.56
1.62
1.68
Radiello
(MQ/m3)
Start Date Time
Stop Date Time
Period
Location
Compound
GC:
Missing
Values =
1/2 MDL
(MQ/m3)
GC:
Missing
values =
MDL
(Mg/m3)
GC:
Missing
values =
NGC
(Mg/m3)
(continued)
fe3
-------
Table 4-21. Comparison of Online GC to Radiello Results by Week (cont.)
-1
a
_i
o
"o"
1 =
¦D U)
ri S _l
o i i
aj 5 _.
"o"
1 =
¦D U)
o "z
"o"
1 =
¦D U)
o "z
o§a
GC:
GC:
GC:
¦C
"5 II
.c Q
|S
> II
¦= CD
g 2
* II
Radiello
(Mg/m3)
Missing
Values =
1/2 MDL
Missing
values =
MDL
Missing
values =
NGC
Start Date Time
Stop Date Time
Period
Location
Compound
IS 8
TO 3
1 s
1 S
3 «
U >
c >
(Mg/m3)
(Mg/m3)
(Mg/m3)
3 5
u >
3 TO
U >
Ratio (onlii
Calculated
Values = 1,
Ratio (onlii
Calculated
values = M
Ratio (onlii
Calculated
values = N
O £
£|
a. s
O .E
£|
a. s
O .E
£|
a. s
8/17/2011 21:34
8/24/2011 20:44
Hartmanl
422First
Tetrachioroethene
0.21
0.505163
0.511078
0.506379
83%
84%
83%
2.41
2.43
2.41
8/24/2011 20:47
8/31/2011 20:10
Hartmanl
422First
Tetrachioroethene
0.11
0.284921
0.284921
0.284921
89%
89%
89%
2.59
2.59
2.59
8/31/2011 20:12
9/7/2011 20:10
Hartmanl
422First
Tetrachioroethene
0.43
0.460549
0.460549
0.460549
7%
7%
7%
1.07
1.07
1.07
9/7/2011 20:12
9/14/2011 21:57
Hartmanl
422First
Tetrachioroethene
0.77
0.768224
0.768224
0.768224
0%
0%
0%
1.00
1.00
1.00
9/14/2011 21:59
9/21/2011 21:50
Hartmanl
422First
Tetrachioroethene
0.33
0.546832
0.546832
0.546832
49%
49%
49%
1.66
1.66
1.66
9/21/2011 21:53
9/28/2011 20:08
Hartmanl
422First
Tetrachioroethene
0.27
0.504122
0.504122
0.504122
60%
60%
60%
1.87
1.87
1.87
9/28/2011 20:11
10/6/2011 20:57
Hartmanl
422First
Tetrachioroethene
0.43
0.611999
0.611999
0.611999
35%
35%
35%
1.42
1.42
1.42
12/7/2011 23:13
12/14/2011 21:20
Hartman2
420BaseS
Tetrachioroethene
0.41
0.762772
0.762772
0.762772
60%
60%
60%
1.86
1.86
1.86
12/9/2011 17:57
12/15/2011 20:46
Hartman2
420BaseS
Tetrachioroethene
0.34
0.835366
0.835366
0.835366
84%
84%
84%
2.46
2.46
2.46
12/14/2011 21:21
12/22/2011 22:26
Hartman2
420BaseS
Tetrachioroethene
0.16
0.280202
0.304444
0.266737
55%
62%
50%
1.75
1.90
1.67
12/22/2011 22:28
12/28/2011 21:48
Hartman2
420BaseS
Tetrachioroethene
0.14
NGC
NGC
NGC
NGC
NGC
NGC
NGC
NGC
NGC
12/28/2011 21:50
1/4/2012 22:13
Hartman2
420BaseS
Tetrachioroethene
0.11
0.072088
0.078681
0.066222
-42%
-33%
-50%
0.66
0.72
0.60
1/4/2012 22:19
1/11/2012 22:19
Hartman2
420BaseS
Tetrachioroethene
0.2
0.307347
0.307347
0.307347
42%
42%
42%
1.54
1.54
1.54
1/11/2012 20:01
1/18/2012 20:01
Hartman2
420BaseS
Tetrachioroethene
0.19
0.298824
0.298824
0.298824
45%
45%
45%
1.57
1.57
1.57
1/18/2012 20:11
1/25/2012 20:56
Hartman2
420BaseS
Tetrachioroethene
0.21
0.39
0.39
0.39
60%
60%
60%
1.86
1.86
1.86
12/7/2011 22:51
12/14/2011 21:09
Hartman2
420First
Tetrachioroethene
0.31
0.597327
0.597327
0.597327
63%
63%
63%
1.93
1.93
1.93
12/9/2011 17:41
12/15/2011 20:43
Hartman2
420First
Tetrachioroethene
0.25
0.789146
0.789146
0.789146
104%
104%
104%
3.16
3.16
3.16
12/14/2011 21:42
12/22/2011 22:12
Hartman2
420First
Tetrachioroethene
0.09
0.2337
0.2457
0.226224
89%
93%
86%
2.60
2.73
2.51
12/22/2011 22:15
12/28/2011 21:43
Hartman2
420First
Tetrachioroethene
0.1
NGC
NGC
NGC
NGC
NGC
NGC
NGC
NGC
NGC
12/28/2011 21:45
1/4/2012 21:53
Hartman2
420First
Tetrachioroethene
0.081
0.013407
0.013407
0.013407
-143%
-143%
-143%
0.17
0.17
0.17
1/4/2012 21:55
1/11/2012 21:55
Hartman2
420First
Tetrachioroethene
0.17
0.183673
0.183673
0.183673
8%
8%
8%
1.08
1.08
1.08
(continued)
fe3
-------
Table 4-21. Comparison of Online GC to Radiello Results by Week (cont.)
-1
a
_i
o
"o"
1 =
¦D U)
ri S _l
o i i
a> 5 _.
"o"
1 =
¦D U)
o "z
"o"
1 =
¦D U)
o "z
o§a
GC:
GC:
GC:
¦C
"5 II
.c Q
|S
> II
¦= CD
g 2
* II
Radiello
(Mg/m3)
Missing
Values =
1/2 MDL
Missing
values =
MDL
Missing
values =
NGC
Start Date Time
Stop Date Time
Period
Location
Compound
IS 8
TO 3
1 s
1 S
3 «
U >
c >
(Mg/m3)
(Mg/m3)
(Mg/m3)
3 5
u >
3 TO
U >
Ratio (onlii
Calculated
Values = 1,
Ratio (onlii
Calculated
values = M
Ratio (onlii
Calculated
values = N
O £
£|
a. s
O .E
£|
a. s
O .E
£|
a. s
1/11/201219:56
1/18/201219:56
Hartman2
420First
Tetrachioroethene
0.14
0.240686
0.240686
0.240686
53%
53%
53%
1.72
1.72
1.72
1/18/2012 20:01
1/25/2012 20:50
Hartman2
420First
Tetrachioroethene
0.15
0.198229
0.198229
0.198229
28%
28%
28%
1.32
1.32
1.32
12/7/2011 22:10
12/14/2011 20:41
Hartman2
422BaseS
Tetrachioroethene
0.93
1.81
1.816
1.822222
64%
65%
65%
1.95
1.95
1.96
12/9/2011 17:33
12/15/2011 20:35
Hartman2
422BaseS
Tetrachioroethene
0.9
1.862195
1.869512
1.877778
70%
70%
70%
2.07
2.08
2.09
12/14/2011 20:42
12/22/2011 23:16
Hartman2
422BaseS
Tetrachioroethene
0.62
1.550495
1.562376
1.569697
86%
86%
87%
2.50
2.52
2.53
12/22/2011 23:18
12/28/2011 21:37
Hartman2
422BaseS
Tetrachioroethene
0.61
NGC
NGC
NGC
NGC
NGC
NGC
NGC
NGC
NGC
12/28/2011 21:40
1/4/201219:56
Hartman2
422BaseS
Tetrachioroethene
0.64
1.14382
1.150562
1.15
56%
57%
57%
1.79
1.80
1.80
1/4/201219:58
1/11/201219:58
Hartman2
422BaseS
Tetrachioroethene
0.71
1.305102
1.305102
1.305102
59%
59%
59%
1.84
1.84
1.84
1/11/201219:50
1/18/201219:50
Hartman2
422BaseS
Tetrachioroethene
1.1
2.179208
2.185149
2.195
66%
66%
66%
1.98
1.99
2.00
1/18/201219:44
1/25/2012 20:45
Hartman2
422BaseS
Tetrachioroethene
1.1
2.136082
2.142268
2.152083
64%
64%
65%
1.94
1.95
1.96
1/25/2012 20:47
2/1/2012 20:24
Hartman2
422BaseS
Tetrachioroethene
0.81
1.557647
1.575294
1.586667
63%
64%
65%
1.92
1.94
1.96
2/1/2012 20:32
2/8/2012 20:03
Hartman2
422BaseS
Tetrachioroethene
0.57
1.065795
1.072614
1.071149
61%
61%
61%
1.87
1.88
1.88
2/8/2012 20:04
2/15/201218:19
Hartman2
422BaseS
Tetrachioroethene
0.65
1.265215
1.265215
1.265215
64%
64%
64%
1.95
1.95
1.95
12/7/2011 21:19
12/14/2011 20:27
Hartman2
422First
Tetrachioroethene
0.48
0.991188
0.991188
0.991188
69%
69%
69%
2.06
2.06
2.06
12/9/2011 17:12
12/15/2011 20:08
Hartman2
422First
Tetrachioroethene
0.47
1.066341
1.066341
1.066341
78%
78%
78%
2.27
2.27
2.27
12/14/2011 20:28
12/22/2011 22:52
Hartman2
422First
Tetrachioroethene
0.32
0.646768
0.664949
0.648229
68%
70%
68%
2.02
2.08
2.03
12/22/2011 22:54
12/28/2011 21:26
Hartman2
422First
Tetrachioroethene
0.27
NGC
NGC
NGC
NGC
NGC
NGC
NGC
NGC
NGC
12/28/2011 21:29
1/4/201219:23
Hartman2
422First
Tetrachioroethene
0.3
0.400787
0.421011
0.393837
29%
34%
27%
1.34
1.40
1.31
1/4/201219:24
1/11/201219:24
Hartman2
422First
Tetrachioroethene
0.35
0.572449
0.572449
0.572449
48%
48%
48%
1.64
1.64
1.64
1/11/201219:38
1/18/201219:38
Hartman2
422First
Tetrachioroethene
0.54
0.927647
0.933529
0.930891
53%
53%
53%
1.72
1.73
1.72
1/18/201219:23
1/25/2012 20:34
Hartman2
422First
Tetrachioroethene
0.53
0.888557
0.894742
0.891563
51%
51%
51%
1.68
1.69
1.68
(continued)
fe3
-------
Table 4-21. Comparison of Online GC to Radiello Results by Week (cont.)
-1
a
_i
o
"o"
1 =
¦D U)
ri S _l
o i i
aj 5 _.
"o"
1 =
¦D U)
o "z
"o"
1 =
¦D U)
o "z
o§a
GC:
GC:
GC:
¦C
"5 II
.c Q
|S
> II
¦= CD
g 2
* II
Radiello
(Mg/m3)
Missing
Values =
1/2 MDL
Missing
values =
MDL
Missing
values =
NGC
Start Date Time
Stop Date Time
Period
Location
Compound
IS 8
TO 3
1 s
1 S
3 «
U >
c >
(Mg/m3)
(Mg/m3)
(Mg/m3)
3 5
u >
3 TO
U >
Ratio (onlii
Calculated
Values = 1,
Ratio (onlii
Calculated
values = M
Ratio (onlii
Calculated
values = N
O £
£|
a. s
O .E
£|
a. s
O .E
£|
a. s
1/25/2012 20:37
2/1/2012 20:05
Hartman2
422First
Tetrachioroethene
0.42
0.644571
0.644571
0.644571
42%
42%
42%
1.53
1.53
1.53
2/1/2012 20:13
2/8/201219:57
Hartman2
422First
Tetrachioroethene
0.33
0.460449
0.460449
0.460449
33%
33%
33%
1.40
1.40
1.40
2/8/2012 20:00
2/15/201218:14
Hartman2
422First
Tetrachioroethene
0.32
0.437412
0.437412
0.437412
31%
31%
31%
1.37
1.37
1.37
12/7/2011 22:33
12/14/2011 22:00
Hartman2
Outside
Tetrachioroethene
0.23
0.425545
0.425545
0.425545
60%
60%
60%
1.85
1.85
1.85
12/9/2011 18:17
12/15/2011 21:12
Hartman2
Outside
Tetrachioroethene
0.21
0.629277
0.629277
0.629277
100%
100%
100%
3.00
3.00
3.00
12/14/2011 22:01
12/22/2011 21:49
Hartman2
Outside
Tetrachioroethene
0.094
0.492198
0.492198
0.492198
136%
136%
136%
5.24
5.24
5.24
12/22/2011 21:53
12/28/2011 21:59
Hartman2
Outside
Tetrachioroethene
0.091
NGC
NGC
NGC
NGC
NGC
NGC
NGC
NGC
NGC
12/28/2011 22:01
1/4/2012 21:37
Hartman2
Outside
Tetrachioroethene
0.074
0.141099
0.167473
0.12
62%
77%
47%
1.91
2.26
1.62
1/4/2012 21:39
1/11/2012 21:39
Hartman2
Outside
Tetrachioroethene
0.18
0.274796
0.274796
0.274796
42%
42%
42%
1.53
1.53
1.53
1/11/2012 20:13
1/18/2012 20:13
Hartman2
Outside
Tetrachioroethene
0.067
0.077723
0.077723
0.077723
15%
15%
15%
1.16
1.16
1.16
1/18/2012 20:32
1/25/2012 21:09
Hartman2
Outside
Tetrachioroethene
0.08
0.245104
0.251354
0.241368
102%
103%
100%
3.06
3.14
3.02
1/25/2012 21:12
2/1/2012 21:02
Hartman2
Outside
Tetrachioroethene
0.11
0.274706
0.274706
0.274706
86%
86%
86%
2.50
2.50
2.50
2/1/2012 21:07
2/8/2012 20:08
Hartman2
Outside
Tetrachioroethene
0.13
0.262045
0.262045
0.262045
67%
67%
67%
2.02
2.02
2.02
2/8/2012 20:10
2/15/201217:58
Hartman2
Outside
Tetrachioroethene
0.071
0.184235
0.184235
0.184235
89%
89%
89%
2.59
2.59
2.59
12/19/2012 23:30
12/26/201215:49
Hartman3
420BaseS
Tetrachioroethene
0.17
0.26644
0.26644
0.26644
44%
44%
44%
1.57
1.57
1.57
12/19/2012 23:33
12/26/201215:51
Hartman3
420BaseS
Tetrachioroethene
0.24
0.26644
0.26644
0.26644
10%
10%
10%
1.11
1.11
1.11
12/26/201215:50
1/2/2013 21:30
Hartman3
420BaseS
Tetrachioroethene
0.16
0.227258
0.227258
0.227258
35%
35%
35%
1.42
1.42
1.42
1/2/2013 21:32
1/9/2013 20:24
Hartman3
420BaseS
Tetrachioroethene
0.25
0.255719
0.255719
0.255719
2%
2%
2%
1.02
1.02
1.02
1/9/2013 20:26
1/16/201319:37
Hartman3
420BaseS
Tetrachioroethene
0.35
0.326374
0.326374
0.326374
-7%
-7%
-7%
0.93
0.93
0.93
1/16/201319:39
1/23/2013 21:11
Hartman3
420BaseS
Tetrachioroethene
0.36
0.473815
0.481106
0.476449
27%
29%
28%
1.32
1.34
1.32
1/23/2013 21:13
1/30/201318:12
Hartman3
420BaseS
Tetrachioroethene
1
0.944917
0.944917
0.944917
-6%
-6%
-6%
0.94
0.94
0.94
(continued)
fe3
-------
Table 4-21. Comparison of Online GC to Radiello Results by Week (cont.)
-1
a
_i
o
"o"
1 =
¦D U)
ri S _l
o i i
aj 5 _.
"o"
1 =
¦D U)
o "z
"o"
1 =
¦D U)
o "z
o§a
GC:
GC:
GC:
¦C
"5 II
.c Q
|S
> II
¦= CD
g 2
* II
Radiello
(Mg/m3)
Missing
Values =
1/2 MDL
Missing
values =
MDL
Missing
values =
NGC
Start Date Time
Stop Date Time
Period
Location
Compound
IS 8
TO 3
1 s
1 S
3 «
U >
c >
(Mg/m3)
(Mg/m3)
(Mg/m3)
3 5
u >
3 TO
U >
Ratio (onlii
Calculated
Values = 1,
Ratio (onlii
Calculated
values = M
Ratio (onlii
Calculated
values = N
O £
£|
a. s
O .E
£|
a. s
O .E
£|
a. s
1/23/2013 21:18
1/30/201318:18
Hartman3
420BaseS
Tetrachioroethene
1
0.944917
0.944917
0.944917
-6%
-6%
-6%
0.94
0.94
0.94
1/30/201318:14
2/6/2013 0:59
Hartman3
420BaseS
Tetrachloroethene
0.36
0.349814
0.349814
0.349814
-3%
-3%
-3%
0.97
0.97
0.97
1/30/201318:20
2/6/20131:03
Hartman3
420BaseS
Tetrachioroethene
0.34
0.349814
0.349814
0.349814
3%
3%
3%
1.03
1.03
1.03
2/6/20131:00
2/13/201319:50
Hartman3
420BaseS
Tetrachloroethene
0.29
0.330543
0.330543
0.330543
13%
13%
13%
1.14
1.14
1.14
2/6/20131:04
2/13/201319:54
Hartman3
420BaseS
Tetrachioroethene
0.28
0.330543
0.330543
0.330543
17%
17%
17%
1.18
1.18
1.18
2/13/201319:52
2/20/2013 21:35
Hartman3
420BaseS
Tetrachloroethene
0.67
0.814083
0.814083
0.814083
19%
19%
19%
1.22
1.22
1.22
2/13/201319:56
2/20/2013 21:39
Hartman3
420BaseS
Tetrachloroethene
0.64
0.814083
0.814083
0.814083
24%
24%
24%
1.27
1.27
1.27
3/6/2013 20:20
3/14/2013 22:53
Hartman3
420BaseS
Tetrachloroethene
0.28
0.335229
0.335229
0.335229
18%
18%
18%
1.20
1.20
1.20
3/6/2013 20:24
3/14/2013 23:04
Hartman3
420BaseS
Tetrachloroethene
0.27
0.335229
0.335229
0.335229
22%
22%
22%
1.24
1.24
1.24
12/19/2012 23:26
12/26/201215:45
Hartman3
420First
Tetrachloroethene
0.13
0.248347
0.248347
0.248347
63%
63%
63%
1.91
1.91
1.91
12/26/201215:46
1/2/2013 21:20
Hartman3
420First
Tetrachloroethene
0.12
0.193949
0.193949
0.193949
47%
47%
47%
1.62
1.62
1.62
1/2/2013 21:22
1/9/2013 20:19
Hartman3
420First
Tetrachloroethene
0.18
0.18147
0.18147
0.18147
1%
1%
1%
1.01
1.01
1.01
1/9/2013 20:22
1/16/201319:32
Hartman3
420First
Tetrachloroethene
0.21
0.263481
0.263481
0.263481
23%
23%
23%
1.25
1.25
1.25
1/16/2013 19:34
1/23/2013 21:05
Hartman3
420First
Tetrachloroethene
0.22
0.202046
0.202046
0.202046
-9%
-9%
-9%
0.92
0.92
0.92
1/23/2013 21:07
1/30/201318:07
Hartman3
420First
Tetrachloroethene
0.73
0.8585
0.8585
0.8585
16%
16%
16%
1.18
1.18
1.18
1/30/2013 18:09
2/6/2013 0:51
Hartman3
420First
Tetrachloroethene
0.2
0.114152
0.122486
0.1084
-55%
-48%
-59%
0.57
0.61
0.54
2/6/2013 0:54
2/13/2013 19:44
Hartman3
420First
Tetrachloroethene
0.16
0.401728
0.401728
0.401728
86%
86%
86%
2.51
2.51
2.51
2/13/2013 19:46
2/20/2013 21:29
Hartman3
420First
Tetrachloroethene
0.34
0.288124
0.288124
0.288124
-17%
-17%
-17%
0.85
0.85
0.85
3/6/2013 20:11
3/14/2013 22:47
Hartman3
420First
Tetrachloroethene
0.22
0.261475
0.261475
0.261475
17%
17%
17%
1.19
1.19
1.19
12/19/2012 23:20
12/26/2012 15:39
Hartman3
422BaseS
Tetrachloroethene
0.09
0.87325
0.87325
0.87325
163%
163%
163%
9.70
9.70
9.70
12/26/201215:39
1/2/2013 21:13
Hartman3
422BaseS
Tetrachloroethene
0.68
0.902784
0.902784
0.902784
28%
28%
28%
1.33
1.33
1.33
(continued)
fe3
-------
Table 4-21. Comparison of Online GC to Radiello Results by Week (cont.)
-1
a
_i
o
"o"
1 =
¦D U)
ri S _l
o i i
a> 5 _.
"o"
1 =
¦D U)
o "z
"o"
1 =
¦D U)
o "z
o§a
GC:
GC:
GC:
¦C
"5 II
.c Q
|S
> II
¦= CD
g 2
* II
Radiello
(Mg/m3)
Missing
Values =
1/2 MDL
Missing
values =
MDL
Missing
values =
NGC
Start Date Time
Stop Date Time
Period
Location
Compound
IS 8
TO 3
1 s
1 S
3 «
U >
c >
(Mg/m3)
(Mg/m3)
(Mg/m3)
3 5
u >
3 TO
U >
Ratio (onlii
Calculated
Values = 1,
Ratio (onlii
Calculated
values = M
Ratio (onlii
Calculated
values = N
O £
£|
a. s
O .E
£|
a. s
O .E
£|
a. s
1/2/2013 21:15
1/9/2013 20:13
Hartman3
422BaseS
Tetrachloroethene
1.2
1.458409
1.458409
1.458409
19%
19%
19%
1.22
1.22
1.22
1/9/2013 20:15
1/16/2013 19:24
Hartman3
422BaseS
Tetrachloroethene
1.2
1.239126
1.239126
1.239126
3%
3%
3%
1.03
1.03
1.03
1/16/201319:27
1/23/2013 20:58
Hartman3
422BaseS
Tetrachloroethene
1.8
2.545342
2.545342
2.545342
34%
34%
34%
1.41
1.41
1.41
1/23/2013 21:01
1/30/201318:01
Hartman3
422BaseS
Tetrachloroethene
4.1
4.94923
4.94923
4.94923
19%
19%
19%
1.21
1.21
1.21
1/30/201318:03
2/6/2013 0:41
Hartman3
422BaseS
Tetrachloroethene
2.7
3.791853
3.791853
3.791853
34%
34%
34%
1.40
1.40
1.40
2/6/2013 0:43
2/13/201319:38
Hartman3
422BaseS
Tetrachloroethene
0.98
1.075208
1.075208
1.075208
9%
9%
9%
1.10
1.10
1.10
2/13/201319:41
2/20/2013 21:22
Hartman3
422BaseS
Tetrachloroethene
2.2
2.383663
2.383663
2.383663
8%
8%
8%
1.08
1.08
1.08
3/6/2013 20:05
3/14/2013 22:41
Hartman3
422BaseS
Tetrachloroethene
1.9
2.358331
2.358331
2.358331
22%
22%
22%
1.24
1.24
1.24
12/19/2012 23:12
12/26/201215:31
Hartman3
422First
Tetrachloroethene
0.32
0.448193
0.448193
0.448193
33%
33%
33%
1.40
1.40
1.40
12/26/201215:33
1/2/2013 21:01
Hartman3
422First
Tetrachloroethene
0.3
0.450698
0.450698
0.450698
40%
40%
40%
1.50
1.50
1.50
1/2/2013 21:05
1/9/2013 20:02
Hartman3
422First
Tetrachloroethene
0.62
0.678085
0.678085
0.678085
9%
9%
9%
1.09
1.09
1.09
1/9/2013 20:04
1/16/201319:12
Hartman3
422First
Tetrachloroethene
0.57
0.585404
0.585404
0.585404
3%
3%
3%
1.03
1.03
1.03
1/16/2013 19:13
1/23/2013 20:47
Hartman3
422First
Tetrachloroethene
0.98
1.116247
1.116247
1.116247
13%
13%
13%
1.14
1.14
1.14
1/23/2013 20:48
1/30/2013 17:51
Hartman3
422First
Tetrachloroethene
1.6
1.991854
1.991854
1.991854
22%
22%
22%
1.24
1.24
1.24
1/30/2013 17:52
2/6/2013 0:21
Hartman3
422First
Tetrachloroethene
1.2
1.658681
1.658681
1.658681
32%
32%
32%
1.38
1.38
1.38
2/6/2013 0:24
2/13/2013 19:28
Hartman3
422First
Tetrachloroethene
0.34
0.524513
0.524513
0.524513
43%
43%
43%
1.54
1.54
1.54
2/13/2013 19:31
2/20/2013 21:10
Hartman3
422First
Tetrachloroethene
0.76
0.871925
0.871925
0.871925
14%
14%
14%
1.15
1.15
1.15
3/6/201319:44
3/14/2013 22:28
Hartman3
422First
Tetrachloroethene
0.76
0.973867
0.973867
0.973867
25%
25%
25%
1.28
1.28
1.28
12/19/2012 23:42
12/26/2012 15:54
Hartman3
Outside
Tetrachloroethene
0.1
2.867612
2.867612
2.867612
187%
187%
187%
28.68
28.68
28.68
12/26/2012 16:00
1/2/2013 21:53
Hartman3
Outside
Tetrachloroethene
0.095
1.594169
1.594169
1.594169
178%
178%
178%
16.78
16.78
16.78
1/2/2013 21:55
1/9/2013 20:40
Hartman3
Outside
Tetrachloroethene
0.15
0.17637
0.17637
0.17637
16%
16%
16%
1.18
1.18
1.18
(continued)
fe3
-------
Table 4-21. Comparison of Online GC to Radiello Results by Week (cont.)
-1
a
_i
o
"o"
1 =
¦D U)
ri S _l
o i i
aj 5 _.
"o"
1 =
¦D U)
o "z
"o"
1 =
¦D U)
o "z
o§a
GC:
GC:
GC:
¦C
"5 II
.c Q
|S
> II
¦= CD
g 2
* II
Radiello
(Mg/m3)
Missing
Values =
1/2 MDL
Missing
values =
MDL
Missing
values =
NGC
Start Date Time
Stop Date Time
Period
Location
Compound
IS 8
TO 3
1 s
1 S
3 «
U >
c >
(Mg/m3)
(Mg/m3)
(Mg/m3)
3 5
u >
3 TO
U >
Ratio (onlii
Calculated
Values = 1,
Ratio (onlii
Calculated
values = M
Ratio (onlii
Calculated
values = N
O £
£|
a. s
O .E
£|
a. s
O .E
£|
a. s
1/9/2013 20:42
1/16/201319:57
Hartman3
Outside
Tetrachioroethene
0.13
0.268352
0.268352
0.268352
69%
69%
69%
2.06
2.06
2.06
1/16/201319:59
1/23/2013 21:29
Hartman3
Outside
Tetrachioroethene
0.095
0.154371
0.154371
0.154371
48%
48%
48%
1.62
1.62
1.62
1/23/2013 21:32
1/30/201318:32
Hartman3
Outside
Tetrachioroethene
0.14
0.234515
0.234515
0.234515
50%
50%
50%
1.68
1.68
1.68
1/30/201318:34
2/6/20131:27
Hartman3
Outside
Tetrachioroethene
0.095
0.168551
0.168551
0.168551
56%
56%
56%
1.77
1.77
1.77
2/6/20131:29
2/13/2013 20:03
Hartman3
Outside
Tetrachioroethene
0.095
0.388508
0.388508
0.388508
121%
121%
121%
4.09
4.09
4.09
2/13/2013 20:06
2/20/2013 21:51
Hartman3
Outside
Tetrachioroethene
0.095
1.506992
1.506992
1.506992
176%
176%
176%
15.86
15.86
15.86
3/6/2013 20:42
3/14/2013 23:15
Hartman3
Outside
Tetrachioroethene
0.08
0.192146
0.192146
0.192146
82%
82%
82%
2.40
2.40
2.40
8/17/2011 22:55
8/24/2011 21:21
Hartmanl
420BaseS
Chloroform
0.13
0.200157
0.200157
0.200157
42%
42%
42%
1.54
1.54
1.54
8/24/2011 21:22
8/31/2011 20:51
Hartmanl
420BaseS
Chloroform
0.19
0.250638
0.250638
0.250638
28%
28%
28%
1.32
1.32
1.32
8/31/2011 20:52
9/7/2011 20:34
Hartmanl
420BaseS
Chloroform
0.34
0.33628
0.33628
0.33628
-1%
-1%
-1%
0.99
0.99
0.99
9/7/2011 20:36
9/14/2011 23:09
Hartmanl
420BaseS
Chloroform
0.26
0.297065
0.297065
0.297065
13%
13%
13%
1.14
1.14
1.14
9/14/2011 23:11
9/21/2011 22:23
Hartmanl
420BaseS
Chloroform
0.25
0.681553
0.681553
0.681553
93%
93%
93%
2.73
2.73
2.73
9/21/2011 22:25
9/28/2011 21:09
Hartmanl
420BaseS
Chloroform
0.089
0.45143
0.45143
0.45143
134%
134%
134%
5.07
5.07
5.07
9/28/2011 21:12
10/6/2011 21:41
Hartmanl
420BaseS
Chloroform
0.34
0.63844
0.666218
0.646583
61%
65%
62%
1.88
1.96
1.90
8/17/2011 22:36
8/24/2011 21:14
Hartmanl
420First
Chloroform
0.18
0.163392
0.163392
0.163392
-10%
-10%
-10%
0.91
0.91
0.91
8/24/2011 21:16
8/31/2011 20:44
Hartmanl
420First
Chloroform
0.24
0.194262
0.194262
0.194262
-21%
-21%
-21%
0.81
0.81
0.81
8/31/2011 20:46
9/7/2011 20:27
Hartmanl
420First
Chloroform
0.16
0.236445
0.236445
0.236445
39%
39%
39%
1.48
1.48
1.48
9/7/2011 20:29
9/14/2011 22:48
Hartmanl
420First
Chloroform
0.21
0.283872
0.283872
0.283872
30%
30%
30%
1.35
1.35
1.35
9/14/2011 22:49
9/21/2011 22:18
Hartmanl
420First
Chloroform
0.26
0.694797
0.694797
0.694797
91%
91%
91%
2.67
2.67
2.67
9/21/2011 22:20
9/28/2011 20:58
Hartmanl
420First
Chloroform
0.094
0.573214
0.573214
0.573214
144%
144%
144%
6.10
6.10
6.10
9/28/2011 21:00
10/6/2011 21:32
Hartmanl
420First
Chloroform
0.22
0.702489
0.723322
0.711292
105%
107%
106%
3.19
3.29
3.23
(continued)
fe3
-------
Table 4-21. Comparison of Online GC to Radiello Results by Week (cont.)
-1
a
_i
o
"o"
1 =
¦D U)
ri S _l
o i i
a> 5 _.
"o"
1 =
¦D U)
o "z
"o"
1 =
¦D U)
o "z
o§a
GC:
GC:
GC:
¦C 04
"5 II
.c Q
|S
> II
¦= CD
g 2
* II
Radiello
(Mg/m3)
Missing
Values =
1/2 MDL
Missing
values =
MDL
Missing
values =
NGC
Start Date Time
Stop Date Time
Period
Location
Compound
IS 8
TO 3
1 s
1 S
3 «
U >
c >
(Mg/m3)
(Mg/m3)
(Mg/m3)
3 5
u >
3 TO
U >
Ratio (onlii
Calculated
Values = 1,
Ratio (onlii
Calculated
values = M
Ratio (onlii
Calculated
values = N
O £
£|
a. s
O .E
£|
a. s
O .E
£|
a. s
8/17/2011 22:17
8/24/2011 20:58
Hartmanl
422BaseS
Chloroform
0.17
0.225866
0.225866
0.225866
28%
28%
28%
1.33
1.33
1.33
8/24/2011 21:00
8/31/2011 20:24
Hartmanl
422BaseS
Chloroform
0.099
0.161171
0.161171
0.161171
48%
48%
48%
1.63
1.63
1.63
8/31/2011 20:26
9/7/2011 20:20
Hartmanl
422BaseS
Chloroform
0.49
0.431763
0.431763
0.431763
-13%
-13%
-13%
0.88
0.88
0.88
9/7/2011 20:22
9/14/2011 22:27
Hartmanl
422BaseS
Chloroform
0.71
0.72271
0.72271
0.72271
2%
2%
2%
1.02
1.02
1.02
9/14/2011 22:29
9/21/2011 22:02
Hartmanl
422BaseS
Chloroform
0.38
1.094928
1.094928
1.094928
97%
97%
97%
2.88
2.88
2.88
9/21/2011 22:05
9/28/2011 20:39
Hartmanl
422BaseS
Chloroform
0.22
0.768588
0.768588
0.768588
111%
111%
111%
3.49
3.49
3.49
9/28/2011 20:42
10/6/2011 21:18
Hartmanl
422BaseS
Chloroform
0.34
0.90011
0.921236
0.917761
90%
92%
92%
2.65
2.71
2.70
8/17/2011 21:34
8/24/2011 20:44
Hartmanl
422First
Chloroform
0.15
0.129456
0.129456
0.129456
-15%
-15%
-15%
0.86
0.86
0.86
8/24/2011 20:47
8/31/2011 20:10
Hartmanl
422First
Chloroform
0.12
0.100367
0.100367
0.100367
-18%
-18%
-18%
0.84
0.84
0.84
8/31/2011 20:12
9/7/2011 20:10
Hartmanl
422First
Chloroform
0.46
0.23895
0.23895
0.23895
-63%
-63%
-63%
0.52
0.52
0.52
9/7/2011 20:12
9/14/2011 21:57
Hartmanl
422First
Chloroform
0.59
0.449548
0.449548
0.449548
-27%
-27%
-27%
0.76
0.76
0.76
9/14/2011 21:59
9/21/2011 21:50
Hartmanl
422First
Chloroform
0.23
0.507337
0.507337
0.507337
75%
75%
75%
2.21
2.21
2.21
9/21/2011 21:53
9/28/2011 20:08
Hartmanl
422First
Chloroform
0.14
0.422027
0.422027
0.422027
100%
100%
100%
3.01
3.01
3.01
9/28/2011 20:11
10/6/2011 20:57
Hartmanl
422First
Chloroform
0.3
0.514956
0.514956
0.514956
53%
53%
53%
1.72
1.72
1.72
12/7/2011 23:13
12/14/2011 21:20
Hartman2
420BaseS
Chloroform
0.3
0.583333
0.583333
0.583333
64%
64%
64%
1.94
1.94
1.94
12/9/2011 17:57
12/15/2011 20:46
Hartman2
420BaseS
Chloroform
0.26
0.602273
0.602273
0.602273
79%
79%
79%
2.32
2.32
2.32
12/14/2011 21:21
12/22/2011 22:26
Hartman2
420BaseS
Chloroform
0.12
0.71
0.71
0.71
142%
142%
142%
5.92
5.92
5.92
12/22/2011 22:28
12/28/2011 21:48
Hartman2
420BaseS
Chloroform
0.15
NGC
NGC
NGC
NGC
NGC
NGC
NGC
NGC
NGC
12/28/2011 21:50
1/4/2012 22:13
Hartman2
420BaseS
Chloroform
0.1
0.350989
0.697143
0.44
111%
150%
126%
3.51
6.97
4.40
1/4/2012 22:19
1/11/2012 22:19
Hartman2
420BaseS
Chloroform
0.14
0.343163
0.521735
0.336042
84%
115%
82%
2.45
3.73
2.40
1/11/2012 20:01
1/18/2012 20:01
Hartman2
420BaseS
Chloroform
0.14
0.301863
0.336176
0.29663
73%
82%
72%
2.16
2.40
2.12
1/18/2012 20:11
1/25/2012 20:56
Hartman2
420BaseS
Chloroform
0.11
0.319789
0.341895
0.317753
98%
103%
97%
2.91
3.11
2.89
(continued)
fe3
-------
Table 4-21. Comparison of Online GC to Radiello Results by Week (cont.)
-1
a
_i
o
"o"
1 =
¦D U)
ri S _l
o i i
aj 5 _.
"o"
1 =
¦D U)
o "z
"o"
1 =
¦D U)
o "z
o§a
GC:
GC:
GC:
¦C
"5 II
.c Q
|S
> II
¦= CD
g 2
* II
Radiello
(Mg/m3)
Missing
Values =
1/2 MDL
Missing
values =
MDL
Missing
values =
NGC
Start Date Time
Stop Date Time
Period
Location
Compound
IS 8
TO 3
1 s
1 S
3 «
U >
c >
(Mg/m3)
(Mg/m3)
(Mg/m3)
3 5
u >
3 TO
U >
Ratio (onlii
Calculated
Values = 1,
Ratio (onlii
Calculated
values = M
Ratio (onlii
Calculated
values = N
O £
£|
a. s
O .E
£|
a. s
O .E
£|
a. s
12/7/2011 22:51
12/14/2011 21:09
Hartman2
420First
Chloroform
0.22
0.380952
0.380952
0.380952
54%
54%
54%
1.73
1.73
1.73
12/9/2011 17:41
12/15/2011 20:43
Hartman2
420First
Chloroform
0.18
0.4
0.4
0.4
76%
76%
76%
2.22
2.22
2.22
12/14/2011 21:42
12/22/2011 22:12
Hartman2
420First
Chloroform
0.067
0.325
0.325
0.325
132%
132%
132%
4.85
4.85
4.85
12/22/2011 22:15
12/28/2011 21:43
Hartman2
420First
Chloroform
0.12
NGC
NGC
NGC
NGC
NGC
NGC
NGC
NGC
NGC
12/28/2011 21:45
1/4/2012 21:53
Hartman2
420First
Chloroform
0.092
0.321978
0.598901
0.215789
111%
147%
80%
3.50
6.51
2.35
1/4/2012 21:55
1/11/2012 21:55
Hartman2
420First
Chloroform
0.12
0.313673
0.477959
0.281538
89%
120%
80%
2.61
3.98
2.35
1/11/201219:56
1/18/201219:56
Hartman2
420First
Chloroform
0.11
0.301765
0.418431
0.277647
93%
117%
86%
2.74
3.80
2.52
1/18/2012 20:01
1/25/2012 20:50
Hartman2
420First
Chloroform
0.087
0.29617
0.378085
0.279722
109%
125%
105%
3.40
4.35
3.22
12/7/2011 22:10
12/14/2011 20:41
Hartman2
422BaseS
Chloroform
0.86
1.081
1.081
1.081
23%
23%
23%
1.26
1.26
1.26
12/9/2011 17:33
12/15/2011 20:35
Hartman2
422BaseS
Chloroform
0.8
1.096296
1.096296
1.096296
31%
31%
31%
1.37
1.37
1.37
12/14/2011 20:42
12/22/2011 23:16
Hartman2
422BaseS
Chloroform
0.61
0.797895
0.834737
0.850588
27%
31%
33%
1.31
1.37
1.39
12/22/2011 23:18
12/28/2011 21:37
Hartman2
422BaseS
Chloroform
0.46
NGC
NGC
NGC
NGC
NGC
NGC
NGC
NGC
NGC
12/28/2011 21:40
1/4/201219:56
Hartman2
422BaseS
Chloroform
0.71
0.973596
0.977528
0.980682
31%
32%
32%
1.37
1.38
1.38
1/4/201219:58
1/11/201219:58
Hartman2
422BaseS
Chloroform
0.68
0.951656
0.951656
0.951656
33%
33%
33%
1.40
1.40
1.40
1/11/201219:50
1/18/201219:50
Hartman2
422BaseS
Chloroform
0.73
1.074851
1.078317
1.0821
38%
39%
39%
1.47
1.48
1.48
1/18/201219:44
1/25/2012 20:45
Hartman2
422BaseS
Chloroform
0.69
1.017216
1.020825
1.024167
38%
39%
39%
1.47
1.48
1.48
1/25/2012 20:47
2/1/2012 20:24
Hartman2
422BaseS
Chloroform
0.51
0.81
0.820294
0.823939
45%
47%
47%
1.59
1.61
1.62
2/1/2012 20:32
2/8/2012 20:03
Hartman2
422BaseS
Chloroform
0.5
0.667273
0.67125
0.67092
29%
29%
29%
1.33
1.34
1.34
2/8/2012 20:04
2/15/201218:19
Hartman2
422BaseS
Chloroform
0.61
0.764167
0.764167
0.764167
22%
22%
22%
1.25
1.25
1.25
12/7/2011 21:19
12/14/2011 20:27
Hartman2
422First
Chloroform
0.39
0.639375
0.639375
0.639375
48%
48%
48%
1.64
1.64
1.64
12/9/2011 17:12
12/15/2011 20:08
Hartman2
422First
Chloroform
0.36
0.648533
0.648533
0.648533
57%
57%
57%
1.80
1.80
1.80
(continued)
fe3
-------
Table 4-21. Comparison of Online GC to Radiello Results by Week (cont.)
-1
a
_i
o
"o"
1 =
¦D U)
ri S _l
o i i
aj 5 _.
"o"
1 =
¦D U)
o "z
"o"
1 =
¦D U)
o "z
o§a
GC:
GC:
GC:
¦C
"5 II
.c Q
|S
> II
¦= CD
g 2
* II
Radiello
(Mg/m3)
Missing
Values =
1/2 MDL
Missing
values =
MDL
Missing
values =
NGC
Start Date Time
Stop Date Time
Period
Location
Compound
IS 8
TO 3
1 s
1 S
3 «
U >
(Mg/m3)
(Mg/m3)
(Mg/m3)
3 5
u >
3 TO
U >
Ratio (onlii
Calculated
Values = 1,
Ratio (onlii
Calculated
values = M
Ratio (onlii
Calculated
values = N
O £
£|
a. s
O .E
£|
a. s
O .E
£|
a. s
12/14/2011 20:28
12/22/2011 22:52
Hartman2
422First
Chloroform
0.26
0.533019
0.533019
0.533019
69%
69%
69%
2.05
2.05
2.05
12/22/2011 22:54
12/28/2011 21:26
Hartman2
422First
Chloroform
0.22
NGC
NGC
NGC
NGC
NGC
NGC
NGC
NGC
NGC
12/28/2011 21:29
1/4/201219:23
Hartman2
422First
Chloroform
0.3
0.491573
0.629213
0.583333
48%
71%
64%
1.64
2.10
1.94
1/4/201219:24
1/11/201219:24
Hartman2
422First
Chloroform
0.22
0.516939
0.531224
0.524043
81%
83%
82%
2.35
2.41
2.38
1/11/201219:38
1/18/201219:38
Hartman2
422First
Chloroform
0.36
0.552549
0.55598
0.554554
42%
43%
43%
1.53
1.54
1.54
1/18/201219:23
1/25/2012 20:34
Hartman2
422First
Chloroform
0.31
0.537835
0.537835
0.537835
54%
54%
54%
1.73
1.73
1.73
1/25/2012 20:37
2/1/2012 20:05
Hartman2
422First
Chloroform
0.26
0.520286
0.520286
0.520286
67%
67%
67%
2.00
2.00
2.00
2/1/2012 20:13
2/8/2012 19:57
Hartman2
422First
Chloroform
0.25
0.395843
0.395843
0.395843
45%
45%
45%
1.58
1.58
1.58
2/8/2012 20:00
2/15/2012 18:14
Hartman2
422First
Chloroform
0.28
0.420235
0.420235
0.420235
40%
40%
40%
1.50
1.50
1.50
12/7/2011 22:33
12/14/2011 22:00
Hartman2
Outside
Chloroform
0.13
0.511667
0.511667
0.511667
119%
119%
119%
3.94
3.94
3.94
12/9/2011 18:17
12/15/2011 21:12
Hartman2
Outside
Chloroform
0.12
0.569
0.569
0.569
130%
130%
130%
4.74
4.74
4.74
12/14/2011 22:01
12/22/2011 21:49
Hartman2
Outside
Chloroform
0.051
0.612
0.612
0.612
169%
169%
169%
12.00
12.00
12.00
12/22/2011 21:53
12/28/2011 21:59
Hartman2
Outside
Chloroform
0.087
NGC
NGC
NGC
NGC
NGC
NGC
NGC
NGC
NGC
12/28/2011 22:01
1/4/2012 21:37
Hartman2
Outside
Chloroform
0.093
0.35
0.7
NGC
116%
153%
NGC
3.76
7.53
NGC
1/4/2012 21:39
1/11/2012 21:39
Hartman2
Outside
Chloroform
0.1
0.323878
0.538163
0.282632
106%
137%
95%
3.24
5.38
2.83
1/11/2012 20:13
1/18/2012 20:13
Hartman2
Outside
Chloroform
0.09
0.317525
0.52198
0.271905
112%
141%
101%
3.53
5.80
3.02
1/18/2012 20:32
1/25/2012 21:09
Hartman2
Outside
Chloroform
0.075
0.345313
0.392708
0.344578
129%
136%
128%
4.60
5.24
4.59
1/25/2012 21:12
2/1/2012 21:02
Hartman2
Outside
Chloroform
0.075
0.374706
0.374706
0.374706
133%
133%
133%
5.00
5.00
5.00
2/1/2012 21:07
2/8/2012 20:08
Hartman2
Outside
Chloroform
0.1
0.337955
0.337955
0.337955
109%
109%
109%
3.38
3.38
3.38
2/8/2012 20:10
2/15/201217:58
Hartman2
Outside
Chloroform
0.1
0.33119
0.339524
0.330732
107%
109%
107%
3.31
3.40
3.31
(continued)
fe3
-------
Table 4-21. Comparison of Online GC to Radiello Results by Week (cont.)
-1
a
_i
o
"o"
1 =
¦D U)
ri S _l
o i i
a> 5 _.
"o"
1 =
¦D U)
o "z
"o"
1 =
¦D U)
o "z
o§a
GC:
GC:
GC:
¦C
"5 II
.c Q
|S
> II
¦= CD
g 2
* II
Radiello
(Mg/m3)
Missing
Values =
1/2 MDL
Missing
values =
MDL
Missing
values =
NGC
Start Date Time
Stop Date Time
Period
Location
Compound
IS 8
TO 3
1 s
1 S
3 «
U >
c >
(Mg/m3)
(Mg/m3)
(Mg/m3)
3 5
u >
3 TO
U >
Ratio (onlii
Calculated
Values = 1,
Ratio (onlii
Calculated
values = M
Ratio (onlii
Calculated
values = N
O £
£|
a. s
O .E
£|
a. s
O .E
£|
a. s
12/19/2012 23:30
12/26/201215:49
Hartman3
420BaseS
Chloroform
0.14
0.348638
0.690305
0.2928
85%
133%
71%
2.49
4.93
2.09
12/19/2012 23:33
12/26/201215:51
Hartman3
420BaseS
Chloroform
0.075
0.348638
0.690305
0.2928
129%
161%
118%
4.65
9.20
3.90
12/26/201215:50
1/2/2013 21:30
Hartman3
420BaseS
Chloroform
0.13
0.35
0.7
NGC
92%
137%
NGC
2.69
5.38
NGC
1/2/2013 21:32
1/9/2013 20:24
Hartman3
420BaseS
Chloroform
0.13
0.348713
0.691266
0.2895
91%
137%
76%
2.68
5.32
2.23
1/9/2013 20:26
1/16/201319:37
Hartman3
420BaseS
Chloroform
0.16
0.348979
0.661745
0.3404
74%
122%
72%
2.18
4.14
2.13
1/16/201319:39
1/23/2013 21:11
Hartman3
420BaseS
Chloroform
0.19
0.363157
0.633611
0.40789
63%
108%
73%
1.91
3.33
2.15
1/23/2013 21:13
1/30/201318:12
Hartman3
420BaseS
Chloroform
0.16
0.351707
0.636591
0.359175
75%
120%
77%
2.20
3.98
2.24
1/23/2013 21:18
1/30/201318:18
Hartman3
420BaseS
Chloroform
0.16
0.351707
0.636591
0.359175
75%
120%
77%
2.20
3.98
2.24
1/30/201318:14
2/6/2013 0:59
Hartman3
420BaseS
Chloroform
0.14
0.349288
0.658591
0.34388
86%
130%
84%
2.49
4.70
2.46
1/30/201318:20
2/6/20131:03
Hartman3
420BaseS
Chloroform
0.14
0.349288
0.658591
0.34388
86%
130%
84%
2.49
4.70
2.46
2/6/20131:00
2/13/201319:50
Hartman3
420BaseS
Chloroform
0.11
0.348683
0.678872
0.326733
104%
144%
99%
3.17
6.17
2.97
2/6/20131:04
2/13/201319:54
Hartman3
420BaseS
Chloroform
0.15
0.348683
0.678872
0.326733
80%
128%
74%
2.32
4.53
2.18
2/13/201319:52
2/20/2013 21:35
Hartman3
420BaseS
Chloroform
0.14
0.354646
0.690063
0.4615
87%
133%
107%
2.53
4.93
3.30
2/13/201319:56
2/20/2013 21:39
Hartman3
420BaseS
Chloroform
0.12
0.354646
0.690063
0.4615
99%
141%
117%
2.96
5.75
3.85
3/6/2013 20:20
3/14/2013 22:53
Hartman3
420BaseS
Chloroform
0.092
0.349513
0.686785
0.3366
117%
153%
114%
3.80
7.47
3.66
3/6/2013 20:24
3/14/2013 23:04
Hartman3
420BaseS
Chloroform
0.096
0.349513
0.686785
0.3366
114%
151%
111%
3.64
7.15
3.51
12/19/2012 23:26
12/26/2012 15:45
Hartman3
420First
Chloroform
0.075
0.35
0.7
NGC
129%
161%
NGC
4.67
9.33
NGC
12/26/201215:46
1/2/2013 21:20
Hartman3
420First
Chloroform
0.12
0.35
0.7
NGC
98%
141%
NGC
2.92
5.83
NGC
1/2/2013 21:22
1/9/2013 20:19
Hartman3
420First
Chloroform
0.14
0.35
0.7
NGC
86%
133%
NGC
2.50
5.00
NGC
1/9/2013 20:22
1/16/201319:32
Hartman3
420First
Chloroform
0.12
0.350142
0.69285
0.3568
98%
141%
99%
2.92
5.77
2.97
1/16/201319:34
1/23/2013 21:05
Hartman3
420First
Chloroform
0.12
0.349995
0.683329
0.3499
98%
140%
98%
2.92
5.69
2.92
(continued)
fe3
-------
Table 4-21. Comparison of Online GC to Radiello Results by Week (cont.)
-1
a
_i
o
"o"
1 =
¦D U)
r; S _l
o i i
aj 5 _.
"o"
1 =
¦D U)
O J.
"o"
1 =
¦D U)
O J.
o§a
GC:
GC:
GC:
¦C 04
"5 II
.c Q
|S
> II
¦= CD
g 2
* II
Radiello
(Mg/m3)
Missing
Values =
1/2 MDL
Missing
values =
MDL
Missing
values =
NGC
Start Date Time
Stop Date Time
Period
Location
Compound
IS 8
TO 3
1 s
1 S
3 «
U >
c >
(Mg/m3)
(Mg/m3)
(Mg/m3)
3 5
u >
3 TO
U >
Ratio (onlii
Calculated
Values = 1,
Ratio (onlii
Calculated
values = M
Ratio (onlii
Calculated
values = N
O £
£|
a. s
O .E
£|
a. s
O .E
£|
a. s
1/23/2013 21:07
1/30/201318:07
Hartman3
420First
Chloroform
0.12
0.350661
0.668843
0.357275
98%
139%
99%
2.92
5.57
2.98
1/30/201318:09
2/6/2013 0:51
Hartman3
420First
Chloroform
0.12
0.347683
0.681017
0.30135
97%
140%
86%
2.90
5.68
2.51
2/6/2013 0:54
2/13/201319:44
Hartman3
420First
Chloroform
0.075
0.35
0.7
NGC
129%
161%
NGC
4.67
9.33
NGC
2/13/201319:46
2/20/2013 21:29
Hartman3
420First
Chloroform
0.12
0.35
0.7
NGC
98%
141%
NGC
2.92
5.83
NGC
3/6/2013 20:11
3/14/2013 22:47
Hartman3
420First
Chloroform
0.06
0.35
0.7
NGC
141%
168%
NGC
5.83
11.67
NGC
12/19/2012 23:20
12/26/201215:39
Hartman3
422BaseS
Chloroform
0.26
0.38025
0.613583
0.44075
38%
81%
52%
1.46
2.36
1.70
12/26/201215:39
1/2/2013 21:13
Hartman3
422BaseS
Chloroform
0.29
0.49425
0.62725
0.582661
52%
74%
67%
1.70
2.16
2.01
1/2/2013 21:15
1/9/2013 20:13
Hartman3
422BaseS
Chloroform
0.54
0.690777
0.690777
0.690777
25%
25%
25%
1.28
1.28
1.28
1/9/2013 20:15
1/16/2013 19:24
Hartman3
422BaseS
Chloroform
0.6
0.789715
0.789715
0.789715
27%
27%
27%
1.32
1.32
1.32
1/16/2013 19:27
1/23/2013 20:58
Hartman3
422BaseS
Chloroform
0.85
1.537021
1.537021
1.537021
58%
58%
58%
1.81
1.81
1.81
1/23/2013 21:01
1/30/2013 18:01
Hartman3
422BaseS
Chloroform
0.77
1.328185
1.328185
1.328185
53%
53%
53%
1.72
1.72
1.72
1/30/2013 18:03
2/6/2013 0:41
Hartman3
422BaseS
Chloroform
0.85
1.474558
1.474558
1.474558
54%
54%
54%
1.73
1.73
1.73
2/6/2013 0:43
2/13/2013 19:38
Hartman3
422BaseS
Chloroform
0.2
0.467058
0.698191
0.694672
80%
111%
111%
2.34
3.49
3.47
2/13/2013 19:41
2/20/2013 21:22
Hartman3
422BaseS
Chloroform
0.28
0.409804
0.511888
0.434429
38%
59%
43%
1.46
1.83
1.55
3/6/2013 20:05
3/14/2013 22:41
Hartman3
422BaseS
Chloroform
0.22
0.393565
0.584475
0.445844
57%
91%
68%
1.79
2.66
2.03
12/19/2012 23:12
12/26/201215:31
Hartman3
422First
Chloroform
0.18
0.346621
0.663288
0.314525
63%
115%
54%
1.93
3.68
1.75
12/26/2012 15:33
1/2/2013 21:01
Hartman3
422First
Chloroform
0.22
0.362552
0.642552
0.41276
49%
98%
61%
1.65
2.92
1.88
1/2/2013 21:05
1/9/2013 20:02
Hartman3
422First
Chloroform
0.37
0.4194
0.545996
0.458727
13%
38%
21%
1.13
1.48
1.24
1/9/2013 20:04
1/16/201319:12
Hartman3
422First
Chloroform
0.32
0.425666
0.537368
0.461134
28%
51%
36%
1.33
1.68
1.44
1/16/2013 19:13
1/23/2013 20:47
Hartman3
422First
Chloroform
0.43
0.686704
0.693847
0.693719
46%
47%
47%
1.60
1.61
1.61
1/23/2013 20:48
1/30/2013 17:51
Hartman3
422First
Chloroform
0.4
0.593417
0.593417
0.593417
39%
39%
39%
1.48
1.48
1.48
(continued)
fe3
-------
Table 4-21. Comparison of Online GC to Radiello Results by Week (cont.)
-1
a
_i
o
"o"
1 =
¦D U)
r; s _!
f i i
a> 5 _.
"o"
1 =
¦D U)
o "z
"o"
1 =
¦D U)
o "z
o§a
GC:
GC:
GC:
¦C
"5 II
.c Q
|S
> II
¦= CD
g 2
* II
Radiello
(Mg/m3)
Missing
Values =
1/2 MDL
Missing
values =
MDL
Missing
values =
NGC
Start Date Time
Stop Date Time
Period
Location
Compound
IS 8
TO 3
1 s
1 S
3 «
U >
(Mg/m3)
(Mg/m3)
(Mg/m3)
3 5
u >
3 TO
U >
Ratio (onlii
Calculated
Values = 1,
Ratio (onlii
Calculated
values = M
Ratio (onlii
Calculated
values = N
O £
£|
a. s
O .E
£|
a. s
O .E
si
a. s
1/30/201317:52
2/6/2013 0:21
Hartman3
422First
Chloroform
0.39
0.694886
0.694886
0.694886
56%
56%
56%
1.78
1.78
1.78
2/6/2013 0:24
2/13/201319:28
Hartman3
422First
Chloroform
0.07
0.383526
0.680696
0.572113
138%
163%
156%
5.48
9.72
8.17
2/13/2013 19:31
2/20/2013 21:10
Hartman3
422First
Chloroform
0.16
0.350648
0.693356
0.3811
75%
125%
82%
2.19
4.33
2.38
3/6/201319:44
3/14/2013 22:28
Hartman3
422First
Chloroform
0.11
0.349662
0.655116
0.347343
104%
142%
104%
3.18
5.96
3.16
12/19/2012 23:42
12/26/2012 15:54
Hartman3
Outside
Chloroform
0.08
0.489826
0.774709
1.101563
144%
163%
173%
6.12
9.68
13.77
12/26/2012 16:00
1/2/2013 21:53
Hartman3
Outside
Chloroform
0.075
0.424082
0.724082
0.868571
140%
162%
168%
5.65
9.65
11.58
1/2/2013 21:55
1/9/2013 20:40
Hartman3
Outside
Chloroform
0.12
0.35
0.7
NGC
98%
141%
NGC
2.92
5.83
NGC
1/9/2013 20:42
1/16/2013 19:57
Hartman3
Outside
Chloroform
0.075
0.347365
0.660906
0.3247
129%
159%
125%
4.63
8.81
4.33
1/16/201319:59
1/23/2013 21:29
Hartman3
Outside
Chloroform
0.075
0.361402
0.678069
0.469725
131%
160%
145%
4.82
9.04
6.26
1/23/2013 21:32
1/30/201318:32
Hartman3
Outside
Chloroform
0.075
0.34572
0.640039
0.3231
129%
158%
125%
4.61
8.53
4.31
1/30/201318:34
2/6/20131:27
Hartman3
Outside
Chloroform
0.075
0.360212
0.702072
0.7891
131%
161%
165%
4.80
9.36
10.52
2/6/20131:29
2/13/2013 20:03
Hartman3
Outside
Chloroform
0.075
0.353025
0.650194
0.370038
130%
159%
133%
4.71
8.67
4.93
2/13/2013 20:06
2/20/2013 21:51
Hartman3
Outside
Chloroform
0.075
0.42005
0.68255
0.6302
139%
160%
157%
5.60
9.10
8.40
3/6/2013 20:42
3/14/2013 23:15
Hartman3
Outside
Chloroform
0.065
0.349014
0.692764
0.2948
137%
166%
128%
5.37
10.66
4.54
Note: NGC = No GC data available for comparison
-------
Section 4—Results and Discussion: OA Checks of Individual Data Sets
Basement
i°°
TO
~0 9
O
§0 6
O
£0.3
a>
"roO 0
GC Period 1
GC Concentration
Variable
Chloroform
Tetrachloroethene
Figure 4-2. XY Comparison plot of Radiello and GC indoor air concentration measurements
(pg/m3), Hartman 1 sampling period.
4-30
-------
Section 4—Results and Discussion: OA Checks of Individual Data Sets
GC Period 2
0.9 H
0.6
0.3
io.o
CD
~0.9
<1>
o
§0-6
O
= 0.3
0)
T3
^00
a:
0_9
0.6
0.3-
o.o-
Basement
i
if
First Floor
l 1 1 1 1— i r
* +
t
MI
i—I 1 1 1 r
r-O
O
rsj
K5
Oi
Variable
•+~ Chloroform
Tetrachloroethene
GC Concentration
Figure 4-3, XY Comparison plot of Radiello and GC indoor air concentration measurements
(pg/m3), Hartman 2 sampling period.
4-31
-------
Section 4—Results and Discussion: QA Checks of Individual Data Sets
The effect of these differences can also be visualized in the time series. In the time series plots (such as
Figures 4-4 through 4-7), the individual GC measurements occurring approximately every 2 hours are
shown as faint grey dots. Orange, green, and blue bars represent the weeklong averages of those GC
measurements after applying different treatments to time periods when there was no signal recorded on
the GC. When the treatment of the no signal data was immaterial, a single green bar marks the average
GC results. The individual occasions when no signal was detected on the GC are shown as red hash marks
just above the X axis. The calculated detection/reporting limit of the online GC (see Section 4.3.2) is
shown as a dark grey line bisecting the graph. The concentration measured by the passive Radiello
sampler exposed for 1 week is shown as a lavender bar. From these Figures 4-4 through 4-7, it is
apparent that the weekly average GC measurements and passive sampler measurements move in parallel
trends and the online GC results almost always exceed the passive sampler results. The agreement of
temporal trends is best when the indoor air concentrations are relatively high (where the vast majority of
the GC runs identified a peak and concentrations were >0.5 |_ig/m3). This provides qualitative confidence
that the high spikes seen in the online GC data likely reflect real events of vapor intrusion. As expected,
the weekly averaged data are less "jagged" than the data collected every 2 hours.
Other studies have also generally showed that online GC results and/or fixed laboratory TO-15 sample
results are generally slightly higher than those obtained with passive samplers under low concentration
ambient or indoor air conditions (Odencrantz et al., 2008; Lutes et al., 2010b; Allen et al., 2007).
Despite the substantial differences between the absolute values for either compound measured by the two
methods, when the data are examined in terms of the ratio of concentrations on the first floor to
concentrations measured in the basement, there is reasonably close agreement between the two
instruments. Correlation between the two methods is better for Hartman 2 (Figure 4-3) and Hartman 3
(discussed in next section).
For brevity, a full set of plots of the correlation of the online GC to the weeklong passive samples at all
locations is appended (Appendix B).
4-32
-------
Section 4—Results and Discussion: OA Checks of Individual Data Sets
Hartman2 - Chloroform
422BaseS
15
0 6-
11
-—r
oe
w
&
&
, ^ Missing values ^ Missing values a Missing Values Dar,i(sMft
in GO = 1/2 MDL in GC = MDL in GC = NA «aaieno
GC
Figure 4-4. Time series comparison of field GC and passive sampling data: 422 basement,
Hartman Period 2, chloroform. Horizontal gray line is calculated GC reporting limit.
Red hash marks on y-axis indicate missing values.
4-33
-------
Section 4—Results and Discussion: OA Checks of Individual Data Sets
1 5"
co
E
?1 0
o
"cu
I—
c
CD
O
c
o
O 0 5
Hartman2 - Chloroform
&
422First
&
&
$
&
Missing values _ Missing values —. Missing Values pariipiin
in GC = 1/2 MDL in GC = MDL in GC = NA «aaieno
GC
Figure 4-5. Time series comparison of field GC and passive sampling data: 422 first floor,
Hartman Period 2, chloroform. Horizontal gray line is calculated GC reporting limit.
Red hash marks on y-axis indicate missing values.
4-34
-------
Section 4—Results and Discussion: OA Checks of Individual Data Sets
Hartman2 - Tetrachloroethene
422BaseS
&
<(>
&
4
&
&
<£>
Missing values _ Missing values _ Missing Values _ . (
in GC = 1/2 MOL in GC = MOL in GC = NA naaieiio
GC
Figure 4-6. Time series comparison of field GC arid passive sampling data: 422 basement,
Hartman Period 2, PCE. Horizontal gray line is calculated GC reporting limit. Red
hash marks on y-axis indicate missing values.
4-35
-------
Section 4—Results and Discussion: OA Checks of Individual Data Sets
2 51
20-
1.5-
E
3
c
o
"?5
c 1.0
-------
Section 4—Results and Discussion: QA Checks of Individual Data Sets
4.3.6.2 Hartman Period 3
The agreement between the field GC and the Radiello data is generally, although not always, better during
this period (Table 4-21, Figure 4-8). A majority of the comparisons are within the stated accuracy
objective of 40% RPD. All PCE cases where both instruments found the concentration to be >0.5 j^ig/m3
met this accuracy objective. There were very few cases where the Radiello indoor chloroform exceeded
the >0.7 |_ig/m3 stated MDL for the GC. Those cases showed RPDs between 50 and 60%. Overall the
correlation between the methods is strong, especially at higher concentration levels (Figure 4-8).
We also plotted the time series of the weekly average field GC results against the passive sampler results
(Figures 4-9 through 4-12). As was seen for Hartman 2 a parallel movement is seen between the two data
sets. Weeks that exhibited high peaks in the online GC (and thus had high average concentrations) were
also high in the passive sampler results.
In cases where the concentrations registered a peak on the GC (where the amount of missing data was
low), the temporal agreement between the field GC and the Radiello was good. This suggests that the
peaks above the detection limit observed by the field GC likely reflect real events of vapor intrusion.
4.3.7 Overall Assessment of Online GC Data
Several overarching assessments can be reached regarding these data sets:
¦ Agreement with other methods/instruments is best for the first 4 weeks of period Hartman Period
1 and for the entirety of Hartman Period 2 and Hartman Period 3. The later portion of Hartman
lappears to contain a substantial high bias.
¦ Agreement is better for the higher concentrations (those well above the MDL) as would be
expected. Thus, the agreement is best in the winter data sets (Hartman Periods 2 and 3). This also
suggests that agreement is generally best in the 422 basement where the concentrations are
highest.
¦ Because of the biases exhibited in the Hartman Period 1 data (most likely due to the lower VOC
concentrations exhibited in the summer months), data analysis results for this period should be
considered less reliable than those drawn from Hartman Periods 2 and 3.
4-37
-------
Section 4—Results and Discussion: OA Checks of Individual Data Sets
Variable
-*¦ Chloroform
Tetrachloroethene
Period 3.
GC Period 3
|o-L
4
GC Concentration
Basement
First Floor
Figure 4-8. XY Plot of field GC vs. passive sampler data, Hartman
4-38
-------
Section 4—Results and Discussion: OA Checks of Individual Data Sets
Hartman3 - Chloroform
422BaseS
_ J= 2
o
CO
i—
c
0)
Q
c
o
O
¦4
1141
^ <$> ^ ^ ^ ^
Missing values _ Missing values _ Missing Values D ri ..
in GC - 1/2 MDL in GO = MOL in GO - NA Kaa e' °
GC
Figure 4-9, Time series comparison of field GC and passive sampling data: 422 basement,
Hartman Period 3, chloroform. Horizontal gray line is calculated GC reporting limit.
Red hash marks on y-axis indicate missing values.
4-39
-------
Section 4—Results and Discussion: OA Checks of Individual Data Sets
HartmanS - Chloroform
422First
0 9-
06-
c
&
&
,*F
&
&
Missing values _ Missing values __ Missing Values
in GC - 1/2 MDL in GC = MDL in GC = NA
Radiello
GC
Figure 4-10, Time series comparison of field GC and passive sampling data: 422 first floor,
Hartman Period 3, chloroform. Horizontal gray line is calculated GC reporting limit.
Red hash marks on y-axis indicate missing values.
4-40
-------
Section 4—Results and Discussion: OA Checks of Individual Data Sets
CO
g 10
o
CD
i—
C
0>
o
c
o
o
&
HartmariS - Tetrachloroethene
422BaseS
—
J!w"
Hpi # »
!
i
i '
^ ^
Missing values _ Missing values _ Missing Values D^iPim
in GC =1/2 MDL in GC = MOL in GC = IMA Kaaieuo
GO
Figure 4-11. Time series comparison of field GC and passive sampling data: 422 Basement,
Hartman Period 3, PCE. Horizontal gray line is calculated GC reporting limit. Red
hash marks on y-axis indicate missing values.
4-41
-------
Section 4—Results and Discussion: OA Checks of Individual Data Sets
Hartman3 - Tetrachloroethene
4
CO
?3
c
o
ro
— 2
c
i
1 4
1' ' 1
1 1
i *
^ ^ ^ ^ ^ ^
oe0 ^ ^ ^ ^
Missing values Missing values Missing Values 0
In GC = 172 MDL in GC = MDL in GC = NA Kaaieiio
GC
Figure 4-12. Time series comparison of field GC and passive sampling data: 422 first floor,
Hartman Period 3, PCE. Horizontal gray line is calculated GC reporting limit. Red
hash marks on y-axis indicate missing values (none in this case).
4-42
-------
Section 4—Results and Discussion: OA Checks of Individual Data Sets
4.4 Radon
4.4.1 Indoor Air: Comparison of Electrets Field, ARCADIS to Charcoal Analyzed by
U.S. EPA R&IE National Laboratory
Three comparisons were made between electrets and charcoal canisters. Charcoal canisters were provided
and analyzed by EPA's Radiation and Indoor Environments National Laboratory Center for Indoor
Environments in Las Vegas, Nevada. ARCADIS collected charcoal canister samples and electret samples.
Electrets were obtained from Rad Elec (Frederick, Maryland) and read by ARCADIS on site before and
after deployment. The charcoal canisters were used as a QC check on three separate occasions:
January 19, 2011, to January 26, 2011, April 27, 2011, to May 4, 2011, and December 28, 2011, to
January 4, 2012. A further test with charcoal canisters occurred on June 19, 2013, through June 26, 2013,
with results pending analysis of the canisters by EPA. Charcoal canisters (plus duplicates) were placed at
indoor locations and the ambient locations that were routinely being used for electret monitoring. When
the results were received, the sample plus its duplicate were averaged together to obtain a result for the
location. This was then compared with the electret result for that location and time period.
For the first occasion, the relative percentage difference between the two methods was 20% or less
(Table 4-22). The maximum absolute difference was 0.63 pCi. A relative percentage difference could not
be calculated for the ambient, which was below the detection limit with the charcoal method (BDL).
On the second occasion, five of six comparisons showed a relative percentage difference of 20% or less
and four of the six comparisons were within 0.5 pCi/L of each other (Table 4-23).
The exceptions were 422 basement north and 420 basement south, which were within 0.9 pCi/L of each
other. The ambient was again BDL by the charcoal method, as would have been predicted from the
electret data.
For the third occasion, December 28, 2011, to January 4, 2012, the absolute difference between the
methods is at or below 0.3 pCi/L and RPD is <6% for all samples (Table 4-24). The ambient charcoal
sample was below the detection limit and that detection limit was equal to the ambient value reported by
the electret method.
Table 4-22. Comparison between Electrets and Charcoal Canisters at the 422/420 EPA House
from January 19-26, 2011
Sample
Location
Electret Rn
Charcoal Rn
Charcoal
Absolute
(pCi/L)
(pCi/L)
Average
Difference (pCi/l)
RPD (%)
422First
5.14
4.8
4.7
0.44
6.84%
422First
4.6
422BaseN
8.44
8
8.4
0.04
5.35%
422BaseN
00
CO
420First
1.68
1.7
1.65
0.03
-1.18%
420First
1.6
420BaseN
3.98
3.3
3.35
0.63
18.68%
420BaseN
3.4
Ambient
0.03
<0.5
<0.5
Ambient
<0.5
4-43
-------
Section 4—Results and Discussion: OA Checks of Individual Data Sets
Table 4-23. Comparison of Electret and Charcoal Canister Data from April 27 to May 4, 2011
Charcoal
Canister
Average
Radon Activity
(pCi/l)
Charcoal
Canister
Radon Activity
(pCi/l)
Absolute
Difference
(pCi/l)
Location
Electret Data
(pCi/l)
RPD (%)
Ambient
0.47
<0.5
Ambient Dup
<0.5
422 First
2.72
2.8
2.6
0.12
4.51%
422 First Dup
2.4
422 Base S
7.39
7.3
7
0.39
5.42%
422 Base S Dup
6.7
422 Base N
7.14
6.3
6.05
0.905
13.92%
422 Base N Dup
6.77
5.8
420 First
0.98
1.3
1.4
-0.42
-35.29%
420 First Dup
1.5
420 Base S
4.58
3.8
3.75
0.83
19.93%
420 Base S Dup
3.7
420 Base N
4.48
4.2
3.95
0.53
12.57%
420 Base N Dup
3.7
Field blank
NA
<0.5
Field blank
NA
<0.5
NA = Not Available
Table 4-24. Comparison of Charcoal and Electret Radon December 28, 2011, to January 4, 2012
Radon
Activity
(pCi/l)
Charcoal
Average
(pCi/l)
Absolute
Difference
(pCi/L)
Canister ID
Location
Electrets
(pCi/L)
RPD (%)
877138
3.1
3.2
420BaseN
3.34
-0.2
-5.86%
877113
3.2
420BaseN Dup
877137
2.8
2.8
420BaseS
2.72
0.0
1.10%
877115
2.7
420BaseS Dup
877133
1.1
1.1
420First
1.09
0.0
-3.74%
877107
1.0
420First Dup
877139
10.0
10.0
422BaseN
10.22
-0.3
-2.67%
877136
9.9
422BaseN Dup
10.35
877128
9.6
9.5
422BaseS
9.57
-0.1
-0.73%
877111
9.4
422BaseS Dup
877108
4.8
4.8
422First
4.86
-0.1
-2.29%
877140
4.7
422First Dup
877110
5.0
5.2
4220ffice
4.92
0.2
4.57%
877131
5.3
4220ffice Dup
877130
<0.5
Ambient
0.5
NA
NA
NA = Not Available
4-44
-------
Section 4—Results and Discussion: OA Checks of Individual Data Sets
Figure 4-13 shows the correlations from Tables 4-22 to 4-24 in graphical form.
u
10
T n-
2011-2012 Charcoal/Electret Comparison
¦—J1 O
U
c
W ®
O
p
B
¦ JwHXptP|ieirt
4 lid D*ptoy-*Hl
—J:IWnf
o 4
J
c
0
c
> j
! 1
e s io
Electret |pd/L)
12
Figure 4-13. Correlation between radon measured using the electret and charcoal methods.
4.4.2 Comparision of Average of Real-Time AlphaGUARD to Electrets and Charcoal
Canisters
Stationary AlphaGUARD units provided by EPA were used for real-time monitoring of indoor air radon
at two locations (422 basement north and 422 office (2nd floor). Several comparisons were made between
the stationary AlphaGUARD data and electrets located nearby (at 422 basement north at first and both
422 basement north and 422 office later).
The first comparison took place over several weeks between March 30, 2011, and May 18, 2011
(Table 4-25). The absolute difference ranged from -0.04 pCi/L to 1.44 pCi/L. The relative percentage
difference ranged from 0.50% to 26.04%.
Table 4-25. Comparison between 422 Basement N AlphaGUARDs and Electrets from March 30,
2011, and May 18, 2011
AlphaGUAR
D Reading
(pCi/l)
Absolute
Difference
(pCi/L)
Relative
Percentage
Difference
Date Range
Electret
(pCi/l)
Electret
Dup(pCi/l)
Electret Ave
(pCi/L)
03/30-04/07
6.18
6.30
4.98
5.64
0.54
9.14%
04/07-04/13
5.90
4.94
5.87
5.41
0.50
8.76%
04/13-04/20
8.41
6.97
7.83
7.40
1.01
12.78%
04/20-04/27
6.25
4.04
5.58
4.81
1.44
26.04%
04/27-05/04
6.92
7.14
6.77
6.96
-0.04
-0.50%
05/04-05/11
4.66
2.93
4.50
3.72
0.95
22.57%
05/11-05/18
6.15
5.81
6.01
5.91
0.24
3.98%
For the second comparison, which occurred from August 3, 2011, to October 6, 2011, in the 422
basement north location, the absolute difference ranged from -1.11 pCi/L to 2.42 pCi/L. The relative
percentage difference ranged from -40.18% to 30.76% (Table 4-26).
4-45
-------
Section 4—Results and Discussion: OA Checks of Individual Data Sets
Table 4-26. Comparison of Real-Time AlphaGUARD to Integrated Electret August through October
End Date/
Time
Rn (pCi/L) A
Guard
(averaged
over a week)
Average of
Duplicate
Electrets
(pCi/L)
Rn (pCi/L)
Electrets 422
Base N
Rn (pCi/L)
Electrets 422
Base N Dup
Absolute
Difference
(pCi/L)
Relative
Percentage
Difference
8/3/2011
6.85
6.85
5.14
6.00
0.85
13.26%
8/10/2011
7.24
7.25
6.79
7.02
0.22
3.09%
8/17/2011
8.38
7.53
7.20
7.37
1.02
12.91%
8/24/2011
3.84
3.48
3.00
3.24
0.60
16.93%
8/31/2011
2.21
2.17
4.46
3.32
-1.11
-40.18%
9/7/2011
4.34
4.52
1.84
3.18
1.16
30.76%
9/14/2011
6.09
5.68
5.44
5.56
0.53
9.16%
9/21/2011
8.69
8.03
7.84
7.94
0.75
9.05%
9/28/2011
12.51
11.67
11.44
11.56
0.96
7.97%
10/6/2011
10.33
7.83
7.99
7.91
2.42
26.53%
During the third comparison, electrets, the AlphaGUARD, and the charcoal canisters were compared from
December 28, 2011, to January 4, 2012. Only the 422 office and 422 basement north were compared by
all three methods during this time. The absolute difference between the canisters and AlphaGUARD
ranged from -0.05 pCi/L to 0.15pCi/L, and the absolute difference between the electrets and
AlphaGUARD ranged from -0.08pCi/L to 0.29pCi/L. The relative percentage difference between
canisters and AlphaGUARD ranged from -0.50% to 2.96%, and the relative percentage difference
between electrets and AlphaGUARD ranged from -1.61% to 2.81% (Table 4-27).
Table 4-27. Comparison of Real-Time AlphaGUARD to Integrated Electret Measurements
December 28, 2011, to January 4, 2012
Location
Canister Radon Activity
(pCi/l)
Dup Canister Radon
Activity (pCi/L)
Canister Average (pCi/L)
Electret (pCi/L)
Electret Dup (pCi/L)
Electret Average (pCi/L)
422 Base N
AlphaGUARD
Approximation (pCi/L)
Absolute Difference
between Canisters and
AlphaGUARDs (pCi/L)
Absolute Difference
between Electrets and
AlphaGUARDs (pCi/L)
Relative Percent
Difference between
Canisters and
AlphaGUARD
Relative Percent
Difference between
Electrets and
AlphaGUARD
422BaseN
10.00
9.90
9.95
10.22
10.35
10.29
10.00
-0.05
0.29
-0.50%
2.81%
422 Office
5.00
5.30
5.15
4.92
5.00
0.15
-0.08
2.96%
-1.61%
The fourth comparison occurred between January 4, 2012, and March 1, 2012, for both the 422 office and
422 basement north locations. The absolute difference between 422 basement north AlphaGUARDs and
electrets ranged from -0.52 pCi/L to 1.79 pCi/L, and the absolute difference between 422 office
AlphaGUARDs and electrets ranged from 0.05 pCi/L to 0.77 pCi/L. The relative percentage difference
for 422 basement north ranged from -5.95% to 26.15%, and the relative percentage difference for the 422
office ranged from 1.05% to 17.68% (Table 4-28).
4-46
-------
Section 4—Results and Discussion: OA Checks of Individual Data Sets
Table 4-28. Comparison of Real-Time AlphaGUARD to Integrated Electret Measurements
January through March 2012
Date Range
422 Base N
AlphaGUARD
Reading (pCi/L)
Office AlphaGUARD
Reading (pCi/L)
422 Base N Electret
(pCi/L)
422 Base N Dup
Electret (pCi/L)
422 Base N Electret
Average (pCi/L)
Office Electret
(pCi/L)
Absolute Difference
between 422 Base N
AlphaGUARDs and
Electrets (pCi/L)
Absolute Difference
between Office
AlphaGUARDs and
Electrets (pCi/L)
422 Base N Relative
Percent Difference
Office Relative
Percent Difference
7-01/04/12
10
5
10.22
10.35
10.29
4.92
-0.29
0.08
-2.81%
1.61%
01/04/12-01/11/12
8.78
4.69
9.05
9.11
9.08
4.56
0.30
0.13
-3.36%
2.81%
01/11/12-01/18/12
9.73
5.09
9.34
9.73
9.54
4.88
0.19
0.21
2.02%
4.21%
01/18/12-01/25/12
8.52
4.79
7.83
7.98
7.91
4.74
0.61
0.05
7.49%
1.05%
01/25/12-02/01/12
7.71
4.46
8.24
8.03
8.14
4.15
-0.43
0.31
-5.36%
7.20%
02/01/12-02/08/12
8.68
4.78
8.60
8.62
8.61
4.58
0.06
0.20
0.81%
4.27%
02/08/12-02/15/12
8.44
4.80
8.28
7.47
7.88
4.41
0.56
0.39
6.93%
8.47%
02/15/12-02/22/12
7.74
4.3
6.08
5.82
5.95
3.68
1.79
0.62
26.15%
15.54%
02/22/12-03/01/12
8.48
4.74
9.00
9.00
9.00
3.97
-0.52
0.77
-5.95%
17.68%
The fifth comparison covers the time period from the week of January 2, 2013, through March 6, 2013
(Table 4-29). It compares the stationary AlphaGUARDs and electrets at both the 422 basement north and
the 422 office. The normal and duplicate electrets at the 422 basement north location are averaged. The
agreement was within 12% RPD when the mitigation system was in a passive mode and the radon
concentrations were above the EPA action level. The portion of the comparison that corresponded with
the mitigation on period (February 6 through April 24, 2013) showed much greater RPDs. However, the
paired results during these weeks are within +/- 0.7 pCi/1. The high RPDs are due to the tiny absolute
value of the radon present as indicated by both methods. This suggests that results below 1.5 pCi/1 may
have a higher percentage uncertainty.
Table 4-29.
Comparison of Real-Time AlphaGUARD to Integrated Electret Measurements
January through March 2013
Week
Start Date
422
Basement
North
Alpha-
GUARD
(pCi/L)
422
Basement
North Ave
Electret
(pCi/L)
422 Office
Alpha-
GUARD
(pCi/L)
422 Office
Electret
(pCi/L)
Absolute
Difference
422
Basement
North
Alpha-
GUARD
and
Electrets
(pCi.L)
Absolute
Difference
422 Office
Alpha-
GUARD
and
Electrets
(pCi.L)
Relative
Percentage
Difference
422
Basement
North
Alpha-
GUARD
and
Electrets
(%)
Relative
Percentage
Difference
422 Office
Alpha-
GUARD
and
Electrets
(%)
01/02/13
8.0
8.7
4.3
4.5
-0.7
-0.2
-8.50
-3.88
01/09/13
8.4
9.4
4.4
4.7
-1.0
-0.3
-11.02
-6.59
01/16/13
00
CO
9.5
4.6
4.6
-0.7
0.0
-7.65
0.65
01/23/13
8.3
8.2
3.9
4.0
0.2
-0.1
1.82
-2.28
(continued)
4-47
-------
Section 4—Results and Discussion: OA Checks of Individual Data Sets
Table 4-29. Comparison of Real-Time AlphaGUARD to Integrated Electret Measurements
January through March 2013
Week
Start Date
422
Basement
North
Alpha-
GUARD
(pCi/L)
422
Basement
North Ave
Electret
(pCi/L)
422 Office
Alpha-
GUARD
(pCi/L)
422 Office
Electret
(pCi/L)
Absolute
Difference
422
Basement
North
Alpha-
GUARD
and
Elect rets
(pCi.L)
Absolute
Difference
422 Office
Alpha-
GUARD
and
Electrets
(pCi.L)
Relative
Percentage
Difference
422
Basement
North
Alpha-
GUARD
and
Electrets
(%)
Relative
Percentage
Difference
422 Office
Alpha-
GUARD
and
Electrets
(%)
01/30/13
9.4
9.2
5.0
4.7
0.3
0.3
2.70
5.76
02/06/13
1.3
0.6
0.8
0.2
0.7
0.6
68.04
116.83
02/13/13
0.4
0.5
0.3
0.1
-0.1
0.2
-26.09
100.00
02/20/13
0.4
0.5
0.3
0.1
-0.1
0.3
-22.22
142.86
03/06/13
0.3
0.5
0.2
0.0
-0.2
0.2
-44.16
147.83
4.4.3 Quality Assurance Checks of Electrets
QC was performed on the electret reader and on the chambers holding the electrets. The QC check on the
reader was performed by placing reference electrets within the reader each week to measure any deviation
from the standard. The standard reference electrets were of 0 V, 245 V, and 250 V. Over the duration of
the project, the readings on the 0 V electret fluctuated but stayed within 4 V of its nominal value. The
245 V electret, with only two exceptions stayed within 20 V of its stated value. It steadily declined over
the duration of the project, hitting a low before slowly rising toward the end of the project. The 250 V
electret stayed within 6 V of its nominal value, showing a slight decline toward the end of the project.
To check for drift within the electret chambers, a normal electret was placed in a closed electret chamber
each week and then read on the voltage meter to measure any change in the voltage from the previous
week's readings. This would indicate any deviation caused by the chambers. Near the beginning of the
project, this electret dropped an average of 5 V/4 weeks or 1.25 V per week. The rate was even lower in
the second half of the project to a drop of 5 V/30 weeks or 0.16 V per week. These rates of drift are
insignificant because the actual observed voltage change at the indoor sampling locations was typically
25 V per week or more.
4.5 On-Site Weather Station vs. National Weather Service (NWS)
A VantageVue weather station from Davis Instruments was installed at the 422/420 house. Because it was
not safe to mount the station directly on the peak of the roof, it was mounted on vertical rods raised to the
approximate peak elevation from the edge of the second story roof. The trees near the house, especially to
the north, are quite tall, equal to or higher than the weather station. Branches extend close to the house on
the northwest corner. The house is much taller than the neighboring building to the east. There is also a
neighboring two-story residential structure to the northeast, approximately 30 to 40 ft away. A seven-
story commercial structure is approximately 150 ft southwest of the studied duplex. Essentially, the only
side completely free from all air current obstructions is the southern side, which borders 28th Street
(Figure 4-14).
4-48
-------
Section 4—Results and Discussion: OA Checks of Individual Data Sets
! Tin rti( It** ii** lilt
^ 3rf rsnrl | PioCi^lriJ Mcdr 0*i - a * WON ""
Figure 4-14. Aerial view of study house, showing potential influences on wind velocity,
red arrow indicates study house.
From roughly mid-October of 2012 through mid-January of 2013, the 422 house weather station would
periodically stop reporting data in the early morning hours, for roughly 15 minutes to 2 hours, and then
restart. Eventually, it was determined that this was attributable to a weakness in the solar-recharged
battery in the exterior weather sensor. When weather conditions were safe enough, Ping's Tree Service
was called on January 15, 2013, to use a bucket truck to change the sensor's battery. Changing the battery
solved the problem.
A 3-month comparison between the house weather station data and NWS data was made from January 1,
2013, to March 31, 2013, as a QC check. Three parameters were compared: temperature, relative
humidity, and wind speed. For temperature, the data from the two weather stations match well, only
differing by an average of 2 degrees F (Figure 4-15). Relative humidity at both weather stations differed
by an average of -4% (Figure 4-16). House wind speed and that of the NWS differed by an average of ~6
mph; the airport weather station was generally higher. This difference is likely due to the local NWS
station being at the Indianapolis International Airport. The KIND weather station is located in the middle
of the runways at the Indianapolis Airport approximately 500 meters from the nearest building. Thus, the
readings obtained at the house are probably a better representation of the wind speeds that directly
impinge on the house (Figure 4-17).
4-49
-------
Section 4—Results and Discussion: OA Checks of Individual Data Sets
422 House Weather Station Temperature versus NWS Temperature
QJ
~o
60
55
50
45
40
35
30
25
20
15
10
~
t
t
¦
1
~
t
~
il
t. "
¦
~
¦ ¦
J
¦
~
¦
¦
A
w
¦
¦
v"
~
jt'
¦
¦
/¦
y-
t
m
¦
~
¦ i
r
~ ~
¦ \
,¦
~
fa 1
i
i,
i
>
't
¦
¦
¦
~ WSTemperature Ave
¦ NWS Temperature Ave
Date
Figure 4-15. Comparison of National Weather Service Indianapolis temperature data to
weather station at 422 East 28th Street.
422 House Weather Station Relative Humidity versus NWS Relative
Humidity
90
2- 80
T3
E 70
3
& 50
¦
1
~
~
9L
f 1
*
ft
~
¦ ¦
V»
*
¦ i
¦
¦ i
¦
~ i
1
¦ft
ft
ft
\r
V
~
>
¦
~
*
p ~
~ •
• ¦ 1
¦~l
~
S"
~ ~
• ¦
ft
~
ft
~
I
~
b™
~
¦
¦
¦
¦
~
» .
ft
~
~
*
>
~
~
¦
¦
~ WS Relative Humidity
¦ NWS Relative Humidity
Date
Figure 4-16. Comparison of National Weather Service Indianapolis relative humidity to weather
station at 422 East 28th Street.
4-50
-------
Section 4—Results and Discussion: OA Checks of Individual Data Sets
Q.
£
TJ
OJ
-------
Section 4—Results and Discussion: OA Checks of Individual Data Sets
Table 4-30. Groundwater (5 mL)—EPA Field Blank Summary
Number of Field Blanks
Mean
Blank
Cone,
(ng)
% of Field
Blanks with
Detections
Std.
Dev.
(ng)
RL
(ng)
MDL
(ng)
Min
(ng)
Max
(ng)
Analyzed
Cone. >
RL
RL>Conc.
> MDL
Benzene
25
1.4
11
0
5
45
1.8
0.9
1.4
4.6
Chloroform
25
10
11
0
0
0
10
NA
10
10
cis-1,2-DCE
25
13
11
0
0
0
13
NA
13
13
PCE
25
14
11
0
0
0
13
NA
17
17
Toluene
25
14
11
0
0
0
10
NA
10
10
TCE
25
17
11
0
0
0
13
NA
13
13
NA = Not Applicable
Table 4-31. Groundwater (5 mL)—EPA Laboratory Blank Summary—TO-17
Number of Field Blanks
% of Field
Blanks
with
Detections
Mean
Blank
Cone,
(ng)
Std.
Dev.
(ng)
RL
(ng)
MDL
(ng)
Min
(ng)
Max
(ng)
Analyzed
Cone.
> RL
RL>Conc
. > MDL
Benzene
25
1.4
17
0
8
47
1.6
0.7
1.4
3.5
Chloroform
25
10
17
0
3
18
11
1.9
10
14
cis-1,2-DCE
25
13
17
0
0
0
13
NA
13
13
PCE
25
14
17
0
0
0
14
NA
14
14
Toluene
25
14
17
0
0
0
14
NA
14
14
TCE
25
17
17
0
0
0
17
NA
17
17
NA = Not Applicable
Table 4-32. Groundwater (25 mL)—EPA Field Blank Summary—TO-17
Number of Field Blanks
% of Field
Blanks
with
Detections
Mean
Blank
Cone,
(ng)
Std.
Dev.
(ng)
RL
(ng)
MDL
(ng)
Min
(ng)
Max
(ng)
Analyzed
Cone.
> RL
RL>Conc
. > MDL
Benzene
13
1.2
6
0
0
0
1.2
NA
1.2
1.2
Chloroform
13
1.3
6
0
0
0
1.3
NA
1.3
1.3
cis-1,2-DCE
13
1.7
6
0
0
0
1.7
NA
1.7
1.7
PCE
13
1.2
6
0
1
17
7.1
NA
1.2
7.1
Toluene
13
1.1
6
0
2
33
1.7
0.04
1.1
1.8
TCE
13
1.6
6
0
0
0
1.6
NA
1.6
1.6
NA = Not Applicable
4-52
-------
Section 4—Results and Discussion: OA Checks of Individual Data Sets
Table 4-33. Groundwater (25 mL)—EPA Laboratory Blank Summary—TO-17
Number of Field Blanks
% of Field
Blanks
Mean
Blank
Std.
RL
(ng)
MDL
(ng)
Analyzed
Cone.
> RL
RL>Conc
. > MDL
with
Detections
Cone,
(ng)
Dev.
(ng)
Min
(ng)
Max
(ng)
Benzene
13
1.2
10
0
1
10
1.6
0.7
1.2
3.5
Chloroform
13
1.3
10
1
0
10
14
NA
1.3
14
cis-1,2-DCE
13
1.7
10
0
5
50
3.0
2.2
1.7
7.0
PCE
13
1.2
10
0
0
0
1.2
NA
1.2
1.2
Toluene
13
1.1
10
0
1
10
2.1
NA
1.1
2.1
TCE
13
1.6
10
0
1
10
5.5
NA
1.6
5.5
NA = Not Applicable
4.6.2 Surrogate Recoveries
To monitor analytical efficiency, 200 ng of dibromofluoromethane, l,4-dichloroethane-d4, and toluene-d8
were added into each QC and field sample with the vapor phase internal standard mix during sample
analysis. Field surrogates were not included in the scope of this project. The recoveries were evaluated
against laboratory limits of 70 to 130%. Most surrogate recoveries met the laboratory criterion, and
summary statistics are presented in Tables 4-34 and 4-35.
Table 4-34. EPA Groundwater (5 mL) Surrogate Recovery Summary
Parameter
Dibromofluoromethane
Result
1,4-dichloroethane-d4
Result
Toluene-d8
Results
Number of surrogate recoveries
measured
111
111
111
Average recovery (%R)
105
95
98
Standard deviation (%R)
10
4
8
Minimum recovery (%R)
79
85
83
Maximum recovery (%R)
131
106
117
Table 4-35. EPA Groundwater (25 mL) Surrogate Recovery Summary
Parameter
Dibromofluoromethane
Result
1,4-dichloroethane-d4
Result
Toluene-d8
Results
Number of surrogate recoveries
measured
105
105
105
Average recovery (%R)
98
94
98
Standard deviation (%R)
8
5
5
Minimum recovery (%R)
77
82
86
Maximum recovery (%R)
115
113
108
4-53
-------
Section 4—Results and Discussion: QA Checks of Individual Data Sets
4.7 Groundwater Analysis—Pace Laboratories
On two occasions, split groundwater samples were submitted to Pace Laboratories. The intent of this
work was to provide an independent check on the groundwater analyses and also to evaluate whether air
transport to the EPA laboratory was having an impact on the chloroform results. A total of seven samples
and two trip blanks were submitted.
The surrogate recovery limits were:
¦ Dibromofluoromethane: 83-123%
¦ 4-Bromoflurobenzene: 72-135%
¦ Toluene-d8: 81-114%
All samples were well within these limits.
There were no detections in either the trip blank or method blank for either sample batch.
Results of the LCS were well within the stated acceptance limits:
¦ PCE: 57-125%
¦ Chloroform: 73-122%.
In summary all reported QA/QC parameters were in control for these two batches.
4.8 Database
4.8.1 Checks on Laboratory Reports
Throughout the project, the ARCADIS project manager briefly reviewed laboratory reports as they were
received from the VOC analytical laboratories. The primary focus of these checks was on blanks and
ambient samples as a sampling performance indicator as well as the general consistency and
reasonableness of the trends in reported concentrations for the primary analytes: PCE and chloroform.
The ARCADIS project manager also performed a manual review of the electrets radon computations in
the spreadsheet used for those calculations. He also reviewed that data set regularly and interacted with
the field scientist collecting this data when any anomalous results were observed.
The lead analyst (from Hartmann Environmental Geosciences), an ARCADIS principal scientist, and an
RTI scientist were all involved in reviewing the online GC calculations. For suspect values QC checks
performed included calibration checks and chromatogram reviews
4.8.2 Database Checks
An Access database was developed and used to compile results for VOCs (TO-17, TO-15, and passive
indoor air) and radon in indoor air and soil gas (electret and AlphaGUARD).
The following QC checks were performed on this database:
¦ The ARCADIS field scientist responsible for the majority of the field sampling performed a
check of the reports received from laboratories against his own records. He checked for the
following: approximate number of each sample type (to determine what reports were still
pending) and a line-by-line check of the sample times, dates, and sample numbers of each
4-54
-------
Section 4—Results and Discussion: QA Checks of Individual Data Sets
sample type. The assignment of sample locations was also reviewed. Notes of any discrepancies
and corrections were sent to the ARCADIS database manager.
¦ During the initial portions of the project, the ATL technical director manually prepared an Excel
spreadsheet from laboratory reports comparing the results of passive samplers exposed at the
same location for multiple durations and calculating percentage bias. The ARCADIS project
manager then used that spreadsheet to spot check the calculations of percentage bias performed
in the database. After correcting for slight differences in the percentage bias formula used,
excellent agreement was found. This indicates that, at least for the calculations spot checked,
both the calculation and the importation of the underlying concentration data from electronic
deliverable files into the database are being performed correctly.
¦ During the initial portions of the project, the ATL technical director manually prepared an Excel
spreadsheet of indoor air VOC results from laboratory reports. The Excel spreadsheet was used
to prepare temporal trend plots of indoor concentrations for key analytes for the first 18 weeks of
the project before the Access database was fully implemented. The ARCADIS project manager
then confirmed that the essential features of these temporal trend plots (such as range of
concentrations and overall temporal trends) were consistent between these plots and similar plots
generated from the Access database. This indicates for this period that the importation of the
underlying concentration data from electronic deliverable files into the database is being
performed correctly.
¦ The ARCADIS project manager provided to the database manager a design document for the
reports to be generated, including definitions of key formulas and variables. The design
document was prepared based on the project objectives in the QAPP. As database reports were
prepared, the ARCADIS project manager reviewed their format and content and requested
changes as necessary.
¦ The ARCADIS project manager and database manager both spot checked the transfer of the
NERL results for groundwater into the database.
¦ The ARCADIS Project manager and RTI statistical intern both reviewed the data sets for
outliers, queried them and addressed any problems identified.
¦ Database reports were run to identify samples that were collected but for which data was not
received. These samples were investigated and often determined to be due to problems that
occurred in the analytical laboratory. These lost samples were notated in the project database.
4.9 Air Exchange Rate Measurements
In this report we present the results of air exchange rate measurements made on three occasions not
presented in our previous report (EPA 2012a). A total of 10 primary samples were analyzed and reported.
In each round we conducted one duplicate measurement, the relative percent difference of the tracer
measurement were:
¦ October 13-14, 2011: 0.9%
¦ October 18-19, 2011: 4.7%
¦ April 2013: 8%
Two trip blanks were analyzed in October 2013. Both trip blanks yielded between 1.9 and 2.0 picoliters
(pi) of PMCH and no reported PDCH. The PMCH concentration in the trip blank for the fan test on
condition could have significantly influenced the measurement of the air exchange rate on the first floor
because the concentration in that sample was only 6.21 pi. No blank correction was performed. If a blank
correction had been performed, the increase in air exchange rate under the fan-on condition discussed in
Section 5 would have been more dramatic. In the other cases, the blank concentration was <15% of the
4-55
-------
Section 4—Results and Discussion: QA Checks of Individual Data Sets
concentration in the primary sample and thus would have had little influence on the calculated air
exchange rates (AERs).
One trip blank was analyzed in April 2013 and showed 3 pi of PMCH and 9 pi of PDCH. The blank
concentration was <15% of the concentration in the primary sample and thus would have had little
influence on the calculated AERs.
4-56
-------
Section 5—Subslab Depressurization Mitigation System Monitoring Results
Table of Contents
5.0 Subslab Depressurization Mitigation System Monitoring Results 5-1
5.1 Differential Pressure and Mitigation System Flow 5-1
5.1.1 Radon System Design Standards for Differential Pressure 5-1
5.1.2 Differential Pressure Monitoring of this SSD System 5-3
5.1.3 Mitigation System Flow 5-21
5.2 Radon Monitoring: Hourly and Weekly Time Scales 5-21
5.3 VOC Monitoring During Mitigation Testing 5-30
5.3.1 Descriptive Statistics 5-33
5.3.2 Effect of Mitigation System Status on Indoor Air VOC Levels 5-45
5.3.3 Discussion 5-46
5.4 Stack Gas Monitoring 5-47
5.4.1 Is Stack Gas an Indicator of System Performance in Protecting Indoor Air? 5-47
5.4.2 Air Exchange Rate Measurements 5-50
5.4.3 Stack Gas Measurements to Define Flux to Structure 5-53
List of Figures
5-1. Subslab vs. basement differential pressure: 422 side during mitigation testing 5-19
5-2. Subslab vs. basement differential pressure: 420 side during mitigation testing 5-20
5-3. Deep soil gas vs. shallow soil gas differential pressure during mitigation testing 5-21
5-4. Basement vs. upstairs differential pressure: 422 side during mitigation testing 5-22
5-5. Basement vs. exterior differential pressure: 422 side during mitigation testing 5-22
5-6. Stack gas flow velocity from SSD system 5-23
5-7. Real-time radon monitoring: 422 basement 5-23
5-8. Real-time radon monitoring: 422 second floor 5-24
5-9. Weekly integrated radon (electret) during mitigation testing 5-24
5-10. Passive sampler monitoring of PCE during mitigation testing 5-31
5-11. Passive sampler monitoring of chloroform during mitigation period 5-31
5-12. Indoor air PCE, real-time monitoring during mitigation testing 5-32
5-13. Boxplots of mitigation effect on indoor air concentrations 5-46
5-14. Stack gas monitoring during mitigation te sting: chloroform 5-48
5-15. 422 first floor versus stack gas chloroform concentrations: mitigation on 5-48
5-16. Stack gas monitoring during mitigation testing: PCE 5-49
5-17. Stack gas versus 422 first floor PCE concentrations: mitigation on 5-49
5-i
-------
Section 5—Subslab Depressurization Mitigation System Monitoring Results
List of Tables
5-1. Subslab vs. Basement Differential Pressures Measured with Handheld Micromanometer
at Permanent SSPs (negative pressure indicates flow out of building; yellow indicates
mitigation off) 5-4
5-2. Wall Port vs. Basement Differential Pressure Measured with Handheld Micromanometer
(negative pressure indicates flow out of building) 5-8
5-3. Shallow Interior SGP (6 ft bis) vs. Basement Differential Pressure Measured with
Handheld Micromanometer (negative pressure indicates flow out of building) 5-9
5-4. Shallow Exterior SGP (3.5 ft and 6 ft bis) vs. Basement Differential Pressure Measured
with a Handheld Micromanometer (negative pressure indicates flow out of building) 5-10
5-5. Deep Interior SGP (9 ft and 13 ft bis) vs. Basement Differential Pressure Measured by
ARCADIS with Handheld Micromanometer (negative pressure indicates flow out of
building) 5-12
5-6. Deep Exterior SGP (9 ft and 13 ft bis) vs. Basement Differential Pressure Measured by
ARCADIS with Handheld Micromanometer (negative pressure indicates flow out of
building) 5-15
5-7. Comparison of Setra Continuous Sensor Differential Pressure vs. Airdata Multimeter
ADM-870 with SSD System Operating: December 29, 2012 (yellow shaded data reflects
an "off scale" response on the Setra) 5-20
5-8. Electret Radon Descriptive Statistics by Mitigation and Heating Status (pCi/L) 5-25
5-9. Indoor Air Radon Descriptive Statistics by Mitigation and Heating Status: From
Stationary Real Time AlphaGUARD (pCi/L) 5-26
5-10. Indoor Radon Descriptive Statistics—Individual Locations by Mitigation and Heating
Status: Electret Data (pCi/L) 5-26
5-11. Descriptive Statistics: Radon in Subslab and Wall Ports by Individual Location and
Mitigation and Heating Status (pCi/L) 5-28
5-12. Radon Descriptive Statistics by Location Type and Mitigation and Heating Status (pCi/L) 5-30
5-13. Descriptive Statistics of Weekly Passive VOC Measurements ((ig/m3) in Indoor Air by
Mitigation Status and Heating Use (yellow indicates statistics during active mitigation) 5-33
5-14. Distribution of Concentrations ((.ig/ni3) by VOC and Mitigation and Heating Status:
Indoor Air, Week-Long Passive Samples (yellow indicates statistics during active
mitigation) 5-34
5-15. Descriptive Statistics of Indoor VOC Concentrations ((ig/m3) During Mitigation Testing
by Location and Mitigation and Heating Status (yellow indicates statistics during active
mitigation) 5-35
5-16. Descriptive Statistics: Average Subslab and Wall Port VOC Concentrations (|ig/m3) by
Mitigation and Heating Status (yellow indicates statistics during active mitigation) 5-39
5-17. Distribution of Subslab and Wall Port VOC Concentrations ((ig/m3) by Mitigation and
Heating Status (yellow indicates statistics during active mitigation) 5-40
5-18. Descriptive Statistics of Subslab and Wall Port VOC Concentrations ((.ig/ni3) by Location
and Mitigation and Heating Status(yellow indicates statistics during active mitigation) 5-41
5-19. April/May 2011 Air Exchange Rate Measurement Results 5-51
5-ii
-------
Section 5—Subslab Depressurization Mitigation System Monitoring Results
5-20. September 2011 Air Exchange Rate Measurement Results 5-51
5-21. October 2011 Air Exchange Measurement Results (during and after fan testing) 5-52
5-22. April 2013 Air Exchange Measurement Results (During Mitigation) 5-52
5-23. National Survey of Air Exchange Rates, Reprinted from the EPA Exposure Factor
Handbook (U.S. EPA, 2011) 5-52
5-23. National Survey of Air Exchange Rates, Reprinted from the EPA Exposure Factor
Handbook (U.S. EPA, 2011) (continued) 5-53
5-24. Stack Gas Discharge Measurements During Mitigation 5-54
5-iii
-------
Section 5—Subslab Depressurization Mitigation System Monitoring Results
5-iv
-------
Section 5—Subslab Depressurization Mitigation System Monitoring Results
5.0 Subslab Depressurization Mitigation System Monitoring Results
Installation and testing of the subslab depressurization (SSD) mitigation system is described in Section
3.2. After the system was installed and tested, a program of SSD mitigation system monitoring was
carried out to investigate the ability of the system to control radon and VOC levels in both active (fan on)
and passive (fan off) modes. In addition, the SSD mitigation system was installed with valves in the
depressurization lines to simulate a house without a mitigation system in place. The SSD mitigation
system had a single fan that served the entire duplex with a total of four extraction points. As described in
Section 3.2, SSD mitigation monitoring involved measuring radon and VOC levels with the SSD
mitigation system in active mode (fan on), passive mode (fan off, valves open), and completely off (fan
off, valves closed). To simplify data interpretation, the on/passive/off mitigation switches were always
conducted on a Wednesday, which was the day when new week-long integrated radon and VOC samples
were begun.
5.1 Differential Pressure and Mitigation System Flow
5.1.1 Radon System Design Standards for Differential Pressure
U.S. EPA's (1993a) most extensive guidance for radon SSD systems states a standard in terms of inches
of water column (in. WC), which is also known as inches of water gauge (in. WG):
Were the system to function solely by the primary mechanism discussed earlier, i.e., by
maintaining a measurable depressurization in the soil everywhere that it contacts the
foundation, a soil depressurization of about 0.015 in. WG, measured during mild
weather, would nominally be required to ensure that subsequent cold weather and winds
would rarely depressurize the house sufficiently to overwhelm the system. If exhaust
appliances were off during the measurement, the soil would nominally have to be
depressurized by an additional 0.01 to 0.02 in. WG to ensure that the system would not
be overwhelmed when these appliances were turned on. However, some experience
suggests that the other mechanisms mentioned earlier, including soil gas dilution and
perhaps air-barrier shielding, can come into play to varying degrees, depending upon the
circumstances. These other mechanisms could explain why good radon reductions are
often achieved by SSD systems even in cases where portions of the sub-slab are only
marginally depressurized, to an extent far less than the nominally required 0.025 to 0.035
in. WG.
U.S. EPA (1993a) goes on to describe in detail that the 0.025 to 0.035 in. WG criteria is meant to take
into account the typical maximum depressurization potentially produced by building HVAC systems of
0.02 in WG attributable to central furnace fans, clothes dryers, and exhaust fans. U.S. EPA (1993) further
notes that achieving a particular numerical target for depressurization "may in fact not be necessary" and
that a less stringent standard of 0.001 to 0.002 in WG can be applied if measured under worst case
conditions:
Depending upon site-specific factors, there may not necessarily be a significant impact
on long-term average indoor concentration if the pressure differential across some
portion of the slab is occasionally reversed by operation of these exhaust fans. Moreover,
since SSD seems to work by mechanisms in addition to soil depressurization (in
particular, by soil gas dilution), it may in fact not be necessary to guarantee that the sub-
slab depressurizations being established by the system are greater at every sub-slab
location than every potential basement depressurization that the system may ever
encounter. However, where the SSD system can reasonably be designed to provide sub-
slab depressurizations of about 0.01 to 0.02 in. WG everywhere during cold weather with
5-1
-------
Section 5—Subslab Depressurization Mitigation System Monitoring Results
the appliances off, in order to ensure that the system will essentially never be
overwhelmed, it is advisable to do so.
And elsewhere:
Where slab pressure measurements are made during cold weather with exhaust
appliances on - i.e., with the system experiencing its worst-case challenge - any
measurable sub-slab depressurization should be sufficient (0.001 to 0.002 in. WG)
U.S. EPA's 1994 guidance for Radon Prevention in the Design and Construction of Schools and Other
Large Buildings states, "A minimum subslab pressure of-0.002 in the water column (WC) is required at
all test holes for an effective ASD system."
U.S. EPA (1993a) also identifies backdrafting as a potential risk of systems with a high degree of
depressurization. U.S. EPA (1993a) explicitly describes the 0.025 to 0.035 in. WG design goal as
conservative:
If sub-slab depressurizations being created by a SSD system were being measured during
mild weather with exhaust appliances off, the conservative rule of thumb would thus be
that the system should be designed to maintain a depressurization of at least 0.015 in.
WG everywhere to avoid being overwhelmed by the stack effect when cold weather
arrives. In addition, to avoid being overwhelmed by the incremental basement
depressurization created when exhaust appliances are turned on during cold weather, the
SSD system should nominally maintain an additional sub-slab depressurization of up to
0.01 to 0.02 in. WG, as discussed previously in Section 2.3.1b. Thus, ideally, sub-slab
depressurizations measured during mild weather with appliances off should total about
0.025 to 0.035 inch WG everywhere in order to ensure that the system will never be
overwhelmed during cold weather with the appliances on.
But as re-iterated several places in this document, this target depressurization is usually
a very conservative design goal. Commonly, sub-slab depressurizations much less than
these ideal targets will still provide satisfactory SSD performance. Thus, an expensive
upgrade of a SSD system in an attempt to achieve these high depressurizations is often
unnecessary. However, where the SSD system can reasonably be designed to achieve
such depressurizations, it is probably advisable to do so.
Furthermore, this conservative maximum basement depressurization of0.025 to 0.035 in.
WG due to thermal and appliance effects is thought to be high for many cases In
addition, the upper end of the range assumes that the major depressurizing appliances
are operating during the coldest weather; among these appliances, whole-house and attic
fans will in fact not be operated in cold weather, and clothes driers will be operated only
intermittently. Combustion appliances in the basement would backdraft if
depressurizations as great as 0.035 in. WG were actually maintained for any extended
period.
Fourteen years later, this same numerical criterion was restated in the much briefer ITRC VI guidance
document (2007) without reprinting the detailed discussion of the basis for the recommendation:
Active SSD systems are the most reliable, cost effective, and efficient technique for
controlling vapor intrusion in the majority of cases, which concentration reductions in
the 90%-99% range (USEPA 1993b) and 99.5% or greater in carefully designed and
installed systems (Folkes 2002). Subslab depressurization in the range of0.025-0.035
inches H2O is generally sufficient to maintain downward pressure gradients (USEPA
1993b).
5-2
-------
Section 5—Subslab Depressurization Mitigation System Monitoring Results
SSD mitigation systems with very similar hardware can provide indoor air quality benefits through two
separate mechanisms described as subslab depressurization and subslab ventilation. These mechanisms
are described in U.S. EPA (2008):
The hardware used in sub-slab ventilation (SSV) systems and sub-slab depressurization
(SSD) systems is similar. The two names describe the different mechanisms through
which the system can be effective in keeping soil gas contaminants out of the building.
When the surrounding soil has a relatively high permeability, the fan pulls large
quantities of air (largely from the atmosphere) down through the soil thus diluting the
contaminant in the sub-slab region resulting in reduced entry into the building. This
mechanism predominates in a sub-slab ventilation system. It is important to ensure that
openings in the slab and foundation are adequately sealed to prevent large quantities of
conditioned indoor air being pulled into the mitigation system. Sealing as part of SSD
system installation is discussed in EPA 1993b, section 4.7 and in NYSDOH, 2006, section
4.3.1. When the soil is much less permeable, less air flows and the fan generates a larger
negative pressure in the subslab region (thus sub-slab depressurization occurs). The
result is a larger negative pressure gradient across the slab. The system works because
the negative pressure gradient ensures that the flow is in the direction from indoors to the
soil and dilution of sub-slab gases is less important in this SSD case. In extreme cases of
low permeability and low flows, it may be necessary to specify a special blower to ensure
that adequate pressure gradients are generated.
Thus, a system operating in an SSV mode would be expected to show substantial reduction in subslab
concentration but relatively low differential pressures across the slab. A system operating in an SSD mode
would show little reduction in subslab concentration but substantial and sustained pressure differential
across the slab.
5.1.2 Differential Pressure Monitoring of this SSD System
After SSD system installation, the mitigation subcontractor conducted tests typical of a commercially
installed residential SSD mitigation system. They tested differential pressure across the SSD system using
a portable micromanometer at a series of 10 temporary pressure monitoring points. As indicated in the
tabulated data in Table 3-8, the differential pressure at 8 of 10 of these locations immediately following
the October installation met and sometimes substantially exceeded the most conservative EPA
depressurization criteria 0.025 to 0.035 in. WC (6-9 Pa). All 10 of the monitoring points substantially
exceeded the 0.002 in. WC (0.5 Pa) criterion that was considered applicable here because the testing
occurred in mid-October and there are no exhaust appliances in the duplex.
After initial testing, we monitored the U-tube micromanometer supplied with the system, which would be
the tool that a homeowner would use to verify that a residential installation of SSD was functional.
U-tube manometers connected to each leg of the mitigation system were routinely monitored during on
periods to determine whether pressures remained constant. At no time during the monitoring did the
pressures deviate from the norm (0.3 in. WC on the 422 side, 0. 25 in. WC on the 420 side).
We subsequently conducted several additional rounds of vacuum influence monitoring using a separate
handheld micromanometer at the permanent subslab ports (Table 5-1), wall ports (WPs) (Table 5-2), and
both shallow (Tables 5-3 and 5-4) and deep (Tables 5-5 and 5-6) interior and exterior soil gas ports. Note
that yellow highlighted rows in these tables represent periods when the mitigation system was off (valve
off or in passive mode).
Testing using the permanent SSPs (Table 5-1) and the conventionally constructed soil gas ports just
below the slab (Table 5-3) generally indicated that the system was functioning well in maintaining
5-3
-------
Section 5—Sub slab Depressnrization Mitigation System Monitoring Results
vacuum influence across the slab in the desired direction (away from the building). A few exceptions
were observed:
¦ at SSP-4 on February 22, 2013 (relatively strong driving force into the building of 0.12 to 0.13 in.
WC)
¦ at SSP-7 on December 29, 2012 (pressures fluctuating strongly from +0.18 to -0.10 in. WC)
¦ at SGP11-6 on April 21, 2013 (pressures fluctuating strongly from +0.2225 to -0.26 in. WC)
Table 5-1. Subslab vs. Basement Differential Pressures Measured with Handheld
Micromanometer at Permanent SSPs (negative pressure indicates flow out of
building; yellow indicates mitigation off)
Negative3 End of
Manometer
Positive3 End of
Mitigation
System On/Off
Date
Time
Manometer
Reading (in. WC)
12/7/2012
11:19
Basement
SSP-1
0.0028
Off
12/7/2012
11:20
Basement
SSP-1
0.0036
Off
12/7/2012
11:21
Basement
SSP-1
0.0042
Off
11/8/2012
12:18
Basement
SSP-1
-0.2297
On
11/14/2012
18:39
Basement
SSP-1
-0.2247
On
11/14/2012
18:40
Basement
SSP-1
-0.2280
On
11/14/2012
18:41
Basement
SSP-1
-0.2287
On
12/29/2012
13:14
Basement
SSP-1
-0.2135
On
12/29/2012
13:15
Basement
SSP-1
-0.2154
On
12/29/2012
13:16
Basement
SSP-1
-0.2126
On
2/22/2013
12:59
Basement
SSP-1
-0.1258
On
2/22/2013
13:00
Basement
SSP-1
-0.1386
On
2/22/2013
13:01
Basement
SSP-1
-0.1214
On
2/22/2013
13:02
Basement
SSP-1
-0.1341
On
2/22/2013
13:03
Basement
SSP-1
-0.1333
On
4/20/2013
15:05
Basement
SSP-1
-0.2124
On
4/20/2013
15:05
Basement
SSP-1
-0.2176
On
4/20/2013
15:05
Basement
SSP-1
-0.2179
On
4/22/2013
15:02
Basement
SSP-1
-0.2206
On
4/22/2013
15:02
Basement
SSP-1
-0.2170
On
4/22/2013
15:03
Basement
SSP-1
-0.2166
On
12/7/2012
11:38
Basement
SSP-2
0.0038
Off
12/7/2012
11:39
Basement
SSP-2
0.0039
Off
12/7/2012
11:40
Basement
SSP-2
0.0038
Off
12/29/2012
13:34
Basement
SSP-2
-0.0299
On
12/29/2012
13:35
Basement
SSP-2
-0.0293
On
12/29/2012
13:35
Basement
SSP-2
-0.0319
On
2/22/2013
12:31
Basement
SSP-2
-0.0317
On
2/22/2013
12:32
Basement
SSP-2
-0.0325
On
2/22/2013
12:33
Basement
SSP-2
-0.0311
On
2/22/2013
12:34
Basement
SSP-2
-0.0318
On
2/22/2013
12:35
Basement
SSP-2
-0.0324
On
4/20/2013
15:06
Basement
SSP-2
-0.0394
On
(continued)
5-4
-------
Section 5—Sub slab Depressnrization Mitigation System Monitoring Results
Table 5-1. Subslab vs. Basement Differential Pressures Measured with Handheld
Micromanometer at Permanent SSPs (negative pressure indicates flow out of
building; yellow indicates mitigation off) (cont.)
Negative3 End of
Manometer
Positive3 End of
Manometer
Mitigation
System On/Off
Date
Time
Reading (in. WC)
4/20/2013
15:07
Basement
SSP-2
-0.0399
On
4/20/2013
15:07
Basement
SSP-2
-0.0386
On
4/22/2013
15:04
Basement
SSP-2
-0.0417
On
4/22/2013
15:04
Basement
SSP-2
-0.0424
On
4/22/2013
15:05
Basement
SSP-2
-0.0421
On
12/7/2012
11:57
Basement
SSP-3
-0.0007
Off
12/7/2012
11:58
Basement
SSP-3
0.0000
Off
12/7/2012
11:59
Basement
SSP-3
0.0002
Off
12/29/2012
16:53
Basement
SSP-3
-0.1079
On
12/29/2012
16:53
Basement
SSP-3
-0.1082
On
12/29/2012
16:53
Basement
SSP-3
-0.1037
On
2/22/2013
14:13
Basement
SSP-3
-0.1122
On
2/22/2013
14:14
Basement
SSP-3
-0.1109
On
2/22/2013
14:15
Basement
SSP-3
-0.1113
On
2/22/2013
14:20
Basement
SSP-3
-0.1123
On
2/22/2013
14:21
Basement
SSP-3
-0.1098
On
4/20/2013
15:15
Basement
SSP-3
-0.1139
On
4/20/2013
15:16
Basement
SSP-3
-0.1133
On
4/20/2013
15:16
Basement
SSP-3
-0.1138
On
4/22/2013
15:12
Basement
SSP-3
-0.1118
On
4/22/2013
15:12
Basement
SSP-3
-0.1136
On
4/22/2013
15:13
Basement
SSP-3
-0.1141
On
12/7/2012
11:34
Basement
SSP-4
0.0038
Off
12/7/2012
11:35
Basement
SSP-4
0.0041
Off
12/7/2012
11:36
Basement
SSP-4
0.0031
Off
12/29/2012
13:31
Basement
SSP-4
-0.0291
On
12/29/2012
13:31
Basement
SSP-4
-0.0408
On
12/29/2012
13:32
Basement
SSP-4
-0.0451
On
2/22/2013
13:44
Basement
SSP-4
0.1247
On
2/22/2013
13:45
Basement
SSP-4
0.1311
On
2/22/2013
13:46
Basement
SSP-4
0.1315
On
2/22/2013
13:47
Basement
SSP-4
0.1315
On
2/22/2013
13:48
Basement
SSP-4
0.1347
On
(continued)
5-5
-------
Section 5—Sub slab Depressnrization Mitigation System Monitoring Results
Table 5-1. Subslab vs. Basement Differential Pressures Measured with Handheld
Micromanometer at Permanent SSPs (negative pressure indicates flow out of
building; yellow indicates mitigation off) (cont.)
Negative3 End of
Manometer
Positive3 End of
Manometer
Mitigation
System On/Off
Date
Time
Reading (in. WC)
4/20/2013
15:09
Basement
SSP-4
-0.0531
On
4/20/2013
15:09
Basement
SSP-4
-0.0534
On
4/20/2013
15:10
Basement
SSP-4
-0.0515
On
4/22/2013
15:07
Basement
SSP-4
-0.0539
On
4/22/2013
15:07
Basement
SSP-4
-0.0545
On
4/22/2013
15:07
Basement
SSP-4
-0.0542
On
12/7/2012
12:03
Basement
SSP-5
0.0004
Off
12/7/2012
12:04
Basement
SSP-5
0.0009
Off
12/7/2012
12:05
Basement
SSP-5
0.0009
Off
12/29/2012
16:54
Basement
SSP-5
-0.0213
On
12/29/2012
16:54
Basement
SSP-5
-0.0183
On
12/29/2012
16:55
Basement
SSP-5
-0.0337
On
2/22/2013
15:08
Basement
SSP-5
-0.0208
On
2/22/2013
15:09
Basement
SSP-5
-0.0200
On
2/22/2013
15:10
Basement
SSP-5
-0.0224
On
2/22/2013
15:11
Basement
SSP-5
-0.0206
On
2/22/2013
15:12
Basement
SSP-5
-0.0210
On
4/20/2013
15:19
Basement
SSP-5
-0.0265
On
4/20/2013
15:19
Basement
SSP-5
-0.0261
On
4/20/2013
15:19
Basement
SSP-5
-0.0261
On
4/22/2013
15:15
Basement
SSP-5
-0.0259
On
4/22/2013
15:15
Basement
SSP-5
-0.0266
On
4/22/2013
15:16
Basement
SSP-5
-0.0265
On
12/7/2012
11:48
Basement
SSP-6
0.0001
Off
12/7/2012
11:49
Basement
SSP-6
0.0002
Off
12/7/2012
11:50
Basement
SSP-6
0.0004
Off
12/29/2012
16:49
Basement
SSP-6
-0.0350
On
12/29/2012
16:49
Basement
SSP-6
-0.0379
On
12/29/2012
16:50
Basement
SSP-6
-0.0351
On
2/22/2013
14:02
Basement
SSP-6
-0.0388
On
2/22/2013
14:03
Basement
SSP-6
-0.0387
On
2/22/2013
14:04
Basement
SSP-6
-0.0399
On
2/22/2013
14:05
Basement
SSP-6
-0.0373
On
(continued)
5-6
-------
Section 5—Sub slab Depressnrization Mitigation System Monitoring Results
Table 5-1. Subslab vs. Basement Differential Pressures Measured with Handheld
Micromanometer at Permanent SSPs (negative pressure indicates flow out of
building; yellow indicates mitigation off) (cont.)
Negative3 End of
Manometer
Positive3 End of
Manometer
Mitigation
System On/Off
Date
Time
Reading (in. WC)
2/22/2013
14:06
Basement
SSP-6
-0.0401
On
4/20/2013
15:13
Basement
SSP-6
-0.0403
On
4/20/2013
15:14
Basement
SSP-6
-0.0387
On
4/20/2013
15:14
Basement
SSP-6
-0.0412
On
4/22/2013
15:10
Basement
SSP-6
-0.0420
On
4/22/2013
15:11
Basement
SSP-6
-0.0434
On
4/22/2013
15:11
Basement
SSP-6
-0.0434
On
12/7/2012
12:00
Basement
SSP-7
-0.0005
Off
12/7/2012
12:01
Basement
SSP-7
-0.0010
Off
12/7/2012
12:02
Basement
SSP-7
-0.0006
Off
11/8/2012
12:27
Basement
SSP-7
0.1726
On
11/14/2012
18:44
Basement
SSP-7
-0.1145
On
11/14/2012
18:45
Basement
SSP-7
-0.1144
On
11/14/2012
18:46
Basement
SSP-7
-0.1131
On
12/29/2012
17:07
Basement
SSP-7
0.1491
On
12/29/2012
17:07
Basement
SSP-7
0.1889
On
12/29/2012
17:07
Basement
SSP-7
-0.1031
On
12/29/2012
17:08
Basement
SSP-7
-0.1046
On
12/29/2012
17:09
Basement
SSP-7
-0.1045
On
12/29/2012
17:09
Basement
SSP-7
-0.1066
On
12/29/2012
17:10
Basement
SSP-7
-0.1044
On
2/22/2013
14:21
Basement
SSP-7
-0.1102
On
2/22/2013
14:22
Basement
SSP-7
-0.1093
On
2/22/2013
14:23
Basement
SSP-7
-0.1105
On
2/22/2013
14:24
Basement
SSP-7
-0.1099
On
2/22/2013
14:25
Basement
SSP-7
-0.1115
On
4/20/2013
15:17
Basement
SSP-7
-0.1143
On
4/20/2013
15:17
Basement
SSP-7
-0.1129
On
4/20/2013
15:17
Basement
SSP-7
-0.1129
On
4/22/2013
15:13
Basement
SSP-7
-0.1129
On
4/22/2013
15:14
Basement
SSP-7
-0.1113
On
4/22/2013
15:14
Basement
SSP-7
-0.1120
On
aln Tables 5-1 through 5-6, the "negative end" is the low pressure manometer port and the "positive end" is the high pressure
manometer port.
5-7
-------
Section 5—Sub slab Depressnrization Mitigation System Monitoring Results
Table 5-2. Wall Port vs. Basement Differential Pressure Measured with Handheld
Micromanometer(negative pressure indicates flow out of building)
Negative End of
Manometer
Positive End of
Manometer
Mitigation
System On/Off
Date
Time
Reading (in. WC)
2/22/2013
12:25
Basement
WP-1
-0.0007
On
2/22/2013
12:26
Basement
WP-1
-0.0010
On
2/22/2013
12:28
Basement
WP-1
-0.0012
On
2/22/2013
12:29
Basement
WP-1
-0.0014
On
2/22/2013
12:30
Basement
WP-1
-0.0008
On
4/21/2013
18:27
Basement
WP-1
-0.0031
On
4/21/2013
18:27
Basement
WP-1
-0.0044
On
4/21/2013
18:28
Basement
WP-1
-0.0043
On
2/22/2013
12:51
Basement
WP-2
-0.0024
On
2/22/2013
12:52
Basement
WP-2
-0.0027
On
2/22/2013
12:53
Basement
WP-2
-0.0043
On
2/22/2013
12:54
Basement
WP-2
-0.0037
On
2/22/2013
12:55
Basement
WP-2
-0.0039
On
4/21/2013
18:33
Basement
WP-2
-0.0034
On
4/21/2013
18:34
Basement
WP-2
-0.0060
On
4/21/2013
18:34
Basement
WP-2
-0.0059
On
2/22/2013
13:33
Basement
WP-3
0.0063
On
2/22/2013
13:34
Basement
WP-3
0.0070
On
2/22/2013
13:35
Basement
WP-3
0.0052
On
2/22/2013
13:36
Basement
WP-3
0.0076
On
2/22/2013
13:37
Basement
WP-3
0.0054
On
4/21/2013
18:47
Basement
WP-3
-0.0002
On
4/21/2013
18:47
Basement
WP-3
0.0000
On
4/21/2013
18:48
Basement
WP-3
0.0000
On
2/22/2013
14:08
Basement
WP-4
-0.0051
On
2/22/2013
14:09
Basement
WP-4
-0.0063
On
2/22/2013
14:10
Basement
WP-4
-0.0033
On
2/22/2013
14:11
Basement
WP-4
-0.0051
On
2/22/2013
14:12
Basement
WP-4
-0.0035
On
4/21/2013
18:50
Basement
WP-4
-0.0090
On
4/21/2013
18:51
Basement
WP-4
-0.0084
On
4/21/2013
18:51
Basement
WP-4
-0.0073
On
(continued)
5-8
-------
Section 5—Sub slab Depressnrization Mitigation System Monitoring Results
Table 5-3. Shallow Interior SGP (6 ft bis) vs. Basement Differential Pressure Measured with
Handheld Micromanometer (negative pressure indicates flow out of building)
Negative End of
Manometer
Positive End of
Manometer
Mitigation
System On/Off
Date
Time
Reading (in. WC)
2/22/2013
12:38
Basement
SGP10-6
-0.0425
On
2/22/2013
12:39
Basement
SGP10-6
-0.0434
On
2/22/2013
12:40
Basement
SGP10-6
-0.0423
On
2/22/2013
12:41
Basement
SGP10-6
-0.0418
On
2/22/2013
12:42
Basement
SGP10-6
-0.0427
On
4/21/2013
18:29
Basement
SGP 10-6
-0.0537
On
4/21/2013
18:29
Basement
SGP 10-6
-0.0542
On
4/21/2013
18:30
Basement
SGP 10-6
-0.0535
On
11/8/2012
12:28
Basement
SGP11-6
-0.0768
On
2/22/2013
14:25
Basement
SGP 11-6
-0.1446
On
2/22/2013
14:26
Basement
SGP 11-6
-0.1593
On
2/22/2013
14:27
Basement
SGP 11-6
-0.1835
On
2/22/2013
14:28
Basement
SGP 11-6
-0.2157
On
2/22/2013
14:29
Basement
SGP 11-6
-0.2276
On
4/21/2013
18:52
Basement
SGP 11-6
0.2225
On
4/21/2013
18:52
Basement
SGP11-6
0.0910
On
4/21/2013
18:53
Basement
SGP11-6
0.0046
On
4/21/2013
18:53
Basement
SGP11-6
0.0650
On
4/21/2013
18:54
Basement
SGP11-6
0.0486
On
4/21/2013
18:54
Basement
SGP11-6
-0.1308
On
4/21/2013
18:55
Basement
SGP11-6
-0.2641
On
4/21/2013
18:55
Basement
SGP11-6
-0.2458
On
4/22/2013
15:17
Basement
SGP11-6
-0.0781
On
4/22/2013
15:17
Basement
SGP11-6
-0.0786
On
4/22/2013
15:17
Basement
SGP11-6
-0.0781
On
4/22/2013
15:18
Basement
SGP11-6
-0.0783
On
4/22/2013
15:18
Basement
SGP 11-6
-0.0778
On
2/22/2013
14:56
Basement
SGP12-6
-0.0463
On
2/22/2013
14:57
Basement
SGP12-6
-0.0792
On
2/22/2013
14:58
Basement
SGP12-6
-0.1263
On
2/22/2013
14:59
Basement
SGP12-6
-0.1610
On
2/22/2013
15:00
Basement
SGP12-6
-0.1937
On
4/21/2013
18:59
Basement
SGP 12-6
-0.0213
On
4/21/2013
18:59
Basement
SGP 12-6
-0.0214
On
(continued)
5-9
-------
Section 5—Sub slab Depressnrization Mitigation System Monitoring Results
Table 5-3. Shallow Interior SGP (6 ft bis) vs. Basement Differential Pressure Measured with
Handheld Micromanometer (negative pressure indicates flow out of building) (cont.)
Negative End of
Manometer
Positive End of
Manometer
Mitigation
System On/Off
Date
Time
Reading (in. WC)
4/21/2013
19:00
Basement
SGP 12-6
-0.0196
On
11/8/2012
12:21
Basement
SGP8-6
-0.1719
On
2/22/2013
13:03
Basement
SGP8-6
-0.1633
On
2/22/2013
13:04
Basement
SGP8-6
-0.1636
On
2/22/2013
13:05
Basement
SGP8-6
-0.1631
On
2/22/2013
13:06
Basement
SGP8-6
-0.1647
On
2/22/2013
13:07
Basement
SGP8-6
-0.1647
On
4/21/2013
18:35
Basement
SGP8-6
-0.1690
On
4/21/2013
18:36
Basement
SGP8-6
-0.1702
On
4/21/2013
18:36
Basement
SGP8-6
-0.1710
On
2/22/2013
13:24
Basement
SGP9-6
-0.0471
On
2/22/2013
13:25
Basement
SGP9-6
-0.0488
On
2/22/2013
13:26
Basement
SGP9-6
-0.0472
On
2/22/2013
13:27
Basement
SGP9-6
-0.0481
On
2/22/2013
13:28
Basement
SGP9-6
-0.0489
On
4/21/2013
18:41
Basement
SGP9-6
-0.0580
On
4/21/2013
18:42
Basement
SGP9-6
-0.0580
On
4/21/2013
18:42
Basement
SGP9-6
-0.0579
On
Table 5-4. Shallow Exterior SGP (3.5 ft and 6 ft bis) vs. Basement Differential Pressure Measured
with a Handheld Micromanometer (negative pressure indicates flow out of building)
Negative End of
Manometer
Positive End of
Mitigation
System On/Off
Date
Time
Manometer
Reading (in. WC)
2/20/2013
13:57
Basement
SGP2-3.5
-0.0035
On
2/20/2013
13:58
Basement
SGP2-3.5
-0.0088
On
2/20/2013
13:59
Basement
SGP2-3.5
-0.0027
On
2/20/2013
14:00
Basement
SGP2-3.5
-0.0012
On
2/20/2013
14:01
Basement
SGP2-3.5
0.0108
On
4/23/2013
14:27
Basement
SGP2-3.5
-0.0084
On
4/23/2013
14:28
Basement
SGP2-3.5
-0.0016
On
4/23/2013
14:28
Basement
SGP2-3.5
-0.0048
On
2/20/2013
14:02
Basement
SGP2-6
0.0032
On
2/20/2013
14:03
Basement
SGP2-6
0.0016
On
2/20/2013
14:04
Basement
SGP2-6
-0.0054
On
(continued)
5-10
-------
Section 5—Sub slab Depressnrization Mitigation System Monitoring Results
Table 5-4. Shallow Exterior SGP (3.5 ft and 6 ft bis) vs. Basement Differential Pressure Measured
by ARCADIS with Handheld Micromanometer (negative pressure indicates flow out of
building) (cont.)
Negative End of
Manometer
Positive End of
Manometer
Mitigation
System On/Off
Date
Time
Reading (in. WC)
2/20/2013
14:05
Basement
SGP2-6
0.0019
On
2/20/2013
14:06
Basement
SGP2-6
0.0049
On
4/23/2013
14:29
Basement
SGP2-6
0.0009
On
4/23/2013
14:30
Basement
SGP2-6
-0.0131
On
4/23/2013
14:30
Basement
SGP2-6
0.0073
On
2/21/2013
12:20
Basement
SGP3-3.5
0.0564
On
2/21/2013
12:21
Basement
SGP3-3.5
-0.4021
On
2/21/2013
12:22
Basement
SGP3-3.5
-0.7826
On
2/21/2013
12:23
Basement
SGP3-3.5
-1.0090
On
2/21/2013
12:24
Basement
SGP3-3.5
-0.9012
On
4/24/2013
13:04
Basement
SGP3-3.5
1.1480
On
4/24/2013
13:04
Basement
SGP3-3.5
0.2347
On
4/24/2013
13:05
Basement
SGP3-3.5
-0.0053
On
2/21/2013
12:28
Basement
SGP3-6
-0.0077
On
2/21/2013
12:29
Basement
SGP3-6
-0.2139
On
2/21/2013
12:30
Basement
SGP3-6
-0.1523
On
2/21/2013
12:31
Basement
SGP3-6
-0.0588
On
2/21/2013
12:32
Basement
SGP3-6
-0.1876
On
4/24/2013
13:05
Basement
SGP3-6
0.1260
On
4/24/2013
13:06
Basement
SGP3-6
0.0659
On
4/24/2013
13:06
Basement
SGP3-6
-0.0082
On
4/23/2013
14:46
Basement
SGP4-3.5
-0.0461
On
4/23/2013
14:47
Basement
SGP4-3.5
-0.0472
On
4/23/2013
14:47
Basement
SGP4-3.5
-0.0713
On
2/21/2013
13:10
Basement
SGP5-6
-0.0045
On
2/21/2013
13:11
Basement
SGP5-6
-0.0091
On
2/21/2013
13:12
Basement
SGP5-6
0.0277
On
2/21/2013
13:13
Basement
SGP5-6
-0.0129
On
2/21/2013
13:14
Basement
SGP5-6
0.0042
On
4/24/2013
13:12
Basement
SGP5-6
-0.0273
On
4/24/2013
13:12
Basement
SGP5-6
-0.0321
On
4/24/2013
13:12
Basement
SGP5-6
-0.0388
On
(continued)
5-11
-------
Section 5—Sub slab Depressnrization Mitigation System Monitoring Results
Table 5-4. Shallow Exterior SGP (3.5 ft and 6 ft bis) vs. Basement Differential Pressure Measured
by ARCADIS with Handheld Micromanometer (negative pressure indicates flow out of
building) (cont.)
Negative End of
Manometer
Positive End of
Manometer
Mitigation
System On/Off
Date
Time
Reading (in. WC)
2/21/2013
14:41
Basement
SGP6-6
-0.0783
On
2/21/2013
14:42
Basement
SGP6-6
0.1263
On
2/21/2013
14:43
Basement
SGP6-6
0.1289
On
2/21/2013
14:44
Basement
SGP6-6
-0.1172
On
2/21/2013
14:45
Basement
SGP6-6
-0.1588
On
4/24/2013
13:22
Basement
SGP6-6
0.1805
On
4/24/2013
13:23
Basement
SGP6-6
0.0933
On
4/24/2013
13:23
Basement
SGP6-6
0.0137
On
Table 5-5. Deep Interior SGP (9 ft and 13 ft bis) vs. Basement Differential Pressure Measured by
ARCADIS with Handheld Micromanometer (negative pressure indicates flow out of
building)
Negative End of
Manometer
Positive End of
Manometer
Mitigation
System On/Off
Date
Time
Reading (in. WC)
2/20/2013
14:23
Basement
SGP7-9
-0.4040
On
2/20/2013
14:24
Basement
SGP7-9
-0.0439
On
2/20/2013
14:25
Basement
SGP7-9
-0.0595
On
2/20/2013
14:26
Basement
SGP7-9
-0.0674
On
2/20/2013
14:27
Basement
SGP7-9
-0.0608
On
4/23/2013
14:40
Basement
SGP7-9
-0.2718
On
4/23/2013
14:40
Basement
SGP7-9
-0.2725
On
4/23/2013
14:41
Basement
SGP7-9
-0.2784
On
2/20/2013
14:28
Basement
SGP7-13
-0.3053
On
2/20/2013
14:29
Basement
SGP7-13
-0.1442
On
2/20/2013
14:30
Basement
SGP7-13
-0.2961
On
2/20/2013
14:31
Basement
SGP7-13
-0.2985
On
2/20/2013
14:32
Basement
SGP7-13
-0.3074
On
4/23/2013
14:42
Basement
SGP7-13
-0.3162
On
4/23/2013
14:43
Basement
SGP7-13
-0.3713
On
4/23/2013
14:43
Basement
SGP7-13
-0.3536
On
11/8/2012
12:22
Basement
SGP8-9
-0.0999
On
2/22/2013
13:19
Basement
SGP8-9
-0.0927
On
2/22/2013
13:20
Basement
SGP8-9
-0.0911
On
2/22/2013
13:21
Basement
SGP8-9
-0.0911
On
2/22/2013
13:22
Basement
SGP8-9
-0.0956
On
2/22/2013
13:23
Basement
SGP8-9
-0.0946
On
(continued)
5-12
-------
Section 5—Sub slab Depressnrization Mitigation System Monitoring Results
Table 5-5. Deep Interior SGP (9 ft and 13 ft bis) vs. Basement Differential Pressure Measured by
ARCADIS with Handheld Micromanometer (negative pressure indicates flow out of
building) (cont.)
Negative End of
Manometer
Positive End of
Manometer
Mitigation
System On/Off
Date
Time
Reading (in. WC)
4/21/2013
18:37
Basement
SGP8-9
-0.1038
On
4/21/2013
18:38
Basement
SGP8-9
-0.1059
On
4/21/2013
18:38
Basement
SGP8-9
-0.1026
On
11/8/2012
12:22
Basement
SGP8-13
-0.0477
On
2/22/2013
13:08
Basement
SGP8-13
-0.0408
On
2/22/2013
13:09
Basement
SGP8-13
-0.0379
On
2/22/2013
13:10
Basement
SGP8-13
-0.0428
On
2/22/2013
13:11
Basement
SGP8-13
-0.0411
On
2/22/2013
13:12
Basement
SGP8-13
-0.0394
On
4/21/2013
18:39
Basement
SGP8-13
-0.0594
On
4/21/2013
18:39
Basement
SGP8-13
-0.0603
On
4/21/2013
18:39
Basement
SGP8-13
-0.0611
On
2/22/2013
13:28
Basement
SGP9-9
-0.0492
On
2/22/2013
13:29
Basement
SGP9-9
-0.0471
On
2/22/2013
13:30
Basement
SGP9-9
-0.0479
On
2/22/2013
13:31
Basement
SGP9-9
-0.0501
On
2/22/2013
13:32
Basement
SGP9-9
-0.0482
On
4/21/2013
18:43
Basement
SGP9-9
-0.0580
On
4/21/2013
18:43
Basement
SGP9-9
-0.0588
On
4/21/2013
18:43
Basement
SGP9-9
-0.0588
On
2/22/2013
13:38
Basement
SGP9-13
-0.0178
On
2/22/2013
13:39
Basement
SGP9-13
-0.0139
On
2/22/2013
13:40
Basement
SGP9-13
-0.0138
On
2/22/2013
13:41
Basement
SGP9-13
-0.0171
On
2/22/2013
13:42
Basement
SGP9-13
-0.0128
On
4/21/2013
18:44
Basement
SGP9-13
-0.0328
On
4/21/2013
18:44
Basement
SGP9-13
-0.0285
On
4/21/2013
18:45
Basement
SGP9-13
-0.0299
On
2/22/2013
12:42
Basement
SGP 10-9
0.2429
On
2/22/2013
12:43
Basement
SGP10-9
-0.0314
On
2/22/2013
12:44
Basement
SGP 10-9
-0.0347
On
2/22/2013
12:45
Basement
SGP 10-9
-0.0375
On
2/22/2013
12:46
Basement
SGP 10-9
-0.0365
On
4/21/2013
18:30
Basement
SGP10-9
-0.0465
On
4/21/2013
18:31
Basement
SGP10-9
-0.0471
On
4/21/2013
18:31
Basement
SGP10-9
-0.0445
On
2/22/2013
12:46
Basement
SGP10-13
-0.0195
On
2/22/2013
12:47
Basement
SGP10-13
-0.0269
On
(continued)
5-13
-------
Section 5—Sub slab Depressnrization Mitigation System Monitoring Results
Table 5-5. Deep Interior SGP (9 ft and 13 ft bis) vs. Basement Differential Pressure Measured by
ARCADIS with Handheld Micromanometer (negative pressure indicates flow out of
building; yellow indicates mitigation off) (cont.)
Negative End of
Manometer
Positive End of
Manometer
Mitigation
System On/Off
Date
Time
Reading (in. WC)
2/22/2013
12:48
Basement
SGP10-13
-0.0197
On
2/22/2013
12:49
Basement
SGP10-13
-0.0228
On
2/22/2013
12:50
Basement
SGP10-13
-0.0267
On
4/21/2013
18:32
Basement
SGP10-13
-0.0362
On
4/21/2013
18:32
Basement
SGP10-13
-0.0370
On
4/21/2013
18:32
Basement
SGP10-13
-0.0375
On
11/8/2012
12:29
Basement
SGP 11-9
-0.0652
On
12/7/2012
11:52
Basement
SGP 11-9
-0.0005
Off
12/7/2012
11:53
Basement
SGP 11-9
-0.0002
Off
12/7/2012
11:54
Basement
SGP 11-9
-0.0001
Off
12/7/2012
11:55
Basement
SGP 11-9
0.0000
Off
12/7/2012
11:56
Basement
SGP 11-9
-0.0006
Off
12/29/2012
16:47
Basement
SGP 11-9
-0.0579
On
12/29/2012
16:47
Basement
SGP 11-9
-0.0603
On
12/29/2012
16:48
Basement
SGP 11-9
-0.0556
On
2/22/2013
14:30
Basement
SGP11-9
-0.2417
On
2/22/2013
14:31
Basement
SGP11-9
-0.2678
On
2/22/2013
14:32
Basement
SGP11-9
-0.2689
On
2/22/2013
14:33
Basement
SGP11-9
-0.2657
On
2/22/2013
14:34
Basement
SGP11-9
-0.2690
On
4/21/2013
18:56
Basement
SGP 11-9
-0.0664
On
4/21/2013
18:56
Basement
SGP11-9
-0.0645
On
4/21/2013
18:56
Basement
SGP11-9
-0.0637
On
11/8/2012
12:30
Basement
SGP11-13
-0.0380
On
2/22/2013
14:51
Basement
SGP11-13
-0.0382
On
2/22/2013
14:52
Basement
SGP11-13
-0.0310
On
2/22/2013
14:53
Basement
SGP11-13
-0.0283
On
2/22/2013
14:54
Basement
SGP11-13
-0.0355
On
2/22/2013
14:55
Basement
SGP11-13
-0.0362
On
4/21/2013
18:57
Basement
SGP11-13
-0.0416
On
4/21/2013
18:58
Basement
SGP11-13
-0.0443
On
4/21/2013
18:58
Basement
SGP11-13
-0.0427
On
2/22/2013
15:00
Basement
SGP 12-9
-0.0194
On
2/22/2013
15:01
Basement
SGP 12-9
-0.0192
On
2/22/2013
15:02
Basement
SGP 12-9
-0.0198
On
2/22/2013
15:03
Basement
SGP 12-9
-0.0180
On
2/22/2013
15:04
Basement
SGP 12-9
-0.0171
On
(continued)
5-14
-------
Section 5—Sub slab Depressnrization Mitigation System Monitoring Results
Table 5-5. Deep Interior SGP (9 ft and 13 ft bis) vs. Basement Differential Pressure Measured by
ARCADIS with Handheld Micromanometer (negative pressure indicates flow out of
building) (cont.)
Negative End of
Manometer
Positive End of
Manometer
Mitigation
System On/Off
Date
Time
Reading (in. WC)
4/21/2013
19:00
Basement
SGP 12-9
-0.0244
On
4/21/2013
19:01
Basement
SGP 12-9
-0.0230
On
4/21/2013
19:01
Basement
SGP 12-9
-0.0215
On
2/22/2013
15:04
Basement
SGP12-13
-0.0065
On
2/22/2013
15:05
Basement
SGP12-13
-0.0196
On
2/22/2013
15:06
Basement
SGP12-13
-0.0168
On
2/22/2013
15:07
Basement
SGP12-13
-0.1048
On
2/22/2013
15:08
Basement
SGP12-13
-0.0979
On
4/21/2013
19:02
Basement
SGP12-13
-0.0303
On
4/21/2013
19:02
Basement
SGP12-13
-0.0320
On
4/21/2013
19:02
Basement
SGP12-13
-0.0292
On
Table 5-6. Deep Exterior SGP (9 ft and 13 ft bis) vs. Basement Differential Pressure Measured by
ARCADIS with Handheld Micromanometer (negative pressure indicates flow out of
building)
Negative End of
Manometer
Positive End of
Manometer
Mitigation
System On/Off
Date
Time
Reading (in. WC)
2/20/2013
13:29
Basement
SGP1-13
-0.0124
On
2/20/2013
13:30
Basement
SGP1-13
-0.0153
On
2/20/2013
13:31
Basement
SGP1-13
-0.0337
On
2/20/2013
13:32
Basement
SGP1-13
-0.0149
On
2/20/2013
13:33
Basement
SGP1-13
-0.0146
On
4/23/2013
14:25
Basement
SGP1-13
-0.0293
On
4/23/2013
14:25
Basement
SGP1-13
-0.0273
On
4/23/2013
14:25
Basement
SGP1-13
-0.0321
On
2/20/2013
13:18
Basement
SGP 1-9
0.1209
On
2/20/2013
13:19
Basement
SGP 1-9
-0.2483
On
2/20/2013
13:20
Basement
SGP 1-9
-0.0130
On
2/20/2013
13:21
Basement
SGP 1-9
-0.0100
On
2/20/2013
13:22
Basement
SGP 1-9
-0.0125
On
4/23/2013
14:23
Basement
SGP 1-9
-0.1319
On
4/23/2013
14:24
Basement
SGP 1-9
-0.1363
On
4/23/2013
14:24
Basement
SGP 1-9
-0.1954
On
2/20/2013
14:06
Basement
SGP2-13
-0.0089
On
2/20/2013
14:07
Basement
SGP2-13
0.0000
On
2/20/2013
14:08
Basement
SGP2-13
0.0018
On
2/20/2013
14:09
Basement
SGP2-13
-0.0169
On
2/20/2013
14:10
Basement
SGP2-13
-0.0134
On
(continued)
5-15
-------
Section 5—Sub slab Depressnrization Mitigation System Monitoring Results
Table 5-6. Deep Exterior SGP (9 ft and 13 ft bis) vs. Basement Differential Pressure Measured by
ARCADIS with Handheld Micromanometer (negative pressure indicates flow out of
building) (cont.)
Negative End of
Manometer
Positive End of
Manometer
Mitigation
System On/Off
Date
Time
Reading (in. WC)
4/23/2013
14:32
Basement
SGP2-13
-0.0316
On
4/23/2013
14:32
Basement
SGP2-13
-0.0240
On
4/23/2013
14:33
Basement
SGP2-13
-0.0141
On
2/20/2013
13:24
Basement
SGP2-9
-0.0025
On
2/20/2013
13:25
Basement
SGP2-9
-0.0079
On
2/20/2013
13:26
Basement
SGP2-9
-0.0080
On
2/20/2013
13:27
Basement
SGP2-9
-0.0084
On
2/20/2013
13:28
Basement
SGP2-9
-0.0032
On
4/23/2013
14:31
Basement
SGP2-9
-0.2818
On
4/23/2013
14:31
Basement
SGP2-9
-0.3829
On
4/23/2013
14:31
Basement
SGP2-9
-0.4583
On
2/21/2013
12:37
Basement
SGP3-13
-0.0253
On
2/21/2013
12:38
Basement
SGP3-13
-0.0191
On
2/21/2013
12:39
Basement
SGP3-13
-0.0168
On
2/21/2013
12:40
Basement
SGP3-13
-0.0310
On
2/21/2013
12:41
Basement
SGP3-13
-0.0194
On
4/24/2013
13:08
Basement
SGP3-13
-0.0240
On
4/24/2013
13:08
Basement
SGP3-13
-0.0250
On
4/24/2013
13:08
Basement
SGP3-13
-0.0343
On
2/21/2013
12:32
Basement
SGP3-9
3.3110
On
2/21/2013
12:33
Basement
SGP3-9
2.6090
On
2/21/2013
12:34
Basement
SGP3-9
2.3070
On
2/21/2013
12:35
Basement
SGP3-9
2.0440
On
2/21/2013
12:36
Basement
SGP3-9
2.0850
On
4/24/2013
13:06
Basement
SGP3-9
-0.0804
On
4/24/2013
13:07
Basement
SGP3-9
-0.1652
On
4/24/2013
13:07
Basement
SGP3-9
-0.1840
On
2/21/2013
12:53
Basement
SGP4-13
-0.0601
On
2/21/2013
12:54
Basement
SGP4-13
-0.0490
On
2/21/2013
12:55
Basement
SGP4-13
-0.0388
On
2/21/2013
12:56
Basement
SGP4-13
-0.0351
On
2/21/2013
12:57
Basement
SGP4-13
-0.0351
On
4/23/2013
14:48
Basement
SGP4-13
-0.0375
On
4/23/2013
14:48
Basement
SGP4-13
-0.0413
On
4/23/2013
14:48
Basement
SGP4-13
-0.0570
On
2/21/2013
12:48
Basement
SGP4-9
-0.8140
On
2/21/2013
12:49
Basement
SGP4-9
-0.4402
On
2/21/2013
12:50
Basement
SGP4-9
-0.3766
On
(continued)
5-16
-------
Section 5—Sub slab Depressnrization Mitigation System Monitoring Results
Table 5-6. Deep Exterior SGP (9 ft and 13 ft bis) vs. Basement Differential Pressure Measured by
ARCADIS with Handheld Micromanometer (negative pressure indicates flow out of
building) (cont.)
Negative End of
Manometer
Positive End of
Manometer
Mitigation
System On/Off
Date
Time
Reading (in. WC)
2/21/2013
12:51
Basement
SGP4-9
-0.3957
On
2/21/2013
12:52
Basement
SGP4-9
-0.3744
On
2/21/2013
13:21
Basement
SGP5-13
-0.0231
On
2/21/2013
13:22
Basement
SGP5-13
-0.0289
On
2/21/2013
13:23
Basement
SGP5-13
-0.0189
On
2/21/2013
13:24
Basement
SGP5-13
-0.0433
On
2/21/2013
13:25
Basement
SGP5-13
-0.0338
On
4/24/2013
13:16
Basement
SGP5-13
-0.0137
On
4/24/2013
13:16
Basement
SGP5-13
-0.1725
On
4/24/2013
13:16
Basement
SGP5-13
-0.1781
On
2/21/2013
13:16
Basement
SGP5-9
-0.0200
On
2/21/2013
13:17
Basement
SGP5-9
-0.0205
On
2/21/2013
13:18
Basement
SGP5-9
-0.0181
On
2/21/2013
13:19
Basement
SGP5-9
-0.0095
On
2/21/2013
13:20
Basement
SGP5-9
-0.0120
On
4/24/2013
13:14
Basement
SGP5-9
-0.0314
On
4/24/2013
13:15
Basement
SGP5-9
-0.0361
On
4/24/2013
13:15
Basement
SGP5-9
-0.0315
On
2/21/2013
14:58
Basement
SGP6-13
-0.0688
On
2/21/2013
14:59
Basement
SGP6-13
-0.0403
On
2/21/2013
15:00
Basement
SGP6-13
-0.0144
On
2/21/2013
15:01
Basement
SGP6-13
-0.0350
On
2/21/2013
15:02
Basement
SGP6-13
-0.0256
On
4/24/2013
13:25
Basement
SGP6-13
-0.0259
On
4/24/2013
13:25
Basement
SGP6-13
-0.0240
On
4/24/2013
13:25
Basement
SGP6-13
-0.0228
On
2/21/2013
14:53
Basement
SGP6-9
-0.1667
On
2/21/2013
14:54
Basement
SGP6-9
-0.1361
On
2/21/2013
14:55
Basement
SGP6-9
-0.1582
On
2/21/2013
14:56
Basement
SGP6-9
-0.2463
On
2/21/2013
14:57
Basement
SGP6-9
-0.0993
On
4/24/2013
13:24
Basement
SGP6-9
0.1724
On
4/24/2013
13:24
Basement
SGP6-9
-0.3384
On
4/24/2013
13:24
Basement
SGP6-9
-0.5029
On
2/20/2013
14:28
Basement
SGP7-13
-0.3053
On
2/20/2013
14:29
Basement
SGP7-13
-0.1442
On
2/20/2013
14:30
Basement
SGP7-13
-0.2961
On
(continued)
5-17
-------
Section 5—Sub slab Depressnrization Mitigation System Monitoring Results
Table 5-6. Deep Exterior SGP (9 ft and 13 ft bis) vs. Basement Differential Pressure Measured by
ARCADIS with Handheld Micromanometer (negative pressure indicates flow out of
building) (cont.)
Negative End of
Manometer
Positive End of
Manometer
Mitigation
System On/Off
Date
Time
Reading (in. WC)
2/20/2013
14:31
Basement
SGP7-13
-0.2985
On
2/20/2013
14:32
Basement
SGP7-13
-0.3074
On
4/23/2013
14:42
Basement
SGP7-13
-0.3162
On
4/23/2013
14:43
Basement
SGP7-13
-0.3713
On
4/23/2013
14:43
Basement
SGP7-13
-0.3536
On
2/20/2013
14:23
Basement
SGP7-9
-0.4040
On
2/20/2013
14:24
Basement
SGP7-9
-0.0439
On
2/20/2013
14:25
Basement
SGP7-9
-0.0595
On
2/20/2013
14:26
Basement
SGP7-9
-0.0674
On
2/20/2013
14:27
Basement
SGP7-9
-0.0608
On
4/23/2013
14:40
Basement
SGP7-9
-0.2718
On
4/23/2013
14:40
Basement
SGP7-9
-0.2725
On
4/23/2013
14:41
Basement
SGP7-9
-0.2784
On
This testing would not have been required for residential SSD systems in many jurisdictions but may have
been conducted in some cases. For example, ITRC (2007) states the following in its section on operation,
maintenance, and monitoring of SSD mitigation systems:
Suction field extension testing may be warranted if manometer readings indicate reduced
suction levels or indoor air tests show increasing trends.
Vacuum influence monitoring through a series of wall ports (Table 5-2), shallow exterior soil gas ports
(Table 5-4), deep interior soil gas ports (Table 5-5), and deep exterior soil gas ports (Table 5-6) was also
conducted in this research project. Such monitoring is not a feature of normal residential SSD system
operation. The results at the wall ports (Table 5-3) showed that the vacuum influence was weak (which is
reasonable given that the extraction ports were beneath the floor and the wall ports are closer to the
surface of the soil). Three of the wall ports showed weak influence but in the desired direction. WP-3,
located on the south wall of the 422 basement, was the exception, showing several readings indicative of
weak driving forces into the basement.
The differential pressure between the shallow exterior ports and the basement (Table 5-4) was much less
consistent. This result would be expected because the SSD system is not designed to depressurize the area
outside of the building footprint.
The differential pressure between the deep interior ports (9 and 13 ft bis = 3 and 6 ft below the basement
floor) is consistent and shows that the driving force is moving out of the building (Table 5-5). There was
only a single exception to this pattern at SGP 10-9, which could have been an artifact because it was the
first reading at that location on that day. The differential pressure between the deep exterior ports and the
basement was also generally negative, indicating a driving force out of the building (Table 5-6).
ARCADIS monitored differential pressure continuously at five locations (methods described in Section
3.3.6). Such continuous monitoring would rarely be performed on a residential SSD system in current
practice but is more common in evaluating commercial building systems. As shown in Figure 5-1, the
5-18
-------
Section 5—Subslab Depressurization Mitigation System Monitoring Results
subslab vs. basement differential pressure monitored near the center of the 422 building but 7 feet from
the nearest extraction point consistently showed depressurization to some degree during active SSD
operation. The exception is a period from October 22 to 25, when readings of approximately 0.5 Pa
greater pressurization in the subslab than in the building were observed. However, during those 3 days, no
anomalous radon results were seen in continuous monitoring.
422 Subslab versus Basement Setra
(positive values indicate greater pressuriiation ol the subsfab and thus flow toward the basement)
60
40
20
"(5*
Q.
— 0
01
i/i
Irt -20
0!
-40
-60
-80
• 422SV8
Itfcif km On
—^•Miti^itiori toisfrt
fM
N
INI
a
m
m
a
m
m
•HI
*4
*4
m
trti
'
¦—
m
rJ
r*
IM
O
N
n
*-1
s
*N|
¦H
r\t
«H
CMl
¦»—
m
o
M
m
Date
Figure 5-1. Subslab vs. basement differential pressure: 422 side during mitigation testing.
In most cases, readings at or beyond the design capacity of the micromanometer +15 Pa to -15 Pa (0.06
to -0.06 in. WC) were observed. According to our discussions with the manufacturer of the Setra sensors,
when the pressure goes above the designed range, the values should be considered semi-quantitative. The
Setra sensor may also at times give a constant result of-15 Pa to indicate the pressure is off scale. We
evaluated this over-range performance on the Setra sensors by comparing it with the Airdata Multimeter
ADM-870 handheld micromanometer, which has a greater design range. These results show that the two
instruments agree well up to approximately double the differential pressure design range of the Setra
sensor which is 0.12 in. WC. However, an essentially constant reading of -0.058 in. WC (-14.3 Pa) was
recorded when the Airdata instrument found a differential pressure more than three times the design range
of tlie Setra (Table 5-7)—0.213 in. WC (-53 Pa).
In general, the subslab vs. basement differential pressure on the 422 side of the duplex responded as a
square wave to the turning on and off of the SSD system. However, unexpectedly, the relaxation of the
vacuum in the first off cycle was gradual over as long as 10 days from November 16 to 26, 2012
(Figure 5-1). On the 420 side of the duplex (Figure 5-2), the subslab vs. basement differential pressure
responded as a square wave to the turning on and off of the SSD system with the exception of two time
periods when vacuum control was apparently lost and the driving force swung toward the building. These
occurred on December 24-28, 2012, and April 16, 2013. These dates corresponded to a major blizzard and
a major storm event from the evening of the April 16th until the late night. The storm produced much rain,
and the Fall Creek stream gauge read over 2,000 cf/s.
5-19
-------
Section 5—Subslab Depressurization Mitigation System Monitoring Results
Table 5-7. Comparison of Setra Continuous Sensor Differential Pressure vs. Airdata Multimeter
ADM-870 with SSD System Operating: December 29, 2012 (yellow shaded data reflects
an "off scale" response on the Setra)
0.002941
0.001885
0.001885
0.001885
0.0013
0.0012
0.0012
422 Basement vs. Upstairs
-0.05761
-0.05761
-0.05814
-0.05814
-0.2135
-0.2154
-0.2126
422 Subslab vs. Basement
0.124052
0.125373
0.125638
0.12511
0.1248
0.1229
0.1268
422 Deep Soil Gas vs.
Shallow Soil Gas
-0.01028
-0.01107
-0.00922
-0.00922
-0.0189
-0.0147
-0.0139
422 Basement vs. Exterior
-0.05791
-0.05791
-0.05853
-0.05815
-0.0579
-0.0603
-0.0556
420 Subslab vs. Basement
Location
Airdata ADM-870 Replicate
Readings (in WC) Taken within
about 2 min per Location
Setra data, Replicate Readings
Bracketing the Airdata Data (in WC);
Data Points at 14-min. Intervals
re
—
3
VI
Ol
420 Setra Pressure Data
(positive values indicate higher pressure in subslab than the basement, thus ftow toward the basement)
16
12
8
4
0
-4
-8
-12
-16
439 bolr*
¦Miiifation Qm
-MnigniaPi Pa&»ve
¦Smrw
¦M
fM
N
-------
Section 5—Subslab Depressurization Mitigation System Monitoring Results
Differential pressure monitoring also indicates that during mitigation the driving force is generally from
the 13 ft bis soil gas depth to a 6 ft bis (subslab) soil gas depth (Figure 5-3). With some temporary
exceptions, this suggests that the mitigation system is drawing in soil gas from above the water table,
which could enhance contaminant migration toward the structure. This effect has been previously
hypothesized as a reason to not overpower SSD systems (Lutes, 2010b).
40
30
422 Deep Soil Gas versus Shallow Soil Gas Setra
{positive values indicate a greater pressure in the deep soil gas relative to shallow}
20
ra
ev
oj
10
-10
-20
!
Hfi
p
trp
ii id
F#
M
If tj§
llli Mi
ty
5-1
*!!¦
I
f:
&
P i
i i
i
s
a -
i
•jHji
"•* i I»
J ;
L-f •«
*
I
4}
r
v-3
%
Q
t
Lj
ispf
Ml
y.
¦> *
j *
> «
r*f
c'*s
ifc
j
i "1
i !
! ''T;
i
iI \
*w *
1 $
"r» .J-
* i *
¦
It
i c
Ft
H
n
4
1 I
i
i
1 *
¦ 1In
t
u
Ui
422 D5G55G
¦Mitigation On
-MifigAfiofi Pa&lvt
CM
«N
rJ
r4
a
3
'—.
Date
n
(T»
Figure 5-3. Deep soil gas vs. shallow soil gas differential pressure during mitigation testing.
The SSD system has little or no effect on the driving force from the basement to the upstairs within the
structure (Figure 5-4). There is relatively little driving force between these zones of the house, most
likely because there is little resistance to flow between floors as a result of poor air sealing. The basement
vs. exterior differential pressure (Figure 5-5) shows some variability, including a sharp drop off in late
November 2012, but no clear correlation of that variability to mitigation status.
5.1.3 SSD Mitigation System Flow
Flow through the SSD system discharge stack is relatively consistent between 1,540 and 1,819 fpm when
the SSD system is on (Figure 5-6). As expected, when the SSD system was in the passive mode, flows
were much lower and variable in direction.
5.2 Radon Monitoring: Hourly and Weekly Time Scales
As expected, radon concentrations show a near immediate substantial drop when the SSD system is
turned on and quickly return to high (premitigation) concentrations when the SSD system is turned off.
This is shown for the 422 side of the duplex based on continuous AlphaGUARD data, which were
collected on 10-min intervals (Figures 5-7 and 5-8). This is shown for all monitored interior locations
with weekly integrated electret samples (Figure 5-9). The weekly data include an ambient location and
show that during active SSD operation radon concentrations in the interior closely approached ambient
levels.
5-21
-------
Section 5—Subslab Depressurization Mitigation System Monitoring Results
422 Basement versus Upstairs Setra
(positive values indicate pressuriutlon of the basement relative to the upstairs)
m bvu
'Mitigation Ch
Mingiiiwi I'iwrw
w
-1
-1
Figure 5-4. Basement vs. upstairs differential pressure: 422 side during mitigation testing.
3
in
in
QJ
422 Basement versus Exterior Setra
(positive values indicate that the basement pressure is higher than the pressure in exterior air)
40
30
— 20
re
a.
10
0
-10
-20
3 ^ H
'• '3 H
^•3
* . *
* B ,*
* E *
if
»4 1 *
"•J* :1
M I
ii ¦
I" f '
* ¦ *
f
.] i
* •
:U
1
m 'J
•
i«
> .*
< ¦
1 a1
I* JH«
¦ *
iti
t
' ii'-iS" J
•
* • 1
*
j.
* %_
1 Js
i Ai
Itt
»
SJ »1
* a .
1 !
c
: ' i
rj!l n
1 1*
* A
t 1
1 1 *¦
v »
•«
* *
i
«
•
i
• i
i •
f
; #
»-4
©
Si
m
Date
¦ «2BVI
On
Mitigation Passive
iM
N
Figure 5-5. Basement vs. exterior differential pressure: 422 side during mitigation testing.
5-22
-------
Section 5—Subslab Depressurization Mitigation System Monitoring Results
Stack Gas Flow Rates
2000
1500
§. 1000
Mart Velocity
0 5Q0 ftwirve
"qj Orv
II
-500
n n n m
H H H H
^ mi*
HM fM H H
O fM M rn
N H H N fl
-h © o uare o o ©
Figure 5-6. Stack gas flow velocity from SSD system
422 Basement Continuous Alphaguard
Downstairs Alphaguard
•Mitigation On
'Mitigation Passive
Figure 5-7. Real-time radon monitoring: 422 basement.
5-23
-------
Section 5—Subslab Depressurization Mitigation System Monitoring Results
422 Second Floor Office Continuous Alphaguard
• Upstair* Alphaguard
Of*
Mutation Passfcv
Figure 5-8. Real-time radon monitoring: 422 second floor ,
Electret Radon with Mitigation On/Off Cycles
is
is
14
12
Zj 10
s>
C
CC 6
4
2
0
-2
0
in
•
o «
\
<
I
•
• °
• I
c
A
o
«
0
¦
0
j
i
u
° a
i
~
B
o
O
0
B
o
0
~ r
UI V
a a
i
OB
" o
0
a °
0
.
-!
li
a
t
i Q
I a
_ I
i
0
-
h
11
i
(
\\
ti
i Sf
li
a H 1
! &a. -
-
I
N
rsi
M
CM
1M
ffl
m
«
m
H
H
H
¦-i
w-i
H
H
H
H
H
crv
U3
rl
0%
m
3
o
«
INI
N
UN
N
f-l
T-l
HI
PS
o
3
at
o
fl
IM
H
fsl
rt
rr
in
o
rl
H
a
O
O
a
a
422 KnwH
41] fetr N CHip
422B«*
-------
Section 5—Sub slab Depressnrization Mitigation System Monitoring Results
Operating the SSD system in a passive mode provided little benefit for radon. It is possible that this
reflects the design of the SSD system, which involved an exterior stack. SSD systems intended to be
primarily run in a passive mode are frequently designed with a stack running up through the center of the
structure to maximize stack effects.
Descriptive statistics for the electret measurements are presented in Table 5-8. The descriptive statistics
have been broken up into four mitigation status categories defined as follows:
¦ Not installed—data collected prior to installing the SSD system
¦ Off—data collected after the SSD system was installed and first operated, but with the SSD
system powered off and valves off as a test
¦ Passive—data collected after the SSD system was installed and first operated, but with the SSD
system powered off but with the valves open
¦ On—data collected with the SSD system powered on and valves open.
The descriptive statistics have also been broken up into two heating status categories:
¦ Off—includes data from the 420 side where the heating system was not installed, as well as
summer data for both sides of the duplex
¦ On—heating system in operation (during the heating season, on the 422 side this was the normal
state). Weekly duration data are coded as "on" for the 420 side if the heating system was on the
422 side, this is reasonable because thermocouple data suggest that the 420 side stays above
ambient temperatures in winter due to heat leakage.
It should be noted that not all possible combinations of mitigation and heating status were tested. For
example, passive mitigation was not tested outside of the heating season.
Table 5-8. Electret Radon Descriptive Statistics by Mitigation and Heating Status (pCi/L)
Number
Samples
Mitigation
Heating
Mean
SD
CV
geoMean
geoSD
geoCV
Not installed
Off
221
4.70
3.54
0.75
2.66
8.28
3.11
Not installed
On
284
5.07
3.36
0.66
3.99
2.15
0.54
Off
On
27
4.39
2.84
0.65
3.55
2.02
0.57
Passive
On
49
4.85
2.68
0.55
4.09
1.87
0.46
On
Off
14
0.66
0.56
0.85
0.50
2.19
4.37
On
On
91
0.47
0.33
0.70
0.28
6.54
23.50
The data presented in Table 5-8 include all the indoor sampling locations and show very similar
arithmetic mean radon concentrations prior to SSD system installation, with the SSD system completely
off and with the SSD system in the passive mode. The SSD system substantially reduces the indoor radon
concentration both within the heating season and outside of the heating season. The reduction was
approximately 91% (comparing SSD on during the heating season to SSD not installed during the heating
season). The ability to achieve and measure a greater radon reduction was probably limited by the
ambient concentration of radon (i.e., SSD systems should not be expected to reduce concentrations in
indoor air below ambient air levels). The operating mitigation systems achieved concentrations well
below the EPA recommended action level of 4 pCi/L. EPA states that "reducing radon levels below 2
pCi/L is difficult" (U.S. EPA, 2012g), so the SSD system performs very well for radon. Table 5-9 shows
nearly identical trends when the data from the real time stationary AlphaGUARD measurements of radon
5-25
-------
Section 5—Sub slab Depressnrization Mitigation System Monitoring Results
are tabulated. For example, the mean at the 422 basement north location was reduced by 93% with the
SSD system turned on during the heating season. Table 5-10 shows that very similar trends also hold in
the electret data at each location within the duplex.
Table 5-9. Indoor Air Radon Descriptive Statistics by Mitigation and Heating Status: From
Stationary Real Time AlphaGUARD (pCi/L)
Number
Samples
Location 1
Mitigation
Heating
Mean
SD
cv
geoMean
geoSD
geoCV
422BaseN
Not installed
Off
29,472
7.18
3.98
0.55
5.23
3.75
0.72
422BaseN
Not installed
On
42,125
7.13
3.22
0.45
6.12
2.11
0.35
422BaseN
Off
Off
382
9.62
3.60
0.37
8.91
1.50
0.17
422BaseN
Off
On
6,479
7.83
3.16
0.40
6.56
2.90
0.44
422BaseN
Passive
On
6,483
7.93
2.34
0.30
7.49
1.54
0.21
422BaseN
On
Off
2,304
0.78
2.20
2.84
0.06
44.43
751.27
422BaseN
On
On
15,543
0.47
0.91
1.95
0.05
43.31
828.75
4220ffice
Not installed
Off
29,645
2.65
2.12
0.80
1.60
4.60
2.87
4220ffice
Not installed
On
41,857
3.36
2.02
0.60
2.41
3.18
1.32
4220ffice
Off
Off
381
2.36
1.69
0.72
1.64
2.66
1.62
4220ffice
Off
On
6,480
3.38
1.98
0.59
2.30
4.22
1.84
4220ffice
Passive
On
6,443
4.04
1.65
0.41
3.51
2.02
0.57
4220ffice
On
Off
2,304
0.37
0.83
2.24
0.06
24.29
414.78
4220ffice
On
On
15,544
0.33
0.50
1.54
0.08
19.72
251.68
Table 5-10. Indoor Radon Descriptive Statistics—Individual Locations by Mitigation and Heating
Status: Electret Data (pCi/L)
Number
Samples
Location 1
Mitigation
Heating
Mean
SD
CV
geoMean
geoSD
geoCV
420BaseN
Not installed
Off
44
4.58
2.01
0.44
3.28
5.43
1.65
420BaseN
Not installed
On
40
3.31
1.40
0.42
3.03
1.55
0.51
420BaseN
Off
On
4
2.82
0.86
0.30
2.70
1.44
0.53
420BaseN
Passive
On
7
4.19
1.41
0.34
3.93
1.51
0.39
420BaseN
On
Off
2
0.64
0.50
0.79
0.53
2.42
4.59
420BaseN
On
On
13
0.47
0.18
0.39
0.44
1.46
3.32
420BaseS
Not installed
Off
43
5.64
2.21
0.39
5.23
1.49
0.28
420BaseS
Not installed
On
40
4.24
2.02
0.48
3.85
1.55
0.40
420BaseS
Off
On
4
3.37
0.78
0.23
3.30
1.27
0.38
420BaseS
Passive
On
7
4.07
1.59
0.39
3.73
1.63
0.44
420BaseS
On
Off
2
1.37
1.41
1.03
0.94
3.68
3.90
420BaseS
On
On
13
0.56
0.15
0.27
0.54
1.44
2.68
(continued)
5-26
-------
Section 5—Sub slab Depressnrization Mitigation System Monitoring Results
Table 5-10. Indoor Radon Descriptive Statistics—Individual Locations by Mitigation and Heating
Status: Electret Data (pCi/L) (cont.)
Number
Samples
Location 1
Mitigation
Heating
Mean
SD
cv
geoMean
geoSD
geoCV
420First
Not installed
Off
42
2.68
1.23
0.46
1.99
5.03
2.53
420First
Not installed
On
40
1.28
0.67
0.52
1.12
1.74
1.56
420First
Off
On
4
1.25
0.48
0.39
1.15
1.67
1.45
420First
Passive
On
7
1.93
0.84
0.43
1.73
1.74
1.01
420First
On
Off
2
0.46
0.52
1.12
0.28
4.56
16.12
420First
On
On
13
0.24
0.18
0.76
0.11
9.59
83.79
422BaseN
Not installed
Off
30
6.37
3.67
0.58
3.90
8.05
2.07
422BaseN
Not installed
On
53
7.31
1.90
0.26
7.01
1.36
0.19
422BaseN
Off
On
3
7.70
3.10
0.40
7.18
1.62
0.23
422BaseN
Passive
On
7
8.11
2.19
0.27
7.80
1.38
0.18
422BaseN
On
Off
2
0.67
0.03
0.04
0.67
1.04
1.56
422BaseN
On
On
13
0.60
0.20
0.33
0.57
1.41
2.47
422BaseS
Not installed
Off
30
7.04
6.41
0.91
3.85
8.55
2.22
422BaseS
Not installed
On
53
9.07
3.78
0.42
8.66
1.31
0.15
422BaseS
Off
On
4
8.82
1.79
0.20
8.66
1.26
0.15
422BaseS
Passive
On
7
8.27
2.27
0.27
7.90
1.43
0.18
422BaseS
On
Off
2
0.74
0.17
0.22
0.73
1.25
1.72
422BaseS
On
On
13
0.73
0.18
0.25
0.71
1.27
1.78
422First
Not installed
Off
29
2.44
2.09
0.86
0.50
38.29
76.72
422First
Not installed
On
53
3.76
1.25
0.33
3.46
1.59
0.46
422First
Off
On
4
3.85
0.96
0.25
3.75
1.32
0.35
422First
Passive
On
7
3.52
0.91
0.26
3.38
1.40
0.41
422First
On
Off
2
0.41
0.20
0.49
0.39
1.68
4.35
422First
On
On
13
0.50
0.25
0.51
0.45
1.55
3.45
4220ffice
Not installed
Off
3
2.87
0.41
0.14
2.85
1.16
0.41
4220ffice
Not installed
On
5
4.16
0.36
0.09
4.15
1.09
0.26
4220ffice
Off
On
4
3.78
1.34
0.35
3.54
1.57
0.44
4220ffice
Passive
On
7
3.83
1.08
0.28
3.65
1.43
0.39
4220ffice
On
Off
2
0.32
0.16
0.51
0.30
1.71
5.75
4220ffice
On
On
13
0.15
0.55
3.62
0.03
29.03
1,108.10
Outside
Not installed
Off
29
0.42
0.67
1.62
0.12
22.87
185.41
Outside
Not installed
On
45
0.18
1.14
6.49
0.06
29.96
488.54
(continued)
5-27
-------
Section 5—Sub slab Depressnrization Mitigation System Monitoring Results
Table 5-10. Indoor Radon Descriptive Statistics—Individual Locations by Mitigation and Heating
Status: Electret Data (pCi/L) (cont.)
Number
Samples
Location 1
Mitigation
Heating
Mean
SD
cv
geoMean
geoSD
geoCV
Outside
Off
On
4
0.17
0.35
2.06
0.04
55.95
1,420.99
Outside
Passive
On
5
0.62
0.24
0.38
0.60
1.38
2.32
Outside
On
Off
2
0.67
0.21
0.31
0.65
1.37
2.10
Outside
On
On
12
0.30
0.29
0.99
0.21
2.35
11.38
Table 5-11 shows that in all cases the concentration of radon was reduced by the SSD system operation
in the subslab sampling ports, wall ports, and shallowest interior soil gas ports. Table 5-12 shows that this
effect is, on average, about a 60% reduction in the subslab sampling ports and 80% in the wall ports.
Comparing this result with the reductions observed in indoor air suggests that the SSD system is
operating at this duplex to reduce radon in indoor air through two mechanisms—both diluting the air
beneath the slab with lower concentration air (presumably atmospheric) as well as reversing the pressure
differential across the slab.
Table 5-11. Descriptive Statistics: Radon in Subslab and Wall Ports by Individual Location and
Mitigation and Heating Status (pCi/L)
Number
Samples
Location 1
Mitigation
Heating
Mean
SD
CV
geoMean
geoSD
geoCV
SGP8-6
Not installed
Off
27
1,277
213
0.17
1,258
1.19
0.00
SGP8-6
Not installed
On
56
1,321
190
0.14
1,306
1.17
0.00
SGP8-6
Off
On
1
911
NA
NA
911
NA
NA
SGP8-6
On
On
7
386
100
0.26
375
1.28
0.00
SGP8-6
Passive
On
3
1,223
86
0.07
1,221
1.07
0.00
SGP9-6
Not installed
Off
31
1,696
129
0.08
1,691
1.08
0.00
SGP9-6
Not installed
On
61
1,598
212
0.13
1,581
1.17
0.00
SGP9-6
Off
On
1
1,349
NA
NA
1,349
NA
NA
SGP9-6
On
On
7
248
80
0.32
240
1.31
0.01
SGP9-6
Passive
On
2
1,566
323
0.21
1,549
1.23
0.00
SSP-1
Not installed
Off
26
776
323
0.42
642
2.27
0.00
SSP-1
Not installed
On
62
929
199
0.21
909
1.24
0.00
SSP-1
Off
On
1
749
NA
NA
749
NA
NA
SSP-1
On
On
7
531
97
0.18
524
1.18
0.00
SSP-1
Passive
On
3
545
195
0.36
524
1.40
0.00
SSP-2
Not installed
Off
11
1,179
158
0.13
1,169
1.14
0.00
SSP-2
Not installed
On
10
984
518
0.53
555
5.33
0.01
SSP-2
Off
On
1
1,268
NA
NA
1,268
NA
NA
SSP-2
On
Off
1
181
NA
NA
181
NA
NA
SSP-2
On
On
6
227
66
0.29
219
1.32
0.01
SSP-2
Passive
On
2
1,338
42
0.03
1,338
1.03
0.00
SSP-3
Not installed
Off
7
551
364
0.66
378
3.24
0.01
(continued)
5-28
-------
Section 5—Sub slab Depressnrization Mitigation System Monitoring Results
Table 5-11. Descriptive Statistics: Radon in Subslab and Wall Ports by Individual Location and
Mitigation and Heating Status (pCi/L) (cont.)
Number
Samples
Location 1
Mitigation
Heating
Mean
SD
CV
geoMean
geoSD
geoCV
SSP-3
Off
Off
1
1,086
NA
NA
1,086
NA
NA
SSP-3
On
Off
8
248
35
0.14
246
1.16
0.00
SSP-3
Passive
Off
2
331
86
0.26
325
1.30
0.00
SSP-4
Not installed
Off
32
1,996
158
0.08
1,990
1.08
0.00
SSP-4
Not installed
On
60
1,850
423
0.23
1,688
1.93
0.00
SSP-4
Off
On
1
1,854
NA
NA
1,854
NA
NA
SSP-4
On
On
7
733
22
0.03
733
1.03
0.00
SSP-4
Passive
On
2
1,896
170
0.09
1,892
1.09
0.00
SSP-5
Not installed
Off
75
1,245
127
0.10
1,238
1.12
0.00
SSP-5
Off
Off
1
1,178
NA
NA
1,178
NA
NA
SSP-5
On
Off
8
195
56
0.29
189
1.29
0.01
SSP-5
Passive
Off
1
1,341
NA
NA
1,341
NA
NA
SSP-6
Not installed
Off
74
1,585
269
0.17
1,552
1.26
0.00
SSP-6
Off
Off
1
1,827
NA
NA
1,827
NA
NA
SSP-6
On
Off
8
321
124
0.39
291
1.68
0.01
SSP-6
Passive
Off
2
1,634
128
0.08
1,631
1.08
0.00
SSP-7
Not installed
Off
91
495
388
0.78
300
3.41
0.01
SSP-7
Off
Off
1
1,214
NA
NA
1,214
NA
NA
SSP-7
On
Off
8
263
62
0.24
255
1.32
0.01
SSP-7
Passive
Off
2
219
175
0.80
181
2.47
0.01
WP-1
Not installed
Off
21
213
84
0.39
194
1.61
0.01
WP-1
Not installed
On
45
245
155
0.64
194
2.21
0.01
WP-1
Off
On
2
173
29
0.17
172
1.19
0.01
WP-1
On
Off
1
50
NA
NA
50
NA
NA
WP-1
On
On
8
46
9
0.19
45
1.20
0.03
WP-1
Passive
On
2
165
11
0.06
165
1.07
0.01
WP-2
Not installed
Off
24
76
34
0.44
68
1.69
0.03
WP-2
Not installed
On
44
37
28
0.74
31
1.84
0.06
WP-2
Off
On
2
22
7
0.30
22
1.36
0.06
WP-2
On
Off
1
28
NA
NA
28
NA
NA
WP-2
On
On
8
23
3
0.12
23
1.12
0.05
WP-2
Passive
On
2
29
8
0.28
28
1.33
0.05
WP-3
Not installed
Off
24
82
39
0.47
73
1.66
0.02
WP-3
Not installed
On
50
78
45
0.57
67
1.78
0.03
WP-3
Off
On
2
181
137
0.76
153
2.33
0.02
WP-3
On
Off
1
289
NA
NA
289
NA
NA
(continued)
5-29
-------
Section 5—Sub slab Depressnrization Mitigation System Monitoring Results
Table 5-11. Descriptive Statistics: Radon in Subslab and Wall Ports by Individual Location and
Mitigation and Heating Status (pCi/L) (cont.)
Number
Samples
Location 1
Mitigation
Heating
Mean
SD
CV
geoMean
geoSD
geoCV
WP-3
On
On
7
122
31
0.25
119
1.25
0.01
WP-3
Passive
On
2
124
37
0.29
121
1.35
0.01
WP-4
Not installed
Off
86
50
34
0.67
40
2.00
0.05
WP-4
Off
Off
2
17
2
0.13
17
1.14
0.07
WP-4
On
Off
9
9
6
0.67
7
1.97
0.28
WP-4
Passive
Off
2
25
2
0.07
25
1.07
0.04
Table 5-12. Radon Descriptive Statistics by Location Type and Mitigation and Heating Status
(pCi/L)
Location
Type
Number
Samples
Mitigation
Heating
Mean
SD
CV
geoMean
geoSD
geoCV
Subslab
Not installed
Off
316
1,129
583
0.52
839
2.79
0.00
Subslab
Not installed
On
132
1,352
571
0.42
1,160
2.09
0.00
Subslab
Off
Off
4
1,326
338
0.25
1,298
1.26
0.00
Subslab
Off
On
3
1,290
553
0.43
1,207
1.58
0.00
Subslab
Passive
Off
7
816
689
0.84
534
2.96
0.01
Subslab
Passive
On
7
1,158
631
0.54
989
1.90
0.00
Subslab
On
Off
33
254
86
0.34
240
1.42
0.01
Subslab
On
On
20
510
219
0.43
454
1.71
0.00
Wall
Not installed
Off
155
81
69
0.86
59
2.26
0.04
Wall
Not installed
On
139
119
129
1.08
74
2.71
0.04
Wall
Off
Off
2
17
2
0.13
17
1.14
0.07
Wall
Off
On
6
126
102
0.81
83
3.04
0.04
Wall
Passive
Off
2
25
2
0.07
25
1.07
0.04
Wall
Passive
On
6
106
65
0.61
83
2.36
0.03
Wall
On
Off
12
37
81
2.18
13
3.73
0.29
Wall
On
On
23
61
45
0.74
48
2.01
0.04
5.3 VOC Monitoring During Mitigation Testing
SSD mitigation reduced the indoor air concentration of the primary VOCs at the site PCE and chloroform
but not as dramatically or consistently as the reduction seen for radon. As shown in Figures 5-10 and
5-11 in the week immediately after installation, the SSD system appeared to have reduced the VOC
concentrations to ambient levels. The concentrations then rose over the next 2 weeks of operation. During
the two subsequent operational periods—December 12, 2012, to December 29, 2012 and February 6,
2013, to April 24, 2013—concentrations were reduced compared with the unmitigated periods but did not
reach ambient concentrations again until late in April 2013 when temperatures had moderated.
5-30
-------
Section 5—Subslab Depressurization Mitigation System Monitoring Results
100
Indoor Air PCE, Mitigation Period
z* 10
DO
3
C
o
c
Of
e
o
u
0-1
o Q ° Uv I
0 H • D © ¦
SyS.iBSo•
o SI "S
i*A e qb
O 4»Fir«m
¦ 4»&is#MPCf
4»HafteSPC(
• 4?2 Am KF
¦ 4?2&«#NFCE
c waiyspci
—Ambiwt PCE
^^•>M4tigaAlon On
MUifrtlort P*»ru»
iow Ivr n(
fan Vn1ifi|{
ov* M*«; 4"
0.01
m
m
O
©
Date
o
CM
*N
O
5
o
Figure 5-10. Passive sampler monitoring of PCE during mitigation testing.
indoor Air Chloroform, Mitigation Period
so
fiJD
3
C
Q
C
<2J
« 0.5
c
o
u
0.05
•
¦
sE
® 1 a ~
„ -
¦ * ©
© ©
k-1
V"
1
~
~ H ~
i'
op '
loo
9^
&
»
K
*® ®S
o
iISs
~
~ B
¦I
~ •
~ ~ D
m ~
i,0l
8
*©
«
¦Hi
cj
H
rt
N
s
o
H
iH
rH
*-§
m
a s
Date
O
rl
IM
Q
try
O
o
a
o
4*0 CHOI
4KJ £Uw M ana*
4?0 Bam S CHUJ
42J Hfrt CMO i
*il to!^OIO>
4£lfc*fr«$atGI
> Arret icn t QICI3
¦WigelidnOn
- MUfMbn P«>Mf
¦Vww Evtnt
¦ I §n T-r%t>nf
-VkwMp* V
»5*wwM«fc"
Figure 5-11. Passive sampler monitoring of chloroform during mitigation period.
5-31
-------
Section 5—Subslab Depressurization Mitigation System Monitoring Results
Real-time GC data were also available for a portion of the mitigation testing (see Figure 5-12 for PCE).
This shows two distinct upward spikes during the February 6 through April period when the SSD system
was on. These are prominent in the 422 basement south GC plot. One occurred on February 16 and 17
(with a secondary peak on February 19) and the other on March 12-15 with a secondary peak on March
17. Brief snow events were noted on February 15, February 19, and March 12. Both peaks were also
associated with west northwest to west winds which as we will show in Section 9 are associated with high
differential pressures and radon concentrations (see Figure 9-14). February 16 and 17 had rather low
temperatures (average 23° F both days) as compared to the surrounding days. However during both GC
peaks the 422 subslab versus basement differential pressure remained at -15 Pa and the 420 subslab
versus basement differential pressure showed very little deflection remaining at -13 to -15 Pa.
pi
E
c
o
1
0.1
10
1
0.1
co 10
1
c
o
o
O 0.1
10
1
420First
422First
Hartman3
Tetrachloroethene
420BaseS
422BaseS
Mitigation Off
Mitigation On
Passive Mitigation
Flags
EE Missing
— U
Figure 5-12. Indoor air PCE, real-time monitoring during mitigation testing.
5-32
-------
Section 5—Sub slab Depressnrization Mitigation System Monitoring Results
5.3.1 Descriptive Statistics
Descriptive statistics (Table 5-13) show a 68% reduction in the mean chloroform concentration in indoor
air with the SSD system turned on during the heating season. The corresponding reduction in mean PCE
with active mitigation in the heating season was 61%. These reductions are both less than those achieved
for radon in this house (about 91%) and substantially less than the 99% reduction generally considered to
be possible for SSD systems operating in high initial concentration vapor intrusion situations (U.S. EPA,
2008).
Table 5-13. Descriptive Statistics of Weekly Passive VOC Measurements (|jg/m3) in Indoor Air by
Mitigation Status and Heating Use (yellow indicates statistics during active
mitigation)
Number
Samples
Variable
Mitigation
Heating
Mean
SD
cv
geoMean
geoSD
geoCV
Chloroform
Not installed
Off
135
0.21
0.13
0.61
0.18
1.78
10.00
Chloroform
Not installed
On
235
0.41
0.50
1.23
0.26
2.45
9.34
Chloroform
Off
On
20
0.36
0.27
0.74
0.28
2.09
7.54
Chloroform
Passive
On
43
0.40
0.31
0.76
0.31
2.09
6.75
Chloroform
On
Off
14
0.13
0.05
0.35
0.13
1.40
10.96
Chloroform
On
On
90
0.13
0.07
0.52
0.11
1.59
14.19
Tetrachloroethene
Not installed
Off
135
0.39
0.22
0.56
0.34
1.72
5.04
Tetrachloroethene
Not installed
On
235
1.36
2.77
2.05
0.53
3.34
6.33
Tetrachloroethene
Off
On
20
1.24
1.08
0.87
0.87
2.45
2.83
Tetrachloroethene
Passive
On
43
1.40
1.35
0.97
0.88
2.78
3.17
Tetrachloroethene
On
Off
14
0.22
0.10
0.43
0.21
1.43
6.90
Tetrachloroethene
On
On
90
0.53
0.52
0.98
0.36
2.41
6.76
The distribution of the VOC data (Table 5-14) shows that the primary effect of the SSD system was to
cut off the highest end of the distribution (90th and 95th percentiles). The trend of the active mitigation
improving indoor air, but to a lesser extent than would have been predicted from radon, holds at all of the
individual monitoring locations (Table 5-15). An explanation for this surprising finding is that the VOC
concentrations in the area immediately outside the building envelope (subslab and wall ports) were
increased by the SSD system, especially for PCE (Table 5-16). Because the concentration of radon
decreased substantially in these areas, this suggests that air is being drawn into the area around the
building envelope at least in part from a zone of high VOC concentration and lower radon concentration.
This effect is seen most dramatically in the higher concentration portion of the distribution for PCE. For
example, during the heating season, the 75th percentile of the subslab data with the mitigation on exceeds
the 95th percentile with the mitigation off (Table 5-17). This trend holds for most, although not all,
subslab and wall ports (Table 5-18). In some cases, the increases are dramatic; for example, PCE
increased 875% at SSP-3, 575% at SSP-4, and 2,000% at SSP-7. Given the observed pressure
differentials, this most likely indicates that the mitigation system is drawing VOCs in from near the water
table. Such an effect has been previously hypothesized (Lutes, 2010b) but not to our knowledge published
in detail.
5-33
-------
Section 5—Sub slab Depressnrization Mitigation System Monitoring Results
Table 5-14. Distribution of Concentrations (ng/m3) by VOC and Mitigation and Heating Status:
Indoor Air, Week-Long Passive Samples (yellow indicates statistics during active
mitigation)
Number
Samples
5th
Percen-
tile
10th
Percen-
tile
25th
Percen-
tile
50th
Percen-
tile
75th
Percen-
tile
90th
Percen-
tile
95th
Percen-
tile
Variable
Mitigation
Heating
Chloroform
Not installed
Off
135
0.07
0.09
0.11
0.18
0.26
0.37
0.47
Chloroform
Not installed
On
235
0.07
0.09
0.13
0.23
0.50
0.93
1.20
Chloroform
Off
On
20
0.12
0.12
0.16
0.19
0.50
0.78
0.85
Chloroform
Passive
On
43
0.12
0.12
0.14
0.32
0.53
0.78
0.99
Chloroform
On
Off
14
0.08
0.09
0.10
0.13
0.15
0.21
0.22
Chloroform
On
On
90
0.07
0.07
0.07
0.12
0.16
0.21
0.25
Tetrachloroethene
Not installed
Off
135
0.14
0.17
0.24
0.33
0.52
0.69
0.87
Tetrachloroethene
Not installed
On
235
0.11
0.15
0.22
0.40
0.92
3.64
6.63
Tetrachloroethene
Off
On
20
0.22
0.33
0.36
1.00
1.63
2.77
3.44
Tetrachloroethene
Passive
On
43
0.16
0.22
0.45
0.93
1.75
3.52
4.08
Tetrachloroethene
On
Off
14
0.14
0.16
0.17
0.19
0.24
0.28
0.36
Tetrachloroethene
On
On
90
0.09
0.10
0.20
0.31
0.71
1.20
1.63
5-34
-------
Table 5-15. Descriptive Statistics of Indoor VOC Concentrations (pg/m3) During Mitigation Testing by Location and Mitigation and
Heating Status (yellow indicates statistics during active mitigation)
Number
Samples
Location 1
Variable
Mitigation
Heating
Mean
SD
cv
geoMean
geoSD
geoCV
420BaseN
Chloroform
Not installed
Off
22
0.20
0.10
0.52
0.17
1.78
10.45
420BaseN
Chloroform
Not installed
On
37
0.18
0.12
0.68
0.15
1.76
11.41
420BaseN
Chloroform
Off
On
3
0.18
0.02
0.09
0.18
1.09
6.20
420BaseN
Chloroform
On
Off
2
0.11
0.04
0.40
0.10
1.51
14.47
420BaseN
Chloroform
On
On
13
0.10
0.04
0.44
0.09
1.45
16.26
420BaseN
Chloroform
Passive
On
7
0.21
0.10
0.51
0.19
1.61
8.69
420BaseN
Tetrachloroethene
Not installed
Off
22
0.31
0.12
0.39
0.29
1.46
5.05
420BaseN
Tetrachloroethene
Not installed
On
37
0.82
1.94
2.36
0.32
2.94
9.09
420BaseN
Tetrachloroethene
Off
On
3
0.65
0.48
0.74
0.55
1.99
3.63
420BaseN
Tetrachloroethene
On
Off
2
0.20
0.06
0.28
0.20
1.33
6.80
420BaseN
Tetrachloroethene
On
On
13
0.30
0.19
0.65
0.25
1.86
7.49
420BaseN
Tetrachloroethene
Passive
On
7
0.69
0.50
0.73
0.51
2.45
4.79
420BaseS
Chloroform
Not installed
Off
25
0.18
0.08
0.47
0.16
1.66
10.53
420BaseS
Chloroform
Not installed
On
44
0.18
0.17
0.95
0.15
1.87
12.83
420BaseS
Chloroform
Off
On
5
0.16
0.02
0.13
0.16
1.13
7.22
420BaseS
Chloroform
On
Off
4
0.15
0.05
0.38
0.14
1.43
10.34
420BaseS
Chloroform
On
On
25
0.10
0.04
0.36
0.10
1.41
14.40
420BaseS
Chloroform
Passive
On
8
0.25
0.16
0.62
0.21
1.80
8.38
420BaseS
Tetrachloroethene
Not installed
Off
25
0.35
0.12
0.35
0.33
1.43
4.38
420BaseS
Tetrachloroethene
Not installed
On
44
0.92
2.29
2.50
0.36
2.85
7.84
420BaseS
Tetrachloroethene
Off
On
5
0.61
0.35
0.58
0.54
1.77
3.30
420BaseS
Tetrachloroethene
On
Off
4
0.23
0.06
0.27
0.22
1.31
5.93
420BaseS
Tetrachloroethene
On
On
25
0.30
0.17
0.57
0.26
1.81
6.98
420BaseS
Tetrachloroethene
Passive
On
8
0.85
0.55
0.64
0.65
2.38
3.66
(continued)
-------
Table 5-15. Descriptive Statistics of Indoor VOC Concentrations (pg/m3) During Mitigation Testing by Location and Mitigation and
Heating Status (yellow indicates statistics during active mitigation) (cont.)
Number
Samples
Location 1
Variable
Mitigation
Heating
Mean
SD
cv
geoMean
geoSD
geoCV
420First
Chloroform
Not installed
Off
22
0.16
0.06
0.40
0.14
1.56
10.86
420First
Chloroform
Not installed
On
37
0.22
0.26
1.18
0.15
2.09
13.50
420First
Chloroform
Off
On
3
0.12
0.00
0.00
0.12
1.00
8.33
420First
Chloroform
On
Off
2
0.13
0.07
0.51
0.12
1.71
13.94
420First
Chloroform
On
On
13
0.09
0.03
0.35
0.09
1.38
15.86
420First
Chloroform
Passive
On
7
0.17
0.09
0.53
0.16
1.58
10.20
420First
Tetrachloroethene
Not installed
Off
22
0.24
0.10
0.42
0.22
1.53
6.95
420First
Tetrachloroethene
Not installed
On
37
1.34
3.89
2.90
0.27
4.06
14.82
420First
Tetrachloroethene
Off
On
3
0.38
0.30
0.78
0.32
2.06
6.47
420First
Tetrachloroethene
On
Off
2
0.19
0.01
0.07
0.19
1.08
5.68
420First
Tetrachloroethene
On
On
13
0.21
0.11
0.51
0.19
1.64
8.86
420First
Tetrachloroethene
Passive
On
7
0.44
0.29
0.64
0.36
2.14
6.03
422BaseN
Chloroform
Not installed
Off
22
0.24
0.14
0.60
0.20
1.91
9.73
422BaseN
Chloroform
Not installed
On
37
0.48
0.30
0.62
0.40
1.89
4.71
422BaseN
Chloroform
Off
On
3
0.61
0.15
0.24
0.59
1.29
2.17
422BaseN
Chloroform
On
Off
2
0.12
0.03
0.24
0.12
1.27
10.72
422BaseN
Chloroform
On
On
13
0.14
0.06
0.41
0.13
1.56
11.76
422BaseN
Chloroform
Passive
On
7
0.63
0.30
0.48
0.58
1.55
2.68
422BaseN
Tetrachloroethene
Not installed
Off
22
0.49
0.21
0.42
0.45
1.59
3.57
422BaseN
Tetrachloroethene
Not installed
On
37
1.52
2.34
1.54
0.84
2.57
3.06
422BaseN
Tetrachloroethene
Off
On
3
2.07
1.19
0.58
1.85
1.77
0.95
422BaseN
Tetrachloroethene
On
Off
2
0.21
0.04
0.20
0.21
1.23
5.90
422BaseN
Tetrachloroethene
On
On
13
0.79
0.41
0.52
0.64
2.16
3.35
422BaseN
Tetrachloroethene
Passive
On
7
2.23
1.45
0.65
1.72
2.36
1.37
(continued)
-------
Table 5-15. Descriptive Statistics of Indoor VOC Concentrations (pg/m3) During Mitigation Testing by Location and Mitigation and
Heating Status (yellow indicates statistics during active mitigation) (cont.)
Number
Samples
Location 1
Variable
Mitigation
Heating
Mean
SD
cv
geoMean
geoSD
geoCV
422BaseS
Chloroform
Not installed
Off
22
0.28
0.17
0.61
0.23
1.89
8.10
422BaseS
Chloroform
Not installed
On
40
0.71
0.41
0.58
0.59
1.87
3.14
422BaseS
Chloroform
Off
On
3
0.82
0.05
0.06
0.82
1.06
1.29
422BaseS
Chloroform
On
Off
2
0.17
0.08
0.50
0.16
1.68
10.59
422BaseS
Chloroform
On
On
13
0.21
0.09
0.45
0.19
1.71
9.17
422BaseS
Chloroform
Passive
On
7
0.76
0.36
0.47
0.69
1.64
2.39
422BaseS
Tetrachloroethene
Not installed
Off
22
0.65
0.25
0.38
0.59
1.58
2.65
422BaseS
Tetrachloroethene
Not installed
On
40
2.38
3.61
1.51
1.30
2.63
2.03
422BaseS
Tetrachloroethene
Off
On
3
2.87
1.16
0.40
2.71
1.51
0.56
422BaseS
Tetrachloroethene
On
Off
2
0.36
0.21
0.59
0.33
1.87
5.72
422BaseS
Tetrachloroethene
On
On
13
1.26
0.83
0.66
0.87
2.96
3.41
422BaseS
Tetrachloroethene
Passive
On
7
2.93
1.90
0.65
2.29
2.26
0.99
422First
Chloroform
Not installed
Off
22
0.22
0.14
0.66
0.18
1.76
9.61
422First
Chloroform
Not installed
On
40
0.67
0.92
1.37
0.39
2.53
6.43
422First
Chloroform
Off
On
3
0.41
0.02
0.05
0.41
1.05
2.59
422First
Chloroform
On
Off
2
0.12
0.03
0.24
0.12
1.27
10.72
422First
Chloroform
On
On
13
0.13
0.05
0.39
0.12
1.49
12.38
422First
Chloroform
Passive
On
7
0.42
0.22
0.53
0.38
1.63
4.30
422First
Tetrachloroethene
Not installed
Off
22
0.33
0.22
0.67
0.28
1.73
6.11
422First
Tetrachloroethene
Not installed
On
40
1.17
1.80
1.55
0.61
2.77
4.55
422First
Tetrachloroethene
Off
On
3
1.26
0.31
0.25
1.23
1.28
1.04
422First
Tetrachloroethene
On
Off
2
0.14
0.04
0.26
0.13
1.30
9.82
422First
Tetrachloroethene
On
On
13
0.50
0.28
0.56
0.41
2.10
5.10
422First
Tetrachloroethene
Passive
On
7
1.31
0.87
0.66
1.04
2.22
2.14
(continued)
-------
Table 5-15. Descriptive Statistics of Indoor VOC Concentrations (pg/m3) During Mitigation Testing by Location and Mitigation and
Heating Status (yellow indicates statistics during active mitigation) (cont.)
Number
Samples
Location 1
Variable
Mitigation
Heating
Mean
SD
cv
geoMean
geoSD
geoCV
Outside
Chloroform
Not Installed
Off
22
0.09
0.03
0.29
0.09
1.36
15.98
Outside
Chloroform
Not Installed
On
41
0.08
0.02
0.28
0.08
1.30
16.99
Outside
Chloroform
Off
On
3
0.08
0.00
0.00
0.08
1.00
13.33
Outside
Chloroform
On
Off
2
0.09
0.01
0.08
0.09
1.08
12.04
Outside
Chloroform
On
On
13
0.09
0.03
0.30
0.08
1.30
15.38
Outside
Chloroform
Passive
On
7
0.12
0.05
0.45
0.11
1.51
14.08
Outside
Tetrachloroethene
Not installed
Off
22
0.16
0.08
0.50
0.15
1.46
9.95
Outside
Tetrachloroethene
Not installed
On
41
0.12
0.05
0.39
0.12
1.47
12.62
Outside
Tetrachloroethene
Off
On
3
0.11
0.03
0.24
0.11
1.25
11.57
Outside
Tetrachloroethene
On
Off
2
0.12
0.10
0.81
0.10
2.50
25.15
Outside
Tetrachloroethene
On
On
13
0.13
0.05
0.41
0.12
1.45
12.04
Outside
Tetrachloroethene
Passive
On
7
0.20
0.15
0.74
0.17
1.76
10.23
O'
V*
Cr-
ET
b
£
I
O'
3
O'
#
o
§
S
3'
Co
£
-------
Table 5-16. Descriptive Statistics: Average Subslab and Wall Port VOC Concentrations (pg/m3) by Mitigation and Heating Status
(yellow indicates statistics during active mitigation)
i
U>
Location
Type
Number
Samples
Variable
Mitigation
Heating
Mean
SD
cv
geoMean
geoSD
geoCV
Subslab
Chloroform
Not installed
Off
244
22.73
51.84
2.28
6.93
3.56
0.51
Subslab
Chloroform
Not installed
On
115
147.86
80.55
0.54
110.47
2.80
0.03
Subslab
Chloroform
Off
Off
10
20.50
19.62
0.96
15.39
2.04
0.13
Subslab
Chloroform
Passive
On
3
207.37
245.97
1.19
52.06
17.25
0.33
Subslab
Chloroform
Passive
Off
4
2.10
0.00
0.00
2.10
1.00
0.48
Subslab
Chloroform
On
Off
32
51.10
54.24
1.06
19.27
5.61
0.29
Subslab
Chloroform
On
On
24
85.32
122.98
1.44
21.81
7.48
0.34
Subslab
Tetrachloroethene
Not installed
Off
244
54.82
57.97
1.06
33.40
2.80
0.08
Subslab
Tetrachloroethene
Not installed
On
115
230.47
197.25
0.86
182.27
2.17
0.01
Subslab
Tetrachloroethene
Off
Off
10
64.17
59.40
0.93
38.61
3.06
0.08
Subslab
Tetrachloroethene
Passive
Off
4
21.91
8.89
0.41
20.26
1.63
0.08
Subslab
Tetrachloroethene
Passive
On
3
492.63
420.54
0.85
162.14
13.69
0.08
Subslab
Tetrachloroethene
On
Off
32
154.54
278.40
1.80
42.33
5.08
0.12
Subslab
Tetrachloroethene
On
On
24
359.46
489.24
1.36
61.93
12.45
0.20
Wall
Chloroform
Not installed
Off
107
12.33
68.16
5.53
4.04
2.26
0.56
Wall
Chloroform
Not installed
On
116
5.35
10.13
1.90
3.69
1.86
0.50
Wall
Chloroform
Off
Off
1
4.00
NA
NA
4.00
NA
NA
Wall
Chloroform
Passive
Off
1
2.10
NA
NA
2.10
NA
NA
Wall
Chloroform
Passive
On
3
2.10
0.00
0.00
2.10
1.00
0.48
Wall
Chloroform
On
Off
6
2.10
0.00
0.00
2.10
1.00
0.48
Wall
Chloroform
On
On
18
14.11
20.08
1.42
5.35
3.99
0.75
Wall
Tetrachloroethene
Not installed
Off
107
13.02
74.46
5.72
4.88
1.95
0.40
Wall
Tetrachloroethene
Not installed
On
116
14.63
50.70
3.47
5.73
2.33
0.41
Wall
Tetrachloroethene
Off
Off
1
1.80
NA
NA
1.80
NA
NA
Wall
Tetrachloroethene
Passive
Off
1
1.75
NA
NA
1.75
NA
NA
Wall
Tetrachloroethene
Passive
On
3
46.47
72.40
1.56
12.03
8.93
0.74
Wall
Tetrachloroethene
On
Off
6
2.19
1.08
0.49
2.04
1.46
0.71
Wall
Tetrachloroethene
On
On
18
60.82
150.56
2.48
7.03
7.58
1.08
&
O'
S
V*
Cr-
ET
b
£
I
O'
s
o'
f
c<3
o
§
S
3'
£
-------
Table 5-17. Distribution of Subslab and Wall Port VOC Concentrations (ng/m3) by Mitigation and Heating Status (yellow indicates
statistics during active mitigation)
Location
Type
Variable
Mitigation
Heating
Number
Samples
5th
Percentile
10th
Percentile
25th
Percentile
50th
Percentile
75th
Percentile
90th
Percentile
95th
Percentile
Subslab
Chloroform
Not installed
Off
244
3
3
3
3
11
75
139
Subslab
Chloroform
Not installed
On
115
6
58
82
140
210
266
273
Subslab
Chloroform
Off
Off
10
10
10
10
10
15
56
58
Subslab
Chloroform
Passive
On
3
16
30
71
140
310
412
446
Subslab
Chloroform
Passive
Off
4
2
2
2
2
2
2
2
Subslab
Chloroform
On
Off
32
2
2
2
28
82
129
159
Subslab
Chloroform
On
On
24
2
2
2
42
135
187
216
Subslab
Tetrachloroethene
Not installed
Off
244
7
9
17
31
75
120
180
Subslab
Tetrachloroethene
Not installed
On
115
91
105
150
190
240
336
457
Subslab
Tetrachloroethene
Off
Off
10
10
10
17
28
128
140
140
Subslab
Tetrachloroethene
Passive
Off
4
12
14
19
23
26
29
30
Subslab
Tetrachloroethene
Passive
On
3
78
148
359
710
735
750
755
Subslab
Tetrachloroethene
On
Off
32
6
7
11
32
95
456
613
Subslab
Tetrachloroethene
On
On
24
2
2
5
104
640
907
1336
Wall
Chloroform
Not installed
Off
107
3
3
3
3
3
8
15
Wall
Chloroform
Not installed
On
116
2
3
3
3
3
7
13
Wall
Chloroform
Off
Off
1
4
4
4
4
4
4
4
Wall
Chloroform
Passive
Off
1
2
2
2
2
2
2
2
Wall
Chloroform
Passive
On
3
2
2
2
2
2
2
2
Wall
Chloroform
On
Off
6
2
2
2
2
2
2
2
Wall
Chloroform
On
On
18
2
2
2
2
23
43
55
Wall
Tetrachloroethene
Not installed
Off
107
4
4
4
4
4
5
8
Wall
Tetrachloroethene
Not installed
On
116
4
4
4
4
5
10
24
Wall
Tetrachloroethene
Off
Off
1
2
2
2
2
2
2
2
Wall
Tetrachloroethene
Passive
Off
1
2
2
2
2
2
2
2
Wall
Tetrachloroethene
Passive
On
3
2
3
5
8
69
106
118
Wall
Tetrachloroethene
On
Off
6
2
2
2
2
2
3
4
Wall
Tetrachloroethene
On
On
18
2
2
2
2
43
129
288
O'
V*
Cr-
ET
b
£
I
O'
s
o'
f
c<3
o
§
S
3'
£
-------
Table 5-18. Descriptive Statistics of Subslab and Wall Port VOC Concentrations (pg/m3) by Location and Mitigation and Heating Status
(yellow indicates statistics during active mitigation)
Number
Samples
Location 1
Variable
Mitigation
Heating
Mean
SD
CV
geoMean
geoSD
geoCV
SSP-1
Chloroform
Not installed
Off
35
74
91.30
1.23
37
3.32
0.09
SSP-1
Chloroform
Not installed
On
67
121
57.72
0.48
104
1.94
0.02
SSP-1
Chloroform
Off
Off
2
58
2.12
0.04
57
1.04
0.02
SSP-1
Chloroform
On
On
8
43
47.49
1.11
20
4.74
0.23
SSP-1
Chloroform
Passive
On
1
140
NA
NA
140
NA
NA
SSP-1
Tetrachloroethene
Not installed
Off
35
115
47.51
0.41
102
1.92
0.02
SSP-1
Tetrachloroethene
Not installed
On
67
292
237.05
0.81
244
1.74
0.01
SSP-1
Tetrachloroethene
Off
Off
2
140
0.00
0.00
140
1.00
0.01
SSP-1
Tetrachloroethene
On
On
8
192
225.55
1.18
120
2.68
0.02
SSP-1
Tetrachloroethene
Passive
On
1
760
NA
NA
760
NA
NA
SSP-2
Chloroform
Not installed
Off
2
8
4.56
0.55
8
1.80
0.24
SSP-2
Chloroform
Not installed
On
3
4
2.19
0.50
4
1.59
0.39
SSP-2
Chloroform
Off
Off
2
10
0.00
0.00
10
1.00
0.10
SSP-2
Chloroform
On
On
8
5
6.82
1.51
3
2.27
0.81
SSP-2
Chloroform
Passive
On
1
2
NA
NA
2
NA
NA
SSP-2
Tetrachloroethene
Not installed
Off
2
12
2.51
0.21
12
1.23
0.10
SSP-2
Tetrachloroethene
Not installed
On
3
5
1.88
0.35
5
1.39
0.27
SSP-2
Tetrachloroethene
Off
Off
2
10
0.21
0.02
10
1.02
0.10
SSP-2
Tetrachloroethene
On
On
8
3
1.84
0.62
3
1.74
0.67
SSP-2
Tetrachloroethene
Passive
On
1
8
NA
NA
8
NA
NA
SSP-3
Chloroform
Not installed
Off
9
8
8.52
1.10
6
2.14
0.38
SSP-3
Chloroform
Off
Off
2
10
0.00
0.00
10
1.00
0.10
SSP-3
Chloroform
On
Off
8
65
29.80
0.46
58
1.65
0.03
SSP-3
Chloroform
Passive
Off
1
2
NA
NA
2
NA
NA
(continued)
-------
Table 5-18. Descriptive Statistics of Subslab and Wall Port VOC Concentrations (pg/m3) by Location and Mitigation and Heating Status
(yellow indicates statistics during active mitigation) (cont.)
Number
Samples
Location 1
Variable
Mitigation
Heating
Mean
SD
CV
geoMean
geoSD
geoCV
SSP-3
Tetrachloroethene
Not installed
Off
9
21
21.38
1.02
16
2.09
0.13
SSP-3
Tetrachloroethene
Off
Off
2
18
2.83
0.16
18
1.17
0.07
SSP-3
Tetrachloroethene
On
Off
8
184
208.14
1.13
83
4.51
0.05
SSP-3
Tetrachloroethene
Passive
Off
1
22
NA
NA
22
NA
NA
SSP-4
Chloroform
Not installed
Off
15
101
69.30
0.69
64
3.72
0.06
SSP-4
Chloroform
Not installed
On
45
198
82.20
0.42
151
3.03
0.02
SSP-4
Chloroform
Off
Off
2
15
0.00
0.00
15
1.00
0.07
SSP-4
Chloroform
On
On
8
209
142.97
0.68
182
1.66
0.01
SSP-4
Chloroform
Passive
On
1
480
NA
NA
480
NA
NA
SSP^
Tetrachloroethene
Not installed
Off
15
188
91.43
0.49
124
4.02
0.03
SSP^
Tetrachloroethene
Not installed
On
45
153
32.88
0.21
150
1.25
0.01
SSP-4
Tetrachloroethene
Off
Off
2
125
7.07
0.06
125
1.06
0.01
SSP-4
Tetrachloroethene
On
On
8
884
493.99
0.56
770
1.78
0.00
SSP-4
Tetrachloroethene
Passive
On
1
710
NA
NA
710
NA
NA
SSP-5
Chloroform
Not installed
Off
67
11
29.72
2.77
5
2.52
0.55
SSP-5
Chloroform
Off
Off
2
10
0.00
0.00
10
1.00
0.10
SSP-5
Chloroform
On
Off
8
11
12.54
1.13
5
3.68
0.68
SSP-5
Chloroform
Passive
Off
1
2
NA
NA
2
NA
NA
SSP-5
Tetrachloroethene
Not installed
Off
67
44
29.66
0.67
34
2.18
0.06
SSP-5
Tetrachloroethene
Off
Off
2
28
4.24
0.15
28
1.16
0.04
SSP-5
Tetrachloroethene
On
Off
8
25
8.83
0.36
23
1.44
0.06
SSP-5
Tetrachloroethene
Passive
Off
1
32
NA
NA
32
NA
NA
SSP-6
Chloroform
Not installed
Off
70
5
12.84
2.46
4
1.74
0.48
SSP-6
Chloroform
On
Off
8
15
28.67
1.92
4
4.26
0.95
SSP-6
Chloroform
Passive
Off
1
2
NA
NA
2
NA
NA
(continued)
-------
Table 5-18. Descriptive Statistics of Subslab and Wall Port VOC Concentrations (pg/m3) by Location and Mitigation and Heating Status
(yellow indicates statistics during active mitigation) (cont.)
Number
Samples
Location 1
Variable
Mitigation
Heating
Mean
SD
CV
geoMean
geoSD
geoCV
SSP-6
Tetrachloroethene
Not installed
Off
70
37
27.57
0.74
29
2.00
0.07
SSP-6
Tetrachloroethene
On
Off
8
64
160.04
2.50
12
4.58
0.39
SSP-6
Tetrachloroethene
Passive
Off
1
10
NA
NA
10
NA
NA
SSP-7
Chloroform
Not installed
Off
46
6
5.00
0.86
5
1.72
0.36
SSP-7
Chloroform
On
Off
8
114
55.98
0.49
98
1.95
0.02
SSP-7
Chloroform
Passive
Off
1
2
NA
NA
2
NA
NA
SSP-7
Tetrachloroethene
Not installed
Off
46
17
11.93
0.71
13
1.99
0.15
SSP-7
Tetrachloroethene
On
Off
8
345
450.57
1.30
141
4.91
0.03
SSP-7
Tetrachloroethene
Passive
Off
1
24
NA
NA
24
NA
NA
WP-1
Chloroform
Not installed
Off
17
4
2.59
0.69
3
1.48
0.43
WP-1
Chloroform
Not installed
On
41
3
1.02
0.31
3
1.35
0.42
WP-1
Chloroform
On
On
6
6
8.63
1.54
3
2.67
0.85
WP-1
Chloroform
Passive
On
1
2
NA
NA
2
NA
NA
WP-1
Tetrachloroethene
Not installed
Off
17
4
0.64
0.16
4
1.25
0.31
WP-1
Tetrachloroethene
Not installed
On
41
4
0.71
0.17
4
1.23
0.29
WP-1
Tetrachloroethene
On
On
6
2
0.00
0.00
2
1.00
0.57
WP-1
Tetrachloroethene
Passive
On
1
2
NA
NA
2
NA
NA
WP-2
Chloroform
Not installed
Off
18
4
2.06
0.56
3
1.39
0.40
WP-2
Chloroform
Not installed
On
39
6
14.80
2.53
4
1.85
0.51
WP-2
Chloroform
Off
Off
1
4
NA
NA
4
NA
NA
WP-2
Chloroform
On
On
6
6
9.94
1.61
3
2.81
0.88
WP-2
Chloroform
Passive
On
1
2
NA
NA
2
NA
NA
WP-2
Tetrachloroethene
Not installed
Off
18
7
11.49
1.65
5
1.82
0.37
WP-2
Tetrachloroethene
Not installed
On
39
4
0.43
0.10
4
1.13
0.26
WP-2
Tetrachloroethene
Off
Off
1
2
NA
NA
2
NA
NA
(continued)
-------
Table 5-18. Descriptive Statistics of Subslab and Wall Port VOC Concentrations (pg/m3) by Location and Mitigation and Heating Status
(yellow indicates statistics during active mitigation) (cont.)
Number
Samples
Location 1
Variable
Mitigation
Heating
Mean
SD
cv
geoMean
geoSD
geoCV
WP-2
Tetrachloroethene
On
On
6
2
1.31
0.57
2
1.53
0.73
WP-2
Tetrachloroethene
Passive
On
1
8
NA
NA
8
NA
NA
WP-3
Chloroform
Not installed
Off
14
12
20.72
1.80
6
2.64
0.44
WP-3
Chloroform
Not installed
On
36
7
9.49
1.34
4
2.30
0.52
WP-3
Chloroform
On
On
6
31
26.66
0.87
15
4.87
0.32
WP-3
Chloroform
Passive
On
1
2
NA
NA
2
NA
NA
WP-3
Tetrachloroethene
Not installed
Off
14
7
8.41
1.16
6
1.80
0.32
WP-3
Tetrachloroethene
Not installed
On
36
38
87.50
2.33
11
3.55
0.32
WP-3
Tetrachloroethene
On
On
6
178
228.42
1.28
95
3.41
0.04
WP-3
Tetrachloroethene
Passive
On
1
130
NA
NA
130
NA
NA
WP-4
Chloroform
Not installed
Off
58
18
91.99
5.19
4
2.57
0.63
WP-4
Chloroform
On
Off
6
2
0.00
0.00
2
1.00
0.48
WP-4
Chloroform
Passive
Off
1
2
NA
NA
2
NA
NA
WP-4
Tetrachloroethene
Not installed
Off
58
19
100.87
5.34
5
2.20
0.44
WP-4
Tetrachloroethene
On
Off
6
2
1.08
0.49
2
1.46
0.71
WP-4
Tetrachloroethene
Passive
Off
1
2
NA
NA
2
NA
NA
O'
V*
Cr-
ET
b
£
I
O'
3
O'
#
o
§
S
3'
Co
£
-------
Section 5—Subslab Depressurization Mitigation System Monitoring Results
5.3.2 Effect of Mitigation System Status on Indoor Air VOC Levels
Figure 5-13 compares the distributions of PCE and chloroform concentrations from weekly Radiello
samples by indoor air sampling location and mitigation status. The SSD system appears to reduce the
variability of the indoor air concentrations (as shown by the smaller green boxes for mitigation on in the
figure) and at all locations, the distributions VOCs in indoor air were lower with mitigation on. To test the
significance of this difference, we first investigated whether the populations (mitigation on and mitigation
off by location and compound) are log-normally distributed, have the same variance, and are independent
from one another.
As shown in Appendix C, all 24 sampling locations/mitigation status/compound combinations tested had
the same variance (based on an F-test) and only six of the 24 failed a Shapiro-Wilk test of log-normality.
With respect to the assumption of independence, we believe that VOC concentrations from consecutive
weeks are not autocorrelated due to the known air exchange rate and results from other published
research, but significant autocorrelation beyond a week was found in our data analysis in some cases (see
Chapter 10). However, because the data being examined in this section do not span an entire year, the
data cannot be detrended and this may contribute to the observed autocorrelation. If the data were
autocorrelated across weeks, the data in each of the two populations considered for each comparison (the
two populations are mitigation "On" observations and mitigation "Off' observations) would be more
similar among themselves than truly randomly chosen observations from each population would be.
That being said, the results are quite convincing. Using a two sided two sample t-test to test the difference
between the log-concentrations with mitigation "On" and "Off' with the null hypothesis that the
difference between the two populations is 0 (that is to say that the null hypothesis is that mitigation has no
effect) provides p-values for that hypothesis that were well below 0.05 for all mitigation
status/location/compound cases tested, with the highest p-value observed being 0.019. Additional details
on this analysis and these results can be found in Appendix C.
5-45
-------
Section 5—Subslab Depressurization Mitigation System Monitoring Results
Effect of Mitigation
on
Indoor Air VOC Concentration
1?
0 1 3
00
1§
o 011
O g
420
*6
422
%
(alt)
CD
Q)
0 1
q)
E
AO^
y *
& -,
-------
Section 5—Subslab Depressurization Mitigation System Monitoring Results
The effect of mitigation on certain particular subslab and wall ports is also shown graphically later in this
document in Figures 6-6, 6-7, and 7-1.
An alternate way of understanding why mitigation could be more effective for radon than VOCs can be
stated as follows:
¦ Radon is formed from the radioactive decay of radium, and then must leave the soil grain and
travel into the fluid filled pore space to produce a risk (a process called emanation). As a first
approximation, the concentration of radon in soil gas is controlled by the emanation rate and the
soil air permeability. The exhalation rate, or the rate of radon release from a soil surface, in turn
depends on the radium concentration of the soil, its moisture content, and temperature (Lewis and
Houle, 2009). However the exhalation rate is not expected to be changed significantly by
increasing the air flow rate through the soil. Therefore assuming a relatively uniform soil profile
with regard to exhalation rate, the concentration of radon in the soil gas may be depleted with
additional flow.
¦ Given its short half-life, radon in soil gas entering a structure must have emanated within a few
days of its entry into and therefore fairly near the house in question.
¦ Subsurface VOCs are usually present as a sorbed or free-phase source that is essentially infinite
over the time span of interest. Increasing airflow rates up to a certain point will increase mass
removal by increasing mass transfer (Le Chatelier's principle). However if the system is diffusion
limited or the soils dry out, further increases in air flow rate may not further increase
volatilization (Rorech, 2001; Thomas 1990).
¦ VOC concentration profiles at sites distant from the point of release, when transport is primarily
through groundwater are typically characterized by decreasing concentration with depth. Because
PCE and TCE are only slowly degraded in aerobic soils (with half-lives over 1 year versus 3.8
days for radon), the PCE entering the structure may have migrated over many weeks or even
years from the source area, and from much greater distances than radon. Thus, enhanced airflow
caused by SSD may draw in higher concentration soil gas from deep soils, but is less likely to
increase the radon concentration.
5.4 Stack Gas Monitoring
5.4.1 Is Stack Gas an Indicator of SSD System Performance in Protecting Indoor Air?
The stack gas week-long integrated chloroform concentration as measured by the Waterloo passive
sampler is highest during periods of active SSD operation as expected (Figure 5-14). The stack gas
concentration has some variability during periods with the mitigation on (8.9 to 33 (.ig/ni'). For
chloroform, the concentration in the stack gas is not a strong predictor of the indoor air concentration at
the 422 first floor location (Figure 5-15) during SSD operation (R2 = 0.26). This could reflect infiltration
of outside air as indoor air levels are within the upper end of the range of outdoor air chloroform levels.
The stack gas week-long integrated PCE concentration as measured by the Waterloo passive sampler is
highest during periods of active SSD operation as expected (Figure 5-16). The stack gas concentration
has high variability during periods with the mitigation on (13 to 98 |_ig/m3). In contrast to the result for
chloroform, there is a good correlation between stack gas and the 422 first floor indoor concentration
(Figure 5-17) during SSD operation (R2 = 0.70).
This would suggest that stack gas monitoring in conjunction with verification of SSD operational status
could provide some information about whether the SSD system is increasing or decreasing concentrations
in the subslab area. Although SSD systems need not decrease subslab concentrations to effectively protect
indoor air, systems that exhibit increased concentration in the subslab area during SSD system operation
5-47
-------
Section 5—Subslab Depressurization Mitigation System Monitoring Results
may require more careful monitoring to ensure that control of flow across the slab is maintained at the
vast majority of times and locations.
35
30
— 25
E
bO
20
0
Stack Gas Chloroform
1- 0 °
§
•jm 15 o
re o
I— Q
O 0
E 0 0
S10
c
5 s
« Sue! Pomm CHCI3
On
—— MitiBjibon P*»vw
a
o° 0 a o o o o
n N m rt m
in i-l i-l H *4
> A k *1
" 8 5 ™ a
° ? 7 ^ ^ s *
g S Date ri jg
¦ INI O ™
Figure 5-14, Stack gas monitoring during mitigation testing: chloroform,
422 First vs. Stack Gas Chloroform: SSD On
y * 0.003891 ~ 0 0642
H" ¦
m Flnt thiorcdonr.
LifiMf (*2i ri»n CMereleitmJ
Figure 5-15, 422 first floor versus stack gas chloroform concentrations: mitigation on.
5-48
-------
Section 5—Subslab Depressurization Mitigation System Monitoring Results
Stack Gas PCE
.E
6JJ
100
90
SO
70
£0
50
40
£
o>
(J
C 30
o
10
0
ts
o
e
Q
V
o
o
0
0
0
o
o
o
O
o ®
•-
o
1 * • M-i
o o o
o SI*: i Itrwlwo 'ihde PL!
Mi?iOn Pjsh w
a
s
z
rs
f*j
a a
I j
^ Date g
CL
<
-------
Section 5—Subslab Depressurization Mitigation System Monitoring Results
5.4.2 Air Exchange Rate Measurements
Air exchange rate measurements were performed using EPA Method IP-4A, which uses passive emitters
and passive samplers known as capillary adsorption tube samplers (CATS) from
¦ April 27 to May 11, 2011 (heating system on)
¦ September 23 to September 29, 2011 (heating system off).10
¦ October 13 to October 14, 2011 (during fan test I11, see description of fan tests in Section 12.2 of
U.S. EPA [2012a], heating system off).
¦ October 18 to October 19, 2011 (after fan testing, heating system off)
¦ April 2 to April 9,2013 (with mitigation system operating, heating system on)
During these periods, the house was operated as usual—with windows closed and the doors opened only
as needed to maintain the house and the sampling equipment.
The emitters were evenly spaced across their respective floors of the 422 side of the duplex:
¦ 10 PDCH emitters in the basement
¦ 10 PMCH emitters on the first floor
¦ 9 PMCH emitters on the second floor (10 in some tests)
No emitters were placed on the 420 side of the duplex, but CATS measurements were made there in the
April/May 2011 round and April 2013 to estimate the amount of airflow between sides of the duplex. The
emitters were deployed on April 22, 2011, to allow the building to come to equilibrium before sampling
and were essentially left in place throughout the measurement periods with one change out to fresh
emitters.
As shown in Table 5-19, the April/May 422 basement air exchange rates showed excellent agreement for
the duplicates (both 0.74/hour). As shown in Table 5-20, the September measurements for the basement
(0.64/hour and 0.72/hour) are slightly more variable but quite similar to the April/May measurements.
The first floor measurements were lower in both measurement periods (0.56 in April/May and 0.55 in
September). The September measurements show a pattern of decreasing air exchange rates up through the
building (basement through second floor office).
Measurements performed in April/May 2012 did not show any detectable crossover of either tracer into
the 420 side of the duplex. The detection limit of the method is approximately 1 pi per sample and the
lowest amount of tracer collected in one of the rooms with the emitters for that tracer present was 126 pi.
So less than 1% of the tracer concentration detected in the 422 zones where it was released was present on
the 420 side of the duplex.
The concentration of the tracer released in the basement (PDCH) was about 20% of the basement
concentration on the first floor. The concentration of the tracer released on the first and second floors
(PMCH) was detected at about 2% of the first floor concentration in the basement. These percentages
suggest that during that measurement period more flow was up from the basement to the first floor,
although some flow did come from the first floor down into the basement.
10Fan testing had ended on September 14 and resumed on October 6. These first two tests were reported previously in section 10 of U.S. EPA
(2012a). All five tests are reported here. There have been minor corrections to the calculations from the September 2011 data set.
nFan tests were attempts to induce worst case vapor intrusion by using box fans, in this case to exhaust air from the basement up the stairway to
the first floor.
5-50
-------
Section 5—Sub slab Depressnrization Mitigation System Monitoring Results
Table 5-19. April/May 2011 Air Exchange Rate Measurement Results
PMCH
Amount
(Pi)
PDCH
Amount
(Pi)
Primary
Tracer
Deployed
Interior
Temperature
(°F)
Duration
of Test
Minutes
Date
Deployed
Date
Collected
CAT ID
Location
Calculated
AER 1/hr
Volume
Ft3
4/27/2011
5/4/2011
11015
30.74
127.51
422 basement
PDCH
61.29
0.74
4,547
10,368
4/27/2011
5/4/2011
8441
28.96
126.67
422 basement
dup
PDCH
61.29
0.74
4,547
10,367
4/27/2011
5/4/2011
779
301.47
25.03
422 first
PMCH
67.82
0.56
9,002
10,364
4/27/2011
5/4/2011
9167
0
0
420 basement
None
58.17
NA
4,547
10,354
4/27/2011
5/4/2011
5273
0
0
420 first
None
61.19
NA
9,002
10,352
4/27/2011
5/4/2011
6963
0.75
0
Travel blank
None
68
NA
0
0
Table 5-20. September 2011 Air Exchange Rate Measurement Results
PMCH
Amount
(Pi)
PDCH
Amount
(Pi)
Primary
Tracer
Deployed
Interior
Temperature
(°F)
Duration
of Test
Minutes
Date
Deployed
Date
Collected
CAT ID
Location
AER 1/hr
Volume
Ft3
9/23/2011
9/29/2011
12621
406.42
28.96
422 office
PMCH
72.416
0.34
9,002
8,594
9/23/2011
9/29/2011
18744
253.51
38.35
422 first
PMCH
72.416
0.55
9,002
8,594
9/23/2011
9/29/2011
18185
5.94
108.79
422 basement
PDCH
67.77
0.72
4,547
8,591
9/23/2011
9/29/2011
9024
4.48
121.27
422 basement
dup
PDCH
67.77
0.64
4,547
8,591
During fan test T the air exchange rates were about a factor of 2 to 4 higher (Table 5-21) in a 1-day
measurement. The increase was actually less than we would have calculated if the observed fan flow
1,224 cfm (Section 12.2.2 of U.S. EPA [2012a]) was completely removing the air from the basement in a
"one pass" sense. It is likely that the air being removed from the basement during the fan tests is being
recirculated within the house. Evidence for this is provided by the nearly equal concentrations of the
PMCH and PDCH tracers found in both the basement and the first floor. Recall that PDCH was released
only in the basement and PMCH on the upper floors. This suggests that there is much more flow both
ways between floors with the fan on.
The 1-day test done after the fan testing was completed (Table 5-21) showed somewhat higher air
exchange rates than seen in the previous longer term tests done without the fan. During that day, exterior
temperatures were in the mid-40s °F. It is possible that a stronger stack effect thus explains the higher air
exchange rate than the April 2011 test (40s to 70s °F) or September 2011 test (mid-5 0s to mid-60s °F).
The April 2013 tests (Table 5-22) were conducted with the SSD system on during a period of very wide-
ranging exterior temperatures (20s to 70s °F). If the SSD system was drawing all the exhausted air from
the basement, then we would expect the AER of the basement to increase by approximately 1 air
exchange per hour. Because the air exchange rates measured with the SSD system on were increased by
somewhat less than 1 air exchange per hour, it is likely (as theory would suggest) that the mitigation
system is drawing air from both the structure and elsewhere in the soil column/atmosphere into the
subslab region and hence out the exhaust pipe.
5-51
-------
Section 5—Sub slab Depressnrization Mitigation System Monitoring Results
Table 5-21. October 2011 Air Exchange Measurement Results (during and after fan testing)
Interior
Temper
ature
(°F)
PMCH
Amount
(Pi)
PDCH
Amount
(Pi)
Primary
Tracer
deployed
Duration
of Test
minutes
Date
Deployed
Date
Collected
CAT ID
Location
AER
1/hr
Volume
ft3
Condi-
tions
10/13/2011
10/14/2011
11067
16.43
4.80
422 Office
PMCH
75.0
1.42
9,002
1,440
Fan test
on
10/13/2011
10/14/2011
18277
6.21
5.08
422 First
PMCH
72.7
3.76
9,002
1,439
Fan test
on
10/13/2011
10/14/2011
2502
4.64
4.49
422 Basement
PDCH
69.0
2.91
4,547
1,440
Fan test
Ion
10/13/2011
10/14/2011
17707
4.51
4.45
422 Basement
Dup
PDCH
69.0
2.94
4,547
1,440
Fan test
on
10002
1.98
0.00
Trip blank
PMCH
71.4
9,002
10/18/2011
10/19/2011
7229
38.92
1.76
422 Office
PMCH
63.8
0.60
9,002.00
1,441.00
Fan off
10/18/2011
10/19/2011
18654
23.30
1.66
422 First
PMCH
62.4
1.06
9,002.00
1,441.00
Fan off
10/18/2011
10/19/2011
15758
2.59
10.27
422 Basement
PDCH
62.2
1.27
4,547.00
1,441.00
Fan off
10/18/2011
10/19/2011
9271
2.70
10.76
422 Basement
Dup
PDCH
62.2
1.21
4,547.00
1,440.00
Fan off
6739
1.90
0.00
Trip blank
62.7
Fan off
Table 5-22. April 2013 Air Exchange Measurement Results (During Mitigation)
PMCH
Amount
(Pi)
PDCH
Amount
(Pi)
Primary
T racer
deployed
Interior
Temperature
(°F)
Duration
of Test
minutes
Date
Deployed
Date
Collected
Volume
ft3
CAT ID
Location
AER 1/hr
4/2/2013
4/9/2013
7247
166.00
21.00
422 Second
Floor Office
PMCH
74.4
1.03
9,002
10,040
4/2/2013
4/9/2013
17946
126.00
16.00
422 First Floor
Center
PMCH
69.7
1.36
9,002
10,033
4/2/2013
4/9/2013
14883
12.00
65.00
422 Basement
Center
PDCH
60.2
1.40
4,547
10,030
4/2/2013
4/9/2013
9304
10.00
60.00
422 Basement
Center Dup
PDCH
60.2
1.52
4,547
10,030
4/2/2013
4/9/2013
15680
0
7
420 1st
PMCH
57.2
9,002
10,022
The measurements of air exchange rate (not during fan tests) are almost all between the 50th and 90th
percentile of the range of Midwestern values compiled in EPA's Exposure Factor Handbook (U.S. EPA,
2011; Table 5-23).
Table 5-23. National Survey of Air Exchange Rates, Reprinted from the EPA Exposure Factor
Handbook (U.S. EPA, 2011)
Summary Statistics for Residential Air Exchange Rates (in AER, 1/hr3), by Region
West
Region
Midwest
Region
Northeast
Region
South
Region
All Regions
Arithmetic mean
0.66
0.57
0.71
0.61
0.63
Arithmetic standard deviation
0.87
0.63
0.60
0.51
0.65
Geometric mean
0.47
0.39
0.54
0.46
0.46
Geometric standard deviation
2.11
2.36
2.14
2.28
2.25
(continued)
5-52
-------
Section 5—Sub slab Depressnrization Mitigation System Monitoring Results
Table 5-23. National Survey of Air Exchange Rates, Reprinted from the EPA Exposure Factor
Handbook (U.S. EPA, 2011) (cont.)
Summary Statistics for Residential Air Exchange Rates (in AER, 1/hra), by Region
West
Region
Midwest
Region
Northeast
Region
South
Region
All Regions
10th percentile
0.20
0.16
0.23
0.16
0.18
50th percentile
0.43
0.35
0.49
0.49
0.45
90th percentile
1.25
1.49
1.33
1.21
1.26
Maximum
23.32
4.52
5.49
3.44
23.32
aAER = ACH = Air exchanges per hour.
Source: Koontz and Rector, 1995, as cited in U.S. EPA (2011), Table 19-24.
5.4.3 Stack Gas Measurements to Define Flux to Structure
From stack velocity measurements (multiple instantaneous) and Waterloo sampler VOC concentrations in
the stack (typically integrated over one week) we calculated estimates of the mass of PCE and chloroform
emitted by the SSD system over time (Table 5-24). Although the values expressed as micrograms may
seem large at first the average emission of PCE can also be expressed as 0.11 grams/day or 0.000010
pounds per hour. Some perspective on these values can be provided by noting that:
¦ Controls are often required on VOC sources with emission rates exceeding 3 pounds per hour and
3.1 tons per year (U.S. EPA, 2008).
¦ Heggie and Stavropoulos (2010) measured a TCE flux rate between 41 and 312 (.ig/nr/h for an
Australian vapor intrusion site. Our average PCE flux converts to 40 (.ig/nr/h.
¦ These discharges are, however, considerably higher than those reported for Altus AFB and Hill
AFB buildings, which expressed as deep soil gas to subslab discharge12 ranged from 0.16 to 467
(ig/day (GSI, 2008).
¦ Without mitigation, we can estimate the discharge of VOCs into the Indianapolis structure as
follows: Assume the 422 basement, with an air exchange rate of 0.74 per hour (Table 5-19) and a
volume of 129 m3. Apply the mean 422 Base South sampling location concentration of 0.65
Hg/m3 (Table 5-16 of this report); this yields a VOC mass discharge of 1,500 (ig/day (0.0015
grams/day) for the half of the duplex that has generally higher concentrations.
This brief analysis suggests that the mass flux and discharge rates measured are reasonable and are likely
increased under mitigation on conditions. This agrees with the finding discussed in Section 6.1.2 that SSD
system operation increases the concentration in subslab soil gas in many of our measurements.
12Reported as "mass flux" in GSI (2008)
5-53
-------
Section 5—Sub slab Depressnrization Mitigation System Monitoring Results
Table 5-24. Stack Gas Discharge Measurements During Mitigation
Stack Discharge Measurement Summary—SSD Mitigation On
VOC Sample Date
Collected
Flow Data Date
Acquired
Stack Discharge of
PCE (|jg/day)
Stack Discharge of
Chloroform (|jg/day)
12/26/2012
12/28/2012
87,000
110,000
12/26/2012
12/28/2012
86,000
110,000
2/20/2013
2/21/2013
160,000
40,000
4/3/2013
4/3/2013
120,000
46,000
4/3/2013
4/3/2013
120,000
46,000
Average
114,000
70,000
Note: PCE and chloroform stack discharges are rounded to 2 significant figures
5-54
-------
Section 6—Results and Discussion:
VOC Concentration Temporal Trends and Relationship to HVAC and Mitigation
Table of Contents
6.0 Results and Discussion: VOC Concentration Temporal Trends and Relationship to HVAC
and Mitigation 6-1
6.1 VOC Seasonal Trends Based on Weekly, Biweekly, and Monthly Measurements for
52+Weeks 6-1
6.1.1 Indoor Air 6-1
6.1.2 Subslab Soil Gas 6-4
6.1.3 Shallow and Deep Soil Gas 6-15
6.2 Radon Seasonal Trends (based on Weekly Measurements) 6-46
6.3 VOC Short-Term Variability (Based on Daily and Hourly VOC Sampling) 6-46
6.3.1 Indoor Air 6-46
6.3.2 Subsurface Soil Gas Data 6-52
6.4 Radon Short-Term Variability (Based on Daily and More Frequent Measurements) 6-59
6.5 Outdoor Climate/Weather Data 6-60
6.5.1 Indianapolis Weather Compared with VOCs and Radon 6-66
List of Figures
6-1. PCE concentrations in indoor and ambient air vs. time (7-day Radiello samples) 6-2
6-2. Chloroform concentrations in indoor and ambient air vs. time (7-day Radiello samples) 6-2
6-3. Benzene concentrations in indoor air 6-3
6-4. Toluene concentrations in indoor air 6-4
6-5. Interior and exterior sampling port locations. Sampling ports sampled by the on-site GC
are shown in red, with parenthetical notes indicating which SGP depths were sampled by
the GC 6-5
6-6a. Plot of subslab chloroform concentrations vs. time (TO-17 data) 6-6
6-6b. Plot of subslab chloroform concentrations vs. time, first intensive sampling period (TO-
17 data) 6-6
6-6c. Plot of subslab chloroform concentrations vs. time mitigation testing period (TO-17
data) 6-7
6-7a. Plot of subslab PCE concentrations vs. time. (TO-17 data) 6-7
6-7b. Plot of subslab PCE concentrations vs. time, first intensive sampling period. (TO-17
data) 6-8
6-7c. Plot of subslab PCE concentrations vs. time, mitigation testing period (TO-17 data) 6-8
6-7d. Plot of subslab PCE concentrations vs. time, mitigation testing period; real time GC 6-9
6-8a. Plot of wall port chloroform concentrations vs. time (method TO-17) 6-11
6-8b. Plot of WP-3 chloroform concentrations vs. time (online GC) 6-12
6-9a. Plot of wall port PCE concentrations vs. time (method TO-17) 6-13
6-9b. Plot of WP-3; PCE concentrations vs. time (online GC) 6-14
6-i
-------
Section 6—Results and Discussion:
VOC Concentration Temporal Trends and Relationship to HVAC and Mitigation
6-10. Chloroform concentrations at subslab and 6-ft soil gas ports directly under the 420 side of
duplex 6-18
6-11. PCE concentrations at 6-ft soil gas ports and subslab immediately below the 420 side of
the duplex 6-19
6-12. Chloroform concentrations at 9-ft soil gas ports below 420 side of the duplex 6-20
6-13. PCE concentrations at soil gas points 9 ft below the 420 side of duplex 6-21
6-14. Chloroform concentrations in soil gas at 13 ft below the 420 side of the duplex 6-22
6-15. PCE concentrations in soil gas at 13 ft below the 420 side of duplex 6-23
6-16. Chloroform concentrations in soil gas at 16.5 ft below the 420 side of duplex 6-24
6-17. PCE concentrations in soil gas at 16.5 ft below the 420 side of the duplex 6-25
6-18. Chloroform concentrations in 6-ft soil gas and subslab ports immediately below the 422
side of the duplex 6-26
6-19. PCE concentrations in 6-ft soil gas ports and subslab ports directly below the 422 side of
the duplex 6-27
6-20. Chloroform concentrations in soil gas port at 9-ft depth below 422 side of duplex 6-28
6-21. PCE concentrations in soil gas at 9 ft below the 422 side of the duplex 6-29
6-22. Chloroform concentrations in soil gas at 13 ft below the 422 side of the duplex 6-30
6-23. PCE concentrations in soil gas at 13 ft below the 422 side of the duplex 6-31
6-24. Chloroform concentrations at 16.5 ft below the 422 side of the duplex 6-32
6-25. PCE concentrations at 16.5 ft below the 422 side of the duplex 6-33
6-26. Chloroform concentrations in exterior soil gas at 3.5 ft bis 6-34
6-27. PCE concentrations in exterior soil gas at 3.5 ft bis 6-35
6-28. Chloroform concentrations in exterior soil gas at 6 ft. bis 6-36
6-29. PCE concentrations in exterior soil gas at 6 ft bis 6-37
6-30. Chloroform concentrations in exterior soil gas at 9 ft bis 6-38
6-31. PCE concentrations in exterior soil gas at 9 ft bis 6-39
6-32. Chloroform concentrations in exterior soil gas at 13 ft bis 6-40
6-33. PCE concentrations in exterior soil gas at 13 ft bis 6-41
6-34. Chloroform concentrations in exterior soil gas at 16.5 ft bis 6-42
6-35. PCE concentrations in exterior soil gas at 16.5 ft bis 6-43
6-36. Subslab PCE concentrations over a 1-week period during the first intensive round 6-44
6-37. Subslab PCE concentrations over a 1-week period during the second intensive round 6-44
6-38. Radon: Weekly time integrated samples (electret) 6-45
6-39. Online GC chloroform indoor air data for 422 first floor 6-46
6-40. Online GC chloroform indoor air data for 422 basement 6-47
6-41. Online GC chloroform indoor air data for 420 first floor 6-47
6-ii
-------
Section 6—Results and Discussion:
VOC Concentration Temporal Trends and Relationship to HVAC and Mitigation
6-42. Online GC chloroform indoor air data for 420 basement 6-48
6-43. Online GC PCE indoor air data for 422 first floor 6-49
6-44. Online GC PCE indoor air data for 422 basement 6-49
6-45. Online GC PCE indoor air data for 420 first floor 6-50
6-46. Online GC PCE indoor air data for 420 basement 6-50
6-47. Online GC subsurface chloroform soil gas data—Phase 1 and Phase 2 6-52
6-48. Online GC subsurface chloroform soil gas data—Phase 1 6-52
6-49. Online GC subsurface chloroform soil gas data—Phase 2 6-53
6-50. Online GC subsurface PCE soil gas data—Phase 1 and Phase 2 6-54
6-51. Online GC subsurface PCE soil gas data—Phase 1 6-54
6-52. Online GC subsurface PCE soil gas data—Phase 2 6-55
6-53. Method 10-17 data for SSP-4 6-55
6-54. Online GC PCE measurements in SSP-4 6-57
6-55. Comparison of online GC measurements of PCE and chloroform in SGP9 at 6 ft 6-57
6-56. Real-time radon levels (422 basement) 2011-2013 6-58
6-57. Real-time radon levels (422, 2nd floor office), 2011-2013 6-59
6-58. Temperature records from the external temperature monitor and the HOBO devices at
seven indoor locations on the 422 and 420 sides of the house 6-61
6-59. Stacked hydrological graph with rainfall in inches (top—green line), depth to water in
feet (middle—red circles), and discharge at Fall Creek in ft3/s (bottom—blue line) 6-62
6-60. Plot of high wind speed for measurement period, wind run and wind speed (average over
measurement period) at 422/420 house over time 6-63
6-61. Weather variables measured inside 422 office (2nd floor) and on roof: a. barometric
pressure (in Hg); b. indoor air density, c. indoor air equilibrium moisture content, d,
indoor percent humidity, f. outdoor percent humidity, g. rain (inches total in measurement
period), h. rain rate—the most intense rainfall during the measurement period in
inches/hour 6-64
6-62. Snow depth vs. time (data are from NCDC records for the Indianapolis International
Airport) 6-65
6-63. GC Phase 2 VOCs at WP-3 compared with 422/420 house external pressure 6-66
6-64. GC Phase 3 VOCs at WP-3 compared with 422/420 house external pressure 6-67
6-65. GC Phase 2 VOCs at WP-3 compared with 422/420 house external wind speed 6-68
6-66. GC Phase 3 VOCs at WP-3 compared with 422/420 house external wind speed 6-69
6-67. GC chloroform at subslab and soil gas ports versus radon from stationary
AlphaGUARDs 6-72
6-68. GC PCE at subslab ports versus radon from stationary AlphaGUARDs 6-73
6-69. GC chloroform in indoor air versus radon from stationary AlphaGUARDs 6-74
6-iii
-------
Section 6—Results and Discussion:
VOC Concentration Temporal Trends and Relationship to HVAC and Mitigation
6-70. GC PCE in indoor air versus radon from stationary AlphaGUARDs 6-75
6-71. GC chloroform concentrations in indoor air, 420 first floor 6-76
6-72. GC chloroform concentrations in indoor air, 420 basement south 6-76
6-73. GC PCE concentrations in indoor air, 420 first floor 6-77
6-74. GC PCE concentrations in indoor air, 420 basement south 6-77
List of Tables
6-1. Frequency of NondetecSamples (%) by Soil Gas Point or Cluster 6-16
6-2. Frequency of Nondetects in TO-17 VOC Data by Soil Gas Sampling Depth 6-16
6-3. Summary Meteorological Data for Central Indiana Note that the symbols "A" and "v"
mean "above" and "below" normal, respectively, and that the weekly values show how
the weekly averages differ from normal (from Scheeringa and Hudson, 2012, 2013) 6-60
6-4. Summary of Meteorological Data During the 2012/2013 Snow and Ice Events 6-78
6-iv
-------
Section 6—Results and Discussion:
VOC Concentration Temporal Trends and Relationship to HVAC and Mitigation
6.0 Results and Discussion: VOC Concentration Temporal Trends
and Relationship to HVAC and Mitigation
6.1 VOC Seasonal Trends Based on Weekly, Biweekly, and Monthly
Measurements for 52+ Weeks
6.1.1 IndoorAir
Figures 6-1 and 6-2 show PCE and chloroform concentrations overtime, respectively, at all six indoor air
monitoring locations, in addition to the ambient location (see Figure 3-10a and b for the placement of
the indoor air sampling racks). PCE concentrations at all six indoor locations follow the same general
trend of starting higher at the beginning of the project, dropping to a low in spring, and rising slightly and
leveling out through the end of the premitigation period. Indoor air sampling was discontinued from
February 2012 to October 2012 because of funding limitations. However, the concentrations in October
2012 before the mitigation system was installed were very similar to those observed in October 2011. The
timing of the spring minimum differed substantially for the unheated side of the duplex (when it occurred
in late March) from the heated side of the duplex (where the minimum was reached in July). The highest
readings were generally found at 422 basement south except during brief periods when first floor
concentrations were higher, which occurred mostly during operation of a basement depressurization fan in
422 (see U.S. EPA, 2012a, Section 12.2).
Highlights of the PCE concentration patterns shown in Figure 6-1 are:
1. Indoor air PCE concentrations during the first period of active mitigation rose to levels not seen
in the duplex since February 2011. The concentrations continued to rise after the mitigation
system was switched into a passive mode, reaching a maximum of 5.7 (.ig/nr1 in November 2012.
Indoor air concentrations higher than 5.7 (ig/m3 had not been observed at the duplex since
January and February 2011.
2. In discussions and comments during conference presentations (Schumacher et al., 2013; Lutes et
al., 2012a, 2012b, 2012c, 2012d) on the earlier PCE data set (through February 2012), questions
were raised about whether the highest PCE concentrations observed in January and February
2011 were artifacts. At the time, the authors offered other lines of evidence (such as the lack of
indoor sources, preparation of the house prior to sampling) as support for the observed levels. The
observation of higher PCE post-mitigation during the winter of 2012 to 2013 does confirm that
the subsurface can yield enough vapor intrusion-derived PCE to account for the January 2011
concentrations. We postulate that VOCs can be moved close to the structure either by a
cumulative stack effect during a severe winter or by an SSD system, at least during its initial
period of operation. How VOC levels will change over time as the mitigation system continues to
operate remains to be seen.
Chloroform concentration patterns (see Figure 6-2) were generally similar to PCE and can be
summarized as follows:
1. Broadly, the six indoor locations show a general concentration decline from a localized maximum
at the beginning of the sampling interval in January 2011. The minimum was reached at the end
of spring on the 422 side of the house (early July), as with PCE. Also similar to PCE's behavior,
the chloroform minimum concentration on the 420 side of the house occurred much earlier in the
year (March).
6-1
-------
Section 6—Results and Discussion:
VOC Concentration Temporal Trends and Relationship to HVAC and Mitigation
Indoor Air PCE
.£
M
£
O
£
II
u
£
O
U
100
10
0.1
0.01
ft
100
10
0.1
0.01
1 Ambient PCE
420 First PCE
420 Base N PCE
420 Base S PCE
422 First PCE
422 Base N PCE
422 Base S PCE
¦Mitigation On
Mitigation Passive
¦Snow Event
¦Fan Testing
¦Snow Max 4"
o
*—1
*—1
*—1
*—1
(N
(N
(N
m
m
M
*—1
*—1
*—1
*—1
*—1
*—1
*—1
*—1
*—i
*—i
*—1
00
10
10
m
*—1
O
00
10
10
*—1
(N
O
*—1
(N
O
*—1
*—1
(N
O
*—1
^
^
^
^
^
_
^
_
*—1
(N
10
0"l
(N
o
*—1
in
00
*—1
O
o
o
*—1
o
o
*—1
O
o
o
Date Collected
Figure 6-1. PCE concentrations in indoor and ambient air vs. time (7-day Radiello samples).
Indoor Air Chloroform
.£
M
£
O
£
II
u
£
O
U
100
10
0.1
0.01
100
10
0.1
0.01
i Ambient CHCI3
420 First CHCI3
420 Base N CHCI3
420 Base S CHCI3
422 First CHCI3
422 Base N CHCI3
422 Base S CHCI3
"Mitigation On
"Mitigation Passive
"Snow Event
"Fan Testing
"Snow Max 4"
"Snow Max 6"
o
*—1
*—1
*—1
*—i
(N
(N
(N
m
m
*—1
*—1
*—1
*—1
*—i
*—1
*—1
*—1
*—i
*—i
*—i
_
_
_
_
^
^
_
_
_
_
_
00
10
V0
m
*—1
O
00
10
V0
*—1
(N
o
*—1
(N
O
*—1
*—1
(N
o
*—i
^
^
^
^
^
_
^
_
*—1
(N
V0
tn
(N
o
*—1
in
00
*—1
O
O
o
*—1
o
o
*—1
O
o
o
Date Collected
Figure 6-2. Chloroform concentrations in indoor and ambient air vs. time (7-day Radiello samples).
6-2
-------
Section 6—Results and Discussion:
VOC Concentration Temporal Trends and Relationship to HVAC and Mitigation
2. The levels at the 422 first floor sampling location rose abruptly to a maximum in March 2011
immediately after the first brief drop in January. During this maximum, the first floor
concentrations exceeded those of even the basement stations. The 422 basement sampling stations
showed a less dramatic rise in this period.
3. Chloroform concentrations reached a minimum in July 2011 and began steadily increasing
thereafter, forming a generally U-shaped curve. The winter 2012 levels more closely approached
their 2011 highs than do the corresponding PCE results.
4. The second maximum concentration for chloroform occurred in October 2011 for the 420
(unheated) locations and was followed by a considerable decline through the winter months. A
second peak occurred later (December 2012) on the 422 (heated) side of the duplex and
concentrations stayed near that maximum until February 2012.
5. The concentrations of chloroform in October 2012 when sampling was restarted after a break
since February 2012 were similar to those observed in October 2011.
With the exception of the elevated chloroform from late February to late March 2011, the highest
chloroform levels were found at 422 basement south, the same station that was generally highest for PCE
(Figures 6-1 and 6-2).
Figures 6-3 and 6-4 show benzene and toluene indoor air concentrations at 422 basement south over
time, along with ambient concentrations of benzene and toluene. Although both benzene and toluene
concentrations are above their action levels (benzene = 0.31(ig/m3; toluene = 0.0052(ig/m3; RSL
Summary Table, Nov., 2011), each tends to trend similarly to its respective ambient concentrations; this is
not the case with PCE or chloroform concentrations, which are almost always considerably higher than in
the ambient air. This suggests that benzene and toluene indoor air concentrations are controlled by
ambient air, not vapor intrusion.
Indoor Air Benzene
100
.£
M
£
O
£
II
u
c
o
u
10
0.1
•
• •
Ba
n
¦
r
i
¦ A
~nil
iii rttjs
J\ m
M
ijr jlZy
ft sw
So
~o
100
10
Ambient Benzene
420 First Benzene
420 Base N Benzene
420 Base S Benzene
422 First Benzene
422 Base N Benzene
422 Base S Benzene
¦Mitigation On
Mitigation Passive
¦Snow Event
¦Fan Testing
¦Snow Max 4"
0.1
o
*—i
00
10
(N
(N
O
ID
O
ID
O
m
(N
0"l
o
^r
o
(N
*—I
O*
o
(N
*—I
00
*—I
o
(D
(N
ID
O
1/1
O
m
*—i
00
o
Date Collected
Figure 6-3. Benzene concentrations in indoor air.
6-3
-------
Section 6—Results and Discussion:
VOC Concentration Temporal Trends and Relationship to HVAC and Mitigation
Indoor Air Toluene
100
.£
M
£
O
£
II
u
£
O
U
10
0.1
o
*—1
*—1
*—1
*—1
(N
(N
(N
m
m
m
*—1
*—1
*—1
*—1
*—1
*—1
*—1
*—1
*—i
*—i
r—i
00
10
10
m
*—1
O
00
10
10
*—1
(N
O
*—1
(N
O
*—1
*—1
(N
O
^
^
^
^
^
^
^
^
*—1
(N
10
0"l
(N
o
*—1
i/l
00
*—1
O
o
o
*—1
o
o
*—1
O
o
o
Date Collected
O 420 First Toluene
¦ 420 Base N Toluene
~ 420 Base S Toluene
• 422 First Toluene
¦ 422 Base N Toluene
~ 422 Base S Toluene
Ambient Toluene
^^•Mitigation On
^^"Mitigation Passive
^^»Snow Event
^^"Fan Testing
^^"Snow Max 4"
^^»Snow Max 6"
^^"Fan Testing
^^"Fan Testing
^^"Snow Max 2"
^^"Snow Max 1"
^^»Snow Max 7"
^^"Snow Max 1/4"
^^"Mitigation On
^^^Mitigation Passive
^^"Mitigation On
^^"Brief Snow
^^•Brief Snow
^^Blce
^^•brief snow
^^•Brief snow
Figure 6-4. Toluene concentrations in indoor air.
6.1.2 Subslab Soil Gas
Subslab sampling ports (SSPs) were placed throughout the basement of both the 420 and 422 sides of the
duplex, as shown in Figure 6-5. Interior soil gas probes (SGPs) are also shown in this figure, with each
probe having multiple sampling ports at 6 ft, 9 ft, 13 ft, and 16.5 ft, with the 6 ft SGPs being at an
equivalent depth bis to the SSPs. Given the low initial concentration of SSP-2 and its nearness to SGP10-
6, SSP-2 was sampled relatively infrequently. On the 420 side of the house are SSP-3 and SSP-5 through
-7 and WP-4. The basements of both sides of the duplex are each divided into thirds in the interior. There
is generally one SSP per basement division, with one section on the 420 side having two. The wall ports
are located on the exterior walls of the duplex. WP-1 and 3 are each located in the centers of the north and
south ends of the 422 basement, and WP-2 is in the center of the east side of the 422 basement. WP-4 is
located in the center of the west wall of the 420 basement (Figure 6-5). The wall ports are approximately
3 ft bis.
Figures 6-6a, 6-6b, and 6-6c (for chloroform) and 6-7a, 6-7b, and 6-7c (PCE) plot VOC concentrations
over time. Figures 6-6a and 6-7a present an overview of subslab TO-17 data, the b versions of these
figures represent intensive sampling periods, and the c sections focus on the mitigation testing period.13
For chloroform, as shown in Figure 6-6a, most of the ports on the unheated 420 side (the various crosses
and the square) were generally stable for most of the duration of the project prior to mitigation. However,
some subslab ports, such as SSP-4 and SSP-7, reached new high concentrations after mitigation began.
Yet within the mitigation testing period there is no clear visible VOC concentration trend as the system is
switched on and off (Figure 6-6c), perhaps because of the limited amount of available data (see also the
discussion of the descriptive statistics for these concentrations in Section 5.3).
13During the normal times, the subslab samples were collected during regular daytime working hours, while the intensive periods involved two
shifts of personnel, allowing up to three samples to be collected, generally early morning, midday, and evening.
6-4
-------
Section 6—Results and Discussion:
VOC Concentration Temporal Trends and Relationship to HVAC and Mitigation
Wm
Storage
Closet
Back
Porch
Back
Porch
SSP 2
Wall Port 1
SSP 6
Open
Walkway
sVv SGP10
Walkway
i—i
SGP 11
SSP 1
11-13 GC)
Port 4
Wall
Port 2
@i
J&
% —SGP 8
(8-9 GC)
O — MW 3
SSP 3
Up to
1st Floor
SGP 4
SSP 7
1st Floor
SGP 9
(9-6 GC)
Open
Walkway
Open
Walkway
Open
Walkway
SSP 5
SSP 4
Wall ¦ 422
Port 3 ¦ Covered
Porch
420
Covered
Porch
SGP 12
Figure 6-5. Interior and exterior sampling port locations. Sampling ports sampled by the on-site
GC are shown in red, with parenthetical notes indicating which SGP depths were
sampled by the GC.
6-5
-------
Section 6—Results and Discussion:
VOC Concentration Temporal Trends and Relationship to HVAC and Mitigation
Subslab Port Chloroform
1000
o o
O • #°f
cPcq°°
£0-
%
K9EM
cmrmmiD
SSP-l CHCI3
SSP-2 CHCI3
SSP-3 CHCI3
SSP-4 CHCI3
SSP-5 CHCI3
SSP-6 CHCI3
SSP-7 CHCI3
¦Mitigation On
¦-Mitigation Passive
¦Mitigation On
Date
Figure 6-6a. Plot of subslab chloroform concentrations vs. time (TO-17 data).
Subslab Chloroform Concentrations 1st Intensive Round
1000
A
zr too m
E
3r
A
A
2
A
A,
A
A *
•
•
¦
B
•
•
•
Concentration
—K 1
• SSP-l CNanlam
1
X
X
x
X
¥
X
¥
x a
* SSP ¦ i C h Ujrol or rn
¦ SSP-J £ihkjrtrfornn
* SSP-l CtitofOlwm
S5P-5 atteraform
SSP-6 Clhkwfwlnrm
$SP>7 CliJoroilorm
p
02-Mar-ll
03-Mar-ll
04-Mar-ll
05-Mar-ll
*—i
ub
o
Date
07-Mar-ll
08-Mar-ll
09-Mar-ll
Figure 6-6b. Plot of subslab chloroform concentrations vs. time, first intensive sampling period
(TO-17 data).
6-6
-------
Section 6—Results and Discussion:
VOC Concentration Temporal Trends and Relationship to HVAC and Mitigation
10000
E
—.
u>
1000
•= 100
to
4->
£
u
c
o
u
10
m
o
Subslab Port Chloroform, Mitigation Period
• •
o
CO
o
o
• ° o
o
O c
°
Q
D
0
o
o
•
•
o
•
o °*
o
• ••
o %
8
o 4
m
*
o • •
oo • o
o • •«
• •
>
(N
(N
Date
(N
O
m
o
^r
o
SSP-l CHCI3
SSP-2 CHCI3
SSP-3 CHCI3
SSP-4 CHCI3
SSP-5 CHCI3
SSP-6 CHCI3
SSP-7 CHCI3
¦Mitigation On
Mitigation Passive
Figure 6-6c. Plot of subslab chloroform concentrations vs. time mitigation testing period (TO-17
data).
10000
E
—.
u>
1000
•= 100
ns
4->
£
0>
u
£
O
u
10
Subslab Port PCE
(N
(N
O
*—I
O
00
o
PM *H
(N cn
° Date °
^r
o
•_
CP °
•
•
8 03
L
fl
? °8 8
• ° «o
• o
•
W*ot
8 o
• ?°®
ot
• •q£.
• ca(
g or
p
o o OO
• o
CD •
/ J •
-------
Section 6—Results and Discussion:
VOC Concentration Temporal Trends and Relationship to HVAC and Mitigation
Subslab PCE Concentrations 1st Intensive Round
10000
j:
M
£
O
1000
•j: 100
as
C
•
n ~
•
~
¦
•
i
¦
~
•
~
*
•
~ A
•
+
X
£
*
£
¥
*
X
*
A
X
A
X
X
1 ~
X
*—I *—I
O O
rM in
*—I
O
IN
*—I
o
IN
o
IN
*—I
o
IN
i
o
IN
*—1
o
IN
• SSP-1PCE
~ SSP-2 PCE
¦ SSP-3 PCE
~ SSP-4 PCE
+ SSP-5 PCE
XSSP-6 PCE
XSSP-7 PCE
no
vo
no
oo
no
Date
Figure 6-7b. Plot of subslab PCE concentrations vs. time, first intensive sampling period. (TO-17
data).
10000
E
—.
u>
to
4->
£
0>
u
c
O
u
1000
•= 100
10
m
o
Subslab Port PCE, Mitigation Period
o o
o
o
. 8
•
° O
0 o
o
m
% oc
0 '
q
8
•
3
0
o o«
• •
•o
• a
om •
•CD t
cM
/
• *1
• •
• •
o o •
• •
•
• •
• • •
• CO
• oc
••
•
•{
•
• o
•
•
(N
(N
(N
O
m
o
Date
^r
o
SSP-l PCE
SSP-2 PCE
SSP-3 PCE
SSP-4 PCE
SSP-5 PCE
SSP-6 PCE
SSP-7 PCE
•Mitigation On
Mitigation Passive
•Mitigation On
Figure 6-7c. Plot of subslab PCE concentrations vs. time, mitigation testing period (TO-17 data).
6-8
-------
Section 6—Results and Discussion:
VOC Concentration Temporal Trends and Relationship to HVAC and Mitigation
Hartman3
Tetrachloroethene
SSP-.2
100^
1 0 5
1 =
3
1003
C
_o
"55
1 0 E
c
• Mitigation Off
• Mitigation On
• Passive Mitigation
Flags
Missing
Figure 6-7d. Plot of subslab PCE concentrations vs. time, mitigation testing period; real time GC.
6-9
-------
Section 6—Results and Discussion:
VOC Concentration Temporal Trends and Relationship to HVAC and Mitigation
Another exception to the general pattern of chloroform stability is vertical alignment of data points on the
plot (indicating concentration variability over a short time period) that occurred during intensive periods
of sampling. This may indicate that there was a diurnal pattern in the subslab sampling that was only
perceptible during the intensive periods (Figures 6-6b and 6-7b).14 Another noteworthy observation on
the 420 side occurred from July 14, 2011, to August 3, 2011, between the time when thieves stole the
house window unit air conditioners (ACs) from both sides of the duplex and when they were replaced on
the 422 side only. Chloroform approached its highest levels on the 420 side during this time. Chloroform
on the 422 side (shown in Figure 6-6a as the circles, diamonds, and triangles) showed a rough sinusoidal
concentration trend over months, although the different ports are somewhat out of phase. These trends
generally show lows during the warmer months (SSP-1 and SSP-4 seem to both reach a minimum in
August/September 2011) and highs during cooler months. It is also notable that the concentration
increases abruptly two orders of magnitude between August 27 and September 8, 2011, a period of time
during which a series of fan tests (coded B and F) intended to simulate the stack effect expected under
winter conditions were conducted (as discussed in Section 12.2 of U.S. EPA [2012a]). Another smaller
rise occurs from September 30 to October 14, 2011. Fan test "I" was conducted from October 6 to
October 14, 2011.
The subslab ports on the 422 side (heated) have higher concentrations of PCE and chloroform than those
on the 420 (unheated) side of the structure. In Figure 6-7a, subslab port PCE concentrations versus time
more prominantly display a simple pattern of high and low concentration changes during warmer and
cooler months, respectively, across the range of ports. Most of the ports on the 420 side of the house and
SSP-4 on the 422 side showed highs during the warmer months and lows during the cooler months. A
notable exception is SSP-1, which showed the opposite PCE concentration trend to all the others.
Essentially, there is much more spatial variability among the subslab ports in winter than in summer.
As occurred with chloroform concentrations, some ports—SSP-3, SSP-6, and SSP-7—reached new high
PCE values after mitigation began (Figure 6-7c). As discussed in the descriptive statistical analysis in
Section 5.3, there was a clear trend in PCE concentrations within the mitigation testing period of higher
subslab PCE concentrations being associated with the mitigation on.
The higher temporal resolution data from the online GC (Figure 6-1 A) shows that two subslab ports on
either end of the 422 side of the duplex were relatively stable in PCE during mitigation testing (SSP-2 and
SSP-4). In contrast, a subslab port near the center of the 420 side of the duplex, SSP-7, showed
approximately two orders of magnitude temporal variability during the mitigation testing period. It
appears that turning the mitigation system on drew VOCs toward that port. However, even during a long
period of having the mitigation system on, more than an order of magnitude of temporal variability was
observed at that port.
Neither compound when graphed for the wall ports (Figure 6-8 nor Figure 6-9) shows the same clear
patterns of highs and lows found during the changing seasons for the subslab ports in Figures 6-6a and 6-
7a. Figure 6-8a plots chloroform concentrations at the four wall ports versus time, and Figure 6-8b
shows more detail of WP-3 based on online GC data. These chloroform levels do not show the same kind
of spike during the period when the ACs were stolen as for the chloroform subslab port. Highs for WP-3
in January through February and September through October 2011 seem to suggest influence of the snow
and ice and fan testing, respectively. The greater temporal flucations of the wall ports as compared with
the subslab ports may be attributable to their more shallow depths (approximately 1.5 ft bis) and their
position through the exterior basement wall, which results in a greater atmospheric influence and lesser
building effects. The detailed data in Figure 6-8b show that WP-3 experienced multiday chloroform
concentration peaks and valleys during all mitigation periods.
14 During normal times, the subslab samples were collected during regular daytime working hours, while the intensive periods involved two
shifts of personnel, allowing up to three samples to be collected, generally early morning, midday, and evening.
6-10
-------
Section 6—Results and Discussion:
VOC Concentration Temporal Trends and Relationship to HVAC and Mitigation
Wall Port Chloroform
1000
E
"So 100
£
o
CO
C
0)
u
c
o
u
10
X
X
~
~
1
~
X
~ '
A A
~
A
A
1 "
A
¦ A
X
¦
a
¦
1
» * H
i
WP-1 CHCI3
WP-2 CHCI3
WP-3 CHCI3
WP-4 CHCI3
•Mitigation On
Mitigation Passive
•Mitigation On
o
*—1
*—1
*—1
*—i
(N
(N
(N
m
m
M
*—1
*—1
*—1
*—1
*—i
*—1
*—1
*—1
*—i
*—i
*—1
^
^
^
^
^
^
^
^
^
^
^
00
10
10
m
*—1
O
00
10
V0
*—1
(N
O
*—1
(N
O
*—1
*—1
(N
o
*—1
*—1
(N
10
tn
(N
o
*—1
in
00
*—1
O
O
o
*—i
o
o
*—1
O
o
o
Date
Figure 6-8a. Plot of wall port chloroform concentrations vs. time (method TO-17).
Figure 6-9a plots PCE concentrations at the four wall ports versus time. The wall port concentrations,
although generally modest, show more variability than the subslab ports. The high concentrations of PCE
in WP-3 at the beginning of the project could be due to the snow and ice capping event during the severe
winter of January and February 2011. Highs in September and October 2011 might be attributable to the
fan testing during that time. Relatively high VOC concentrations at WP-3 were also reached after the
mitigation testing began. This suggests that the SSD system may be drawing VOCs closer to the building
envelope.
Higher temporal resolution data using the online GC were obtained for PCE at WP-3 (Figure 6-9b). This
shows nearly two orders of magnitude of variability in wall port PCE concentrations during the mitigation
testing period, including multiday concentration peaks and valleys during all mitigation periods that are
similar to those observed for chloroform in Figure 6-8b.
6-11
-------
Section 6—Results and Discussion:
VOC Concentration Temporal Trends and Relationship to HVAC and Mitigation
Hartman3
Chloroform
100-
10
E
c
o
TO
i_
C
-A
~ * *
t **
•
t.
*• -
•
. > •
; *'
•
7
*
•
i
*
K 41
?'At * *
f#
• f • t
* •
~ ~ |»
»% .
. *
%
\ ••
ti •
• *
t * # #
* * *
•*.
n ~
~ * *
• •«
1 ~
~ t
*
~
•
* ~
~
1 r 3T*
1 1 %_
;
- <.
#
•
• ;* .
TT
•
!
-
\
•
~
•
fc\ —
t*
" .
A
~
•
* **
*
—»—^
~ •
•
• ~
•
••
~
~
1
•
• ~
1 * +
*
^ —
•
*_ ~
*
V ¦
•
•
1 ~ ~
* t _
*
» *
I*
1 *—
* *
%
•
f
•
• >*
•
«
•
•
•
* •
*
&
•
*
•
*
•
*
•:
•
•
~
V—
•
l>i- ^ i
~
—r
&*.
f I
~ ±1
"pt
{
* «
4t
»
~
* 1 *
#
-
"1
1
—1
VT- t/b Vb t/b \"i> Vb
tfr tfy tfy tfy- tfr tfy
u A ¦ ^ ^ ** **>'' ^ ^
~ Mitigation Off
* Mitigation On
• Passive Mitigation
Flags
EJ
—- Missing
U
06° ^ ^ ^ ^ ^ ^
Figure 6-8b. Plot of WP-3 chloroform concentrations vs. time (online GC).
6-12
-------
Section 6—Results and Discussion:
VOC Concentration Temporal Trends and Relationship to HVAC and Mitigation
Wall Port PCE
1000
E
*5o 100
£
O
CO
C
0)
u
c
o
u
10
X
~
~
~
~
A
~
A
A
k
~
, A
k. ~
X
*
~
~
l
A
~
i
k A
A
A
s
~
~
¦
~
¦
¦
¦
X
'
¦
¦
it
KB ¦ ¦
~ WP-1 PCE
¦ WP-2 PCE
A WP-3 PCE
X WP-4 PCE
Mitigation On
^^¦Mitigation Passive
o
*—1
*—1
*—1
*—i
(N
(N
(N
m
m
m
*—1
*—1
*—1
*—1
*—i
*—1
*—1
*—1
*—i
*—i
*—i
00
10
10
M
*—1
O
00
10
10
*—1
(N
O
*—1
(N
O
*—1
*—1
(N
O
*—i
^
^
^
^
^
^
^
^
*—1
(N
10
tn
(N
o
*—1
i/l
00
*—1
O
o
o
*—1
o
Date
o
*—1
O
o
o
Figure 6-9a. Plot of wall port PCE concentrations vs. time (method TO-17).
6-13
-------
Section 6—Results and Discussion:
VOC Concentration Temporal Trends and Relationship to HVAC and Mitigation
Hartman3
Tetrachloroethene
100-
3
c
o
ro
"E
\ \*>
^ ^
Mitigation Off
Mitigation On
Passive Mitigation
Flags
Missing
Figure 6-9b. Plot of WP-3; PCE concentrations vs. time (online GC).
6-14
-------
Section 6—Results and Discussion:
VOC Concentration Temporal Trends and Relationship to HVAC and Mitigation
6.1.3 Shallow and Deep Soil Gas
A series of 12 nested soil gas ports surround the 420/422 house or originate in the basements of either
side of the duplex (Figure 6-5). The five depths at each of the external nested locations are as follows:
3.5 ft bis, 6 ft bis, 9 ft bis, 13 ft bis, and 16.5 ft bis. Internal to the house are the nested locations notated
SGP8 through 12. Each individual port is notated based on its location and its depth (e.g., SGP1-3.5 for
the 3.5-ft depth at the SGP1 location). At the internal nested locations, there are only four depths; the 3.5-
ft depth is omitted because the basement floor is at ~5 ft bis. The internal soil gas VOC concentration data
are graphed in Figures 6-10 through 6-17 for the 420 side of the duplex and Figures 6-18 to 6-25 for the
422 side of the duplex. External to the house, there are seven nested locations, notated SGP1 through 7
and graphed as Figures 6-26 to 6-35.
Groundwater levels varied throughout the project but remained high enough most of the time to render the
16.5-ft depths inaccessible for soil gas sampling for much of the project. Based on subsurface profiles
from core data described in Section 3.1.1, the predominant subsurface lithology of the nested probe
depths can be described as follows:
• 3.5 ft bis: silt and silty sand with some clay and evidence of fill material (e.g., cinders, ash, coal
fragments, organic material)
• 6 ft bis: transition zone between finer silt and silty sand with clay above over coarser sand below
• 9 ft bis: sand and gravel outwash with some clay
• 13.5 ft bis: sand and gravel coarsening with depth
• 16 ft bis: sand, gravel, some cobbles
Thus, the general stratigraphy under the house is about 6 ft finer grain sediments (from fill or till)
overlying coarse to very coarse glacial outwash deposits (i.e., sand, gravel, and cobbles). Cobbles were
encountered during the drilling of MW-3, just to the south of SSP-1 and SGP-8 on the 422 side of the
duplex. The coarseness of the deeper material at the site is evidenced by the rapidity of the water table
rise after an increase in the gage height at nearby Falls Creek (see Section 11).
Prior to mitigation, VOC concentrations were generally highest in the deepest ports of each cluster and
decrease at shallower depths. This pattern is consistent with expectations for attenuation of vapor
intrusion of VOCs originating from a deep source (whether in the vadose zone or groundwater). This
attenuation pattern appears to be more pronounced for chloroform (frequently two to three orders of
magnitude) than for PCE (generally one order of magnitude).
An analysis of the frequencies of nondetects was performed for each compound by borehole and depth.
Of the boreholes outside the house footprint, only SGP1 (south of the 422 part of the duplex) has less than
a 20% frequency of nondetects for both PCE and chloroform (Table 6-1).
6-15
-------
Section 6—Results and Discussion:
VOC Concentration Temporal Trends and Relationship to HVAC and Mitigation
Table 6-1. Frequency of Nondetectable Samples (%) by Soil Gas Point or Cluster
Location
Percent Nondetect Samples
ID
Chloroform
PCE
TCE
Benzene
Toluene
Hexane
Radon
Soil Gas Probes
SGP1
22
13
74
33
67
79
0
SGP2
40
38
75
32
65
82
0
SGP3
70
83
91
34
78
89
0
SGP4
55
22
90
35
76
87
0
SGP5
45
55
91
37
77
92
0
SGP6
53
42
93
39
79
92
0
SGP7
51
38
91
36
80
90
0
SGP8
8
14
75
35
68
80
0
SGP9
10
6
79
33
71
81
0
SGP10
41
15
92
43
80
91
0
SGP11
20
5
89
34
77
89
0
SGP12
30
8
91
36
81
89
0
Subslab Ports
SSP-1
8
7
55
19
50
67
0
SSP-2
63
50
81
56
50
81
0
SSP-3
30
5
75
50
40
80
0
SSP-4
10
8
65
28
66
80
0
SSP-5
58
9
79
26
68
78
0
SSP-6
70
8
82
28
67
80
0
SSP-7
33
13
78
29
73
80
0
Wall Ports (Basement)
WP-1
82
86
86
32
75
86
0
WP-2
78
83
88
37
78
86
0
WP-3
49
37
75
28
74
82
0
WP-4
75
83
85
28
74
85
0
All of the wall ports have more than 20% nondetects for all compounds (Table 6-2). Nondetects are
infrequent (<20%) in almost all the subslab ports for PCE but more frequent for chloroform and the 420
side of the duplex.
Interestingly, SSP-1 and SSP-4 are consistently detectable (>80%) for benzene as well—in the center and
on the south end of the 422 side of the duplex. Benzene is also consistently detectable at the 16.5-ft depth.
Table 6-2. Frequency of Nondetects in TO-17 VOC Data by Soil Gas Sampling Depth
Depth
bis (ft)
Probe
reikeiii iMunueieki oainpies
Type
Chloroform
PCE
TCE
Benzene
Toluene
Hexane
Radon
3
Wall Port
72
73
84
31
75
85
0
3.5
Soil Gas
85
86
95
39
82
95
0
6
Soil Gas
38
19
79
30
68
80
0
6
Subslab
35
10
71
27
61
76
0
9
Soil Gas
35
23
79
32
71
82
0
13
Soil Gas
13
14
97
50
81
95
0
16.5
Soil Gas
11
20
98
15
84
94
0
bis = below land surface; PCE = tetrachloroethene; TCE = trichloroethene
Note that a depth of 4 is assigned to the wall ports and a depth of 5 is assigned to the subslab ports.
6-16
-------
Section 6—Results and Discussion:
VOC Concentration Temporal Trends and Relationship to HVAC and Mitigation
For the trend in nondetects by depth, we see about what we would expect for a deep vapor intrusion
source; there are fewer nondetects at lower depths. PCE is under 20% nondetects at subslab ports (depth =
5) as mentioned before and from 6 ft down in soil gas ports it is also under 20%. Chloroform is under
20% nondetects only at a depth of 13 ft or deeper. Benzene is under 20% nondetects only at the deepest
depth of 16 ft. No other compounds were consistently detectable. Thus, the shallowest depths (3.5 ft)
were generally the most stable, with little fluctuation because most results were below the detection limit
(Figure 6-25 and Figure 6-26). The 9-ft depths had periods of stability as well (see Figures 6-12, 6-20,
6-30, and 6-31). Notable exceptions to the shallow stability can be found at SGP1 and, to differing
degrees, all of the indoor ports, SGP8 through 12 where the shallow concentrations were higher and thus
less affected by nondetects. At each of those ports, shallow concentrations seem to partially track the
seasonal variations seen in the deeper ports (see Figures 6-13, 6-14, 6-15, 6-23, 6-30, and 6-33). At SGP3
and 4, the deeper ports are often low or stable (see Figures 6-32 through 6-34). At the 6-ft depths, PCE
concentrations on the 420 side (Figure 6-11) rose toward the summer of 2011 and then fell off. A similar
trend is also seen at 9 ft. That trend is not seen in Figure 6-19 for the 422 side where PCE concentrations
in SGP8-6 and SSP-1 showed a decline.
Many of the deeper ports at each location (9 ft through 13 ft, sometimes 16.5 ft) showed what appears to
be a rough cycle responding to seasonal changes (see Figures 6-13, 6-14,6-15, 6-23, 6-30, and 6-33), that
is, VOC concentrations were higher in the cooler months and lower in the warmer months (Figures 6-13,
6-15, 6-20, 6-23, and 6-20). SGP3 and 4 were too diffuse to show much of a trend. SGP1 and 2 showed
the opposite PCE concentration trend, at SGP1-6 (see Figures 6-28 and 6-29).
TO-17 VOC concentrations at some sampling locations (SSP-5, -6, and -7 in Figure 6-6a) varied only
within a narrow range of two or three times over 1 year, suggesting that multiple samples or time-
integrated samples may have limited benefit. However, at some locations (SSP-1 and -4 in Figure 6-6a
and SGP11-9 and 12-9 in Figure 6-12), 10-times changes in soil gas VOC concentrations occur over 1
year, suggesting that there would be significant additional information provided from additional soil gas
sampling rounds at these locations.
Some features among the figures might be attributed to natural or project-related phenomena. Although
samples were taken multiple times per week, and in some cases per day, during the intensive rounds
(yielding as many as >12 successive samples at some locations during a week), there were no discernible
or notable concentration trends in the data. This suggests that there is probably not a strong diurnal
variance in subslab soil gas VOC concentrations at this duplex and that the frequency of sampling (and
thus, the artificial volumetric flow in the subsurface induced by frequent sampling) does not appear to be
significant (for example, see Figures 6-36 and 6-37). High concentrations found at the beginning of the
project but tapering off toward the spring could be due to the period of heavy snow and ice in the very
cold winter of January and February 2011 (for example, see Figures 6-14 and 6-20).
During the mitigation testing period PCE appeared to be gradually depleted at a number of 9- and 13-ft
ports (such as 11-9, 12-9; 6-13, 8-13, 9-13 and 10-13; -see Figures 6-13, 6-15, 6-23 and 6-33). In
contrast, PCE concentrations were increasing in other ports such as 8-9 and 9-9 and chloroform
concentrations increased in 8-13 and 9-13 (Figures 6-21 and 6-22).
6-17
-------
Section 6—Results and Discussion:
VOC Concentration Temporal Trends and Relationship to HVAC and Mitigation
10
Chloroform (420 Side)
SGP11-6
~ ~
%
k * • »
CO
10i
3
£
100 j
^ 3 *
>3 ioi
c
o
2 1001
1 1°l
s 15
o
O 1001
10J
100 i
10s
SGP12-6
*%
*
^ 'S
~ *» » <+* * MH «
—
SSP-3
i
SSP-5
SSP-6
SSP-7
^ &
•
I I
¦vO1
!¦ ~ * 4
Method
• TO-15
* TO-17
• • m
: • ~t:
^ ' 'P
Figure 6-10, Chloroform concentrations at subslab and 6-ft soil gas ports directly under the 420
side of duplex.
6-18
-------
Section 6—Results and Discussion:
VOC Concentration Temporal Trends and Relationship to HVAC and Mitigation
Tetrachloroethene (420 Side)
SGP11-6
AO
$
SSP-3
SSP-5
S 1°i
SSP-7
Method
• TO-15
* TO-17
v* yp s* yP js* ^
•F
Figure 6-11, PCE concentrations at 6-ft soil gas ports and subslab immediately below the 420
side of the duplex.
6-19
-------
Section 6—Results and Discussion:
VOC Concentration Temporal Trends and Relationship to HVAC and Mitigation
100 =
Chloroform (420 Side)
SGP11-9
CO
10 r
£
O)
c
o
CO
100 =
SGP12-9
Method
• TO-17
-------
Section 6—Results and Discussion:
VOC Concentration Temporal Trends and Relationship to HVAC and Mitigation
Tetrachloroethene (420 Side;
SGP11-9
100
r» ».
*
*
SGP12-9
O 100
O
10
«y*
* —
55:
&
^ oO
**
-# h
* *
Method
[I TO-17
Figure 6-13. PCE concentrations at soil gas points 9 ft below the 420 side of duplex.
6-21
-------
Section 6—Results and Discussion:
VOC Concentration Temporal Trends and Relationship to HVAC and Mitigation
Chloroform (420 Side)
SGP11-13
100:
E 10
100
10
* * •
*
w
-•
—
—*
V
—*—*—
*
*
SGP12-13
V
——V
*
* * —
r—— -*+
*
#
V
*
* .
*
*
Method
H TO-17
Figure 6-14. Chloroform concentrations in soil gas at 13 ft below the 420 side of the duplex.
6-22
-------
Section 6—Results and Discussion:
VOC Concentration Temporal Trends and Relationship to HVAC and Mitigation
100-
Tetrachloroethene (420 Side)
CO
£
O)
c
o
CO
o
c
o
O
I SGP11-13
•
#
f *# •#v\
•
v>v
SGP12-13
&
$
&
10-
100-
Method
• TO-17
Figure 6-15. PCE concentrations in soil gas at 13 ft below the 420 side of duplex.
6-23
-------
Section 6—Results and Discussion:
VOC Concentration Temporal Trends and Relationship to HVAC and Mitigation
100:
CO
£ 10:
O)
c
o
CO
o 100
c
o
O
10;
Chloroform (420 Side)
SGP11-165
SGP12-165
Method
• TO-17
Figure 6-16. Chloroform concentrations in soil gas at 16.5 ft below the 420 side of duplex.
6-24
-------
Section 6—Results and Discussion:
VOC Concentration Temporal Trends and Relationship to HVAC and Mitigation
100:
cn
c
o
CO
c
a>
o
c
o
O
10:
I0i
SGP11-16.5
#
+
•
SGP12-16 5
• •
#
Method
* TO-17
Figure 6-17. PCE concentrations in soil gas at 16.5 ft below the 420 side of the duplex.
6-25
-------
Section 6—Results and Discussion:
VOC Concentration Temporal Trends and Relationship to HVAC and Mitigation
1001
10;
100;
10;
n_
CO
£
I
100 =
10i
c
o
ra
100 i
c.
-------
Section 6—Results and Discussion:
VOC Concentration Temporal Trends and Relationship to HVAC and Mitigation
1001
101
Tetrachloroethene (422 Side)
SGP10-6
9
CO
1001
10 =
E
O) 1001
^ 10i
c
o
ra s
c 100 s
-------
Section 6—Results and Discussion:
VOC Concentration Temporal Trends and Relationship to HVAC and Mitigation
10 3
(*>
Q>
^100 =
c
o
ro
£ 10
o>
o
c
o
O
Chloroform (422 Side)
I •
«*
_
SGP10-9
SGP8-9
SGP9-9
Method
~ TO-17
100;
10;
^ ^ ^
&
Figure 6-20. Chloroform concentrations in soil gas port at 9-ft depth below 422 side of duplex.
6-28
-------
Section 6—Results and Discussion:
VOC Concentration Temporal Trends and Relationship to HVAC and Mitigation
Tetrachloroethene (422 Side)
SGP10-9
1°
H
-
1
CO
£
i
w
C
o
100 =
3
CO
u_
4
J
c.
-------
Section 6—Results and Discussion:
VOC Concentration Temporal Trends and Relationship to HVAC and Mitigation
100 =
10 =
CO
£
c
o
CO
-------
Section 6—Results and Discussion:
VOC Concentration Temporal Trends and Relationship to HVAC and Mitigation
10
CO
£
3
c
o
CO
-------
Section 6—Results and Discussion:
VOC Concentration Temporal Trends and Relationship to HVAC and Mitigation
100:
cr>
£
c
o
CO
-------
Section 6—Results and Discussion:
VOC Concentration Temporal Trends and Relationship to HVAC and Mitigation
CO
£
c
o
CO
-------
Section 6—Results and Discussion:
VOC Concentration Temporal Trends and Relationship to HVAC and Mitigation
10 E
i
L.
Chloroform (Outside)
SGP1-3.5
10?
SGP2-3.5
CO
gr 100g
^ 10-
c
o
CO
c
o
O
10!
-------
Section 6—Results and Discussion:
VOC Concentration Temporal Trends and Relationship to HVAC and Mitigation
Tetrachloroethene (Outside)
SGP1-3 5
10 =
101
3.
100 i
O)
c
o
CO
10 §
10 =
CD
c 10s
o
O
10:
SGP2-3.5
SGP3-3 5
SGP4-3 5
SGP5-3 5
SGP6-3 5
SGP7-3 5
Method
• TO-17
^
Figure 6-27. PCE concentrations in exterior soil gas at 3.5 ft bis.
6-35
-------
Section 6—Results and Discussion:
VOC Concentration Temporal Trends and Relationship to HVAC and Mitigation
10
10|
3
E
10;
Chloroform (Outside)
SGP1-6
SGP2-6
SGP3-6
C
o
2 i
C 1°!
o -
o
c ^
o
O 100;
10.
4
SGP5-6
SGP6-6
SGP7-6
Method
• TO-17
10a
:
4
r ,4
V
$
&
Figure 6-28. Chloroform concentrations in exterior soil gas at 6 ft. bis.
6-36
-------
Section 6—Results and Discussion:
VOC Concentration Temporal Trends and Relationship to HVAC and Mitigation
1001
3
*
—
10i
Tetrachloroethene (Outside)
SGP1-6
10 =
SGP2-6
cr>
£
SGP3-6
c
o
CO
SGP5-6
Method
* TO-17
-------
Section 6—Results and Discussion:
VOC Concentration Temporal Trends and Relationship to HVAC and Mitigation
Chloroform (Outside)
SGP1-9
100 g
101
¦
100 i
101
CO
£
1001
10 i
w
C
10J
=
o
(U
L—
c
CD
:
O
C
101
o
3
O
J,
SGP2-9
## %
*
SGP3-9
' ' " V"—1
SGP4-9
SGP5-9
v
• ^ '
SGP6-9
100
10
1 ^
!
-K—•
SGP7-9
10;
*t *
% *
~%
t«
i.
*
rfiS
Method
• TO-15
F TO-17
^ ^
Figure 8-30. Chloroform concentrations in exterior soil gas at 9 ft bis.
6-38
-------
Section 6—Results and Discussion:
VOC Concentration Temporal Trends and Relationship to HVAC and Mitigation
Tetrachloroethene (Outside)
SGP1-9
9
100 i
toi
100
101
CO
C. 10:
o) -i;
C 10 E
O i
CO L
2 10 i
c =
o :
o
100
10;
=
10:
• «
_tt ft
SGP2-9
'>1
SGP3-9
SGP4-9
» ~ »• 4*
SGP5-9
I il*
SGP6-9
y~r.
r—* t
SGP7-9
Method
• TO-15
F TO-17
* *
«# ~ ~ ~
xa°N'1
Figure 8-31. PCE concentrations in exterior soil gas at 9 ft bis.
6-39
-------
Section 6—Results and Discussion:
VOC Concentration Temporal Trends and Relationship to HVAC and Mitigation
I
100 8
10 i
1001
10
c 1001
~ 10 i
co =
-------
Section 6—Results and Discussion:
VOC Concentration Temporal Trends and Relationship to HVAC and Mitigation
100 1
10s
100
101
100 s
10 g
c 100 i
o =
~ 10 =
CO 3
0) 100 =
o
o 10 =
O
1001
Tetrachloroethene (Outside)
SGP1-13
10
>
100-
a
10 j
SGP2-13
SGP3-13
SGP4-13
»
«v,
X
SGP5-13
SGP6-13
SGP7-13
>N° ^
' * •»
*#•
* #
* *
* ~~
,a«- ^ ^
Method
• TO-15
F TO-17
Figure 8-33. PCE concentrations in exterior soil gas at 13 ft bis.
6-41
-------
Section 6—Results and Discussion:
VOC Concentration Temporal Trends and Relationship to HVAC and Mitigation
=
100i
CO
10:
o 10|
-------
Section 6—Results and Discussion:
VOC Concentration Temporal Trends and Relationship to HVAC and Mitigation
100§
3
10 i
Tetrachloroethene (Outside)
SGP1-16 5
100
SGP2-16 5
CO
£
o 10 =
co i
vvV ~C>
Figure 6-35. PCE concentrations in exterior soil gas at 16.5 ft bis.
6-43
-------
Section 6—Results and Discussion:
VOC Concentration Temporal Trends and Relationship to HVAC and Mitigation
Subslab PCE Concentrations 1st Intensive Round
10000
ST 1000
60
+3 100
o
4-J
re
+J
U
c
ai
u
c
o
u
io n
a
~
I
~
i
a
~
a
~
~
+
X
*
*
IT
¥
~XT
SSP-l PCE
~ SSP-2 PCE
¦ SSP-3 PCE
~ SSP-4 PCE
+ SSP-5 PCE
XSSP-6 PCE
X SSP-7 PCE
IN
IN
P0
ro
ro
40 ^
ro ro
Date
00
ro
CT1
no
Figure 6-36. Subslab PCE concentrations over a 1-week period during the first intensive round.
Subslab PCE Concentrations 2nd Intensive Round
10000
m 1000
60
O
'4= 100
ro
cu
u
o
u
10
IT
• SSP-l PCE
~ SSP-2 PCE
¦ SSP-3 PCE
A SSP-4 PCE
+ SSP-5 PCE
XSSP-6 PCE
X SSP-7 PCE
O
IN
no
IN
00
o
IN
IN
00
o
IN
00
IN
00
o
IN
CT1
IN
00
Date
Figure 6-37. Subslab PCE concentrations over a 1-week period during the second intensive round.
6-44
-------
Section 6—Results and Discussion:
VOC Concentration Temporal Trends and Relationship to HVAC and Mitigation
6.2 Radon Seasonal Trends (based on Weekly Measurements)
Please see Section 5.2 of U.S. EPA (2012a) for a complete discussion of this topic based on the 2011—
2012 data sets, and Section 5.2 of the current report for a discussion of the effects of the mitigation
system on radon concentrations. The periodic operations of the mitigation system, while having dramatic
short-term effects in reducing radon levels, appear to have not changed the long-term concentrations
observed in periods when the system was not on (Figure 6-38).
Electret Readings with Mitigation On/Off Cycles
u
Q.
20.00
15.00
10.00
5.00
0.00
-5.00
W
•
422 First
•
422 Base N
o
422 Base N Dup
•
422 Base S
~
420 First
~
420 Base N
~
420 Base S
-
Ambient
-
422 Office
^^•Mitigation On
Mitigation Passive
o
*—I
00
lO
o
00
*—I
o
lO
IN
O
-------
Section 6—Results and Discussion:
VOC Concentration Temporal Trends and Relationship to HVAC and Mitigation
CHCI3 Field GC Indoor Air Data - 422
0.00 '
7/26/2011
9/14/2011
11/3/2011
12/23/2011
2/11/2012
4/1/2012
Date & Time
O 1st 422
Figure 6-39. Online GC chloroform indoor air data for 422 first floor.
Measured chloroform concentrations for 422 basement were generally slightly higher than on the first
floor, ranging from detection level (-0.1 (ig/m3) to -1.7 |ig/nr\ Similar to the first floor, there was a
notable increase in concentration by approximately a factor of 5 starting in September, and concentrations
remained at that level until the end of the program in February (Figure 6-40). Short-term temporal
variations were less than a factor of 3.
Measured chloroform concentrations for 420 first floor (the non-climate controlled part of the house)
ranged from detection level (-0.1 (ig/m3) to -1.0 |ig/nr\ Concentrations were about the same as measured
in 422 first floor for the first phase, but slightly lower than 422 during the second phase and showed less
scatter. Similar to 422, there was an increase in concentration starting in September and continuing into
October (Figure 6-41). Other than these step changes, short-term temporal variations were generally less
than a factor of 2.
Measured chloroform concentrations for 420 basement ranged from -0.3 (ig/m3 to -1.0 (ig/m3
(Figure 6-42). A less distinct step change is seen at this port in late September. Aside from that step
change, short-term temporal variations were generally less than a factor of 2. Values were slightly lower
than values measured in 422 basement especially during the second phase.
6-46
-------
Section 6—Results and Discussion:
VOC Concentration Temporal Trends and Relationship to HVAC and Mitigation
CHCI3 Field GC Indoor Air Data - 422
7/26/2011
9/14/2011 11/3/2011 12/23/2011
Date & Time
2/11/2012
4/1/2012
d Basement 422
Figure 6-40. Online GC chloroform indoor air data for 422 basement.
CHCI3 Field GC Indoor Air Data - 420
3.0 -i
2.5
_ 2.0
m
E
"m
3
O
0.0 H 1 1 1 ' 1 ' ' ' ' 1 ' ' ' ' 1 ' ' ' ' 1 1 ' ' r
7/26/2011 9/14/2011 11/3/2011 12/23/2011 2/11/2012 4/1/2012
Date & Time
o 1st 420
Figure 6-41. Online GC chloroform indoor air data for 420 first floor.
6-47
-------
Section 6—Results and Discussion:
VOC Concentration Temporal Trends and Relationship to HVAC and Mitigation
CHCI3 Field GC Indoor Air Data - 420
3.0
2.5
_ 2.0
1.5
1.0
0.5
0.0 H 1
7/26/2011
9/14/2011
11/3/2011
12/23/2011
2/11/2012
4/1/2012
Date & Time
a Basement 420
Figure 6-42. Online GC chloroform indoor air data for 420 basement.
6.3.1.2 Tetrachloroethylene (PCE)
Measured PCE concentrations for 422 first floor ranged from 0.2 (ig/m3 to -2.2 |ig/nr\ although the vast
majority of values ranged from 0.5 (ig/m3 to 1.0 (ig/m3 (Figure 6-43). Generally, PCE concentrations
were similar for both sampling phases, although there were periods of higher values in the second phase.
Short-term temporal variations in the second phase were up to a factor of 4.
Measured PCE concentrations for 422 basement ranged from -0.3 (ig/m3 to -3.2 |ig/nr\ Short-term
temporal variations in the second phase were up to a factor of 4, similar to the variations seen on the first
floor (Figure 6-44).
Measured PCE concentrations for 420 first floor (the non-climate controlled part of the house) ranged
from detection level (-0.1 (.ig/rn3) to -2.2 (.ig/rn3 (Figure 6-45). Generally, the concentrations were higher
in the first phase with little temporal variation and much greater short-term variation during the second
phase. Temporal variation during the first phase was generally less than a factor of 2, but short-term
temporal variations in the second phase were up to a factor of 10.
Measured PCE concentrations for 420 basement ranged from detection level (-0.1 (.ig/rn3) to -2.2 |ig/nr\
Patterns were similar to those seen on the first floor with little temporal variation during the first phase
(<2x) and higher short-term variations during the second phase of a factor of 10 (Figure 6-46).
6-48
-------
Section 6—Results and Discussion:
VOC Concentration Temporal Trends and Relationship to HVAC and Mitigation
PCE Field GC Indoor Air Data - 422
~
~
~
~
~
t
% *
~
J 1
*
v
ur? * ?
~
7/26/2011 9/14/2011 11/3/2011 12/23/2011 2/11/2012 4/1/2012
Date & Time
Figure 6-43. Online GC PCE indoor air data for 422 first floor.
PCE Field GC Indoor Air Data - 422
LHJ3QDKLD q p——
rm ~~~ ~ ~ ip
ID ~~~ Hill I ~ ~ QD
~j ID D^ PEP ^
~ m p ~ 11 n i
~ ~~
mm
~an
-cnfiS*
7/26/2011
9/14/2011
11/3/2011 12/23/2011
Date & Time
2/11/2012
4/1/2012
n Basement 422
Figure 6-44. Online GC PCE indoor air data for 422 basement.
6-49
-------
Section 6—Results and Discussion:
VOC Concentration Temporal Trends and Relationship to HVAC and Mitigation
PCE Field GC Indoor Air Data - 420
II I
7/26/2011
9/14/2011
11/3/2011 12/23/2011
Date & Time
2/11/2012
4/1/2012
Figure 6-45. Online GC PCE indoor air data for 420 first floor.
PCE Field GC Indoor Air Data - 420
5.0
0.0 H 1
7/26/2011
9/14/2011
11/3/2011
12/23/2011
2/11/2012
n Basement 420
4/1/2012
Figure 6-46. Online GC PCE indoor air data for 420 basement.
6-50
-------
Section 6—Results and Discussion:
VOC Concentration Temporal Trends and Relationship to HVAC and Mitigation
6.3.2 Subsurface Soil Gas Data
Subsurface VOC concentrations were monitored at eight locations with the automated GC:
¦ three subslab locations: SSP-2, SSP-4, and SSP-7
¦ four soil gas locations: SGP2-9 ft, SGP8-9 ft, SGP9-6 ft, and SGP11-13 ft
¦ one location in the wall on the side of the basement (WP-3).
Approximately 600 measurements per location were collected in Phase 1, and approximately 900
measurements per location were collected in Phase 2 at each of these eight locations.
6.3.2.1 Chloroform
The chloroform concentration data from the automated GC for all locations for both sampling phases are
summarized in Figure 6-47 and for the separate phases in Figures 6-48 and 6-49.
In the first phase of the program, chloroform concentrations were relatively constant until approximately
September 13. At that time, the instrument inexplicably stopped and was not restarted until 2 days later on
September 15. Upon restart, there was an abrupt increase in all the chloroform concentrations but not the
PCE concentrations. This shift occurred because of a change in the chloroform baseline definition by the
integration software and is not due to changes in the actual chloroform concentrations.
The following chloroform concentration behaviors were observed in the first phase (Figure 6-48):
¦ Temporal variation is generally less than a factor of 2 at all the sample locations during this phase
except for location WP-3. However, several ports showed what appeared to be low frequency
bimodal behavior. For example, SGP9-6, SGP11-13, SSP-2, SSP-4 show substantial number of
points at a second level offset by more than order of magnitude from the most common level.
¦ At probe WP-3, concentrations show smoothly varying high and low variations of a factor of 3 to
5 times occurring over time scales of several days. WP-3 was the only location to exhibit this
behavior.
In the second phase of the program (Figure 6-49), the following behaviors were observed:
¦ Probe WP-3 continued showing the same oscillations in chloroform concentrations as in the first
phase.
¦ Probes SGP9-6 ft and SSP-4 showed a continual rise in chloroform concentrations throughout the
sampling period, increasing by approximately 2 to 2.5 times above the starting concentration of
the second phase. This same increase in chloroform concentrations at SGP9-6 ft was also
observed in the TO-17 grab soil gas samples as a trend running from late August to December.
This pattern was not seen in the first phase of the program. Despite this large concentration
increase of chloroform soil gas concentrations during this second phase, there was no concurrent
increase in the indoor air concentrations of chloroform measured by the online GC in either the
basement or first floor of unit 422 (Figure 6-39).
¦ SSP-2 had approximately one order of magnitude variation but at relatively low chloroform
concentration levels.
6-51
-------
Section 6—Results and Discussion:
VOC Concentration Temporal Trends and Relationship to HVAC and Mitigation
CHCI3 Field GC Subsurface Air Data
265 1
255 •
245 1
235 1
225 1
215 •
205 1
195 1
185 1
m 175 '
£ 165 •
3 155 •
I 145 '
'jo 135 •
c 125 •
85 '
7/^6/2011
T 1 1 1—
9/14/2011 11/3/2011 12/23/2011 2/11/2012 4/1/2012
Date & Time
~ WP-3 nSGP9-6 OSGP8-9 "SGP2-9 nSGPll-13 0SSP-7 DSSP-4
Figure 6-47. Online GC subsurface chloroform soil gas data—Phase 1 and Phase 2.
CHCI3 Field GC Subsurface Air Data
8/5/2011
8/15/2011
8/25/2011
9/4/2011
9/14/2011 9/24/2011
Date & Time
10/4/2011 10/14/2011 10/24/2011
~ SSP-2 OWP-3 ESGP9-6 ~SGPS-g "SGP2-9 nSGPll-13 ¦ SSP-7 ~ SSP-4
Figure 6-48. Online GC subsurface chloroform soil gas data—Phase 1.
6-52
-------
Section 6—Results and Discussion:
VOC Concentration Temporal Trends and Relationship to HVAC and Mitigation
CHCI3 Field GC Soil Gas Data
11^23/2011 12/3/2011 12/13/2011 12/23/2011 1/2/2012 1/12/2012 1/22/2012 2/1/2012 2/11/2012 2/21/2012
Date & Time
~ SSP-2 nWP-3 ~SGP9-6' © SGP8-9 ¦ SGP2-9 nSGPll-13 S SSP-7 ~ SSP-4
Figure 6-49. Online GC subsurface chloroform soil gas data—Phase 2.
6.3.2.2 Tetrachloroethylene (PCE)
The PCE concentration data from the automated GC for all locations for both sampling phases are
summarized in Figure 6-50 and for the separate phases in Figures 6-51 and 6-52.
In the first phase of the program (Figure 6-51), it appears as if there is a lot of fluctuation in the
subsurface PCE concentrations. However, inspection of the individual locations shows the following:
¦ Probes SGP2-9 ft, SGP8-9 ft, and SGP9-6 ft show only slight temporal variations of 20% to 50%,
except for some very infrequent outliers.
¦ There are two probes that field records suggest may have been inadvertently closed for a period
of time:
SGP11-13 ft 8/29/11 @ 15:16 closed; 9/9/11 between 14:00 and 15:00 opened
SSP-7 ft 8/29/11 @ 15:36 closed; 9/9/11 between 14:00 and 15:00 opened
¦ Probes SSP-4 and SSP-2 also show less than a factor of 2 temporal variation over most of the
sampling period. However, both of these probes contain a group of analyses when the PCE
concentrations dropped rapidly by large amounts and then increased rapidly back to the prior
values (Figure 6-51). The cause for this behavior is not clear. The drop in SSP-2 data occurred at
times that may suggest an effect of the fan tests (discussed in Section 12.2). The SSP-4 drop offs
happen more frequently and do not appear to be caused by the fan tests. The TO-17 data for SSP-
4 PCE over the whole year also did show considerable variability (Figure 6-53). The pattern of
this subslab probe's plot is reminiscent of Johnson et al's (2012) observation of data from another
house: "There are long periods of relative VI activity with sporadic VI inactivity. ".
6-53
-------
Section 6—Results and Discussion:
VOC Concentration Temporal Trends and Relationship to HVAC and Mitigation
PCE Field GC Soil Gas Data
L t E
kjL
G dj
i
¦
7/26/2011
9/14/2011
11/3/2011 12/23/2011
Date & Time
2/11/2012
4/1/2012
Figure 6-50. Online GC subsurface PCE soil gas data—Phase 1 and Phase 2.
PCE Field GC Soil Gas Data
$
Jn
4
~
° 0 ^ 0^
n ~ ~ ^ D
u 1
—* ?
° . Iff
¦i ^
—-'i—\ ¦
$
p mo fe" ^o-
b D1
n-»- A 1
'®9a^ JHfc
7T
8/5/2011 8/15/2011 8/25/2011 9/4/2011 9/14/2011 9/24/2011 10/4/2011 10/14/2011 10/24/2011
Date & Time
¦WP-3 ® SGP9-61 O SG P8-9
°SGP11-13 "SSP-7 OSSP-4
Figure 6-51. Online GC subsurface PCE soil gas data—Phase 1.
6-54
-------
Section 6—Results and Discussion:
VOC Concentration Temporal Trends and Relationship to HVAC and Mitigation
PCE Field GC Subsurface Air Data
Date & Time
Figure 6-52. Online GC subsurface PCE soil gas data—Phase 2
VOC Data for SSP-4 (Method TO17)
S
c
o
u
•
• A
A
•
A
• m
•
A
•
•
•
A.
A
A
~v
• •
A
• A
• M
* •
te *
~Wr —
» A
A J.
* * *
•
0
A*
A A A
•
A
A
• A
• ^
a
t
* ¦«
\ \ \ * \ \
*titaiorcrfbtifittr4
Date
Figure 6-53. Method TG-17 data for SSP-4.
6-55
-------
Section 6—Results and Discussion:
VOC Concentration Temporal Trends and Relationship to HVAC and Mitigation
¦ Probe WP-3 PCE concentrations show repeated high and low variations of a factor of 3 to 5 times
occurring over weekly time scales. These fluctuations are similar to the chloroform variations
seen in this same probe.
In the second phase of the program (Figure 6-52), the following behaviors were observed:
¦ Probes SGP2-9 ft, SGP8-9 ft, SGP9-6 ft, SGP11-13 ft, and SSP-7 show slight temporal variations
of 20% to 50% over the sampling period. Probes SGP2-9 and SGP11-13 have a gradual
downward drift over several months.
¦ Probe SSP-4 is constant within 25% for most of this phase of observation but shows one period of
a rapid drop in PCE concentrations down to near-zero values and then a quick rebound to the pre-
drop values (Figure 6-54). This probe is located very close both spatially and within 18 inches
vertically to probe SGP9-6 ft. SGP9-6 ft had similar PCE concentrations and did not show the
same rapid variations. However, the drop in values is also seen in the method TO-17 samples of
location SSP-4 at other times. This suggests that the behavior at SSP-4 was due to air leakage in
the thin void zone that often exists under concrete slabs (DePersio and Fitzgerald, 1995) and,
thus, had less influence on the SGP9-6 ft probe, which had a wider screened interval.
¦ Probe SSP-2 was characterized by low PCE concentrations that varied up to an order of
magnitude.
¦ Probes SGP11-13 ft and SSP-7 did not show the rapid drop in concentrations seen during the first
phase, suggesting that the behavior in the first phase might indeed be due to valve closure, not
actual variations in the soil gas PCE concentrations as discussed above.
¦ Probe WP-3 continued to show the same oscillations as in the first phase with slightly greater
variations of a factor of 5 to 8 times occurring over time scales of several days. These fluctuations
are similar to the variations in chloroform concentrations seen in this same probe.
¦ The PCE concentrations at locations SGP9-6 ft and SSP-4 decreased slightly over the sampling
period in contrast to the CHC13 concentrations, which showed large increases in these two probes
over the same time period (Figure 6-55 shows data from SGP9-6). This trend was also observed
in the TO-17 sampling of this port during the same time period. This is indicative of different
sources for the chloroform and tetrachloroethylene.
In summary, except for probe WP-3, the regular short time scale (<14 day) temporal variations in PCE
concentrations seen in all the subsurface probes are typically less than an order of magnitude. Probe WP-3
is located closest to the ground surface (~3; bis) so the variations detected might be due to surface
influences. SSP-4 showed long periods at relatively steady elevated concentrations punctuated by short
intervals of dramatically lower concentrations.
Variations in soil gas PCE concentration that were observed at WP-3, and to a lesser extent at SSP-2,
occurred over a period of days, indicating that there is little advantage to collecting 24-hour composite
samples versus instantaneous grab samples at this probe.
6-56
-------
Section 6—Results and Discussion:
VOC Concentration Temporal Trends and Relationship to HVAC and Mitigation
PCE - SSP-4 (Port 13)
170
160
150
140
130
120
110
100
90
80
70
60
50
40
30
20
10
9/14/2011
11/3/2011
12/23/2011
2/11/2012
4/1/2012
Date & Time
Figure 6-54. Online GC PCE measurements in SSP-4.
PCE & CHCI3 - SGP9-6
—
~ PCE ~ CHCI3
11/23/2011 12/3/2011 12/13/2011 12/23/2011 1/2/2012 1/12/2012 1/22/2012 2/1/2012 2/11/2012 2/21/2012
Date & Time
Figure 6-55. Comparison of online GC measurements of PCE and chloroform in SGP9 at 6 ft.
6-57
-------
Section 6—Results and Discussion:
VOC Concentration Temporal Trends and Relationship to HVAC and Mitigation
6,4 Radon Short-Term Variability (Based on Daily and More Frequent
Measurements)
This section discusses the short term variability in indoor radon levels measured by the stationary
AlphaGUARD instruments located in the basement and in the second floor office of the 422 side of the
duplex. For additional discussion of the stationary AlphaGUARD data, please see the discussion in
Section 5.4 of U.S. EPA (2012a) regarding the 2011—2012 data and Section 5.2 of the current report
regarding the effect of mitigation system. Section 5.4 of U.S. EPA (2012a) discussed electret radon in
indoor air and a breakdown of electret and stationary AlphaGUARD data during intensive periods.
Regular sampling with the portable AlphaGUARD was summarized as well.
The stationary AlphaGUARD data set now includes more than 110,000 measurements at each of two
locations. The degree of short-term variability in radon concentration observed when the mitigation
system is not on, but after it began operating, is quite similar to the long-term trend. Dramatic variations
of as much as 15 pCi/L within a few days are common in the basement data set, collected every 10
minutes in indoor air when the mitigation system is not operating (Figure 6-56). During the "mitigation
on" periods the vast majority of the data is confined to a narrower absolute range from -0.5 to 2.3 pCi/L
(note the negative readings are not physically realistic, but likely reflect a small offset error).
The office data set (422 side 2nd floor) has a somewhat smaller range of short-term variation (about 8
pCi/L is typical) but shows a similar response to mitigation (Figure 6-57). During the period of
mitigation the variation is confined to a smaller absolute range, approximately 0 to 2.5 pCi/L.
Downstairs Stationary AlphaGUARD
Downstairs AiphaGUARD
¦Mitigation On
Mitigation Passive
Snow Events
¦Flooding
Figure 6-56. Real-time radon levels (422 basement) 2011-2013.
6-58
-------
Section 6—Results and Discussion:
VOC Concentration Temporal Trends and Relationship to HVAC and Mitigation
Upstairs Stationary Alp ha GUARD
HI
H
H
rM
r%i
r4
N
ny
H
H
H
m
H
¦4
H
s
%
&
o"
OS
i
m
mm
in
Sr
>
7t
Q
O
i-l
o
o
o
d
O
Upstairs AJphaGUARD
•Mitigation On
Mitigation Passive
•Snow Events
•Flooding
Date
Figure 6-57. Real-time radon levels (422, 2nd floor office), 2011-2013.
6.5 Outdoor Climate/Weather Data
External and internal weather parameters were measured at the 422/420 house on a Vantage Vue weather
monitor. Internal temperatures were recorded by HOBO data loggers. Barometric pressure readings were
taken about every 15 minutes by Setra pressure sensors. Data were downloaded from these sources
approximately once per week. Well water levels were measured approximately once per month during the
first portion of the study but then continuously in late 2012 early 2013. The 2011 through 2013 weather
data were presented and analyzed in Section 5.5 of U.S. EPA (2012a). This section thus focuses on the
2012 through 2013 weather situation.
Table 6-3 presents data from monthly weather summaries for 2012 and 2013 published by the Indiana
State Climate Office (Scheeringa, 2012-2013). The 2012 through 2013 project year was very different
from the previous year. The year began with less rain than normal, which quickly led to months of
drought conditions for much of Indiana. Hie summer of 2012 saw heat waves followed by the third
wannest July in Indiana record. However, as the year proceeded to autumn, weather became wetter and
cooler than normal. A wanner than average December finally yielded to winter on December 21, leading
to cooler conditions and even a blizzard the day after Christmas. The winter of 2013 began much wetter
than normal, with wild temperature swings and precipitation in Febniary. In 2012, March had been the
wannest March recorded, but March 2013 was 20 °F colder than the one March 2012, falling 6 °F below
average for the state (Table 5-3 and Scheeringa and Hudson, 2012-2013).
6-59
-------
Table 6-3. Summary Meteorological Data for Central Indiana Note that the symbols "A" and "v" mean "above" and "below" normal,
respectively, and that the weekly values show how the weekly averages differ from normal (from Scheeringa and Hudson,
2012, 2013)
State Ave T
(°F)
Central IN
Ave T (°F)
State Ave
Precipitation
(in)
Central Ave
Precipitation
(in)
Week 1 T ave
(°F)
Week 2 T ave
(°F)
Week 3 T ave
(°F)
Week 4 T ave
(°F)
Month/Year
Special Notes on Central IN
April 2012
53.0, 1.7 A
52.4, 1.6 A
2.32, 1.62v
2.47, 1.44v
Cool, hard freezes, less rain
9 A normal
1 v normal
3 A normal
3 v normal
May 2012
67.8, 5.8 A
67.6, 5.9 A
2.79, 1.61 v
3.19, 1.2 v
Mod drought
10 A normal
normal
3 A normal
16 A normal
June 2012
72.1, 1.2 A
71.9, 1.3 A
1.3, 2.89v
0.86, 3.24v
Heat wave, severe drought
4 v normal
1 A normal
5 A normal
4 A normal
July 2012
80.5, 5.9 A
80.5, 6.2 A
2.45, 1.65v
1.75, 2.51 v
Drought, 3rd warmest July
11 A normal
3 A normal
5 A normal
6 A normal
August 2012
72.3, 0.3 v
71.9, 0.3 v
3.95, 0.17 A
4.47, 0.72 A
Drought broken, windy
3 A normal
3 v normal
5 v normal
1 A normal
September 2012
64.2, 1.5 v
63.6, 1.7 v
5.38, 2.29 A
6.40, 3.42 A
Cool, above normal rain
4 A normal
3 v normal
5 v normal
3 v normal
October 2012
51.7, 2.2 v
51.2, 2.3 v
3.97, 1.07 A
4.44, 1.62 A
Cooler and wetter than normal
4 v normal
6 v normal
3 A normal
5 v normal
November 2012
40.2, 2.2 v
39.8, 2.1 v
1.02, 2.57 v
1.19, 2.45v
Cold, dry, and uneventful
19 v normal
1 v normal
2 A normal
2 v normal
December 2012
37.3, 6.2 A
37.0, 6.3 A
3.22, 0.16 A
3.22, 0.24 A
Warmer and wetter than ave.
13 A normal
5 A normal
10 A normal
7 v normal ?
January 2013
29.7, 3.7 A
28.9, 3.6 A
4.7, 2.26 A
5.18, 2.84 A
Precipitation 210% normal
4 v normal
14 A normal
8 v normal
2 A normal
February 2013
29.8, 0.6 v
29.3, 0.3 v
2.24, 0.04v
2.13, 0.14v
Wild swings in T and precip.
1 v normal
9 A normal
15 v normal
3 v normal
March 2013
34.8, 6 v
34.3, 6.2 v
2.41, 1 v
2.14, 1.14v
20 deg colder than last March
6 v normal
<1 v normal
9 v normal
7 v normal
April 2013
50.9, 0.5 v
50.9, norm
6.45, 2.51 A
7.45, 3.54 A
Wnds and heavy precip.
6 v normal
7 A normal
15 v normal
3 v normal
§
n
n
0
3
rs
ns
s
1
».
o
s
O
2
s
S-
q
3
>3
ST
o'
s
"S'
s
n
q
s
Note that the symbols "A" and "v" mean "above" and "below" normal, respectively, and that the weekly values show how the weekly averages differ from normal (from Scheeringa and Hudson, 2012, 2013)
fe3
rs
O
a
On
>3
o'
s
s
b
>3
si
S'
-------
Section 6—Results and Discussion:
VOC Concentration Temporal Trends and Relationship to HVAC and Mitigation
Figure 6-58 shows the temperature record from the external temperature monitor and HOBO devices
placed at seven indoor locations on the 422 and 420 sides of the house.
100"
100-
-------
Section 6—Results and Discussion:
VOC Concentration Temporal Trends and Relationship to HVAC and Mitigation
The same general trend can be seen in both figures, cycling from the winter lows to the summer highs.
However, indoor temperatures reflect the moderating effects of the thermal mass of the building, HVAC
systems (422), and air conditioning (both sides of the duplex at times).
As stated in Section 3.2.1, the gas-fired furnace was run from November 19, 2010, until June 22, 2011,
then from November 7, 2011, until June 1, 2012, and then from October 31, 2012, through May 22, 2013
on the 422 side only, with no heating unit on the 420 side. Initially, window-mounted ACs ran on both
sides of the duplex from June 29, 2011, until July 12,2011. When the ACs were replaced, they were
replaced on the 422 side only and ran from March 3, 2011, until October 24, 2011. Figures 5-73 and 5-74
show some of the highest temperatures occurring during the period between the AC theft and when they
were replaced on the 422 side, along with higher temperatures on the 420 side where the AC units were
not replaced. The higher temperatures between AC periods could be a result of the solar stack effect,
which may have been driving the higher radon and VOC concentrations observed during that time (see
Section 5.2.1).
The most obvious features of the stacked hydrological graph of Figure 6-59 are the prominent highs in
rainfall and stream discharge, coupled with the high water levels measured during gauging. These highs
align well with the period of heavy snowfall and rain experienced in central Indiana (see Table 6-3). Dips
in stream discharge and the lower depths during well gauging match well with the much hotter drier
summer period.
re 4
<4- H
c
OC 2
6000
oj 4QQ0
CL0
™ 2000
K S
d
Stream Discharge, Average Depth to Water, and Rainfall
* * «
• •
*
%
rJ
©
rsl
cn
F4
LO
o
Date
<
ia JD
0)
15 «
Id O
12
¦ Strum Orvttafg*
• AVW HRf! DTW
- Ftanfnll
Figure 6-59. Stacked hydrological graph with rainfall in inches (top—green line), depth to water in
feet (middle—red circles), and discharge at Fall Creek in ft3/s (bottom—blue line).
6-62
-------
Section 6—Results and Discussion:
VOC Concentration Temporal Trends and Relationship to HVAC and Mitigation
Various indices related to wind speed are shown in Figure 6-60. Generally, wind speeds are lowest in
summer at this house. As shown in Figure 6-61, summer is also the period of least barometric pressure
variation at this house. Figure 6-61 also shows that indoor humidity is much more stable over short time
scales than outdoor humidity. In addition, the bottom graphs on Figure 6-61 include rainfall in
inches/hour and total amounts ("Rain") to show how the high rates correlate with significant total rainfall
amounts. Figure 6-62 shows snow depths in inches. The first winter 2010 to 2011, with a depth of 6
inches, and third winter (2012 to 2013), with a depth of approximately 7 inches, were more severe
regarding snowfall than the second winter (2011 to 2012).
50-
40'
30-
20'
10
0-
10-0-
7.5
5.0-
2.5
o.o-
Hi Speed MPH
Wind.Run, Miles
Wind.Speed. MPH
20-
15-
10-
5-
o-
,ro
a
&
Jo
$>
a
$
&
Figure 6-60. Plot of high wind speed for measurement period, wind run and wind speed (average
over measurement period) at 422/420 house over time.
Wind run is calculated by multiplying the wind speed by the measurement period and summing over time.
6-63
-------
Section 6—Results and Discussion:
VOC Concentration Temporal Trends and Relationship to HVAC and Mitigation
30.4"
30.0-
29.6-
29.2-i
15
10-
5"
60-
40"
20-
100-
75-
50-
25-
1.00
0.75
8:38=
0.00 ^
15-
10-
5-
Bar.., In.Hg
In.Air.Density NA
In.EMC. NA
In.Hum, Percent Humidity
Out.Hum Percent Humidi
Rain. In
t . . ~J~ , *
Mitels
Rain.Rate In/Hr
0-i—^
ii T|r.»k^U rj.arWi it
y>v
&
yJ
-------
Section 6—Results and Discussion:
VOC Concentration Temporal Trends and Relationship to HVAC and Mitigation
Q.
Q
o
c
CO
&
Figure 6-62. Snow depth vs. time (data are from NCDC records for the Indianapolis International
Airport).
6.5.1 Indianapolis Weather Compared with VOCs and Radon
6.5.1.1 Wall Port VOC Concentrations as a Function of Barometric Pressure and Wind
Speed
Sudden peaks and troughs in real-time VOC and radon data appear to have a qualitative relationship to
weather phenomena. Figures 6-63 and 6-64 (from Hartman 2 and 3, respectively) compare PCE and
chloroform data for WP-3 in the 422 basement wall from the GC to barometric pressure changes recorded
on the 422/420 house weather monitor. From theory we would expect a decrease in barometric pressure to
be associated with higher vapor intrusion. Note that the general patterns of peaks and troughs in the wall
port PCE and chloroform data only bear a rough resemblance to some of the highs and lows in the
pressure data. Also note that although a late December 2012 low pressure trough might correspond with a
6-65
-------
Section 6—Results and Discussion:
VOC Concentration Temporal Trends and Relationship to HVAC and Mitigation
high peak in the PCE concentration trend, a similar late February 2012 high PCE concentration does not
show a corresponding pressure drop. This suggests that multiple factors control the PCE concentration,
and because of this complexity, the relationship between barometric pressure and PCE concentration is
not strongly supported by either of the GC phases.
GC Phase 2 at WP-3 PCE and Chloroform Compared with Outdoor Pressure
TS J1
tic
I
c
01
I—
CL
30
29
m
300
c E 200
M Ei6
o £ ioo
U
UJ
u
0
a r
r\ *J
v >
/M
r ^ A
fc1
A, A
iv V
i /AAj
\ / V
u 1
lA A Aj
V \r
A
i f V rtjr
\ f\
V vv
30
A_,V
A I
v J KaA.:VV^w
70
SO
k i0
OUo
JL
O
a
o
o
^ i
3 3
~ £
3
O
3
•WPJGC PCI
* WP-i tit Q403
Outdoor Pmsure
Date
Figure 6-63. GC Phase 2 VOCs at WP-3 compared with 422/420 house external pressure.
6-66
-------
Section §—Results and Discussion:
VOC Concentration Temporal Trends and Relationship to HVAC and Mitigation
GC Phase 3 at WP-3 PCE and Chloroform Compared with Outdoor Pressure
-3 "
lA
irt
O
29
300
200
o 3t ioo
U
U
0
\ A AaA/
ulMV
vy*1
/
. I A Al /I JL
AA/
V [IVf
Aa a aA
\/VVy
H «
I
i /A f\ fi
\r\ wa f
' V
M
i
\
¦i
¦ m
* »
*•
m
* '
l,h
N
"V
y
&
f
3P
O
a
O
o
so CQ o
w s
3
s
3
* WP-3GC PC£
• wmgcchcii
Om Itkw Pwwmw
Date
Figure 6-64. GC Phase 3 VOCs at WP-3 compared with 422/420 house external pressure.
Figures 6-65 and 6-66 (from Hartman 2 and 3, respectively) compare the same GC VOC data for WP-3
with wind speed recorded by the 422/420 house weather station. Again, there is a rough correspondence
between highs and lows within the data sets. And again, as with the pressure data, high wind spikes might
correspond in time to some peaks in PCE and chloroform data, but not in a constant ratio. It is possible
that a relatively continuous dramatic change in wind speed has more of an effect than just a high speed
alone (compare for example January 2012 with December 2012). Modeling studies would suggest that
both the speed and direction of wind would be associated with vapors being driven up on one side of a
building. In this case north or northwesterly winds would be likely to increase concentrations at WP-3.
6-67
-------
Section §—Results and Discussion:
VOC Concentration Temporal Trends and Relationship to HVAC and Mitigation
GC Phase 2 PCE and Chloroform at WP-3 Compared with Wind Speed
JC
Q.
£
*o
&>
i'
a
~o
c
20
10
.2
I ~
11,
UJ
Q.
300
200
100
rnm, *
eh *
• ¦
» • *4
£ •
jJf= Ji
2 IT
.*£ 2 .
-u-v
•S&
w 1
A.
•
L *
^ ^ ^
SQ
3D
&
O
zr
o
o
o
t: 3
n
3 °
3
O
3
• W?-3GCPCC
• VVPJGCCHCI3
• WindSpred
Date
Figure 6-65. GC Phase 2 VOCs at WP-3 compared with 422/420 house external wind speed.
6.5.1.2 Effects of Snow and Ice on Radon and VOCs
It has been hypothesized that frozen soil, snow packs, or ice packs could affect vapor intrusion by
providing a "cap"' that limits interaction between shallow soil gas and the atmosphere (ITRC, 2007). A
snow/ice pack did develop at the Indianapolis site in the winter of 2011 (see previous Figures 3-9, 3-16,
and 6-62). Impervious surface caps have been shown to have a substantial effect on VOC distribution in
soil gas (Schumacher, 2010). The air permeability of a snow layer is a complex function of pore size,
grain size, ice fraction, density (Armstrong, 2008; Bender, 1957; Conway and Abrahamson, 1984). A
recent North Battleford Saskatchewan vapor intrusion study showed little effect on petroleum vapor
intrusion from a foot deep, light snow pack. However, since that study focused on biodegradable VOCs, it
may indicate only that oxygen was successfully able to be delivered to soil microorganisms through the
snow; thus, this study may not predict the effects of snow on recalcitrant VOCs (Hers et al., 2011).
Investigators in that study planned in the future to create an ice lens to see if it had an effect. A successful
field demonstration was conducted at Oak Ridge, TN, of the use of an engineered frozen soil barrier to
hydraulically isolate a volume of radioactive waste to prevent its migration (DOE, 2009).
6-68
-------
Section §—Results and Discussion:
VOC Concentration Temporal Trends and Relationship to HVAC and Mitigation
GC Phase 3 PCE and Chloroform at WP-3 Compared with Wind Speed
JC
Q.
E
20
X3
&>
i'
£L
trt 10
c
.2
I
HM
C
a
y
e
3
LU
CL
300
200
5 100
,* J J JT *
Mr -frr <• ft MB Hto-i
i j
* • » * ~*
£
i i
%
¦i
¦ m
•
* •*
uLi » Ji mmlL
s
%
%
,
•
i ,-V \'^\J
m
• v*
* ¦
p*V
o
zr
O
p1
II
3 °
3 3
3
O
3
&
f
• WP-3 GC PC £
• VVPJGCCHCI3
• WindSpwd
Date
Figure 6-66. GC Phase 3 VOCs at WP-3 compared with 422/420 house external wind speed.
The EPA research team observed a potentially weather-mduced phenomenon in the winter of 2011.
During a period of extensive snow and ice fall, observers noted an increase in VOCs and radon matching
the period of a firm snow and ice cover in the surrounding grounds of the 422/420 house. Unfortunately,
the following winter, this snow/ice period did not repeat itself; however, a mini-intensive sampling period
conducted during a brief snowfall in 2012 did manifest similar characteristics, to a much smaller degree,
to the 2011 snow/ice period.
Figures 6-67 through 6-69 plot GC VOC data and radon data from December 2012 through March 2013.
In each, the radon data are from the 422/420 house stationary AlphaGUARDs and are repeated on each
figure. Each of these figures further compares the GC VOC (chloroform and PCE) data and radon data
with mitigation on/off cycles and visually observed winter weather conditions. Observed winter weather
conditions included long and short periods of snow cover on the 422/420 house yard and periods of the
front and backyards being frozen.
Radon (any of Figures 6-67 through 6-69) shows a dramatic decrease during the mitigation on periods,
but little to no change during periods of passive mitigation (as discussed in Section 5). The upstairs and
downstairs AlphaGUARDs are separated by two floors, yet they show nearly the same general trend in
radon levels (although at lower absolute values for the upstairs AlphaGUARD). This suggests that the
separation between the basement and upstairs HVAC zones and outside air dilution provides little delay
and do not significantly change the pattern of the radon time series.
6-69
-------
Section 6—Results and Discussion:
VOC Concentration Temporal Trends and Relationship to HVAC and Mitigation
Radon fluctuations in indoor air appear to be somewhat related to periods of snow fall or frozen ground.
For example the prominent radon peaks on 1/21 and 1/31 appear to coincide with brief snow events.
Previous studies have suggested that snow cover can block the usual effects of diurnal pressure variations
on radon concentrations (Moses, 1963). However, Moses, Lucas, and Zerbe (1963) concluded that "a
snow cover on the order of four inches or less, probably offers little obstruction to the release of radon
from the soil. Yamazawa (2005) and coworkers,15 however, found that either frozen ground or snow
cover can reduce radon flux from surface soils.
Subslab VOCs (chloroform in Figure 6-67 and PCE in Figure 6-68) appear to be controlled more by the
mitigation on/off cycles than by weather fluctuations. However, WP-3 (purple symbols) does appear to
have highs associated with some periods of snow cover or frozen ground. In Figure 6-66, SGP2-9 (green
symbols) also appears to fluctuate greatly but not associated with any period of snow cover or frozen
ground.
Figures 6-69 and 6-70 show indoor air chloroform and PCE concentration time series, respectively. Both
chloroform and PCE concentrations peak during the mitigation off period, roughly coinciding with two
periods of snow cover and frozen ground (1/21 and 1/31/2013 peaks). It is not clear whether the influence
comes from the snow event, snow cover, frozen ground, or a combination of those factors.
Even during mitigation on cycles, there appears to be a relationship between a rise in chloroform and PCE
concentrations and periods of frozen ground. PCE (Figure 6-68) shows its most dramatic increases in
concentration during periods of snow cover or frozen ground. It is very interesting to note is that VOC
concentrations rose to nearly equivalent levels during periods of snow or frozen ground whether the
mitigation system was on or off. Yet the mitigation system provides very effective protection from radon
even during snow events.
We cannot yet fully explain why the mitigation system provided less protection for VOCs than for radon.
But we note that an extensive literature exists on interactions between organic pollutants (primarily
semi volatile s) and snow/ice packs (see review by Wania et al., 1998). The known effects include
adsorption to liquid layers within the snow/ice, scavenging of VOCs from the atmosphere, as well as
retardation of the air exchange from the surface soil. Snow is modeled as including air filled pore spaces,
liquid water, organic matter, and an air-ice interface (Wania et al., 1998). Aaltonen et al. (2012) found
that certain biogenic volatile organics emanating from soil were trapped by the snow pack. The air
permeability of snow varies greatly among snows of different textures and changes as the snow pack ages
(Conway and Abrahamson, 1984). The State of Alaska vapor intrusion guidance states without citation
that "Caps around a building, such as an asphalt driveway or frozen ground, may reduce volatilization to
outdoor air and increase the concentration of contaminants near the building foundation"16
An EPA review of VOC behavior in soil stated, "Yeates and Nielsen (1987) noted that differences
between winter and summer concentrations occur when the frozen soil acts as a 'lid,' creating higher soil
gas concentrations during winter because release to the atmosphere is inhibited."
Figures 6-71 through 6-74 plot PCE and chloroform in indoor air on the 420 side of the house as
measured by the online GC. The 420 side of the house is unheated, so the conditions there are roughly as
though that half of the duplex were unoccupied. When looking at radon concentrations (in Figures 6-67
through 6-70), it is possible to interpret elevated radon concentrations on the 422 side as being controlled
by increased HVAC activity due to colder temperatures causing the heater to turn on and not snow cover
or frozen ground. This could also be true for the higher VOC concentrations in indoor air (Figures 6-69
15Radon exhalation from a ground surface during a cold snow season
16State of Alaska. Department of Environmental Conservation
6-70
-------
Section 6—Results and Discussion:
VOC Concentration Temporal Trends and Relationship to HVAC and Mitigation
and 6-70). However, the 420 side of the house has no HVAC system and is considerably cooler than the
422 side during the winter (Figure 6-58). Figures 6-71 and 6-72 plot chloroform concentrations of the
first floor and southern basement, respectively. The chloroform data do weakly suggest increased
concentrations during periods of snow or frozen ground, although ambient and indoor air concentrations
are similar, suggesting that outdoor air may be controlling chloroform levels during this period. PCE
concentration data suggest a stronger relationship between increased concentrations and periods of snow
or frozen ground, although, the alignment between events and concentration increases is not perfect
(Figures 6-73 and 6-74).
Each of our data sets (Figures 6-67 through 6-74) suggests a relationship between indoor air quality and
winter weather conditions (snow cover, frozen ground) to a greater or lesser degree. However, those
correlations were not always uniform. Snow events, from a light dusting of less than a quarter of an inch
to a blizzard producing 7 or more inches, seemed to have varying effects. The amount of snow fall and
depth of accumulation did not seem to matter significantly as effects were seen even with no
accumulation, just flurries. These snow flurry effects might suggest that snow events may be a marker for
other meteorological events, like barometric pressure drops and wind shifts, rather than being a direct
physical influence on soil gas concentrations. Alternatively, perhaps the flurries are sufficient to provide
the scavenging followed by melting described by Wania (1998). The winter 2012-2013 snow events and
other meteorological circumstances surrounding them have been summarized in Table 6-4. Some of the
highest observed concentrations for VOCs and radon seemed to occur in conjunction with periods of
frozen ground or when a snow event occurred after the ground was already frozen. But not every period
of frozen ground produced increased concentrations in VOCs and radon at every location (especially for
some of the latter periods of snow and frozen ground in Figures 6-73 and 6-74). Additionally, the peak
radon and VOC concentrations sometimes occurred partway through a period of frozen ground and not at
its beginning (Figures 6-69 and 6-70). Increased radon and VOC concentrations could occur immediately
before, during, or after the snow and frozen ground events. Although the observed behavior at this house
is complex, the literature of air permeability of snow packs and interactions between hydrophobic
organics and snow/ice layers suggests that complex, multifaceted behavior is possible depending on snow
type and its aging after fall. This is consistent with our observation that there appears to be no clear
relationship between snow conditions and indoor air quality at this duplex; there are suggestions of a
relationship in some cases, but not in all. Additional work is needed to understand this issue and the
complex influences of meteorological variables on vapor intrusion processes.
6-71
-------
Section 6—Results and Discussion:
VOC Concentration Temporal Trends and Relationship to HVAC and Mitigation
'Mitigation On
Pa M itif.itied
Gr&wf>£3 freui
Indoor Air Radon and Sub Slab and Soil Gas Port GC CHClj
5 Date 5
Figure 6-67. GC chloroform at subslab and soil gas ports versus radon from stationary
AlphaGUARDs.
6-72
-------
Section 6—Results and Discussion:
VOC Concentration Temporal Trends and Relationship to HVAC and Mitigation
Indoor Air Radon and Sub Stab and Soil Gas Port GC PCE
^ Downstairs fiascm
0 Up-stMs Radon
0 O S0FI.9»C£
£ 0 WMMS
£T 0 SQP11-13KC
5'
3P SGP9-6PCI
3
— # $GP&-9PCi
o
<: A SSP-4 PCE
r— w
Umiif»sien Of*
Ptmvt M
-------
Section 6—Results and Discussion:
VOC Concentration Temporal Trends and Relationship to HVAC and Mitigation
Indoor Air Radon and Indoor Air GC CHCk
3
m
2? 2
K-
N*
M" K-
^ SJ
I-*
§
7?
M-
£
lu
wOate w
hf
yj
-Q 16
W 10
<0 DovwutaifiHn
• UpStJtrjRft
10
9 c O AflibiWtTO
- s
7 jj 9 4?0 First PC£
5 £ O 422 Ftrst PCE
* a
3 5" o 43BB4KSPCE
j G
! > * 422Ba»?KE
o
_j — Mrtiga;«jnt)it
- — p«sj«« MujgatKMt
—— Biif id and miliar
SUKWH
• Brief Sr-sw
Figure 6-69. GC chloroform in indoor air versus radon from stationary AlphaGUARDs.
6-74
-------
Section 6—Results and Discussion:
VOC Concentration Temporal Trends and Relationship to HVAC and Mitigation
Indoor Air Radon and Indoor Air GC PCE
9 Downstairs An
# Upstairs Rn
9 C ° Ambient PCS
e-g
1 & • 420 First PCE
6 %¦
5 a 0 422flfltPCE
* a
o «(JBawSPCJ
i n
l> * iU BawSPCE
0 —
*1 Miration On
— Paiuve M i:;ja lion
=—• $fa* aref and minor
snows
Ground frw«i
• Brief Snow
Date
Figure 6-70. GC PCE in indoor air versus radon from stationary AlphaGUARDs.
6-75
-------
Section 6—Results and Discussion:
VOC Concentration Temporal Trends and Relationship to HVAC and Mitigation
Indoor Air GC CHCI3 at 420 First Floor
10
E
"So
3
c
o
o.i
• .*
• •
:
••
oo
Ambient
• 420 First
^^¦Mitigation On
Passive Mitigation
Blizzard/minor snows
round Frozen
• Brief Snow
Date
Figure 6-71. GC chloroform concentrations in indoor air, 420 first floor.
Indoor Air GC CHCI3 at 420 Basement South
10
o.i
• I
ft
- y
oo
o •
o
O Ambient
O 420 Base S
^¦Mitigation On
^¦Passive Mitigation
^¦Blizzard/minor snows
^¦Ground Frozen
O Brief Snow
o
Date
Figure 6-72. GC chloroform concentrations in indoor air, 420 basement south.
6-76
-------
Section 6—Results and Discussion:
VOC Concentration Temporal Trends and Relationship to HVAC and Mitigation
Indoor Air GC PCE at 420 First
100
ro
E
•A
0.01
Ambient
420 First
¦Mitigation On
Passive Mitigation
Blizzard/minor snows
•Ground Frozen
Brief Snow
Date
Figure 6-73. GC PCE concentrations in indoor air, 420 first floor.
Indoor Air GC PCE at 420 Basement South
• Ambient
• 420BaseS
^^"Mitigation On
Passive Mitigation
Blizzard/minor snows
^^"Ground Frozen
• BriefSnow
*^ooo
M
Figure 6-74. GC PCE concentrations in indoor air, 420 basement south.
6-77
-------
Section 6—Results and Discussion:
VOC Concentration Temporal Trends and Relationship to HVAC and Mitigation
Table 6-4. Summary of Meteorological Data During the 2012/2013 Snow and Ice Events
Time Surrounding
Storm
Average Wind
Speeds MPH
Indoor Air
PCE Peak?
Present?
Indoor Air
Chloroform Peak?
Present?
Date and
Estimated Time
Mitigation
Status
Predominant
Wind Direction
Radon Peak
422 Basement?
Snow Event
Temperature °F
12/20/12 18:00
Minor snow (~1")
On
wsw
12-20
422 basement
ambient
Ambient
Tiny
30-40
12/26/12 6:00
Blizzard and minor
snows
On
NE
10-12
420 first floor
422 basement
Weak
Ambient
Tiny
-30
1/21/13 7:00
Brief snow
Off
WNW
7-10
422 basement
422 first floor
420 basement
422 first floor
422 basement
Strong
16-20
1/25/13 7:00
Minor snow (1")
Off
SE
1-5
422 basement
Moderate
19-21
1/31/13 7:00
Minor snow (1/4")
Off
WNW-W
7-15
422 basement
422 first floor
422 First Floor
422 Basement
Strong
22-24
2/15/13 10:00
Brief snow
On
WNW
4-10
422 basement
420 basement
422 first floor
Ambient
422 basement
None
32-34
2/19/13 8:00
Brief snow
On
WNW
10-17
Ambient
422 basement
420 basement
Ambient
Tiny
24-28
2/22/13 1:00
Ice
On
E
4-10
422 basement
420 basement
28-32
2/27/13 12:30
Brief snow
On
WSW
8-10
None
34-35
3/1/13 1:00
Brief snow
On
WNW
2-4
Weak
422 first floor
422 first floor
422 Basement
31-33
3/6/13 1:00
Brief snow
On
WNW
2-4
422 basement
422 first floor
Tiny
28-30
3/12/13 23:00
Brief snow
On
WNW
6-12
422 basement
422 first floor
420 basement
422 first floor
422 basement
30-34
6-78
-------
Section 7—Results and Discussion: Establishing the Relationship between VOCs and Radon in
Subslab/Subsurface Soil Gas and Indoor Air
Table of Contents
7.0 Results and Discussion: Establishing the Relationship between VOCs and Radon in
Subslab/Subsurface Soil Gas and Indoor Air 7-1
7.1 Previously Reported Tests of Radon as a Semiquantitative VOC Tracer 7-1
7.2 Understanding the Performance of the Radon Tracer During Mitigation Testing 7-1
7.3 Attenuation Factors Derived Using the Radon Tracer 7-4
7.4 Radon Tracer in Statistical Time Series Analysis 7-4
List of Figures
7-1. Comparison of mean concentrations among wall and subslab ports under different
mitigation conditions (heat on data only): radon (top), PCE (middle), and chloroform
(bottom) 7-2
7-2. Long-term trends in radon and VOCs with shading showing mitigation status during
Phase 2 7-5
7-i
-------
Section 7—Results and Discussion: Establishing the Relationship between VOCs and Radon in
Subslab/Subsurface Soil Gas and Indoor Air
7-ii
-------
Section 7—Results and Discussion: Establishing the Relationship between VOCs and Radon in
Subslab/Subsurface Soil Gas and Indoor Air
7.0 Results and Discussion: Establishing the Relationship between
VOCs and Radon in Subslab/Subsurface Soil Gas and Indoor Air
7.1 Previously Reported Tests of Radon as a Semiquantitative VOC Tracer
In Section 2 of this report, we discuss the conditions under which radon is expected to be a useful
semiquantitative tracer of vapor intrusion. The most difficult to satisfy condition at this study site is the
need for radon and VOCs to be similarly distributed in subsurface air just outside the building envelope
and thus for entry routes to be similar.
Our previous report (U.S. EPA, 2012a) contained an extensive discussion of radon-to-VOC correlation in
our 2011-2012 data set leading to the following conclusions:
¦ Radon concentrations in subslab air were much steadier than VOC concentrations, presumably
because the shallow soils themselves were the dominant source of radon and VOCs originated at
a greater depth/distance.
¦ Radon concentrations in indoor air varied over approximately an order of magnitude short-term—
apparently greater short-term variation than was observed for VOCs.
¦ However, with a 1-week integration time, radon had less seasonal variability than VOCs in indoor
air.
¦ Statistical cross-correlation testing found that radon and VOCs were positively cross-correlated at
most indoor air sampling locations (5% critical level). Some cross-correlations of radon and
VOCs were observed at soil gas ports, but these cross-correlations were less consistent/strong.
¦ Radon was clearly a marker for soil gas in this system. Thus, radon was a useful aid to VOC data
interpretation. But long-term radon exposure would not have completely predicted VOC exposure
in this house over all time scales.
7.2 Understanding the Performance of the Radon Tracer During Mitigation
Testing
Additional data on the correlation between radon and VOCs were gathered in 2012-2013, during a period
of time when the mitigation system was being tested by being turned on and off. In Section 5 of this
report, we show that radon does not seem to be a sensitive indicator of mitigation system performance vs.
PCE. The performance of the mitigation system was substantially better and more consistent for radon
than it was for VOCs. We observed a 91 to 93% reduction in indoor radon. Table 5-11 showed that in all
cases the concentration of radon was reduced by the SSD system operation in the subslab sampling ports,
wall ports, and shallowest interior soil gas ports. Table 5-12 showed that this effect is, on average, about
a 60% reduction in the subslab sampling ports and 80% in the wall ports. Comparing this result with the
reductions observed in indoor air suggests that the SSD system is operating at this duplex to reduce radon
in indoor air through two mechanisms—both by diluting the air beneath the slab with lower concentration
air (presumably atmospheric) and reversing the pressure differential across the slab. Our discussion in
Section 5 focused on mitigation performance and causes for that performance. Our discussion in this
section focuses on the performance of the radon tracer technique and why the tracer was or was not
useful.
As shown in Figure 7-1, prior to mitigation (blue bars) the distribution of radon among various subslab
and wall ports was more continuous/smoothly varying than it was for PCE and chloroform.
Concentrations of radon were generally reduced or only slightly increased under the mitigation on
condition (purple bars). PCE and chloroform concentrations increased dramatically at some locations with
7-1
-------
Section 7— Results and Discussion: Establishing the Relationship between VOCs and Radon in
Sub slab/Sub surface Soil Gas and Indoor Air
Radon Concentration in Su&slab and Wall Ports by Mitigation Status
HO I.-3® l .-'3D LUWO LHD jorv
• tpCAJ
PCE Conctiitration In Substeb and W.ill Ports, By Mitigation Status
L-W KC
MO 403 MO *00 MO MB MO LOCO
«%•« KI ItXrCmrtnUlo* W**l
Chloroform Concentrations fcn Subslab and Wall Ports
tfi 100 119 HQ 3W KC lifi UtXl 4V! VE'
Uwm ((mvrtalian
* en ¦ A
Figure 7-1. Comparison of mean concentrations among wall and subslab ports under different
mitigation conditions (heat on data only): radon (top), PCE (middle), and chloroform
(bottom).
7-2
-------
Section 7—Results and Discussion: Establishing the Relationship between VOCs and Radon in
Subslab/Subsurface Soil Gas and Indoor Air
the mitigation on. Under post-mitigation conditions, there were still more dramatic differences in
concentration between ports for PCE than for radon.
The flow of soil gas into the building is expected to be very nonuniform with some areas of soil gas just
outside the building envelope near major cracks/gaps dominating the total flow into the building.
Mitigation may change the particular locations from which the predominant flow is coming. Thus, the
greater homogeneity of the radon concentrations as compared to VOC concentrations can explain why the
percentage improvement in indoor air concentrations is not the same for radon and VOCs.
Similarly, Figure 6-70 showed that several day-long peaks in PCE in indoor air were observed under
mitigation on conditions that were not observed in indoor air radon (neither with the professional grade
AlphaGUARD instrument, integrated electret measurements, nor Safety Siren Consumer grade detector).
Those peaks were also much less prominent for chloroform. We discuss in Section 5 that those PCE
peaks were not necessarily accompanied by a loss of differential pressure control in the center of the
subslab areas where continuous differential pressure observations were being made. A likely explanation
for those two sets of observations is that differential pressure control was lost in only a portion of the
building envelope, but high PCE concentrations were present in the soil gas adjacent to those portions of
the building envelope where pressure control was lost. Although radon would have been expected to have
also been present in the soil gas, the greater variability in soil gas PCE concentrations could cause a more
dramatic PCE peak relative to baseline. The lack of a distinct peak in the radon concentration plots during
this time period provides useful mechanistic information; it suggests that during these PCE peaks the
mitigation system is still effective in substantially limiting the total amount of soil gas entering the
structure.
Fundamentally, we can understand the potential for radon and VOCs to have different behavior in
mitigation systems in terms of travel time of the gas flowing into the extraction points (and subslab area
more generally). We can envision the subslab depressurization system for some purposes as a tiny soil
vapor extraction (SVE) system, although we do not advocate using SSD for mass removal. Gas travel
times to extraction wells in SVE systems can be much larger than the radon half-life of 3.8 days (Falta,
2006). Thus, because the sources of radon and VOCs are generally both diffusion-controlled processes
that originate from sources in the fine soil particles, we can understand that as the SSD continues to
operate and draws in gas with long travel times, the temporal profiles of radon and VOC concentrations
could differ.
As discussed above in our premitigation studies, the relative amount of variability of radon and VOCs
was compared as follows:
¦ Radon concentrations in subslab air were much steadier than VOC concentrations.
¦ Radon concentrations in indoor air varied over approximately an order of magnitude short-term—
apparently greater short-term variation than was observed for VOCs.
¦ However, with a 1-week integration time, radon had less seasonal variability than VOCs in indoor
air.
7-3
-------
Section 7—Results and Discussion: Establishing the Relationship between VOCs and Radon in
Subslab/Subsurface Soil Gas and Indoor Air
Figure 7-2 illustrates long-term trends in VOC and radon concentrations both before the mitigation
testing and during the mitigation testing. Note that radon concentrations respond very quickly to
mitigation system on/off cycles, forming essentially a "square wave" as the mitigation system is shut on
and off. The PCE and chloroform concentration profiles trace a less responsive-looking pattern, with a
minimum occurring during the week after the mitigation system was turned on and maximums, higher
than before mitigation, when the system was shut down in December and in early January. This differs
from the typical pattern seen with soil vapor extraction systems in which extracted concentrations
generally decline with time as the system is operated. For example, Thomson and Flynn (2000) describe a
first extraction stage of high concentration soil gas that is assumed to be in equilibrium with NAPL and a
final stage where extraction is mass transfer limited by gas and aqueous phase diffusion into preferential
gas flow pathways. This may reflect that SVE wells are usually intentionally screened in areas of high soil
gas concentrations, while the SSD extraction points in this case are more distant from the source zone.
Also note that Figure 7-2 shows the behavior of each contaminant in indoor air when compared with
outside ambient air concentrations measured at the duplex. The chemicals fall into two main groups: those
whose indoor air concentrations follow ambient air concentrations very closely (benzene, toluene, and
hexane) and those with indoor concentrations mostly above ambient levels (radon, chloroform, and PCE).
Indoor air concentrations of benzene, toluene, and hexane are clearly derived from outdoor air sources,
probably influenced by the traffic on the busy road south of the duplex. Radon, chloroform, and PCE are
well above ambient levels, although the ambient levels clearly influence chloroform and PCE
concentrations over time to a greater degree than can be observed for radon, probably because outdoor air
concentrations are closer to indoor air concentrations for chloroform and PCE than for radon. Note that
TCE behaves similarly to chloroform and PCE at the higher concentrations observed early in the study
but mirrors the outdoor air concentrations when TCE concentrations decrease to close to ambient later in
the study, although there is a small increase over background when the mitigation system was first shut
off, which could reflect the same causes as the similar increases observed for PCE and chloroform.
7.3 Attenuation Factors Derived Using the Radon Tracer
As shown in Section 8, radon predicts reasonably well the attenuation factor from wall port to indoor air
and from subslab to indoor air. The agreement is much less good for deeper soil gas to indoor air. Thus,
radon serves as a reasonable tracer for the portion of the vapor intrusion process across the slab or
building envelope but is less effective as a tracer for VOC attenuation from deeper media. The reason for
that is also very clear in Section 8; there is substantial reduction in chloroform concentration and PCE
concentration from 13 to 6 ft bis. However, radon concentrations are quite similar at 6, 9 and 13 ft depths.
Thus, radon can be seen as a reasonable surrogate for the portion of the attenuation factor that describes
attenuation across the building envelope (AFbidg as defined in U.S. EPA, 2012c). As might be expected,
radon does a much poorer job of describing the portion of the attenuation factor that is due to subsurface
migration processes (AFSOii).
7.4 Radon Tracer in Statistical Time Series Analysis
In Section 10, we present time series analyses for both radon and VOCs in which we examine potential
correlations to a wide variety of meteorological and soil conditions. For the purposes of the current
section, two primary lessons stand out:
¦ In most of the analyses including mitigation, mitigation was shown to have a significant
beneficial effect for radon at either the 1% or 5% level of significance. In contrast, in the VOC
7-4
-------
Section 7— Results and Discussion: Establishing the Relationship between VOCs and Radon in
Sub slab/Sub surface Soil Gas and Indoor Air
Phase 1
^vvvvvva ^vvv^
Mitigation
Not Yet
Installed
Off
On
Location
420BaseN
420BaseS
-420First
—422BaseN
M422BaseS
— 422First
~422Office
Outside
Figure 7-2. Long-term trends in radon and VOCs with shading showing mitigation status during
Phase 2.
time series results, although mitigation generally appeared beneficial, it was not statistically
significant. This agrees with the other modes of mitigation analysis discussed previously in this
section and in Section 5.
7-5
-------
Section 7—Results and Discussion: Establishing the Relationship between VOCs and Radon in
Subslab/Subsurface Soil Gas and Indoor Air
¦ In the radon time series analysis, many meteorological variables (often a majority of the variables
tested) were found to be statistically correlated to indoor radon concentrations. In contrast, for the
VOCs, a smaller proportion of the variables tested rose to statistical significance. It is possible
that this is an artifact of the sample size. Alternatively, this may be a more formal mathematical
confirmation of two phenomena discussed earlier:
— As described in Section 6.4, radon responds almost immediately to both "mitigation on"
and "mitigation off cycles. After decreasing to near ambient air background (a range of
0 to 2.5 pCi/L) with the mitigation system on, it quickly returns to more elevated levels
and a more dramatic range (4 to 15 pCi/L) when the mitigation system is turned off.
VOCs tended to showed a slower and more variable response to mitigation than radon, at
least during the initial operation of the system described in this report.
— Radon's concentration in indoor air may be more simply defined as a first approximation
based on the amount of transport across the building envelope, while VOCs have a more
heterogeneous spatial and temporal patterns of concentration in soil gas. Therefore,
radon's behavior may be easier to predict using statistical models with a limited number
of variables (in this case, one variable at a time was tested, plus the mitigation on/off
variable).
In essence, radon responded more quickly, cleanly, and repeatedly to changes in subslab depressurization
because the radon entering the house is generated very near the house because radon's relatively short
half-life (3.8 days) does not allow it to travel far from its point of generation before decaying. VOCs, with
their much longer half-lives (generally over a year or longer), have a much longer transport path than
radon, making them more subject to processes along this path and leading to a much wider source area
than is typical for radon, along with a greater medium to long-term (daily to weekly) variability both
during and before mitigation with subslab depressurization.
7-6
-------
Section 8—Results and Discussion: Attenuation of Soil Gas VOCs and Radon
Table of Contents
8.0 Results and Discussion: Attenuation of Soil Gas VOCs and Radon 8-1
8.1 Calculation of Daily Attenuation Factors without Mitigation 8-1
8.1.1 Daily Radon Attenuation Factors without Mitigation 8-5
8.1.2 Daily VOC Attenuation Factors without Mitigation 8-5
8.2 Subslab and Soil Gas to Indoor Air Weekly Attenuation Factors 8-8
8.2.1 Radon Weekly Attenuation Factors 8-8
8.2.2 VOC Weekly Attenuation Factors 8-11
8.3 Effect of Mitigation 8-19
8.3.1 Subslab to Indoor Air Daily Attenuation 8-20
8.3.2 Subslab and Soil Gas to Indoor Air Weekly Attenuation 8-23
List of Figures
8-1. AlphaGUARD concentrations and daily radon attenuation factors (number of samples
indicated below each bar graph pair). Sample pairs represented indicated along x-axis.
SSP-7 is on the 420 side of the building and so may not be representative of subslab
attenuation as the other sample pairs. Sampling period = January 2011-May 2013,
mitigation system off. 8-3
8-2. Daily indoor air and soil gas PCE concentrations used in attenuation factor calculations
(number of samples with heat on and off are indicated below each bar graph pair) 8-4
8-3. Daily PCE attenuation factors (number of samples indicated below each bar graph pair) 8-5
8-4. Daily chloroform concentrations used in attenuation factor calculations (number of
samples indicated below each bar graph pair) 8-6
8-5. Daily chloroform attenuation factors (number of samples indicated below each bar graph
pair) 8-7
8-6. Electret radon concentrations used in weekly attenuation factor calculations. Number of
measurements indicated below each box and whiskers pair 8-9
8-7. AlphaGUARD soil gas radon concentrations (pCi/L) used in weekly attenuation factor
calculations. Number of measurements noted below each box and whiskers pair. Note
that these concentrations represent averages across all samples taken at the same depth
and time (e.g., subslab samples are averages across all subslab points) 8-10
8-8. Weekly radon attenuation factors (number of samples indicated below each bar graph
pair). Note that these attenuation factors were calculated from indoor air and soil gas
concentrations averaged at the same depth and time (as described in Figure 8-7) 8-11
8-9. Radiello PCE concentrations used in attenuation factor calculations. Numbers at the
bottom of each column indicate the number of available readings for unheated (left) and
heated (right) conditions 8-12
8-10. TO-17 PCE concentrations used in attenuation factor calculations 8-13
8-11. Weekly PCE attenuation factors (numbers of samples indicated by numbers beneath each
bar graph pair) 8-14
8-i
-------
Section 8—Results and Discussion: Attenuation of Soil Gas VOCs and Radon
8-12. Radiello chloroform concentrations used as numerators for weekly attenuation factor
calculations 8-15
8-13. TO-17 chloroform concentrations used as denominators for weekly attenuation factor
calculations 8-16
8-14. Weekly chloroform attenuation factors (number of samples indicated below each bar
graph pair) 8-17
8-15. Attenuation of PCE, chloroform, and radon juxtaposed 8-18
8-16. Mitigation status and schedule 8-19
8-17. Radon indoor air and soil gas concentrations used in daily attenuation factor calculations 8-20
8-18. Daily radon attenuation factors 8-21
8-19. Daily PCE attenuation factors 8-22
8-20. Daily chloroform attenuation factors 8-23
8-21. Electret indoor radon concentrations 8-24
8-22. AlphaGUARD subsurface radon concentrations (pCi/L) 8-25
8-23. Weekly radon attenuation 8-26
8-24. Radiello indoor air PCE concentrations 8-27
8-25. TO-17 soil gas PCE concentrations 8-28
8-26. Weekly PCE attenuation 8-29
8-27. Radiello weekly indoor air chloroform concentrations 8-30
8-28. TO-17 soil gas chloroform concentrations 8-31
8-29. Weekly chloroform attenuation factors 8-32
8-30. Attenuation of PCE, chloroform, and radon juxtaposed 8-33
8-ii
-------
Section 8—Results and Discussion: Attenuation of Soil Gas VOCs and Radon
8.0 Results and Discussion: Attenuation of Soil Gas VOCs and
Radon
In this section, we explore the relationship between radon and VOC levels in indoor air and
concentrations measured in subslab and deeper soil gas, which is usually portrayed through the vapor
intrusion attenuation factor. As described in Section 2, the vapor intrusion attenuation factor is the indoor
air concentration divided by a subsurface soil gas concentration at the same time and location. For
example, if subslab soil gas concentrations are 100 times greater in the indoor air concentration measured
at the same time, the subslab has a 100-fold attenuation and the attenuation factor would be 0.01.
The previous report on the first phase of this project (U.S. EPA, 2012a) described and compared the
temporal variability of attenuation factors based on individual paired points and whole side averages for
subslab and deeper soil gas attenuation factors for VOCs and radon. This report updates the U.S. EPA
(2012a) data sets and compares radon and VOC attenuation factors with and without operation of the
subslab depressurization mitigation system discussed in Section 3.
8.1 Calculation of Dally Attenuation Factors without Mitigation
Daily attenuation factors were calculated for radon and VOCs for the entire project period, including the
phases before and after the mitigation system was installed on 10/15/2012, through the end of this
reporting period in early May 2013. To calculate the daily attenuation factors, we averaged subsurface
and indoor air concentrations measured on the same day and divided these indoor air concentration
averages for each day by the nearest (temporally and spatially) subsurface concentrations available. This
would be within a 24-hour range or within 24 hours from each sample point (i.e., no daily attenuation
sample pairs are more than 24 hours apart).
In terms of the lag time between subsurface and indoor air (i.e., how long it takes for a radon or VOC
molecule to move from the surface into indoor air), we believe it is reasonable for a radon or VOC
molecule to travel from the subslab or outside the wall into the basement indoor air in 24 hours or less
and that a daily resolution for attenuation factors is therefore informative. The most interesting daily
attenuation factors are the attenuation from subslab ports and wall ports to basement indoor air. The time
of travel from subslab or wall exterior to basement indoor air should be relatively quick, so "same day"
averaging is appropriate. Although we expect that attenuation from deeper soil gas will take longer,
because the deeper vadose zone is very coarse grained, we think that advective flow is likely operative
and assumed that it would take longer than 24 hours for VOCs in deep soil gas to reach indoor air in the
building.
As described in Section 3, subsurface concentration measurements were taken from the soil gas and
subslab sampling ports using a portable AlphaGUARD sampler for radon and TO-17 active sorbent tubes
for VOCs. These measurements serve as the soil gas concentrations (denominator) for all radon or VOC
attenuation factor calculations, and although they are grab samples, we assumed they are representative of
the average soil gas concentrations for the day or week during which they were taken. In other words, for
both radon and VOCs, we assumed that the subsurface soil gas concentrations, which were TO-17 grab
samples taken approximately weekly, remain relatively constant.
Radon was measured in indoor air using stationary AlphaGUARD monitors as well as electret sampler
badges. The more continuous indoor AlphaGUARD measurements allow the calculation of a daily indoor
air radon concentration for the daily attenuation factors while the electret samplers provide a weekly
physical average attenuation factor that could be used directly. Similarly, VOCs were measured fairly
continuously (every hour or so) in indoor air during the periods of on-site GC operation, and these
8-1
-------
Section 8—Results and Discussion: Attenuation of Soil Gas VOCs and Radon
measurements were used to calculate daily average concentrations for the periods of online GC operation.
The weekly indoor air Radiello sampler concentrations, which are available for the entire project period,
were used to calculate the weekly VOC attenuation factors.
8.1.1 Daily Radon Attenuation Factors without Mitigation
Indoor AlphaGUARD measurements were only taken on the 422 side of the duplex, so daily radon
attenuation factors are only computed for pairs of locations on the 422 side. This allowed us to observe
the effect of heating because only the 422 side was heated. Figure 8-1 compares the distributions of daily
radon attenuation factors with the heat on and off and the mitigation system either off or not installed and
shows the ranges of the indoor air and subslab concentrations used to calculate them. In general,
attenuation is slightly lower when the heat is on, as would be expected given increased stack effect for a
building when the heat is on, which tends to increase indoor VOC concentrations. Also as expected, the
wall port shows much less attenuation (i.e., higher attenuation factors) than the two subslab ports on the
422 side of the building (SSP-2 and SSP-4) because of the lower soil gas (denominator) concentrations in
the wall port due to its stronger connections with the atmosphere. SSP-2 and SSP-4 attenuation factors are
similar because of similar subslab concentrations. SSP-7, which is on the 420 side of the duplex, is
included in the figure and calculations, but because it represents 420 subslab data with 422 indoor air
data, it is not as representative of subslab attenuation as SSP-2 and SSP-4.
8.1.2 Daily VOC Attenuation Factors without Mitigation
The on-site GC data allow daily subslab to indoor air attenuation factors with very high time coverage for
the three periods when the on-site GC was in operation at the site, two periods before and one period after
the mitigation system was installed (see Section 3). The on-site GC has 2 basement sampling points: the
south side of the 420 and 422 basements. As with the radon data, the data also include three subslab
locations: SSP-7 on the 420 side and SSP-2 and SSP-4 on the 422 side. Because of differences in
instrument calibration (see Section 4), the two periods prior to mitigation are treated as disjointed data,
but the two periods are juxtaposed in the following figures because the heat was off for the entirety of the
first period of GC monitoring and on throughout the second period.17
Figure 8-2 provides the range of daily indoor air and soil gas concentrations used to calculate the PCE
attenuation factors and Figure 8-3 provides the range of those attenuation factors for the individual
indoor air/subslab pairs calculated for PCE. Note that for PCE in indoor air, only the 422 basement and
SSP-4 show the increase in PCE concentrations with the heat on, which is as expected because SSP-7 is
on the unheated 420 side of the building and WP-3 is shallow and likely affected by atmospheric air. On
the 420 side, both the basement and SSP-7 show a decrease in VOC levels in the heated (winter) months,
suggesting that perhaps lower temperatures decrease vapor intrusion driving forces. Also, SSP-2, on the
north side of 422, was not expected to show the measured decrease in PCE concentration with the heat on,
suggesting that the subslab area sampled by SSP-2 is not strongly connected to the 422 indoor air
environment and/or there is atmospheric air diluting the subslab concentrations at this point.
17Note that five zero PCE concentration values for the 420 basement were replaced with 0.005 (the smallest positive concentration value at that
location). While the zero values would still yield a finite attenuation value (0), they were replaced with 0.005 so they can be plotted on log-
scaled figures.
8-2
-------
Section 8—Results and Discussion: Attenuation of Soil Gas VOCs and Radon
10 =
1 =
1000 =
o
Cl
o 100e
"O
CO
a:
0.11
0.011
0.001 =
Daijy Radon Attenuation
Indoor AlphaGUARD Concentration (pCi/L)
fct3
7-5
27-32
*24-27
Portable AlphaGUARD Concentration (pCi/L)
7-5
27-32
—*—
'24-27*
Attenuation
4-^
I
&
J^5_
%
27-32
24-27
&
2
0>
&
&
0>
&
&
Heating
• Off
^ On
Figure 8-1. AlphaGUARD concentrations and daily radon attenuation factors (number of
samples indicated below each bar graph pair). Sample pairs represented indicated
along x-axis. Sampling period = January 2011-May 2013, mitigation system off.
As far as attenuation is concerned, the 422 basement / SSP-4 attenuation factor shows a small decrease
(i.e., more attenuation) with heating because the SSP-4 concentration increased more than the 422
basement indoor air concentration with heating on. Also, SSP-2 and WP-3 show very low attenuation
8-3
-------
Section 8—Results and Discussion: Attenuation of Soil Gas VOCs and Radon
(with an attenuation factor around 1), which would be expected if the soil gas at these ports was
connected to the atmosphere.
Similar patterns can be seen for chloroform in Figures 8-4 and 8-5 except that WP-3 does not show the
decrease in chloroform concentrations and increase in attenuation factors with heating as seen for PCE.
As with PCE, the anomalously high attenuation factors for WP-3 and (especially) SSP-2 suggest
atmospheric leakage and that the VOC levels in indoor air is not greatly influenced by soil gas from these
locations.
Daily PCE Concentrations
100
10:
;+ ::
i
to
£
CD
ID
O
CL
+1
0.1 =
:
001
46-69 66-69 63-69 66-66 46-69 63-69
Heating
*Off
$ On
& 4?«
Figure 8-2. Daily indoor air and soil gas PCE concentrations used in attenuation factor
calculations (number of samples with heat on and off are indicated below each bar
graph pair).
8-4
-------
Section 8—Results and Discussion: Attenuation of Soil Gas VOCs and Radon
0.1 :
C
o
^ 0.01
3
C
-------
Section 8—Results and Discussion: Attenuation of Soil Gas VOCs and Radon
100
E
CD
—*
E
o
4—
o
o
rod
o 1-
0.1:
Daily Chloroform Concentrations
*
¦
j 51-51 67-69 62-62 66-66 51-51 67-69
Heating
+ Off
^ On
Figure 8-4. Daily chloroform concentrations used in attenuation factor calculations (number of
samples indicated below each bar graph pair).
8-6
-------
Section 8—Results and Discussion: Attenuation of Soil Gas VOCs and Radon
1 :
o
'n 0.1
c
- aft?
3
Heating
^Off
E^3 On
Figure 8-5. Daily chloroform soil gas to indoor air attenuation factors (number of samples
indicated below each bar graph pair).
8-7
-------
Section 8—Results and Discussion: Attenuation of Soil Gas VOCs and Radon
8.2 Subslab and Soil Gas to Indoor Air Weekly Attenuation Factors
8.2.1 Radon Weekly Attenuation Factors
Week-long attenuation factors for radon were calculated by averaging all subsurface radon measurements
that coincide with an indoor air electret badge exposure period and pairing that average to the electret
measurement. The attenuation between deeper soil gas and indoor air is interesting on a weekly scale
because the soil gas has time to travel from deep in the soil to the indoor air. Given the duration of the
measurement, it also makes sense to average soil gas samples from the same depth. Similarly, basement
samples are lumped together, so the south (420BaseS) and north (420BaseN) samples are all classified as
420 basement (420Base).
Because of the way electrets are read, there were some small negative and 0 values in the indoor air
concentrations. For the sake of calculating attenuation these values were replaced with the smallest non-
negative value in the data series.
As with the daily attenuation figures, the weekly figures depict the data segregated by heat on and off, but
only the 422 side was heated. Figure 8-6 is an overview of all indoor air electret concentrations that were
used to calculate attenuation factors.
As with the daily data, indoor air concentrations are higher during the heating season for the heated 422
side of the duplex, while radon concentrations actually decreased during the colder months on the
unheated 420 side. Figure 8-7 shows the subsurface soil gas radon concentrations measured with the
portable AlphaGUARD instrument. In this case, the shallower soil gas samples (subslab, wall port, 3.5 ft,
and 6 ft) show a decrease in radon concentrations while the deeper soil gas samples (9 ft, 13 ft, and 16.5
ft) show an increase during the heating season. This observation needs additional research because it is
opposite of what might be expected, at least for the 422 side of the house with the increasing stack effect
during the heating season. The same is true with respect to the observation that the difference between
periods of heat on and off in subslab concentrations mirrors the 420 side indoor air rather than the 422
side indoor air. This was not expected as the stack effect created by heating should affect the subslab air.
It is also not immediately obvious that the radon concentrations in soil gas fall off with depth, but the
decrease is greater in the summer months.
Figure 8-8 shows the attenuation factors for the weekly radon samples. Weekly attenuation factors for the
420 and 422 basements are similar, ranging from 0.005 to 0.01 for the subslab and deeper soil gas, and
without any obvious trends with depth. Wall port attenuation factors are higher than the others, but these
results are difficult to evaluate because of the multiple influences (soil gas, atmospheric, building
pressures) on these shallow samples.
-------
Section 8—Results and Discussion: Attenuation of Soil Gas VOCs and Radon
0.1
Electret Concentrations
10=
o
CL
c
o
"O
CO
oc
Az
w
-
B
~
~
Heating
*Off
$On
41-71
21-37
42-74
21-37
&
0
0>
&
t#
Figure 8-6. Electret radon indoor air concentrations used in weekly attenuation factor
calculations. Number of measurements indicated below each box and whiskers pair.
8-9
-------
Section 8—Results and Discussion: Attenuation of Soil Gas VOCs and Radon
AlphaGUARD Soil Gas Radon Concentrations
-1000-
o
Q_
c
o
"D
CD
a:
100
*1
~
• 102-165 113-207 101-118
107-183 95-207 113-182 18-23
Heating
¦ Off
$ On
,o^ .0,$
.. G^"
k . ^
^ J* . *?>
cgr g>» .. <& q£>-
& # # 0^°
Figure 8-7. AlphaGUARD soil gas radon concentrations (pCi/L) used in weekly attenuation factor
calculations. Number of measurements noted below each box and whiskers pair.
Note that these concentrations represent averages across all samples taken at the
same depth and time (e.g., subslab samples are averages across all subslab points).
8-10
-------
Section 8—Results and Discussion: Attenuation of Soil Gas VOCs and Radon
0,1
0.01
0.0011
1 e-04 -
0.1 i
0.01
c 0.001
.2 1 e-04:
TO
3
C
aj -i
e 0.11
< 0.01j
0.001
1 e-04 i
0.1
0.01
0 001 j
1 e-04:
Weekly Radon Atteriaution
420Base
36-G2 34-56 32-70 38-70 38-61 34-40 6-8
420First
i
ft
18-31 17-28 16-35 19-35 19-31 17-20 3-4
422Base
+SI
4
¦
36-62 34-56 32-70 38-70 38-62 34-40 6-8
422First
:--r
z—l K_L_ •—]- =-4-
18-31 17-28 16-35 10-35 19-31 17-20 3-4
1 1 1 1 1 1 1
. V .1
^ ^«rf> , - ,,
6
<3
Denominator Location
Heating
^Off~
On
Figure 8-8. Weekly radon attenuation factors (number of samples indicated below each bar
graph pair). Note that these attenuation factors were calculated from indoor air and
soil gas concentrations averaged at the same depth and time (as described in Figure
8-7).
8.2.2 VOC Weekly Attenuation Factors
The attenuation factors calculated from the Radiello/TO-17 data for VOCs are comparable to the
electret/portable AlphaGUARD attenuation factors calculated for radon in that they represent weekly
indoor air averages against grab subsurface samples taken about once per week. Figure 8-9 shows the
weekly Radiello indoor air sample concentrations used in these calculations, which show a drop off in
8-11
-------
Section 8—Results and Discussion: Attenuation of Soil Gas VOCs and Radon
winter concentrations on the unheated 420 side of the building, which was statistically significant for the
basement (p = 0.04, simple T-test, null hypothesis = there is no decrease in concentration with mitigation)
but not for the first floor (p = 0.04). This could reflect a weaker stack effect in this colder, unheated side
of the building. The heated 422 side showed the expected increase in concentrations during the heated
period, which was statistically significant for both the basement (p = 0.0004) and the first floor (p = 0.01).
Both sides exhibited the higher basement VOC concentrations (than first floor) that would be expected at
a vapor intrusion site.
Radiello PCE Concentrations
io-
CO
£
O)
LU
O
~_
1 -
0,1
Heating
*Off
£$3 On
38-75
18-34
36-71
18-37
&
&
tP
Figure 8-9. Radiello PCE indoor air concentrations used in attenuation factor calculations.
Numbers at the bottom of each column indicate the number of available readings for
unheated (left) and heated (right) conditions.
8-12
-------
Section 8—Results and Discussion: Attenuation of Soil Gas VOCs and Radon
Figure 8-10 shows the TO-17 grab samples taken from the subsurface sampling ports used as a
denominator in calculating the PCE attenuation factors. PCE concentrations tend to drop off with depth
between subslab and the deeper soil gas, with the winter/summer differences not showing any consistent
trends. The much lower concentrations observed in the wall port and shallow exterior soil gas samples are
consistent with and serve to validate the conceptual models of shallow soil gas concentrations modeled
and pictured in Abreu and Johnson (2005) and U.S. EPA (2012d).
TO-17 PCE Concentrations
100
cO
E
CD
LU
O
CL
10
*
*
98-197 98-204 110-208 36-54
97-189 110-204 110-168
Heating
*Off
On
$$> «
.<5
<5
^ ^ ^ ^ ^ ^
<=P
Figure 8-10. TO-17 PCE soil gas concentrations used in indoor air / soil gas attenuation factor
calculations with multiple soil gas sample probes averaged at each depth.
8-13
-------
Section 8—Results and Discussion: Attenuation of Soil Gas VOCs and Radon
Attenuation factors calculated from the weekly data are provided in Figure 8-11. As expected, attenuation
factors are higher (less attenuation) in the winter (during heating) on the 422 side, with summer levels
tending to have less attenuation than winter for the unhealed 420 basement.
0.
0.0
0 00
0.
c 0.0
o
~0.00
cc
c
-------
Section 8—Results and Discussion: Attenuation of Soil Gas VOCs and Radon
Figure 8-12 shows the weekly Radiello indoor air samples used for the weekly attenuation factors. Again
winter VOC levels are statistically higher (p = 6e-l 1 for basement and p = 0.0004 for first floor) due to
furnace operation on the 422 side but about the same (with no statistical difference), summer to winter, on
the unheated 420 side of the duplex.
Radiello Chloroform Concentrations
O
0.1 -
38-75
18-34
36-71
18-37
i
Heating
*Qff
$ On
&
4?
&
0
Figure 8-12. Radiello chloroform indoor air concentrations used as numerators for weekly
attenuation factor calculations.
8-15
-------
Section 8—Results and Discussion: Attenuation of Soil Gas VOCs and Radon
Figure 8-13 shows the TO-17 grab samples taken from the subsurface sampling ports used as a
denominator in calculating the chloroform attenuation factors. Unlike PCE, chloroform concentrations
tend to more or less increase with depth, with winter (heating on) concentrations higher for subslab and 6
ft and 9 ft soil gas samples but about the same as heating off concentrations in the deeper (13 ft and 16.5
ft) soil gas samples.
TO-17 Chloroform Concentrations
co
£ 100
CO
—J
E
o
M—
o
X—
o
O
10:
*
*
•
1
~
f
~
*
__
"
H
h
~
L
_
i
~
#
ii
¦I
+
*
~
ii
#
•
~
•
i
~
*
ll
n
¦i
:
_
—
•
1
-
:
1
i
# #
*
~
98-197 98-204
97-189 110
110-
-204
¦208
110-
36-
¦168
54
*
Heating
*Off
On
Figure 8-13. TO-17 chloroform soil gas concentrations (averaged by depth across soil gas
probes) used as denominators for weekly attenuation factor calculations.
8-16
-------
Section 8—Results and Discussion: Attenuation of Soil Gas VOCs and Radon
Figure 8-14 shows the weekly chloroform attenuation factors for the 420 and 422 first floors and
basements. For the 422 side the attenuation factors tend to decrease with depth from about 0.01 at subslab
to 0.002 at 13 and 16.5 ft, with attenuation going down (higher attenuation factors) in the heated winter.
For the unheated 420 side the attenuation factors show a similar if slightly lower trend with depth, but
tend to show an increase m attenuation (lower attenuation factors) from summer to winter.
Weekly Chloroform Attenaution
1
0.1
0.01
0 001
1 e-04
1
0.1
0.01
o 0.001
ra 1e-04
sz
0) 1
< 0.1
0.01
0.001
1e-04
1
0.1
0.01
0 001
1e-04
420Base
4^ ^
34-67 33-65 34-70 38-70 38-72 38
8-^8
12-18
420First
S * ^ ^ ^
16-31 16-30 16-32 18-32 18-33 18-26 6-9
422Base
32-65 32-62 32-67 36-67 36-68 36-55 12-18
422First
9- 4^4^
4^ jj?is
1:6-34 16-32 16-35 18-35 18-35 18-29 6-9
^
-------
Section 8—Results and Discussion: Attenuation of Soil Gas VOCs and Radon
It is also informative to compare the attenuation of all three compounds. Figure 8-15 juxtaposes the
weekly attenuation factors observed during the study period for radon, chloroform, and PCE for the 420
and 422 basement indoor air concentrations over subslab and near-subslab (6 ft) soil gas. One can see a
strong agreement between the different chemicals and house sides, with subslab attenuation factors
centering on or just below 0.01 and opposite summer-to-winter (heat off to heat on) trends for the heated
(422) and unheated (420) sides of the building.
Sub-Slab to Basement Weekly Attenuation
420 Base
0.1
0.01
0.0011 4
1e-04
34-67 36-62 34-67
+f
422Base
i
* : §
"O
o
3-
32-65 36-62 32-65
c 0.1
o
ro 0.01
o 0.001 a
< 1e-04
34-70 32-70 34-70 32-67 32-70 32-67
C/>
c
cr
CO
Q)
cr
0.1
0.011
0.001
1e-04
i
i ' ' 1
^3 +4"
38-70 38-70 38-70
.
36-67 38-70 36-67
*
CO
2.
O
0)
co
o
\
-------
Section 8—Results and Discussion: Attenuation of Soil Gas VOCs and Radon
8.3 Effect of Mitigation
Since the previous report, we have installed and experimented with a mitigation system underneath the
420/422 duplex. In this section, we will expand upon the attenuation discussion in Section 7 of the
previous report (U.S. EPA, 2012a) by examining the effect of mitigation on VOC attenuation. The
mitigation system has three possible settings—"On" (fan operating), '"Passive" (fan off, system open), and
"Off' (fan off, system valves closed). Figure 8-16 shows how those settings were deployed since the
mitigation system was installed in October. However, as described in Section 5.2, the not-yet-installed,
passive, and off states were determined to be statistically the same and so are treated as "Off' in the
analyses described in this section.
Mitigation Status: [/]
Not Yet
Installed
October 2012
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
January 2013
April 2013
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
—1 1 r -T -
off[7
On
November 2012
1
2
3
4
5
6
7
8
9
10
11
12
13
14
February 2013
i
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
May 2013
18
19
20
21
22
23
24
25
26
27
28
29
30
31
December 2012
March 2013
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
Figure 8-16. Mitigation status and schedule.
8-19
-------
Section 8—Results and Discussion: Attenuation of Soil Gas VOCs and Radon
8.3.1 Subslab to Indoor Air Daily Attenuation
Daily Radon Attenuation Factors
The mitigation system was designed for radon mitigation and it has been proven to be very effective in
this regard (see Section 5). Radon attenuation between the subslab and indoor air responded as expected
to the mitigation system, with dramatic drops in radon concentrations when the system was turned on and
equally rapid increases in radon system when the system was shut off or operated in passive mode. Even
with relatively few indoor air and soil gas sample pairs (Figure 8-17), the effect of the mitigation on
radon attenuation is obvious (Figure 8-18), with over an order of magnitude reduction in radon
concentrations with mitigation on, which is mainly due to the over an order of magnitude reduction in
indoor air radon concentrations (see upper panel of Figure 8-17).
10
Daily Radon Concentration
Indoor AlphaGUARD Concentration
1 =
O
Cl
e
o
TJ
ra
q:
1000
3-7
3-7
4-8
Portable AlphaGUARD Concentration
100
3-7
1
%
&
L#
3-7
)
>
&
&
&
4-8
2
Mitigation
¦ Off
SOn
Figure 8-17. Indoor air (AlphaGUARD) and soil gas (portable AlphaGUARD) radon concentrations
used in daily attenuation factor calculations.
8-20
-------
Section 8—Results and Discussion: Attenuation of Soil Gas VOCs and Radon
0.1 -
C 0-01 d
_o
TO
ZJ
sz
0.001 -
Daily Radon Attenuation Factors
3-7
3-7
ri?'
&
&
&-
4-8
3
t#
Mitigation
Son
^On
Figure 8-18. Daily radon attenuation factors.
Interestingly, at some locations it appears that mitigation also decreases subslab and wall port radon
concentrations (bottom panel of Figure 8-17). Even in locations where the exterior concentration
dropped, the attenuation factor between exterior locations and indoor air is reduced considerably by
mitigation (Figure 8-18).
Daily VOC Attenuation Factors
For daily VOC attenuation factors (Figure 8-19 and 8-20), turning the mitigation system on decreased the
attenuation factors and increased attenuation for all conditions tested. Although the SSP-2 and WP-3
concentrations are suspected low because of dilution, there was an increase in attenuation in both cases
for PCE (Figure 8-19) and chloroform (Figure 8-20). Attenuation factors over 1 in these figures for PCE
and (especially) chloroform in SSP-2 are clearly unrealistic and may be reflecting limited connections
between the subslab area sampled by SSP-2 and indoor air.
8-21
-------
Section 8—Results and Discussion: Attenuation of Soil Gas VOCs and Radon
1 =
0.1 :
s=
_o
03
c
£
< 0.01
0.001 :
Dajly PCE Attenuation Factors
11
+
t
38-57 38-57 38-57 38-57
J
-/J, -Jb
^ ^ & V
„<=£
^ ^ ^
Figure 8-19. Daily PCE attenuation factors.
Mitigation
^Off
^ On
8-22
-------
Section 8—Results and Discussion: Attenuation of Soil Gas VOCs and Radon
10 =
1 :
C
_o
to 0-1 =
=s
c
<1)
0.01 :
0_001 -
Daijy Chloroform Attenuation
1-5
ii
17-19 38
-4I
38-46
Mitigation
^Off
^On
,3
3
Figure 8-20. Daily chloroform attenuation factors.
8.3.2 Subslab and Soil Gas to Indoor Air Weekly Attenuation
The weekly attenuation factors are assigned a mitigation status based on what mitigation status was most
common during the sample's exposure. For example, a weekly sample that was exposed for 7 days, four
of them with the mitigation system off and three of them with the mitigation system in passive mode, is
assigned a mitigation status of "Off" As in previous discussions, weekly attenuation factors were based
weekly indoor air electret concentrations for radon and Radiellos for VOC concentrations, with
subsurface concentrations being represented by portable AlphaGUARD and TO-17 grab samples.
8-23
-------
Section 8—Results and Discussion: Attenuation of Soil Gas VOCs and Radon
Radon
The electret indoor air concentrations before and after mitigation are shown in Figure 8-21, representing
the entire period of electret record available for this study (January 2011 to May 2013).18 The dramatic
reduction of radon in indoor air during mitigation of about an order of magnitude can be seen in the
figure, with all post-mitigation radon concentrations being under the 4 pCi/L target concentration.
Electret Concentrations
10
i
1-
o
Q-
c
o
rs
A3
cr
Mitigation
¦ Off
^On
0 1
14-17
S&
&
&
7-9
12-17
&
&
7-9
~1
Figure 8-21, Electret indoor air radon concentrations.
Subsurface radon concentrations also show a reduction with mitigation (Figure 8-22), but not as great a
reduction as observed for indoor air. The attenuation factors calculated from these pairs of measurements
(Figure 8-23) show that in every case, the lowered indoor air concentration predominates with a decrease
in the attenuation factors (and increase in attenuation) as a result.
I8See mitigation on and off periods in Figure 8-16.
8-24
-------
Section 8—Results and Discussion: Attenuation of Soil Gas VOCs and Radon
1000-
o
Q.
c
0
"O
ro
01
100-
AlphaGUARD Concentrations
IL
I
I*
1
29-34 34-17 34-52 34-52 34-41 23-46
xO?'"1 d'"®' /-Qy*'
^ ,v<^ «?> ^
cp
<=F cp c#
Mitigation
• Off
E3 On
Figure 8-22. AlphaGUARD soil gas radon concentrations (pCi/L) used as denominator in
attenuation factor calculations.
8-25
-------
Section 8—Results and Discussion: Attenuation of Soil Gas VOCs and Radon
0.1
0.01
0.001
1e-04
0.1
0.01
_ 0.001 =
.9. ie-04
0.1 I
< 0.01s
0.001
1e-04
0.1
0.01 4
0.001
1e-04
Weekly Radon Attenaution
420Base
-r
10-12 12-6 12-18 12-18 12-14 8-16
420First
_5-£ _6r.3
6-9
j
AS.
$
422Base
10-12 11-6 11-18 11-18 11-14 7-16
422First
l±) S 3 =^=
* ^ .
5-6 6-3 6-9 6-9
6-8 4-8
^ XT* A 0#
*
-------
Section 8—Results and Discussion: Attenuation of Soil Gas VOCs and Radon
However, note that the number of samples was too low to determine the statistical significance of these
VOC drops for PCE.
Radiello PCE Concentrations
ro
O)
LLj
O
¦
8-18
3-6
6-12
3-6
&
<0* <0?
£
Mitigation
*Off
On
Figure 8-24. Radiello indoor air PCE concentrations.
Figure 8-25 shows that mitigation effects on the subsurface PCE concentrations were not so definitive,
with concentrations actually increasing in a few cases, and decreasing in others. These data suggest
effects on the distribution of subsurface VOCs that are not consistent with radon and not fully understood
8-27
-------
Section 8—Results and Discussion: Attenuation of Soil Gas VOCs and Radon
at this time. However, Figure 8-26, which shows the effect of mitigation on weekly PCE attenuation,
increased attenuation (i.e., decreased the attenuation factor) in every case, although again the sample size
was small.
100
to
E
CO
III
O
Q_
10
TO-17 PCE Concentrations
2-5 3-4
2-5
Tl
3-6 3-6
3-6
.
-------
Section 8—Results and Discussion: Attenuation of Soil Gas VOCs and Radon
1
0.1
0.01
0.001
1
0.1
c 0.01
So 001
03
3
c
© 1
< 0.1
0.01
0.0014
1?
°-1 i
0.01 j
0.001 s
Weekly PCE Atteriaution
4206ase
qg
5-15 8-12
¦
5-tfla 8-18 8-te 8-18
420First
m
2-5 3-4 2-d
3-6
3-^.! 3-6
422Base
TTTIlB
4-10 6-8 4-1Q 6-12 6-12 6-12
422First
3 9
¦.
2-5
3-4
2-1
3-6
3-6
Mitigation
4 Off
E^l On
.-5- ,v G^" A
Denominator Location
-------
Section 8—Results and Discussion: Attenuation of Soil Gas VOCs and Radon
but in this case there was a statistically significant decrease in chloroform with mitigation for the 422
basement (p = 6e-l 1) and the 422 first floor (p = 0.004).
Radiello Chloroform Concentrations
1 -
co
E
CD
—>
E
i—
o
H—
o
o
6
0.1 -
6-12
e§>
3-6
Mitigation
• Off
^ On
Figure 8-27. Radiello weekly indoor air chloroform concentrations.
8-30
-------
Section 8—Results and Discussion: Attenuation of Soil Gas VOCs and Radon
TO-17 Chloroform Concentrations
100-
co
E
o>
mm*
E
o
<+—
o
o
6 10-
H h
13-35 20-28 13-35 20-42 20-42 20-42
90^
..G8,
cp*
/V
<3
.6
»¦
^ # # cp*
-------
Section 8—Results and Discussion: Attenuation of Soil Gas VOCs and Radon
Weekly Chloroform Attenuation
4203ase
0.11
0.011
0.0011 5.15
Q
8-12
~
8-18
8-18^
420First
0.1 i
0.01 i
°0,001 ;
-*-»
03
13
C
0
ti
<
2-5 3-6
2-5
3-4
J=
o.i ^
0.01 I
0.0011 4_-io
6-8
422Base
- t
4-10 6-12
6-12 6-^P
422First
0.1 i
|
0.011
0.001 i
2-5 3-4 2-5
i
3-6 3-6
^ T 1— 1
Mitigation
ft Off
3 On
Denominator Location
Figure 8-29. Weekly chloroform attenuation factors.
In summary. Figure 8-30 compares the weekly attenuation of radon and the two VOCs. In general, one
can see that attenuation increases with mitigation for all three compounds, with less variable results being
identified for radon compared to PCE and for PCE compared to chloroform. However, the influence of
the mitigation system on subslab VGC levels (Figures 8-22, 8-25, and 8-28) should be considered when
evaluating these data as well as whether the attenuation factor is meaningful when an SSD mitigation
system is on. Also, the performance of the mitigation system on VOCs at this site is not fully understood
8-32
-------
Section 8—Results and Discussion: Attenuation of Soil Gas VOCs and Radon
at this point. To better understand VOC attenuation at this site, additional scmtiny of possible VOC entry
routes and subslab pressure field variability is warranted, particularly in the vicinity of subsurface
structures such as the sanitary sewer.
Sub-Slab to Basement Weekly Attenuation
420Base 422Base
0.
0.0
0,00
c
o
% °-
TO
3
g 0-0
ti
<0,00
0.
0.0
0.00
0
D
5-15 10-12 5-15
*
4-10 10-12 4-10
9L
-Q
o
7X
5-15 12
b * e *
4-10 11-18 4-1 b
05
c
9"
w
to
cr
fri. I
~ Hr
8-19 12-
8-18
in
a
to
O)
a>
6-12 11-18 6-12
\o-
&
&
Mitigation
*Off
3 On
Figure 8-30. Attenuation of PCE, chloroform, and radon juxtaposed.
8-33
-------
Section 8—Results and Discussion: Attenuation of Soil Gas VOCs and Radon
8-34
-------
Section 9—Results and Discussion: Determine If Observed Changes in Indoor Air Concentration of
Volatile Organics of Interest are Mechanically Attributable to Changes in Vapor Intrusion
Table of Contents
9.0 Results and Discussion: Determine If Observed Changes in Indoor Air Concentration of
Volatile Organics of Interest are Mechanistically Attributable to Changes in Vapor
Intrusion 9-1
9.1 Large Differential Pressures, Pressure Changes and Meteorological Factors Analysis
with Mitigation Off 9-1
9.1.1 Temperature Effects on Differential Pressure 9-1
9.1.2 Barometric Pressure Effects on Differential Pressure 9-4
9.1.3 Precipitation Effects on Differential Pressure 9-8
9.1.4 Wind Effects on Differential Pressure 9-10
9.1.5 Assessment of Whether High Observed Differential Pressures Could be
Artifactual 9-16
9.1.6 Examination of High-Resolution Time Series Data for Individual Extreme
Differential Pressure Events 9-16
9.2 Influence of Meteorological Conditions on Indoor VOC Concentration; Mitigation
Off 9-44
9.2.1 Temperature Effect on VOCs 9-44
9.2.2 Barometric Pressure Effect on VOCs 9-47
9.2.3 Precipitation Effects on VOCs 9-47
9.2.4 Effect of Wind on VOC Concentrations 9-52
9.3 Summary of Meteorological Effects on Vapor Intrusion—Evidence Presented in
Sections 6 and 9 9-59
List of Figures
9-1. Long-term trend in subslab vs. basement differential pressure (Pa) compared with
exterior temperature and the first derivative of exterior temperature (°F) 9-2
9-2. XY plot of subslab vs. basement differential pressure vs. daily low exterior temperature 9-3
9-3. XY plot of daily low exterior temperature vs. (subslab vs. basement differential pressure) 9-4
9-4. Long term pressure trends in subslab vs. basement differential pressure (Pa) compared
with external barometric pressure (inches) with derivative plots 9-5
9-5. XY plot of external barometric pressure vs. (422 subslab vs. basement differential
pressure) 9-6
9-6. XY plot of barometric pressure drop (per hour) vs. (422 subslab vs. basement differential
pressure) 9-7
9-7. Long term trends in subslab vs. basement pressure (Pa) compared to rainfall and snow
depth (inches) 9-8
9-8. XY graph of total daily rainfall vs. (subslab vs. basement differential pressure) 9-9
9-9. XY graph of total snow depth (inches) vs. differential pressures (Pa) 9-10
9-10. Long-term trends in subslab vs. basement differential pressure (Pa) compared to
maximum wind speed and average wind speed data (mph) 9-11
9-i
-------
Section 9—Results and Discussion: Determine If Observed Changes in Indoor Air Concentration of
Volatile Organics of Interest are Mechanically Attributable to Changes in Vapor Intrusion
9-11. Long-term trend in subslab vs. basement differential pressure (Pa) vs. wind direction
parameters (change in direction, maximum direction, average direction) (degrees) 9-12
9-12. XY plot of daily high wind speed vs. (subslab vs. basement differential pressure) 9-13
9-13. XY plot of daily average wind speed vs. (subslab vs. basement differential pressure) 9-14
9-14. XY plot of wind direction effects on subslab vs. basement differential pressure (upper
plot) and radon concentrations in the 422 basement (lower plot) 9-15
9-15. Extreme Event 1: subslab vs. basement differential pressure (positive difference
indicates flow into the structure) 9-17
9-16. Extreme Event 2: subslab vs. basement differential pressure (positive difference
indicates flow into the structure) 9-18
9-17. Extreme Event 3: subslab vs. basement differential pressure (positive difference
indicates flow into the structure) 9-19
9-18. Extreme Event 4: subslab vs. basement differential pressure (positive difference
indicates flow into the structure) 9-20
9-19. Extreme Event 5: subslab vs. basement differential pressure (positive difference
indicates flow into the structure) 9-21
9-20. Extreme Events 6/7: subslab vs. basement differential pressure (positive difference
indicates flow into the structure) 9-22
9-21. Detailed time series of unusual pressure Event 1 showing barometric pressure changes
(barometric pressure in inches of Hg, differential pressure in Pa) 9-23
9-22. Detailed time series of unusual pressure Event 1 showing precipitation events (rainfall in
inches, snow depth in inches, differential pressure in Pa) 9-24
9-23. Detailed time series of unusual pressure Event 2 showing barometric pressure changes
(barometric pressure in inches of Hg, differential pressure in Pa) 9-25
9-24. Detailed time series of unusual pressure Event 2 showing temperature changes
(differential pressure in Pa, temperature in °F) 9-26
9-25. Detailed time series of unusual pressure Event 3 showing barometric pressure changes
(barometric pressure in inches of Hg, differential pressure in Pa) 9-27
9-26. Detailed time series of unusual pressure Event 3 showing temperature changes
(differential pressure in Pa, temperature in °F) 9-28
9-27. Detailed time series of unusual pressure Event 3 showing wind direction variables (wind
direction-related variables in degrees, differential pressure in Pa) 9-29
9-28. Detailed time series of unusual pressure Event 3 showing PCE and radon 9-30
9-29. Detailed time series of Event 4 showing wind direction variables (wind direction-related
variables in degrees, differential pressure in Pa) 9-32
9-30. Detailed time series of Event 4 showing wind speed variables (wind speed variables in
MPH, differential pressure in Pa) 9-33
9-31. Detailed time series of Event 4 showing temperature variables (differential pressure in
Pa, temperature in °F) 9-34
9-32. Detailed time series of Event 5 showing precipitation event (rainfall in inches, snow
depth in inches, differential pressure in Pa) 9-35
9-ii
-------
Section 9—Results and Discussion: Determine If Observed Changes in Indoor Air Concentration of
Volatile Organics of Interest are Mechanically Attributable to Changes in Vapor Intrusion
9-33. Detailed time series of Event 5 showing wind speed variables (wind direction-related
variables in degrees, differential pressure in Pa) 9-36
9-34. Detailed time series of Event 5 showing wind direction variables 9-37
9-35. Detailed time series of Event 5 showing temperature variables (differential pressure in
Pa, temperature in °F) 9-38
9-36. Detailed time series of Event 5 showing PCE and radon 9-39
9-37. Detailed time series of Event 6/7 showing barometric pressure (barometric pressure in
inches of Hg, differential pressure in Pa) 9-40
9-38. Detailed time series of Event 6/7 showing precipitation (rainfall in inches, snow depth in
inches, differential pressure in Pa) 9-41
9-39. Detailed time series of Event 6/7 showing wind speed variables (wind direction-related
variables in degrees, differential pressure in Pa) 9-42
9-40. Detailed time series of Event 6/7 showing wind direction variables (wind direction-
related variables in degrees, differential pressure in Pa) 9-43
9-41. Detailed time series of Event 6/7 showing radon and PCE 9-44
9-42. XY graph of total heating degree days per week vs. weekly PCE concentration (Radiello
data) 9-45
9-43. XY graph of heating degree days vs. chloroform concentration (weekly Radiello data) 9-46
9-44. XY graph of average weekly barometric pressure vs. PCE concentration 9-47
9-45. XY graph of weekly average snow depth vs. PCE concentration 9-48
9-46. XY graph of weekly average snow depth vs. chloroform concentration 9-49
9-47. XY graph of weekly average snow depth vs. radon concentration 9-50
9-48. XY graph of total weekly rainfall vs. PCE concentration in indoor air 9-51
9-49. XY graph of total weekly rainfall vs. chloroform concentration in indoor air 9-52
9-50. XY graph of wind direction vs. PCE concentration in indoor air in 422 basement 9-53
9-51. XY graph of wind direction vs. chloroform concentration in indoor air in 422 basement 9-54
9-52. XY graph of wind direction vs. radon concentration in indoor air in 422 basement 9-55
9-53. Modeled effect of building wind loads on ground surface and subslab gauge pressure
distribution (adapted from U.S. EPA, 2012d) 9-56
9-54. Modeled effect of building wind load on subslab soil vapor distribution for recalcitrant
and aerobically biodegradable VOCs (adapted from U.S. EPA, 2012d) 9-57
9-55. XY graph of wind speed vs. PCE concentration in indoor air in 422 basement 9-58
9-56. XY graph of wind speed vs. chloroform concentration in indoor air in 422 basement 9-59
List of Tables
9-1. Summary of Qualitative Lines of Evidence for Meteorological Factors Influencing Vapor
Intrusion in This Study 9-60
9-iii
-------
Section 9—Results and Discussion: Determine If Observed Changes in Indoor Air Concentration of
Volatile Organics of Interest are Mechanically Attributable to Changes in Vapor Intrusion
9-iv
-------
Section 9—Results and Discussion: Determine If Observed Changes in Indoor Air Concentration of
Volatile Organics of Interest are Mechanically Attributable to Changes in Vapor Intrusion
9.0 Results and Discussion: Determine If Observed Changes in
Indoor Air Concentration of Volatile Organics of Interest are
Mechanistically Attributable to Changes in Vapor Intrusion
9.1 Large Differential Pressures, Pressure Changes and Meteorological Factors
Analysis with Mitigation Off
Significant vapor intrusion into the indoor air is normally attributable to an advective driving force across
the building envelope—a differential pressure. For this research effort, we gathered a larger high temporal
resolution data set on these differential pressures than on indoor air concentrations of VOCs.19 Thus, we
performed a focused analysis of the events surrounding the periods where unusual differential pressure,
especially subslab to indoor air differential, pressure was evident using the following criteria:
¦ great variation (swings of more than 5 Pa in a short time period) and
¦ the strongest driving forces pushing subslab vapors into the building (>10 Pa).
A change of 5 Pa is generally seen as the maximum change expected in residential situations (U.S. EPA,
1993; Environmental Quality Management, 2004; U.S. EPA, 2012d; Yao, 2010). However, there are
some reports of larger pressure fluctuations of up to 20 Pa in the literature (Environmental Quality
Management, 2004; Lutes, 2010).
We identified seven such extreme variation events over nearly 2 years of data collection (typical time
resolution of 14 minutes) and reviewed them to determine if they could have been artifactual or if they
had the hallmarks of real events. Real causative factors that were examined included extremes in any of
these metrological parameters:
¦ wind velocity gusts
¦ abrupt significant changes in wind direction (approaching 180 degrees)
¦ barometric pressure changes
¦ sudden changes in temperature (outside)
¦ snowfalls
¦ rain events.
We used multiple graphing approaches to examine this data, including:
¦ long time-scale stacked graphs comparing differential pressure trends with time to those of
meteorological parameters
¦ XY graphs of meteorological parameters vs. differential pressure (which were done on the data
after aggregating to 1-day time resolution)
¦ short time-scale stacked graphs comparing differential pressures with time during individual
events to meteorological parameters.
9.1.1 Temperature Effects on Differential Pressure
The seven events are clearly visible in the subslab vs. basement differential pressure plot (Figure 9-1).
They appear to not be distributed randomly over time but to be clustered in midwinter and to occur with
much greater frequency on the days with the lowest daily low temperatures (Figure 9-2). A related trend
19In statistical terms, differential pressure can be viewed as an intermediary variable, one that is part of a causal pathway linking a predictor
variable such as temperature with an outcome such as indoor air concentration.
9-1
-------
Section 9—Results and Discussion: Determine If Observed Changes in Indoor Air Concentration of
Volatile Organics of Interest are Mechanically Attributable to Changes in Vapor Intrusion
is that these high differential pressures occurred on the days that were the coldest overall as measured by
heating degree days (Figure 9-3). The occurrence of strong driving forces into the building during cold
weather is likely mechanistically related to the stack effect. More rarely, however, a strong differential
pressure pushing out of the building was also observed in the daily average differential pressure.2"
However, not all low-temperature days exhibited strong differential pressure driving forces on average,
suggesting that other meteorological parameters may also be important.
SubSlab.Vs.Basement
30-
20-
10
0
-10
100"
75'
50-
25-
0
20"
0-
-20-
if_i
Temp. Out
Temp.Out.dTdt
%v f*v •** ***** ** • *
• * ~* . U:
• * * * *
\
&
,0
¦A
Figure 9-1. Long-term trend in subslab vs. basement differential pressure (Pa) compared with
exterior temperature and the first derivative of exterior temperature (°F).
20The meteorological data were examined, but no consistent common factor could be discerned to explain the 4 days with an average differential
pressure out of the structure of more than -2Pa: 2/6/11, 2/17/12, 1/30/12, and 1/26/13. Nor do those days stand out as unusual in the
AlphaGUARD radon data set.
9-2
-------
Section 9—Results and Discussion: Determine If Observed Changes in Indoor Air Concentration of
Volatile Organics of Interest are Mechanically Attributable to Changes in Vapor Intrusion
Average Daily 422 Subslab vs. Basement Differential Pressure as a
Function of Daily Low Exterior Temperature - Premitigation and
Mitigation Shut Off Data Only
O)
3
u
3
~
o
c
5
_o
LL.
O)
3
in
in
01
a.
~
~ ~
> iU
*55
O
a.
a.
~
~~
~ ~ ~~
oj 5
3
Vt
V
a.
« o
|io
0)
b
10 20 * 30 ~ 40 50 60 70 80 90
~ ~
~
~
Exterior Temperature (F)
Figure 9-2. XY plot of subslab vs. basement differential pressure vs. daily low exterior
temperature.
9-3
-------
Section 9—Results and Discussion: Determine If Observed Changes in Indoor Air Concentration of
Volatile Organics of Interest are Mechanically Attributable to Changes in Vapor Intrusion
Average Daily 422 Subslab vs. Basement Differential Pressure as a
Function of Heating Degree Days- Premitigation and Mitigation Shut Off
Data Only
3
u
3
~
o
C
5
o
LL.
01
3
in
in
01
a.
~
~ ~
> iU
"in
O
a.
"to"
a.
~
~ ~
~ ~ * * *
ij
c
0)
01
£
o
10 a 30 * 40 50 60
~ ~
~
~
Heating Degree Days)
Figure 9-3. XY plot of daily low exterior temperature vs. (subslab vs. basement differential
pressure).
9.1.2 Barometric Pressure Effects on Differential Pressure
The greatest degree of barometric pressure variation was observed in winter (Figure 9-4). However, on a
scale of days, extreme differential pressures occur under a variety of barometric pressure conditions
(Figure 9-5). It might be expected that these extreme differential pressure events would be related to the
rate of barometric pressure change, but this did not appear to be the case (Figure 9-6).
9-4
-------
Section 9—Results and Discussion: Determine If Observed Changes in Indoor Air Concentration of
Volatile Organics of Interest are Mechanically Attributable to Changes in Vapor Intrusion
SubSlab.Vs.Basement
30-
20-
10-
0-
-10-
Bar..
30.4-
30.0-
29.6"
29.2-
SubSlab.Vs.Basement.dPdt
100-
o-
-100-
; *
dPdt
Figure 9-4. Long term pressure trends in subslab vs. basement differential pressure (Pa)
compared with external barometric pressure (inches) with derivative plots.
9-5
-------
Section 9—Results and Discussion: Determine If Observed Changes in Indoor Air Concentration of
Volatile Organics of Interest are Mechanically Attributable to Changes in Vapor Intrusion
Average Daily 422 Subslab vs. Basement Differential Pressure as a
Function of Average Barometric Pressure
Premitigation and Mitigation Shut Off Data Only
iU "
*55
O
a.
a.
~
~ ~
~ | » j ~ ~
0) ^ =
3
V)
01
a.
« 0 -
| 2C.
01
Q
.2 29.4 29.6 29.8 * 30 $0.2 30.4 30.6
~ ~
~
~
Barometric pressure
Figure 9-5. XY plot of external barometric pressure vs. (422 subslab vs. basement differential
pressure).
9-6
-------
Section 9—Results and Discussion: Determine If Observed Changes in Indoor Air Concentration of
Volatile Organics of Interest are Mechanically Attributable to Changes in Vapor Intrusion
Average Daily 422 Subslab vs. Basement Differential Pressure as a
Function of Barometric Pressure Drop
Premitigation and Mitigation Shut Off Data Only
3
~
3
o
i
o
LL.
01
3
~
trt
trt
01
a.
~ ~
V)
o
a.
"to"
a.
~
~ ~
t f .
oj 5
3
in
in
01
a.
« 0 ¦
c
0) 1
0)
5fc
Q
1 ~ 0.05 0.1 0.15 0.2 0.25 0.3
~ ~
~
~
Barometric Pressure Drop per hour
Figure 9-6. XY plot of barometric pressure drop (per hour) against 422 subslab vs. basement
differential pressure.
9-7
-------
Section 9—Results and Discussion: Determine If Observed Changes in Indoor Air Concentration of
Volatile Organics of Interest are Mechanically Attributable to Changes in Vapor Intrusion
9.1.3 Precipitation Effects on Differential Pressure
There was no apparent relationship between rain events and the extreme differential pressure events.
(Figures 9-7 and 9-8). There was no apparent relationship between average differential pressure and
snow depth (Figure 9-9) either.
SubSlab.Vs.Basement
*
i
•
*
*
*
t
i
|«*M
k»
"b
Mp
1 *
ill
ij
f
*
LI.
Rain
1
*
*
*
* *
. *
*
t
*
* *
* At
t :
* *
* •
* *1 •
•
*
••
*
**
i ivi
A
riitiiiAL
~» ± ',
.~ .
' 1
SnowDepth
»
*
K
«• V*
~
*
*
*»#*
r *
•**
.
•
i
*
* f
30"
20-
10-
0-
-10-
1.00"
0.75"
0.50"
0.25-
0.00-
6-
4"
2"
o-
a
&
$
&
nr
fir
r<\
Figure 9-7. Long term trends in subslab vs. basement pressure (Pa) compared to rainfall and
snow depth (inches).
9-8
-------
Section 9—Results and Discussion: Determine If Observed Changes in Indoor Air Concentration of
Volatile Organics of Interest are Mechanically Attributable to Changes in Vapor Intrusion
Average Daily 422 Subslab vs. Basement Differential Pressure as a
Function of Rainfall
Premitigation and Mitigation Shut Off Data Only
25
£
3
U
3
? 20
o
c
5
o
LL.
.2 15
£
3
in
in
£
g 10
*55
£
re
i:
c
O)
0)
5fc
b
-5
~
~
~
~
— ~ ~' * :
}~ * 0.5 1 1.5 2 2.5 3
~
~
~
Rainfall (inches)
Figure 9-8. XY graph of total daily rainfall against subslab vs. basement differential pressure
9-9
-------
Section 9—Results and Discussion: Determine If Observed Changes in Indoor Air Concentration of
Volatile Organics of Interest are Mechanically Attributable to Changes in Vapor Intrusion
20
Average Daily 422 Subslab vs. Basement Differential Pressure as a
Function of Snow Depth
Premitigation and Mitigation Shut Off Data Only
~ t
-10
Snow Depth (inches)
Figure 9-9. XY graph of total snow depth (inches) against differential pressures (Pa).
9.1.4 Wind Effects on Differential Pressure
Based on the long-term time series, it is clear that winter is when these large differential pressure swings
occur, which is also the season when the highest wind speeds are experienced (Figure 9-10). When the
differential pressure time series was plotted against the wind direction-related time series (Figures 9-11
and 9-14), the highest differential pressures are associated with wind directions between 220 and 320
degrees. Surprisingly, the high differential pressure days appear to occur with moderate average wind
speeds and minimal values for the daily high wind speed (Figures 9-12 and 9-13). The clearest
relationship on a scale of days, however, is when high differential pressures that are created by average
winds from 220 to 320 degrees (roughly SW to NW) draw vapors into the building.
9-10
-------
Section 9—Results and Discussion: Determine If Observed Changes in Indoor Air Concentration of
Volatile Organics of Interest are Mechanically Attributable to Changes in Vapor Intrusion
SubS lab. Vs. Basement
Hi. Speed
J/VincL Speed
Figure 9-10. Long-term trends in subslab against basement differential pressure (Pa) compared
to maximum wind speed and average wind speed data (mph).
9-11
-------
Section 9—Results and Discussion: Determine If Observed Changes in Indoor Air Concentration of
Volatile Organics of Interest are Mechanically Attributable to Changes in Vapor Intrusion
SubSlab. Vs. Basement
ihNtom Jk
Chanqe.Dir
¦¦ 00 -
wind.Dir
Figure 9-11. Long-term trend in subslab vs. basement differential pressure (Pa) vs. wind direction
parameters (change in direction, maximum direction, average direction) (degrees).
9-12
-------
Section 9—Results and Discussion: Determine If Observed Changes in Indoor Air Concentration of
Volatile Organics of Interest are Mechanically Attributable to Changes in Vapor Intrusion
Average Daily 422 Subslab vs. Basement Differential Pressure as a
Function of maximum wind speed
Premitigation and Mitigation Shut Off Data Only
3
u
3
~
o
C
5
o
LL.
0)
3
in
wi
01
a.
~
~ ~
> iU "
*55
O
a.
"to"
a.
~
~ ~
~ , ~ ~ ~ ,
0) ^ c
3
in
in
01
a.
« 0 ¦
•'iiillhliHliiHlMl'nv::.- s:
E
0) 1
0)
O
) 10 * 2b 30 40 50
~ ~
~
60
~
wind speed high (mph)
Figure 9-12. XY plot of daily high wind speed against subslab vs. basement differential pressure.
9-13
-------
Section 9—Results and Discussion: Determine If Observed Changes in Indoor Air Concentration of
Volatile Organics of Interest are Mechanically Attributable to Changes in Vapor Intrusion
Average Daily 422 Subslab vs. Basement Differential Pressure as a
Function of Average Wind Speed
Premitigation and Mitigation Shut Off Data Only
3
u
3
~
o
C
5
_o
LL.
01
3
in
wi
01
a.
~
~~
.> iU "
O
a.
"to"
a.
~
~ ~
QJ 5
3
in
in
01
a.
« 0 ¦
~ «
E
0) 1
0)
5fc
b
2 * 4 ~ 6 8 10 12 14
~ ~
~
~
wind speed (mph)
Figure 9-13. XY plot of daily average wind speed against subslab vs. basement differential
pressure.
9-14
-------
Section 9—Results and Discussion: Determine If Observed Changes in Indoor Air Concentration of
Volatile Organics of Interest are Mechanically Attributable to Changes in Vapor Intrusion
Average Daily 422 Subslab vs. Basement Differential Pressure as a
Function of average wind direction
Premitigation and Mitigation Shut Off Data Only
3
ts
3
~
a 20
+J
£
|
3
~
~ ~
O
*ra
~
~ ~
• t •
3
c
it
5
) 50 foo 150 200 * 250 300 350 400
~ ~
~
~
wind direction (degrees)
Average Radon in 422 Basement as a Function of average wind direction
Premitigation and Mitigation Shut Off Data Only
~~
~ t
~
*
~ ~~ ~
~*~~~ ~
4»,» » +
~
~ ~
* ~
~ ~ ~
* i ~~ ~
A
~
*:
\ ~ *
~ a* 4 V* 4 t
~ * ..*v* «:*•*
• ~
~ ~
~~~~
>. .~ ~ ~ bfi .. "Ir
~
~ ~
~ ft f ~ A ~
~~ «* ~
~ ~~ *~ ~ ~
A ~ . ~ A ' ~t* >
~
~
~
» ^ ~ ~
t ~ ~ * *
~ 4
~ ~
~~
~ ~ ~
^ * * *
~
0 50 100 150 200 250 300 350 400
wind direction (degrees)
Figure 9-14. XY plot of wind direction effects on subslab vs. basement differential pressure
(upper plot) and radon concentrations in the 422 basement (lower plot).
9-15
-------
Section 9—Results and Discussion: Determine If Observed Changes in Indoor Air Concentration of
Volatile Organics of Interest are Mechanically Attributable to Changes in Vapor Intrusion
9.1.5 Assessment of Whether High Observed Differential Pressures Could be
Artifactual
We examined multiple lines of evidence to determine if these differential pressures could be artifactual:
¦ The association between the extreme differential pressures and particular temperature ranges and
wind directions discussed above suggests that they are real phenomenon potentially caused by
meteorological events.
¦ The occurrence of these pressure swing events in three winters and the absence in intervening
summers would have been unlikely if they were due to an instrument failure.
¦ The lack of outliers in some of the other meteorological data sets occurring at the same time as
the very high/changeable differential pressures tends to eliminate one possible cause of artifactual
data—power fluctuations or lightning strikes—because all of the instrumentation is located in one
room and shares a common building power supply.
¦ The lack of extreme values in the first derivative of the differential pressure (see Figure 9-4) also
suggests that these results are not artifactual, because an artifact caused by a discrete event (for
example, a technician stepping on or tripping over a hose of the differential pressure instrument)
should be evidenced by an extremely high rate of change in the differential pressure.
¦ When examined in detail, the pattern of the data points taken every 14 minutes shows a
continuous increase or decrease for 10 or more successive measurements. This would not be
characteristic of either random noise or a discrete, inadvertently human-caused event.
¦ We focused our review presented in this section on extremes that occurred in the time period
before the mitigation system was installed, although some extreme events did occur that were not
fully managed by the mitigation system. As an additional step to determine if these extreme
events were artifactual, we reviewed their duration and time of occurrence. We reasoned that
hypothetical artifacts caused by human action should be short in duration and occur primarily
during regular on-site working hours. Artifacts caused by power system disruptions or spikes
should also be of short duration. As discussed in the sections that follow, many of these events
extend over multiple days, which suggests a physical as opposed to artifactual explanation.
¦ As discussed in Section 5.1.2, a separate, handheld differential pressure instrument was also used
to make some observations. On 2 days, fluctuations of subslab:interior differential pressure of
>0.28 inches of water column (70 Pascals) were observed with that instrument.
These lines of evidence taken together strongly suggest that these extreme pressure events are real
physical phenomena.
9.1.6 Examination of High-Resolution Time Series Data for Individual Extreme
Differential Pressure Events
Since:
¦ we concluded in Section 9.1.5 that these extreme pressure events were a real physical phenomena
¦ we expect from theory that high differential pressures toward the structure will lead to high vapor
intrusion flux and
¦ we expect from theory that large changes in differential pressure if sustained over a period of
hours or days should be associated with large changes in indoor air concentrations
¦ we sought to understand the conditions under which these extreme differential pressure events
occurred.
The first extreme event lasts at least 3 days, from Friday February 4 at approximately 4 PM to
approximately 1 PM on Monday February 7, 2011 (Figure 9-15). During that period, two pressure
9-16
-------
Section 9—Results and Discussion: Determine If Observed Changes in Indoor Air Concentration of
Volatile Organics of Interest are Mechanically Attributable to Changes in Vapor Intrusion
IO-
CS
CL
TO
"E
0)
5—
0)
ifc 0-
b
CL
-10"
Figure 9-15. Extreme Event 1: subslab vs. basement differential pressure (positive difference
indicates flow into the structure).
maximums and two pressure minimums were observed with the first maximum and first minimum
showing evidence that the pressure sensor had reached or exceeded its maximum design range (-15 Pa to
15 Pa) for a sustained period (for example, a pressure maximum that included 92 successive data points
over approximately 21 hours). The second maximum and second minimum are much more narrow peaks.
The exact start time of this event is uncertain because of missing data, perhaps related to the ice storm that
occurred on February 1, 2011. During much of the time period of this event (the weekend of February 5
and 6), we have no record of staff working at the facility, suggesting that human causes for the extreme
pressures are unlikely.
The second extreme event lasted fewer than 2 days, from the evening of Monday March 7 to the
afternoon of Tuesday March 8 (Figure 9-16). This was a period of intensive on-site work with staff
working on site about 12 hours a day, which suggested some possibility of an association between the on-
site work and the pressure observations. However as discussed in Section 9.1.5, because similar
differential pressure events occurred during time periods when the site was not occupied, and human
activities were more likely to have caused brief pressure fluctuations rather than long-term sustained
differentials, we do not believe that site work was the likely cause. The end of this extreme event record is
Extreme Event 1
SubSlab.Vs.Basement
•r $
* #
*
* 1 *
mj*
*
•
*
;
*
*
•
*
*
*
*
• •
*
*
* •
»
*
* ****
V
A* \
* * *
: *
• '¦(
•«
. •
4
*
• .
¦ 4
~ . <
£
***
>
¦
* !
<$>
^
9-17
-------
Section 9—Results and Discussion: Determine If Observed Changes in Indoor Air Concentration of
Volatile Organics of Interest are Mechanically Attributable to Changes in Vapor Intrusion
Extreme Event 2
SubS lab. Vs. Basement
*
*
#
* *
*
%
•
t *
* *
* *
*
*
*
t
* *
*
*
*
*
* 1
*
*
~
•
«
*
* •
*
*
*
*
* .
! *
r ;
10
ro 5
(li-
ra
c
D
i_
D
t
b
-------
Section 9—Results and Discussion: Determine If Observed Changes in Indoor Air Concentration of
Volatile Organics of Interest are Mechanically Attributable to Changes in Vapor Intrusion
Extreme Event 3
10-
<0
Q_
c
(Si
1 '
>
•
*
*
#
*
*
*
*
*
*
*
#
*
*
W *
.•:
*
t
+
*
*
"if *• *
a«V
t
• *
*
*
. *
* t
+
* *
i
* *
*
* *
** •
**
+
•
Figure 9-17. Extreme Event 3: subslab vs. basement differential pressure (positive difference
indicates flow into the structure).
9-19
-------
Section 9—Results and Discussion: Determine If Observed Changes in Indoor Air Concentration of
Volatile Organics of Interest are Mechanically Attributable to Changes in Vapor Intrusion
The fourth event (Figure 9-18) lasted just 8 hours, beginning with a brief high positive differential
pressure spike and ending with a period of high negative differential pressure that ended abruptly at
approximately 9:30 PM local time (later than our staff would typically be at the house). From that we
conclude that a human cause was unlikely. There was a period of snow flurries late that day, which, as
discussed Section 10, may be associated with vapor intrusion.
15
10-
cc
CL
TO
c
-------
Section 9—Results and Discussion: Determine If Observed Changes in Indoor Air Concentration of
Volatile Organics of Interest are Mechanically Attributable to Changes in Vapor Intrusion
The fifth event lasted about 5 days (Figure 9-19) and has several peaks and troughs, including three
separate periods when it appears that the instrument went off range below -15 Pa, as well as two briefer
periods that may have been off range at >15 Pa. The long period of extreme values argues against a
human artifactual cause. The occurrence of multiple maxima and minima in differential pressure suggests
that the associated predictor variable(s) should also display multiple maxima and minima at this time
period.
Extreme Event 5
SubSlab.Vs.Basement
rfO
-------
Section 9—Results and Discussion: Determine If Observed Changes in Indoor Air Concentration of
Volatile Organics of Interest are Mechanically Attributable to Changes in Vapor Intrusion
The sixth and seventh events (Figure 9-20) were very close to each other, taking place over about a week.
During that time, there were at least five separate peaks to >10 Pa and three troughs to <-10 Pa. This
event included several periods during which the pressure likely exceeded the sensors" capacity >15 Pa. As
discussed above, the long duration of the events argues against a cause due to human disturbance and
suggests that linked predictor variables should also display multiple maxima and minima.
Extreme Event 6/7
SubSlab.Vs.Basement
\\ \
-------
Section 9—Results and Discussion: Determine If Observed Changes in Indoor Air Concentration of
Volatile Organics of Interest are Mechanically Attributable to Changes in Vapor Intrusion
We examined visually the high resolution metrological data for each event, seeking coincidences in time
between marked changes in metrological parameters and the differential pressure events. Plots that did
not suggest a relationship are not shown.
Event 1 appears to coincide with a substantial drop in barometric pressure and a snowstorm (Figures 9-21
and 9-22). Event 2 appears to begin slightly ahead of a substantial drop in barometric pressure and
increase in temperature (Figures 9-23 and 9-24). Event 3 also appears to begin slightly ahead of a
substantial drop in barometric pressure and to coincide with a rise in temperature as well as a wind shift.
(Figures 9-25 through 9-27). The online GC was available during Event 3 and appears to show some
correlation of rising PCE to positive differential pressure (toward the structure).
30.6
30.3
30.0
29.7
10
0
-10
20
0
-20
-40
5.0
2.5
0.0
-2.5
-5.0
-7.5
Figure 9-21. Detailed time series of unusual pressure Event 1 showing barometric pressure
changes (barometric pressure in inches of Hg, differential pressure in Pa).
Event 1
Bar..
SubSlab.Vs.Basemerit
ifH .
* i
LS
SubSlab.Vs. Basement.dPdt
dPdt
u
¦A
-------
Section 9—Results and Discussion: Determine If Observed Changes in Indoor Air Concentration of
Volatile Organics of Interest are Mechanically Attributable to Changes in Vapor Intrusion
0.06"
0.04-
0.02"
o.oo-
6"
4-
2-
o-
10"
0-
-10-
Figure 9-22. Detailed time series of unusual pressure Event 1 showing precipitation events
(rainfall in inches, snow depth in inches, differential pressure in Pa).
Event 1
Rain
*
*
•
—
«¦»
SnowDepth
*
*
* *
*
SubSlab.Vs.Basement
#
%
--
#
t
%
*
i:
•!
r
M m*
il
* •
¦
q\ \\
9-24
-------
Section 9—Results and Discussion: Determine If Observed Changes in Indoor Air Concentration of
Volatile Organics of Interest are Mechanically Attributable to Changes in Vapor Intrusion
30.3
30.2
30.1
30.0
29.9
29.8
10
5
0
-5
10
0
-10
2
0
-2
Event 2
Bar..
i
%
\
SubSlab.Vs.Basement
,,, j 1
'
* i
j
* * *
} ;;
-i- */
SubSlab.Vs Basement.dPdt
*
„ t
* '
* - j
*
a *y. j^lrr
4
* .
* * *
~ *
#
-
•
* *
*
dPdt
** - ** ,
'k r..y\V
«~ *
* - t */
v* •
V \ *
*
** • *.*
*:•* *
* » H " •
* + • * \
*
*
* *:
*
1 —
1
<$>
<$>
Figure 9-23. Detailed time series of unusual pressure Event 2 showing barometric pressure
changes (barometric pressure in inches of Hg, differential pressure in Pa).
9-25
-------
Section 9—Results and Discussion: Determine If Observed Changes in Indoor Air Concentration of
Volatile Organics of Interest are Mechanically Attributable to Changes in Vapor Intrusion
10-
5-
o-
-5
50"
45-
40-
35-
30"
5.0-
2.5
0.0-
-2.5-
-5.0-
Event 2
SubSlab.Vs. Basement
*—
r
* r
* *
*
*
¦ mm *
t
* *
*
* *
\
*
*
t * *
.** * *
J-
y* *
Temp Out
/V
.
*
* i ^
/
v
- *
*
\
* v
* V. _
*
*
*
J
Temp.Out.dTcit
*4
w*
**
>v
+ * * • •
* - Ll**#* *
* /
— *
* *
- * ** . .*~*¦
«/vv .; .
* *-• * **•
* *
+* ******
*
. ' f
* *
*
* & M
* > f *
* *
*
*
&
&
&
&
&
Figure 9-24. Detailed time series of unusual pressure Event 2 showing temperature changes
(differential pressure in Pa, temperature in °F).
9-26
-------
Section 9—Results and Discussion: Determine If Observed Changes in Indoor Air Concentration of
Volatile Organics of Interest are Mechanically Attributable to Changes in Vapor Intrusion
30.6-
30.4-
30.2"
30.0-
29.8-
10-
0-
-10-
25"
o-
-25-
-50"
-75-
2.5"
o.o-
-2.5-
-5.0-
Figure 9-25. Detailed time series of unusual pressure Event 3 showing barometric pressure
changes (barometric pressure in inches of Hg, differential pressure in Pa).
Event 3
Bar..
SubSlab.Vs.Basement
\ r
;
; .
1 ' ¦ ¦
>
V"
SubSlab.Vs.Basement dPdt
*
l ¦ u
ff.r rna.in,
¦ 1 1
* *
• *
*
dPdt
*
% * *** *v!
* * *•
* * •*. .*1
¦¦ y
*
*
\T- \* \*>
^ ^
9-27
-------
Section 9—Results and Discussion: Determine If Observed Changes in Indoor Air Concentration of
Volatile Organics of Interest are Mechanically Attributable to Changes in Vapor Intrusion
10-
0-
-10-
60"
50-
40-
30-
20-
5-
0-
-5-
-10-
Figure 9-26. Detailed time series of unusual pressure Event 3 showing temperature changes
(differential pressure in Pa, temperature in °F).
Event 3
SubSlab.Vs.Basement
k : *
#
*
_t_J * ,
* * fc
&
* »
Temp.Out
A s~~
' \
. A,
A.
\
Temp.Out.dTdt
*
#
*
»v V
^ * *
* 1 *
¦r*
* v£
7»
*
*
*
*
I l
_\® Al A
b
o*
9-28
-------
Section 9—Results and Discussion: Determine If Observed Changes in Indoor Air Concentration of
Volatile Organics of Interest are Mechanically Attributable to Changes in Vapor Intrusion
Event 3
Chanqe.Dir
100-
-100-
Hi.Dir
300
200
100
SubSlab.Vs Basement
10-
-10-
Wind.Dir
300-
200-
100-
Figure 9-27. Detailed time series of unusual pressure Event 3 showing wind direction variables
(wind direction-related variables in degrees, differential pressure in Pa).
9-29
-------
Section 9—Results and Discussion: Determine If Observed Changes in Indoor Air Concentration of
Volatile Organics of Interest are Mechanically Attributable to Changes in Vapor Intrusion
16
12
8
4
10
0
-10
3.0
2.5
2.0
1.5
1.0
Figure 9-28. Detailed time series of unusual pressure Event 3 showing PCE and radon.
Event 3
Radon
&
SubSlab.Vs.Basement
II
^4
r
\
*:
*
i
Tetrachloroethene
A*
9-30
-------
Section 9—Results and Discussion: Determine If Observed Changes in Indoor Air Concentration of
Volatile Organics of Interest are Mechanically Attributable to Changes in Vapor Intrusion
In reviewing the stacked graphs seeking evidence of a cause of these extreme differential pressure values,
we noted that Event 4 appears to be associated with a significant drop in temperature and a substantial
wind direction shift that was accompanied by a period of calm wind (Figure 9-29 through Figure 9-31).
Others have noted an association between a period of calm and a wind shift.21 Since subslab to interior
differential pressure spatial patterns can vary with wind direction (U.S. EPA, 2012d), this association in
the data is reasonable.
Event 5, which shows multiple peaks and troughs in differential pressure, appears to be related to a series
of weather events. It appears to have been triggered by a light snow (Figure 9-32). The middle
fluctuations of the event appear to track with a period of high and shifty winds, including some gusts over
30 mph (Figures 9-33 and 9-34). As discussed above, exterior winds are known to be associated with the
subslab to interior differential pressure (U.S. EPA, 2012d). The last fluctuation appears to coincide with a
substantial temperature rise (Figure 9-35). Exterior temperature changes are expected to cause changes in
the strength of the stack effect. Event 5 seems to have caused two spikes in the indoor radon
concentration that each reached 15 pCi/1 (Figure 9-36).
21gcc. glendale.edu/fire/Documents/ClassMaterials/s190-2.pdf
9-31
-------
Section 9—Results and Discussion: Determine If Observed Changes in Indoor Air Concentration of
Volatile Organics of Interest are Mechanically Attributable to Changes in Vapor Intrusion
Event 4
Change.Dir
100-
50-
-50-
-100-
Dir
300
200-
100-
SubSlab.Vs. Basement
-5-
-10-
-15-
Wind Dir
300-
200-
100-
Figure 9-29. Detailed time series of Event 4 showing wind direction variables (wind direction-
related variables in degrees, differential pressure in Pa).
9-32
-------
Section 9—Results and Discussion: Determine If Observed Changes in Indoor Air Concentration of
Volatile Organics of Interest are Mechanically Attributable to Changes in Vapor Intrusion
Event 4
Hi Speed
*
*
*
*
* * *
ft ft
ft*
*
* *
*
* * *
*
* *
-V
*
44* *
* * *
44'
*
»
*
44 *
ft * *
** * *
** #
• •
44**
* * 44 *
* # *** * *
*4 4ft *
* 44 4
*
* *
*
144 4
* «
44 44 4 4 4
4 * 4
* 44 #
*
SubSlab.Vs.Basement
*
*
* -
*
*
*
*
*
*
* 4 *
Am* 4
Wind.Speed
*
—
*
4 44 * 4*
4 44
44 * *
444 44
44 ft4 444 444
44
A°>
yS
yS
yS
&
&
Figure 9-30. Detailed time series of Event 4 showing wind speed variables (wind speed variables
in MPH, differential pressure in Pa).
9-33
-------
Section 9—Results and Discussion: Determine If Observed Changes in Indoor Air Concentration of
Volatile Organics of Interest are Mechanically Attributable to Changes in Vapor Intrusion
Event 4
15
SubSlab.Vs.Basemen
t
*
*
•t
*
*
*
*
*
Temp.Out
I ^*** A.
/
\
1
*.
***
\
/
V
" *
'\y
Temp. Out.dTdt
*
*
• *
* * 4
* S
V* *
*
•j, m*
* .V *
*
* ** *
V* *
» * t
•
*
*******
v *
h ^
*
* V
•.*
*
**•
*
*
«**'
V 4 4 4
•• . *
*
*
*
-15
20
15
10
&
&
^ ^ ^ ^
yS
&
&
Figure 9-31. Detailed time series of Event 4 showing temperature variables (differential pressure
in Pa, temperature in °F).
9-34
-------
Section 9—Results and Discussion: Determine If Observed Changes in Indoor Air Concentration of
Volatile Organics of Interest are Mechanically Attributable to Changes in Vapor Intrusion
0.06"
0.04-
0.02"
0.00-
1.00"
0.75"
0.50"
0.25"
o.oo-
10-
o-
-10-
Figure 9-32. Detailed time series of Event 5 showing precipitation event (rainfall in inches, snow
depth in inches, differential pressure in Pa).
Event 5
Rain
*
*
*
h*
* IOI CZ
1 ¦ ¦ ¦ ¦ *•
. - -
*
s
nowDepth
* *
* ~—
SubSlab.Vs. Basement
i. 1 P- . ,
*
*
* *
"r V
"J
y
*
i * *
#
• ¦
•
*
X
•
* •
* * *
*
*
t
—
_ *1
&
&
&
9-35
-------
Section 9—Results and Discussion: Determine If Observed Changes in Indoor Air Concentration of
Volatile Organics of Interest are Mechanically Attributable to Changes in Vapor Intrusion
30
20
10
0
10
0
-10
15
10
5
0
Figure 9-33. Detailed time series of Event 5 showing wind speed variables (wind direction-related
variables in degrees, differential pressure in Pa).
Event 5
Hi.Speed
•>7 *\
.V. t.
. *r
. .. h . «v
.& ** r
r 4
£_
'J r> V
V
V
. -
SubSlab.Vs.Basement
k. ,
•
*
* •
V
t * *
'• S
\ : *
y
* .
*
* ¦
X
t
*
* *
* * *
*
*
i
—
Wind.Speed
*
# ** **
w * #
• * • *
*
* *
** *»+* m
#
*# *
• «* * •
* •
* *
*«*«e t *
* * ***** • • *
«* Mt * * +
** * * *M4
• *
• <¦ * *
¦ +4+ *
* •** « m
•H * t *
* H ««* + » *
M * * * • *
« •
h * * *
* 4*
•
* 4
~
9-36
-------
Section 9—Results and Discussion: Determine If Observed Changes in Indoor Air Concentration of
Volatile Organics of Interest are Mechanically Attributable to Changes in Vapor Intrusion
Event 5
Change. Dir.
100 -
-100 4
Hi.Dir
300-
200-
100"
SubSlab.Vs.Basement
-10-
Wind.Dir
300
200
100
Figure 9-34. Detailed time series of Event 5 showing wind direction variables.
9-37
-------
Section 9—Results and Discussion: Determine If Observed Changes in Indoor Air Concentration of
Volatile Organics of Interest are Mechanically Attributable to Changes in Vapor Intrusion
10-
o-
-10"
60-
50"
40-
30"
6-
4"
2"
0-
-2"
Figure 9-35. Detailed time series of Event 5 showing temperature variables (differential pressure
in Pa, temperature in °F).
Event 5
SubSlab.Vs.Basement
w "\
ft t -
*
*
* *
* *
I. -~s
I I *
y
*
* ¦
.
*
* • I
*
t
—
1
remp.Out
A^vA°-
*
*
*
*
*
>V>x.
*
f
/
Temp.Out.dTdt
—%—
t*
* *
*
/
*
*.
*- . .
>, , a .*
* ~ * #
». * • • t %• i
* V ~ •
* * *T •
»v. . * ¦
**
. % *
~ .
*
. - *
&
&
&
9-38
-------
Section 9—Results and Discussion: Determine If Observed Changes in Indoor Air Concentration of
Volatile Organics of Interest are Mechanically Attributable to Changes in Vapor Intrusion
15
10
5
10
0
-10
2.2
2.0
1.8
1.6
1.4
Figure 9-36. Detailed time series of Event 5 showing PCE and radon.
Event 6/7 is more difficult to interpret. Although we initially numbered these events differently, we
discuss them together here because they occur in short succession. The event shows some association
with substantial barometric pressure changes, with a falling barometer being associated with positive
differential pressure as was seen in other events (and is expected from theory) (Figure 9-37). This event
appears also to be associated with a light snow (Figure 9-38). There also appears to be some coincident
events in the wind speed and direction plots (Figures 9-39 and 9-40). Event 6/7 may have a
corresponding radon and PCE peak, but the data for these parameters are quite noisy during this time
period, so the evidence shown in Figure 9-41 is weak.
Taken together, the detailed examination of these time series for multiple extreme pressure events
suggests that all of these events have some likely relationship to meterological variables. However, the
particular meterological variables causing each pressure event may be different and multiple. This is not
suprising because changes in temperature, barometric pressure, wind speed, and direction are typically
associated with a frontal passage or storm event.
Event 5
Radon
~i*
• * *
-¦W
SubSlab.Vs. Basement
i * r*
* m s
• * . *
4—*-
*
t
Tetrachloroethene
******
yS
Jp
9-39
-------
Section 9—Results and Discussion: Determine If Observed Changes in Indoor Air Concentration of
Volatile Organics of Interest are Mechanically Attributable to Changes in Vapor Intrusion
30.4
30.3
30.2
30.1
30.0
29.9
10
0
-10
40
0
-40
4
2
0
-2
Figure 9-37. Detailed time series of Event 6/7 showing barometric pressure (barometric pressure
in inches of Hg, differential pressure in Pa).
Event 6/7
Bar..
-
\
V /\ / V-^v'
—V
V w
r.
SubSlab.Vs.Basement
! *#2— r
1
I
-J
* *
i_t> * ,'v
s ~
.* —
J \
! m
s
1
£
t
SubSlab.Vs.Basement.dPdt
j
*
~
*
„» *
'¦* j .
*
* «¦
!¦ ^ *
*
*
*
dPdt
•J
&
* *+ •
4
f
.
1 A3
* * * *
*>• A f %
t *v$
'fit
; JL
t*n *
t
* v<
* t .
v* tijr * *
.
* * ;
**
* *•
*
~
J
•
^ ^
K
-------
Section 9—Results and Discussion: Determine If Observed Changes in Indoor Air Concentration of
Volatile Organics of Interest are Mechanically Attributable to Changes in Vapor Intrusion
Event 6/7
0.06-
0.04-
0.02"
0.00-
1.5-
1.0"
0.5-
o.o-
10-
o-
-10
Rain
*
*
*
¦ ¦ ii ~-
C
SnowDepth
~
SubS lab. Vs. Basement
-t
: hjr~
*
}
—
%
* ?
I *
*
*
dM
*
f
I
I
*
4
i
, *
A*
•wt
4
4
r
<$> v**"
.
-------
Section 9—Results and Discussion: Determine If Observed Changes in Indoor Air Concentration of
Volatile Organics of Interest are Mechanically Attributable to Changes in Vapor Intrusion
Event 6/7
20-
10-
o-
10-
o-
-10-
12.5
10.01
7.5
5.0"
2.5-
0.0-
Hi.Speed
*
*
•
~
H *
-> \
*
1
•
. vj • ~ .<*-
• lit'..* t \
* « * * ~
m *m **• • t ~ *
«v *•» vr ,
W *« * * • + **
*
*
*
* P +
•
**
* . •**
** «»•
~ * M4
* +*
«~ *
* «~
.*
•#
• **
A «£.
*
«*
~ •
* *
~ *
•
• HH
HH
~ *
r IV
•• %
M * ** *•
X Jt
•* ~
* * *
t \ *
» •
*' * •
* «M
i • *
% *
•
*
* .
*, ^ . # «T *
1 * •* * ~
• *t 4 *
*•< *~.*
K 49# *•
* »
SubSlab.Vs.Basement
•t
: *jrr-
sr.
%
j
i
) :»•
wfc, * * t
I.
•
*
f K
fii
* f
* w
%
«
•
«
i
\
«
¦
n
V
Vind.Speei
i
*
* .V
~ ** •«
M ~&#
• 4 * *
*
•
**
~
* *• •w
MM !¦ a ¦¦
* *» *
» ¦* «*¦> •
-at- *
• 4*
4# * * *
*
*
+ * !¦
*«
•* *». •>
•
** * **
* * * * • *
mm
*
¦
4* 4
•a* *
• ~
» «»—
*• *
tfb \\ i\b \^>
Figure 9-39. Detailed time series of Event 6/7 showing wind speed variables (wind direction-
related variables in degrees, differential pressure in Pa).
9-42
-------
Section 9—Results and Discussion: Determine If Observed Changes in Indoor Air Concentration of
Volatile Organics of Interest are Mechanically Attributable to Changes in Vapor Intrusion
Event 6/7
Change.Dir
Hi.Dir
SubSlab.Vs. Basement
Wind Jir
Tf
In r
cSb \\ \b
Figure 9-40. Detailed time series of Event 6/7 showing wind direction variables (wind direction-
related variables in degrees, differential pressure in Pa).
9-43
-------
Section 9—Results and Discussion: Determine If Observed Changes in Indoor Air Concentration of
Volatile Organics of Interest are Mechanically Attributable to Changes in Vapor Intrusion
Event 6/7
Radon
i i i
M.
it* * Z
M&A
*
* * jfl
* -r *
* . •
_i *_
*
A $
• * $ '
* %* *f
*
*
2L
tj* _ * - *
V**
* i -
fei. *
* *
f •
SubSlab.Vs. Basement
•*
irr~
\
-
I •
•
*
—
i
•
!
i>:
• *
Tetrachloroethene
*
*
*
•
* * * *
* • * *
. * *
* 1*. ,.
*
*
*
*
1 *
* * *
t *
..T 4
*
«** %
*
* **
*%• *
* *
^ * H **
#
w *
* * *•
*
*
A*
Figure 9-41. Detailed time series of Event 6/7 showing radon and PCE.
9.2 Influence of Meteorological Conditions on Indoor VOC Concentration;
Mitigation Off
In Section 6, we examined this topic based on a review of patterns in the high temporal resolution on-site
GC data and found evidence that peaks in wind speed and snow events were coincident in time with peak
VOC concentrations in WP-3 and indoor air.
9.2.1 Temperature Effect on VOCs
The effect of meteorological variables can also be examined using the weeklong passive Radiello data set.
We visually screened our meteorological variables and present here those with an evident correlation. As
shown in Figure 9-42, there is a visual correlation between PCE and heating degree days, a measure of
9-44
-------
Section 9—Results and Discussion: Determine If Observed Changes in Indoor Air Concentration of
Volatile Or games of Interest are Mechanically Attributable to Changes in Vapor Intrusion
sustained cold weather.2 Because the 422 side had thermostatically controlled temperature in winter, this
result is expected and is similar in appearance to that based on the analysis of the stack effect driving
force provided in Section 10.3 of our previous report (U.S. EPA, 2012a). The analysis according to
heating degree days may be a useful one to practitioners and in public communication because it is an
easily calculated and widely reported parameter.23
20
18
16
14
E
2 12
c
&
1 10
c
s
B
8 a
2
6
Weekly Total Heating Degree Days vs. PCE Concentration (Radiello)
422 Base South; Mitigation Off or Not Installed Data Only
JL.
~
*
~
~
~
~
**
~
•J
J ^ '~ f
~
4-
ft
ffX vtt* ***~# ~ ~~ Vn # ~
~
~
~
V
0 50 100 150 200
Heating Degree Days, per Week
250 300 350
Figure 9-42. XY graph of total heating degree days per week vs. weekly PCE concentration
(Radiello data).
22One heating degree day is the amount of heat required to keep a structure at 65°F when the outside temperature remains one degree below the
65°F threshold for 24 hours.
23The stack effect driving force calculation discussed in U.S. EPA (2012a) is a slightly more complex calculation (stack effect is proportional to
the square root of the indoor outdoor temperature difference divided by the indoor temperature). The data set shown here includes some
additional weeks and uses the heating degree day function, which is calculated using the difference between 65°F and average exterior
temperature for the half-hour time interval, divided by 48 (to correct from hours to days). The degree days for each half hour are then totaled
over the week.
9-45
-------
Section 9—Results and Discussion: Determine If Observed Changes in Indoor Air Concentration of
Volatile Organics of Interest are Mechanically Attributable to Changes in Vapor Intrusion
Similarly, the concentration of chloroform is visually correlated to temperature but shows a somewhat
different curve shape than PCE (Figure 9-43). Four data points that appear above the general trend have
been marked in the graph to indicate that they were collected in January through March 2011 before a
floor drain sealing procedure was implemented (see discussions of this matter in Sections 3.2 and 10.8 of
U.S. EPA [2012a]). The sewer line has been discussed as apotential source of chloroform based on
measurements described in the previous report (U.S. EPA, 2012a), although the sealing procedure appears
to have been effective in blocking any subsequent vapor entry through the drains in the house. However,
as described in the previous report, the contribution of sewer gases to chloroform indoor air
concentrations cannot be ruled out prior to when the floor drain was sealed in March 2011.
Weekly Total Heating Degree Days vs. PCI Concentration (Radielto)
422 Base South; Mitigation Off or Mot Installed Data Only
1.8
1.6
1.4
V2
c
~
I 1
*-•
I
S 0.8
i
o
"S
5 0.6
§
0.2
~
J A A
(
7^Z \
f Before Floor Drain Seated \
~
A. Jk.
* ~
~ *
~ 4 ~«, ~
~ * ~
* ~
~
^ ~ #
~ A * *
~
*
—* *
~ ~
* *
~
~
50
100 ISO 200
Heating Degree Days per Week
250
300
350
Figure 9-43. XY graph of heating degree days vs. chloroform concentration (weekly Radiello
data).
9-46
-------
Section 9—Results and Discussion: Determine If Observed Changes in Indoor Air Concentration of
Volatile Organics of Interest are Mechanically Attributable to Changes in Vapor Intrusion
9.2.2 Barometric Pressure Effect on VOCs
The 11 weeks with the highest PCE concentrations were all characterized by average barometric pressures
in a relatively narrow range from 30.01 to 30.18 inches of mercury. There is no readily apparent direct
physical mechanism to explain why these midrange barometric pressures may be associated with the
highest PCE concentrations. Chloroform did not show the same relationship (graph not shown for
brevity). It is possible that this is a fortuitous result or that this range of barometric pressures may be
serving as a marker for some more causative meteorological variable.
Weekly Barometric Pressure Average vs. PCE Concentration (Radiello)
422 Base South; Mitigation Off or Not Installed Data Only
18 ¦
16 ¦
14 ¦
f
"a 12
&
| 10 ¦
s
u
S
o
H 8
U
Ph
6 ¦
4 ¦
2 ¦
0 ¦
2S
~
~
~
~ ~
~~~ ~ ~ ~ ~
.7 29.8 29.9 30 30.1 30.2 30.3 30.4
barometric pressure average in Hg
Figure 9-44. XY graph of average weekly barometric pressure vs. PCE concentration.
9.2.3 Precipitation Effects on VOCs
A rough relationship is visually evident between snow depth on the ground averaged over the week of
sampling and the PCE concentration in indoor air measured with a passive sampler (Figure 9-45). This
effect was less evident for chloroform (Figure 9-46) and absent for radon (Figure 9-47). In Section 6, we
present data that indicate a partial agreement in time between the occurrence of even light snowstorms
(even those not leaving an accumulation) and peaks in chloroform and PCE on-site GC data as well as
radon.
The relationship between rainfall and PCE concentration in our data set is unclear (Figures 9-48 and
949). Although some of the highest PCE concentrations are associated with low rain weeks, which is
almost surely attributable to those weeks being very cold, the rain sensor on the Vantage Vue instrument
is not designed to melt snow. The chloroform data show no evident correlation to rainfall.
9-47
-------
Section 9—Results and Discussion: Determine If Observed Changes in Indoor Air Concentration of
Volatile Organics of Interest are Mechanically Attributable to Changes in Vapor Intrusion
Weekly Average Snow Depth vs. PCE Concentration (Radiello)
422 Base South; Mitigation Off or Not Installed Data Only
18
16
14
f
"a 12
A
C
£
| io
S
V
u
S
o
H 8
U
Ph
6
4 <
j
~
~
~
~ ~
~
0.5 1 1.5 2 2.5 3 3.5
Snow Depth on Ground (inches)
Figure 9-45. XY graph of weekly average snow depth vs. PCE concentration.
9-48
-------
Section 9—Results and Discussion: Determine If Observed Changes in Indoor Air Concentration of
Volatile Organics of Interest are Mechanically Attributable to Changes in Vapor Intrusion
Weekly Average Snow Depth vs. Chloroform Concentration (Radiello)
422 Base South; Mitigation Off or Not Installed Data Only
~
<
~
1
~ • ~ •
Sm
s
t ~
O 4
U
|
~ ~
; ~
~
«2
2 u-b .
_© <
2 <
U ;
~
* ~
: ~
f
0.5 1 1.5 2 2.5 3 3.5
Snow Depth on Ground (inches)
Figure 9-46. XY graph of weekly average snow depth vs. chloroform concentration.
9-49
-------
Section 9—Results and Discussion: Determine If Observed Changes in Indoor Air Concentration of
Volatile Organics of Interest are Mechanically Attributable to Changes in Vapor Intrusion
Weekly Average Snow Depth vs. Radon Concentration (Electret)
422 Base South; Mitigation Off or Not Installed Data Only
30
25
20
1 io
pj
~~
~ ~
~
Snow Depth on Ground (inches)
Figure 9-47. XY graph of weekly average snow depth vs. radon concentration.
9-50
-------
Section 9—Results and Discussion: Determine If Observed Changes in Indoor Air Concentration of
Volatile Organics of Interest are Mechanically Attributable to Changes in Vapor Intrusion
Weekly Rainfall vs. PCE Concentration (Radiello)
422 Base South; Mitigation Off or Not Installed Data Only
16 ¦
14 ¦
f
"a 12
A
C
£
I io
S <
QJ
(J
s
o
H 8
U
Ph
<
6 ¦
4 ¦
i
~
~
~
~
~
* . ~
~~ ~
0123456789
Rainfall (inches)
Figure 9-48. XY graph of total weekly rainfall vs. PCE concentration in indoor air.
9-51
-------
Section 9—Results and Discussion: Determine If Observed Changes in Indoor Air Concentration of
Volatile Organics of Interest are Mechanically Attributable to Changes in Vapor Intrusion
Weekly Rainfall vs. Chloroform Concentration (Radiello)
422 Base South; Mitigation Off or Not Installed Data Only
Chloroform concentration (jwg/m3)
poop J—1 J—1 J—1 [-
:> NJ CT> 00 H> NJ CT> 0
-A.. 1 A. A. 1 A. A. 1 A.
~
~
f
~
~~ * ~ / * ~ ~
~ ~
~V * ~ ~ ~
» * ~ 4?
~ ~ ^ ~
~ ~ ~ * ~ ~
~a . ~ ~
~ ~ ~ ~
0123456789
Rainfall (inches)
Figure 9-49. XY graph of total weekly rainfall vs. chloroform concentration in indoor air.
9.2.4 Effect of Wind on VOC Concentrations
The six highest weekly PCE concentrations in the 422 basement occurred during average wind weeks that
were generally westerly, ranging from 205 to 296 degrees (Figure 9-50). That pattern of directionality
agrees with the observations of high differential pressures into the 422 basement during winds in a similar
range of direction (see Section 9.1.4), an analysis that used daily averaged data. Chloroform
concentrations; however, are apparently independent of wind direction (Figure 9-51). West winds do
seem to have some effect on increasing radon concentrations (Figure 9-52) as measured by weekly
electret samplers, but the effect is not as dramatic as the effect on PCE. The same trend of west winds
increasing radon concentrations was also seen in the daily AlphaGUARD data (Figure 9-14b).
The observed wind direction effects for PCE and radon agree with the predictions of a 3D numerical
model (Abreu and Johnson, 2005) presented in U.S. EPA (2012d). As seen in Figures 9-53 and 9-54,
VOC concentrations and subslab to indoor differential pressures are expected to increase on the side of
the building opposite to the direction from which the wind is blowing. In this case, the 422 side of the
duplex lies east of the 420 side. Therefore, the observation of higher vapor intrusion on the 422 side of the
duplex under westerly winds fits the model prediction. Note that the modeled wind velocities in U.S. EPA
(2012d) were 5 m/s (11 mph), which is very comparable to monthly average wind speeds in Indianapolis
(4 to 6 m/s).
9-52
-------
Section 9—Results and Discussion: Determine If Observed Changes in Indoor Air Concentration of
Volatile Organics of Interest are Mechanically Attributable to Changes in Vapor Intrusion
Weekly Average Wind Direction vs. PCE Concentration (Radiello)
422 Base South, Mitigation Off or Not Installed Data Only
18 ¦
16 ¦
14 ¦
f
a
| io ¦
c
u
S
o
H 8
u
Ph
6 ¦
4 ¦
2 ¦
<
~
~
~
~ ~
0 50 100 150 200 250 300 350 400
Wind Direction (Degrees)
Figure 9-50. XY graph of wind direction vs. PCE concentration in indoor air in 422 basement.
9-53
-------
Section 9—Results and Discussion: Determine If Observed Changes in Indoor Air Concentration of
Volatile Organics of Interest are Mechanically Attributable to Changes in Vapor Intrusion
Weekly Average Wind Direction vs. Chloroform Concentration (Radiello)
422 Base South, Mitigation Off or Not Installed Data Only
1.8
~
~
~
• • ~
~
~ ~
—~—
~
~
~
~
~
~ H
~~ ~ ~
~
~
~
~~~
~ ~~
~
~ 4
~
~ ~
~
~
~
~
~
1 1-2
0.6
0.4
50 100 150 200 250 300 350 400
Wind Direction (Degrees)
Figure 9-51. XY graph of wind direction vs. chloroform concentration in indoor air in 422
basement.
9-54
-------
Section 9—Results and Discussion: Determine If Observed Changes in Indoor Air Concentration of
Volatile Organics of Interest are Mechanically Attributable to Changes in Vapor Intrusion
Weekly Average Wind Direction vs. Radon Concentration (Electret)
422 Base South, Mitigation Off or Not Installed Data Only
35
30
25
g 20
10
~ • ~~
~ ~
~ ~
~ ++
~
~ **
~ ~
~ ~
50
100
150
200
250
300
350
Wind Direction (Degrees)
Figure 9-52. XY graph of wind direction vs. radon concentration in indoor air in 422 basement.
9-55
-------
Section 9—Results and Discussion: Determine If Observed Changes in Indoor Air Concentration of
Volatile Organics of Interest are Mechanically Attributable to Changes in Vapor Intrusion
Plan View at Ground Surface
Building
Building
Plan View at 0.2 m bgs (sub-slab depth)
N
Pa
— — — Building footprint
Figure 9-53. Modeled effect of building wind loads on ground surface and subslab gauge
pressure distribution (adapted from U.S. EPA, 2012d).
The gauge pressure contour lines are in Pa; negative values reflect over-pressurization and positive values
reflect under-pressurization. Wind velocity was constant at 5 m/s (11 mph).
9-56
-------
Section 9—Results and Discussion: Determine If Observed Changes in Indoor Air Concentration of
Volatile Organics of Interest are Mechanically Attributable to Changes in Vapor Intrusion
Recalcitrant VOC
Aerobically Biodegradable VOC ( X= 0.18 h" )
mg/L
40 20
150
130
110
90
70
50
30
10
1
0.1
0
Building footprint
Figure 9-54. Modeled effect of building wind load on subslab soil vapor distribution for
recalcitrant and aerobically biodegradable VOCs (adapted from U.S. EPA, 2012d).
The vapor concentration contour lines are in mg/L. The source vapor concentration is 160 mg/L. Wind
velocity was constant at 5 m/s (11 mph).
9-57
-------
Section 9—Results and Discussion: Determine If Observed Changes in Indoor Air Concentration of
Volatile Organics of Interest are Mechanically Attributable to Changes in Vapor Intrusion
A more subtle aspect of the data that may or may not be significant is that the radon maximum appears to
occur under winds with more of a northwesterly component (245 to 309 degrees in Figure 9-14b; 264 to
305 degrees in Figure 9-52), while the PCE concentrations reach their maximum under somewhat more
southwesterly conditions of 205 to 296 degrees in Figure 9-50. It is possible that this is attributable to the
distribution of PCE in soil gas toward the southern edge of the building and the more uniform distribution
of radon along the south-north axis (see Sections 5 and 6 of U.S. EPA [2012a] for a discussion of this
spatial data). The 3D model would predict that winds with a southerly component might move more
VOCs toward the center of the building.
No clear, physically rational relationship between wind speed and PCE concentration could be discerned
(Figure 9-55), but a rough relationship of increasing chloroform with increasing wind speed was seen
(Figure 9-56).
Weekly Average Wind Speed vs. PCE Concentration (Radiello)
422 Base South; Mitigation Off or Not Installed Data Only
18 ¦
16 ¦
14 ¦
"8b 12 ¦
3
c
1 io ¦
s
u
S
o
H 8
6 ¦
4 ¦
2 ¦
~
~
~
~ ~
~ f ^
* * ~ * ~ * ~ ~
51234567
Wind Speed (mph)
Figure 9-55. XY graph of wind speed vs. PCE concentration in indoor air in 422 basement.
9-58
-------
Section 9—Results and Discussion: Determine If Observed Changes in Indoor Air Concentration of
Volatile Organics of Interest are Mechanically Attributable to Changes in Vapor Intrusion
Weekly Average Wind Speed vs. Chloroform Concentration (Radiello)
422 Base South; Mitigation Off or Not Installed Data Only
1.6 ¦
1.4 ¦
f 1.2 ¦
"eto
s
«5 1
s
V
u
s0-8'
i-
<2
©
i-
o
2 0.6 ¦
U
0.4 ¦
0.2 ¦
0 ¦
~
~
~ •• • ~
~ ~
~ ~ * % * 4 ~~
~ ~ ~ \ ~
~ 4 ~ ~
A ~ ~ ~ ~ ~
~ ~ /
* ~ ~~~~~~
» ~ ~~
~ ~ ~ ~
51234567
Average Wind Speed (mph)
Figure 9-56. XY graph of wind speed vs. chloroform concentration in indoor air in 422 basement.
9.3 Summary of Meteorological Effects on Vapor Intrusion—Evidence
Presented in Sections 6 and 9
Up to this point, we have analyzed relationships between meteorological variables and vapor intrusion in
several ways in Sections 6 and 9. Before entering into a quantitative time series analysis in Section 10, we
pause to provide the reader with a summary of the more qualitative analyses as Table 9-1. Similarities are
seen, as advective flow considerations would suggest, between the meteorological conditions that drive
high differential pressures and those that are associated with higher indoor and subslab concentrations.
Although this is speculative at this point, the lines of evidence can be used to select particular variables to
focus on in the quantitative analysis described in Section 10.
9-59
-------
Section 9—Results and Discussion: Determine If Observed Changes in Indoor Air Concentration of
Volatile Organics of Interest are Mechanically Attributable to Changes in Vapor Intrusion
Table 9-1. Summary of Qualitative Lines of Evidence for Meteorological Factors Influencing
Vapor Intrusion in This Study
(Blank cells reflect types of analysis not completed for a given parameter)
Cold Exterior
Temperatures
(or substantial
change in
temperatures)
Snow or Ice
Accumulation
on Ground
Rain
Events/
Rainfall
Amount
Barometric
Pressure
Changes
Snowfall
West to
NW Winds
High Wind
Velocity
Apparent Temporal
Association with
VOC Concentrations
in Indoor Air (Section
6, also U.S. EPA,
2012a)
Yes
Yes
Yes
Possibly for
chloroform
Apparent Temporal
Association with
VOC Concentrations
in Wall Ports or
Subslab Ports
(Section 6)
Yes
Yes
Weak
Some
Apparent Temporal
Association with
Large Subslab to
Indoor Differential
Pressure Events
(Section 9.1)
Yes in
some
cases
Yes in some
cases
Yes in some
cases
Yes in a
few cases
Yes in a
few cases
Apparent Trend in
XY Graph of.
Meteorological
Parameter vs.
Subslab/lndoor
Differential Pressure
(Section 9.1 and
U.S. EPA, 2012a )
No
Yes
No
Yes
No
Apparent Trend in
XY Graph of
Meteorological
Parameter vs. VOC
Concentration
(Section 9.2)
Yes for PCE,
not definitive for
chloroform
Yes
No clear
relationship
Not
definitive
Yes for
PCE, No
for
chloroform
No for
PCE, Yes
for
chloroform
9-60
-------
Section 10—Time Series Analysis
Table of Contents
10.0 Time Series Analysis 10-1
10.1 Time Series Analysis of Indoor Radon Data (AlphaGUARD) Aggregated with 1-Day
Time Resolution 10-2
10.2 Correlation between Radon Concentration and Categorical Predictors 10-18
10.3 Correlation between Radon Concentration Time Series for 422 Basement South in
2011-2012 (X422BN-1) and Continuous Predictor Variables 10-19
10.4 Correlation between Radon Concentration Time Series for 422 2nd Floor Office
(2011-2012) and Predictor Variables 10-25
10.5 Correlation between Radon Concentration Time Series for 422 Basement South
(2012-2013) and Predictor Variables 10-32
10.6 Correlation between Radon Concentration Time Series for 422 Office on 2nd Floor
and Predictor Variables 10-37
10.7 Correlation between VOC (Radiello) Time Series and Predictor Variables in 422
Basement South 10-41
10.7.1 Stationarity and Serial Correlation Analysis 10-41
10.7.2 Predictor Variables Modeled and Their Potential for Autocorrelation 10-50
10.7.3 Time Series Analysis Results for 2011-2012 Chloroform Data Set 10-53
10.7.4 Time Series Analysis Results for 2011-2012 PCE Data Set 10-57
10.7.5 Time Series Analysis of 422 Basement South Chloroform Data Set from the
period Sept2012-Apr2013 10-65
10.7.6 Time Series Analysis Results of 422 Basement South PCE Data from Sept 2012
through April 2013 10-71
Addendum 10-78
List of Figures
10-1. Daily time series of radon concentrations (pCi/L) 422 basement north, 2011-2012 with
rolling-averages, and autocorrelation (ACF) and partial autocorrelation function (PACF)
plots 10-4
10-2. Time series of first difference of daily radon concentrations (pCi/L) 422 basement north,
2011-2012 with rolling-averages, and autocorrelation (ACF) and partial autocorrelation
function (PACF) plots 10-5
10-3. Daily time series of radon concentrations (pCi/L) 422 basement north, 2012-2013 with
rolling-averages, and autocorrelation (ACF) and partial autocorrelation function (PACF)
plots 10-6
10-4. Time series of first difference of daily radon concentrations (pCi/L) 422 basement north,
2012-2013 with rolling-averages, and autocorrelation (ACF) and partial autocorrelation
function (PACF) plots 10-7
10-5. Time series of daily radon concentrations: 422 office (2nd floor), 2011-2012 with
rolling-averages, and autocorrelation (ACF) and partial autocorrelation function (PACF)
plots 10-8
10-i
-------
Section 10—Time Series Analysis
10-6. Time series of first difference of daily radon concentrations: 422 office (2nd floor),
2011-2012 with rolling-averages, and autocorrelation (ACF) and partial autocorrelation
function (PACF) plots 10-9
10-7. Time series of radon concentrations: 422 office (2nd floor), 2012-2013 with rolling-
averages, and autocorrelation (ACF) and partial autocorrelation function (PACF) plots 10-10
10-8. Time series of differences of daily radon concentrations: 422 office (2nd floor), 2012-
2013 with rolling-averages, and autocorrelation (ACF) and partial autocorrelation
function (PACF) plots 10-11
10-9. Correlation between radon concentration and predictors for both sites (basement north
and office 2nd floor). Time period 2011-2012. NOTE: See Table 10-2 for "Plain
Language" key of abbreviations used in this figure 10-16
10-10. Correlation between radon concentration and predictors for both sites (basement north
and office 2nd floor). Time period 2012-2013. NOTE: See Table 10-2 for "Plain
Language" key of abbreviations used in this figure 10-17
10-11. Time series plot, ACF and PACF for weekly Radiello chloroform. Location X422
basement south. Time period: Jan 5, 2011-Feb 15, 2012 10-42
10-12. Time series plot, ACF and PACF for first difference of weekly Radiello. Chloroform.
Location X422 basement south. Time period: Jan 5, 2011-Feb 15, 2012 10-43
10-13. Time series plot, ACF and PACF for weekly Radiello PCE. Location X422 basement
south. Time period: Jan 5, 2011-Feb 15, 2012 10-44
10-14. Time series plot, ACF and PACF for weekly chloroform. Location X422 basement south.
Time period: Sept 26, 2012-April 10, 2013 10-46
10-15. Time series plot, ACF and PACF for first difference weekly Radiello CHCI3-2
(X422BaseS_Radiello_Weekly_CHC13). Location X422 basement south. Time period:
Sept 26, 2012-April 10, 2013 10-47
10-16. Time series plot, ACF and PACF for weekly Radiello PCE-2
(X422BaseS_Radiello_Weekly_PCE). Location X422 basement south. Time period: Sept
26, 2012-April 10, 2013 10-48
10-17. Time series plot, ACF and PACF for first difference weekly Radiello PCE-2
(X422BaseS_Radiello_Weekly_PCE). Location X422 basement south. Time period: Sept
26, 2012-April 10, 2013 10-49
10-18. XY plot of weekly average snow depth vs. PCE concentration 10-63
10-19. XY Plot of change in weekly average snow depth vs. PCE concentration 10-63
10-20. XY Plot of weekly average snow depth vs. change in PCE 10-64
List of Tables
10-1. Significant Lags for ACF and PACF for Each Time Series and Regression Model 10-12
10-2. Significant Lag and AR Model by Predictor and Site Location 10-13
10-3. Model Parameters, Standard Errors by Model, Predictor and Time Series: All Radon
Time Periods, Categorical Variables 10-18
10-4. Model Parameters, Standard Errors by Model and Predictor for X422baseN_AG_radon
(X422BN-1): 2011-2012, Lag 1 Models 10-20
10-ii
-------
Section 10—Time Series Analysis
10-5. Model Parameters, Standard Errors by Predictor for X422baseN_AG_radon (X422BN-
1): 2011-1012, Lag 2 Models 10-25
10-6. Model Parameters, Standard Errors by Predictor for X422baseN_AG_radon (X422BN-1),
Lag 4 Models 10-25
10-7. Model Parameters, Standard Errors by Predictor for X422office_2nd_AG_radon
Concentration (X4220F2-1): 2011-2012, No Lag Terms in Model 10-26
10-8. Model Parameters, Standard Errors by Model and Predictor for Time Series Analysis of
Radon in 422 Office: 2011-2012, Lag 1 Models 10-26
10-9. Model Parameters, Standard Errors by Predictor for X422office_2nd_AG_radon
Concentration (X4220F2-1): 2011-2012. Lag 2 Models 10-32
10-10. Model Parameters, Standard Errors by Predictor for X422baseN_AG_radon
(X4220F2-1): 2011-2012. Lag4 Models 10-32
10-11. Model Parameters, Standard Errors by Predictor for 422 Basement Radon: 2012-2013.
No Lag Terms in Model 10-33
10-12. Model Parameters, Standard Errors by Model and Predictor for 422 Basement Radon:
2012-2013 Lag 1 Models 10-33
10-13. Model Parameters, Standard Errors by Predictor for X422office_2nd_AG_radon
Concentration (X4220F2-2): 2012-2013. No Lag Terms in Model 10-37
10-14. Model Parameters, Standard Errors by Model and Predictor for
X422office_2nd_AG_radon Concentration (X4220F2-2): 2012-2013 Lag 1 Models 10-38
10-15. Name, Periodicity, Time Period, and Location of Time Series (Outcome) Considered 10-41
10-16. Transformation and Terms Required by Time Series 10-45
10-17. Continuous Covariates by Time Period 10-50
10-18. Analysis for Outcome First Difference of X422BaseS_Radiello_Weekly_CHCl3.
Variables That Did Not Need Lag Terms. Period Jan 5, 2011-Feb 15, 2012 10-53
10-19. Analysis for Outcome X422BaseS_Radiello_Weekly_CHCl3. Variables that Needed a
lag-1 Term. Period Jan 5. 201 1 Feb 15,2012 10-54
10-20. Analysis for Outcome X422BaseS_Radiello_Weekly_CHCl3. Variables that Needed Lag-
1 and Lag-2 Terms. Period Jan 5, 2011-Feb 15, 2012 10-58
10-21. Time Series Analysis for Outcome First Difference of 422 Basement South PCE
Concentration Variables that Did Not Need Lag Terms. Period Jan 2011 to Feb 2012 10-59
10-22. Analysis for PCE Concentration at 42 Base South, Variables that Needed a Lag-1 Term.
Period Jan 2011 to Feb 2012 10-60
10-23. Analysis for PCE Concentration at 422 Base South Variables that Needed Both Lag-1
Week And Lag-2 Week Terms. Period Jan 2011 to Feb 2012 10-64
10-24. Analysis for First Difference of Chloroform Concentration at 422 Basement South.
Variables that Did Not Need Lag Terms. Period Sept 2012 to April 2013 10-66
10-25. Analysis Chloroform Concentration at 422 Base South. Variables that Needed A Lag-1
One Week Term. Period Sept 2012 to April 2013 10-67
10-26. Analysis for Chloroform Concentration at 422 Base South. Variables Needing Lag-1 And
Lag-2 Week Terms. Period Sept 2012 to April 2013 10-71
10-iii
-------
Section 10—Time Series Analysis
10-27. Analysis for First Difference of422 Base South PCE Concentration. Variables that Did
Not Need Lag Terms. Period Sept 2012 to April 2013 10-72
10-28. Analysis for422 Base South PCE Concentration. Variables that Needed A Lag-1 Week
Term. Period Sept 2012 to April 2013 10-74
10-29. Analysis for PCE Concentration at 422 Base South. Variables Needing Both Lag-1 And
Lag-2 Week Terms. Period Sept 2012 to April 2013 10-78
10-iv
-------
Section 10—Time Series Analysis
10.0 Time Series Analysis
Observations in time series are, in general, time-correlated and thus, not independent of each other.
Modeling time series data using standard modeling approaches (e.g., usual regression analysis) will
produce standard errors estimates that can be wrong and the results of the statistical tests used in
hypothesis testing might be biased, which can affect the conclusions derived from them. We considered in
this analysis only consecutive, evenly spaced observations (i.e., daily or weekly observations). Having
missing and non-evenly spaced data introduces technical complications and requires modifications to the
approaches adopted here.
Given the expected correlation between consecutive observations, a time series regression model usually
includes past and present observations of the outcome of interest (e.g., radon concentration) as well as
other predictors. In statistics, the term "predictor" refers to a variable that is possibly a predictor of the
outcome under study, also known as (aka) the independent variable. Models that include past observations
of the time series are called autoregressive models. Previous values of a time-ordered variable are referred
as lagged terms. The order of the lag of the outcome (aka dependent), or y-variable, in a model
determines the order of the time-series model. For example, if the model only includes the previous
observation (denoted as y(t-l)) and predictors, it will be termed autoregressive model of order 1, first
order autoregressive model or AR(1). A model can include lag terms of the predictors as well.
We conducted a statistical analysis to determine if any of the predictors available were good predictors of
the variability of the outcome (e.g., radon or VOC concentrations). The analysis included the evaluation
of the stationarity of the time series, determination of which autoregressive model to use, and
determination of the lags for the predictor functions. Full and reduced model approaches were used to
evaluate the significance of the reduced models.
A time series is termed "stationary" if the mean, variance, autocorrelation, etc. are all constant overtime.
The Augmented Dickey-Fuller test (ADF) (Said and Dickey, 1984) and the Phillips-Perron Unit Root
Test (PP) (Perron, 1988) test for stationarity were also calculated to formally evaluate stationarity of the
time series. The null hypothesis for the two tests is that the data is non-stationary. Small p-values (p-
values < 0.05) suggest evidence favoring stationarity. It is desirable for a time series to be stationary—it
does not necessarily mean that it is boring; only amenable to analysis. As Nau (2005a) states:
Most statistical forecasting methods are based on the assumption that the time series can be rendered
approximately stationary (i.e., "stationarized") through the use of mathematical transformations. A
stationarized series is relatively easy to predict: you simply predict that its statistical properties will be the
same in the future as they have been in the past!... The predictions for the stationarized series can then
be "untransformed," by reversing whatever mathematical transformations were previously used, to obtain
predictions for the original series. ... Thus, finding the sequence of transformations needed to stationarize
a time series often provides important clues in the search for an appropriate forecasting model.
Another reason for trying to stationarize a time series is to be able to obtain meaningful sample statistics
such as means, variances, and correlations with other variables. Such statistics are useful as descriptors
of future behavior only if the series is stationary. For example, if the series is consistently increasing over
time, the sample mean and variance will grow with the size of the sample, and they will always
underestimate the mean and variance in future periods. And if the mean and variance of a series are not
well-defined, then neither are its correlations with other variables. For this reason you should be cautious
about trying to extrapolate regression models fitted to nonstationary data.
10-1
-------
Section 10—Time Series Analysis
Predictor variables that are significantly associated with the transformed outcome variable are also
associated with the original outcome variable.
10.1 Time Series Analysis of Indoor Radon Data (AlphaGUARD) Aggregated with
1-Day Time Resolution
Four radon time series are analyzed in this section; their time frames of the data and location are
described below.
16. daily radon concentrations collected at 422 basement north between March 31, 2011, to July 23,
2012 (referred to as data set 422BN-1). This data set was constructed by averaging the
AlphaGUARD data for each day.
17. daily radon concentrations collected at 422 basement north between September 7, 2012, and May
20, 2013 (referred to as data set 422BN-2)
18. daily radon concentrations collected at 422 office (2nd floor) between March 31, 2011, to July 23,
2012 (referred to as data set 42202F-1)
19. daily radon concentrations collected at 422 office (2nd floor) between September 7, 2012, and
May 20, 2013 (referred to as data set 42202F-2)
The first step in the statistical analysis was the evaluation of the non-stationarity of the time series itself
(shown as the left panels Figures 10-1,10-3,10-5, and 10-7). We evaluated the stationarity (or non-
stationarity) of the radon time series using the ADF and PP tests. Small p-values suggest that a time series
is stationary. A solution for non-stationarity is to calculate the first differences (y, (outcome
concentration at time t)- (outcome concentration at time t-1)) and to use the difference variable as the
outcome. We calculated the first differences (yt~yt_l) and evaluated their stationarity using ADF and PP
tests. The first difference series (left panels in Figures 10-2,10-4,10-6, and 10-8) resulted in small p-
values and a time series plot more consistent with the stationarity assumption (constant variance and
constant mean). Thus, the first difference series is more suitable for analysis.
The fact that the first differences of our time series are more stationary than the time series themselves
can also be seen visually by comparing the left-hand panels between Figures 10-1 and 10-2; 10-3 and 10-
4, etc. In the case of constant mean, one should be able to have a straight line fit going across all the
points in the series from left to right. A constant variance means that the series oscillates about the mean
within a band of equal size. So if a time series goes up and down around a constant mean but the width of
the max and min values is not constant, then we have non-constant variance.
The second step of the statistical analysis performed was determining the serial correlation. The
autocorrelation function (ACF) plot (shown in the central panel in each of the Figures 10-1 through 10-8)
shows the correlation between a time series and lags of itself. In the ACF plot, the lag 1 (day) is the
correlation between all pairs of two consecutive observations. The broken horizontal blue lines on the
ACF plots correspond to 95% confidence limits. For example in Figure 10-1, we can see that the radon
concentration in the 422 basement on any given day is strongly related to what the radon concentration
was the day before in the 422 basement. This makes intuitive physical sense because research has shown
that radon concentrations can be controlled by weather variables (see Section 2), and the type of weather
experienced from 1 day to the next is not completely random. For example, it is very unlikely that a 10°F
day will be followed by a 90°F day. This autocorrelation is also to be expected because the indoor air
concentrations only change gradually—with an air exchange rate of less than one air exchange per hour
(see Section 12.1 in U.S. EPA [2012a] for air exchange rate measurements). Thus, even if the weather
conditions did radically change, it would take several hours for that change to have its full influence on
the indoor radon concentration.
10-2
-------
Section 10—Time Series Analysis
Partial autocorrelations are shown in the right-most panel in each of the figures from Figures 10-1
through 10-8. A spike in the partial autocorrelation function (PACF) plot is the amount of correlation
between a variable and a lag of itself that is not already explained by correlations at all lower order lags.
For example, partial correlation value for 2 days is the correlation between day 0 and day 2 observations
that is not already explained by including the lag 1-day correlation in a model. The broken horizontal blue
lines on the PACF plots correspond to 95% confidence limits.
Note that if y, (outcome concentration at time t) is correlated with (outcome concentration at time
t-1), and yt_x is equally correlated with yt_2 (outcome concentration at time t-2), then we should also
expect to find correlation between y, and yt_2 . This behavior is called "propagation." Thus, in theory, the
correlation at lag 1 "propagates" to higher order lags, and as a result, some of the statistical significance
(i.e., spikes crossing the 95% confidence bands in the ACF plot) observed in larger autocorrelations in the
ACF plot might be the result of this "propagation."
Figures 10-1 and 10-3 show the time series plots for time series 422BN-1 and 422BN-2, respectively.
The ACF plot show significant autocorrelations for a large number of lags (lag 26 days and lag 14 days,
Figures 10-1 and 10-3, respectively), but, as shown by the PACF plots (right panels in each figure), the
autocorrelations at lag 2 days and above may be the result of the propagation of the autocorrelation at lag
1 day. To determine the true correlation between the outcome and the lag of the outcome after removing
the effect of previous lags, we look at the PACF plots.
In Figure 10-1, the PACF plot for time series 422BN-1 shows significant autocorrelations up to lag 4
days (crossing the blue line either positively or negatively). Thus, the first data set acquired in the 422
basement for radon shows that the radon concentration on any given day is not random; it is related to the
radon concentration on the 4 previous days. This also shows that the time series provides at least the
equivalent of one independent measurement every 5 days. In Figure 10-3, the PACF plot for 422BN-2
shows significance at lag 1 day, suggesting that all the higher order autocorrelations are explained by the
lag 1 day autocorrelation. The spike crossing the blue line in lag 3 could be the result of some random
noise.
Figure 10-3; for the second data set acquired in 422 basement north shows only one significant
autocorrelation at lag 1 day. The spike crossing the blue line in lag 11 could be the result of some random
noise. Figures 10-5 and 10-7 show the time series plots with corresponding ACF and PACF plots for
time series data sets from the 422 second floor office, 42202F-land 42202F-2, respectively. The ACF
plots show significant autocorrelations up to lags 26 days and 14 days for time series 42202F-1 and
42202F-2, respectively. A closer look at the PACF plots (right most panels) suggests that only lag 1 day
was significant for these two time series.
Figures 10-2, 10-4, 10-6, and 10-8 show the ACF and PACF of first difference daily concentrations.
PACF plots do not display any significant autocorrelation, suggesting there is no need to include a lag
variable in a model using the transformed data. In lay terms, this means that although the radon
concentration on Tuesday is dependent on the radon concentration on Monday and the radon
concentration on Wednesday is, in turn, dependent on the radon concentration on Tuesday, the direction
and magnitude of the change in radon that occurs between Monday and Tuesday is not connected to the
direction and magnitude of the change in radon that occurs between Tuesday and Wednesday.
10-3
-------
Daily Time Series of radon concentration (pCi/L). Location: X422 Base N AG
ADF and PP tests pvalues= (0.01 ,0.01)
03/31 05/10 06/19 07/29 09/07 10/17 11/26 01/05 02/14 03/25 05/04 06/13 07/23
Date: 2011-2012
~i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—r
0 1 2 3 4 5 6 7 8 9 11 13 15 17 19 21 23 25 27
Lag
i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—r~
0 1 2 3 4 5 6 7 8 9 10 12 14 16 18 20 22 24 26
Lag
Figure 10-1. Daily time series of radon concentrations (pCi/L) 422 basement north, 2011-2012 with rolling-averages, and
autocorrelation (ACF) and partial autocorrelation function (PACF) plots.
-------
First Difference of Daily Time Series of radon concentration (pCi/L). Location: X422 Base N AG
ADF and PP pvalues = (0.01 , 0.01 )
03/31 05/10 06/19 07/29 09/07 10/17 11/26 01/05 02/14 03/25 05/04 06/13 07/23
Figure 10-2. Time series of first difference of daily radon concentrations (pCi/L) 422 basement north, 2011-2012 with rolling-averages,
and autocorrelation (ACF) and partial autocorrelation function (PACF) plots.
-------
Daily Time Series of radon concentration (pCi/L). Location: X422 Base N AG
ADF and PP tests pvalues = (0.43 ,0.23)
14.68
14.18 "
13.68
13.18
12.68 -
12.18 -
11.68
10.68 "
10.18 _
O 7.68
MM
09/07 10/02
10/27 11/21
I M M M
12/16 01/10 02/04
Date: 2012-2013
rrrrrr
03/01 03/26
04/20 05/15
-1—I—I—I—I—I—I-
0 1 2 3 4 5 6
-1—I—I—I-
8 9 10
l ' i
1 2 3 4 5
~l—I 1 1—
7 8 9 10
Figure 10-3. Daily time series of radon concentrations (pCi/L) 422 basement north, 2012-2013 with rolling-averages, and
autocorrelation (ACF) and partial autocorrelation function (PACF) plots.
-------
First Difference of Daily radon concentration (pCi/L). Location: X422 Base N AG
ADF and PP pvalues = (0.01 ,0.01)
10.98
10.48 -
7.48 -
5.98 "
5.48 -
-0.02
-1.02
-1.52
-3.02
-4.52
-6.02
-6.52
I ! ! ! ! i i ! ! ! i i ! ! ! ! i ! ! ! ! i i ! ! ! i i ! ! ! ! i ! ! ! ! i ! ! ! ! ! i ! ! ! I
09/07 09/27 10/17 11/06 11/26 12/16 01/05 01/25 02/14 03/06 03/26 04/15
Date: 2012-2013
~l r
~n~
~i—i—i—i—i-
2 3 4 5 6
-1 1—I—
8 9 10
~l 1 1 I-
12 3 4
"I 1 1 1 1 1 I-
6 7 8 9 10 12
~l—I—I—I—I—r~
Figure 10-4. Time series of first difference of daily radon concentrations (pCi/L) 422 basement north, 2012-2013 with rolling-averages,
and autocorrelation (ACF) and partial autocorrelation function (PACF) plots.
-------
Daily Time Series of Radon concentration. Location: 422 Office
ADF and PP tests pvalues = (0.01 ,0.01)
O 4.67
03/31 05/10 06/19 07/29 09/07 10/17 11/26 01/05 02/14 03/25 05/04 06/13 07/23
Date: 2011-2012
n—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—r~
0 1 2 3 4 5 6 7 8 9 11 13 15 17 19 21 23 25 27
_LL
r~r
1—I—I—I—I—I—I—I—I—I—I—I—I—I—I—I—I—I—I—I—I—I—I—I—I—I—
0 1 2 3 4 5 6 7 8 9 10 12 14 16 18 20 22 24 26
Figure 10-5. Time series of daily radon concentrations: 422 office (2nd floor), 2011-2012 with rolling-averages, and autocorrelation
(ACF) and partial autocorrelation function (PACF) plots.
-------
First Difference of Daily radon concentration (pCi/L). Location: X422 Base N AG
ADF and PP pvalues = (0.01 ,0.01)
3.82 -
2.82 -
2.32 ~
D.82 -
-0.18 "
-0.68 -
.68 "
2.18 -
3.18 "
03/31 05/10 06/19 07/29 09/07 10/17 11/26 01/05 02/14 03/25 05/04 06/13 07/23
Date: 2011-2012
J_L
tt
~r~
~\—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—r
I'll
n—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—
0 1 2 3 4 5 6 7 8 9 11 13 15 17 19 21 23 25 27 0 1 2 3 4 5 6 7 8 9 10 12 14 16 18 20 22 24 26
Lag
Lag
Figure 10-6. Time series of first difference of daily radon concentrations: 422 office (2nd floor), 2011-2012 with rolling-averages, and
autocorrelation (ACF) and partial autocorrelation function (PACF) plots.
-------
Daily Time Series of Radon concentration. Location: 422 Office
KPSS, ADF and PP tests pvalues = (0.02 , 0.22 ,0.04)
Ji
111111111111111111111111111111111111111111111111111
09/07 10/02 10/27 11/21 12/16 01/10 02/04 03/01 03/26 04/20 05/15
Date: 2012-2013
1—I—I—I—I—I—I—I—I—I—I—I—I—I—I—I—I—I—I—I—I—I—I—r
0 1 2 3 4 5 6 7 8 9 10 12 14 16 18 20 22
n—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—r~
1 2 3 4 5 6 7 8 9 10 12 14 16 18 20 22
o
Figure 10-7. Time series of radon concentrations: 422 office (2nd floor), 2012-2013 with rolling-averages, and autocorrelation (ACF) and
partial autocorrelation function (PACF) plots.
£
>3
-------
Difference Daily Time Series of Radon concentration. Location: 422 Office
ADF and PP pvalues = (0.01 ,0.01)
111111 1111111 1111111 1111111 11111111 1111111 111111
09/07 09/27 10/17 11/06 11/26 12/16 01/05 01/25 02/14 03/06 03/26 04/15
Date: 2012-2013
1—I—I—I—I—I—I—I—I—I—I—I—I—I—I—I—I—I—I—I—I—I—I—r
0 1 2 3 4 5 6 7 8 9 10 12 14 16 18 20 22
~I I I I I I I I I I I I I I I I I I I I
2 3 4 5 6 7 8 9 10 12 14 16 18 20 22
Lag
Figure 10-8. Time series of differences of daily radon concentrations: 422 office (2nd floor), 2012-2013 with rolling-averages, and
autocorrelation (ACF) and partial autocorrelation function (PACF) plots.
-------
Section 10—Time Series Analysis
The PACF plots (right-most panel in Figures 10-1 through 10-8) indicate the number of autoregressive
terms (k) needed to explain the autocorrelation pattern in the time series. If the partial autocorrelation is
significant at lag k and not significant at any higher order lags, then this suggests an autoregressive model
(AR) of order k. Table 10-1 shows a summary of the ACF and PACF analysis and the auto regressive
model of order k (AR(k)) required for the original and the first difference time series. Since the first
differences are stationary, a model with a first difference as the outcome was fit to the data. The shorthand
AR model=0 denotes an autoregressive model for the first difference that does not incorporate any lag
term of the outcome as predictor. For example: if Diff(t) = outcome(t)—outcome(t-l), then a model with
no lag term of the outcome will be Diff(t) = intercept + predictor (t), i.e., outcome (t) -outcome (t-1) =
intercept + predictor (t). The results of this analysis make intuitive sense because the concentration of
radon is expected to be influenced by the amount of vapor intrusion occurring over the past 12 hours,
given the observed air exchange rate (0.3 to 0.8 air exchanges per hour, see Section 10.1 in U.S. EPA
[2012a]).
Table 10-1. Significant Lags for ACF and PACF for Each Time Series and Regression Model
Original Data First Difference
Time Series n . D . . Significant Significant
Variable Name a a er'° Lags Autoregressive Lags
Model Chosen Autoregressive
ACF
PACF
ACF
PACF
Model Chosen
422BN-1
2011-2012
>26
4
AR(4)
1
0
0
422BN-2
2012-2013
14
1
AR(1)
1
0
0
4220N-1
2011-2012
>26
1
AR(1)
1
0
0
4220N-2
2012-2013
14
1
AR(1)
1
0
0
The third step of the statistical analysis is to determine the order of the serial autocorrelation in each of
the continuous predictors. Table 10-2 displays the type of AR model for each continuous predictor. Most
of the continuous predictors required only a lag 1 day term in the model. The significance of the
association between radon concentration and the categorical variables available was studied separately.
Also listed in Table 10-2 are plain language explanations for the predictor variables used in this analysis.
Next, for each continuous variable listed in Table 10-2, a full or saturated model was fit to the time series.
In each time series analysis, the first difference was the dependent or outcome variable modeled because
it passed the test for being stationary. The term "full model" or "saturated model" refers to a model that
includes an intercept, the predictor, all needed lags of the predictor (as specified in Table 10-2), and any
control variable. The larger lag term of the predictor variable in the full or saturated model was dropped
from the list of predictors and the resulting first "reduced" model was fit. The next lag-term was then
dropped from the list of predictors and a second reduced model was fit to the data. This process continued
until all lag terms of the predictors were removed from the previous model.
A control variable is a variable that can affect the association between the dependent variable and other
predictors in the model. In this analysis, the variable Mitigation_status_daily was considered a control
variable since it was expected to have a dramatic effect on the behavior of the vapor intrusion process. It
was included in the model to account for the testing of subslab depressurization. Since mitigation was not
installed in the period 2011 to 2012, none of the models using the 2011 to 2012 data set (422BN-1 and
4220F2-1) included mitigation as a control variable. Mitigation was included as control variable only in
models using data from 2012 and 2013. Thus, for the 2012 to 2013 data, the change in radon
concentration was modeled separately for each predictor variable, but mitigation was always included in
the equation.
10-12
-------
Section 10—Time Series Analysis
For example, the full model for the first predictor in Table 10-2, Air Density Interior for data period
2012-2013, is: Diff(t) = Intercept + Air Density Interior (t) + Air Density Interior (t-l)+ Mitigation(t).
The first reduced model is: Diff (t) = Intercept + Air Density Interior (t) + Mitigation (t); the second
reduced model is Diff(t) = Intercept + Mitigation (t). Models for the same predictor using data period
2011 to 2012 contain the same terms except the variable mitigation is not used. The mitigation was coded
with on =1 and off = 0.
Table 10-2. Significant Lag and AR Model by Predictor and Site Location
Data Period
:2011-2012
Data Period
:2012-2013
Predictor Name
(Plain Language)
Predictor Name
(Abbreviation)
Significant
Lags (days)
AR Model
Chosen
Significant
Lags(days)
AR Model
Chosen
Air density interior
AirDens_422
1
AR(1)
1
AR(1)
Drop in barometric
pressure per hour
Bar_drop_.Hg.hr
0
0
0
0
Barometric pressure in
inches of mercury
Bar_in_Hg
1
AR(1)
1
AR(1)
Cooling degree day
Cool_Degree_Day
1
AR(1)
1
AR(1)
Dew point, interior,
Fahrenheit
Dew_pt_422_F
1
AR(1)
1
AR(1)
Dew point, exterior
Dew_pt_out_F
1
AR(1)
1
AR(1)
Height measured at Fall
Creek stream gauge in feet
Fall_Crk_Gage_ht_ft
1
AR(1)
1
AR(1)
Heating degree days
Heat_Degree_Day
1
AR(1)
1
AR(1)
Exterior heating index—
calculated based on
temperature and humidity
Heat_lndex_F
1
AR(1)
1
AR(1)
Humidity interior
Hum_422_.
1
AR(1)
1
AR(1)
Humidity exterior
Hum_out_.
1
AR(1)
1
AR(1)
Interior heating index
lndoor_Heat_lndex
1
AR(1)
1
AR(1)
Rain (inches) totaled during
observation period
Rain_ln_met
1
AR(1)
0
0
Rain highest rate during
observation, hperiod
RainJPH
0
0
0
0
420 side, subslab vs.
basement differential
pressure
Setra_420ss. base_Pa
1
AR(1)
422 side basement vs.
exterior differential
pressure, pascals
Setra_422base.out_Pa
4
AR(4)
422 side, basement vs.
upstairs differential
pressure, pascals
Setra_422base.upst_Pa
1
AR(1)
422 side, deep vs. shallow
soil gas differential
pressure, pascals
Setra_422SGdp.ss_Pa
1
AR(1)
422 side, subslab vs.
basement differential
pressure, pascals
Setra_422ss. base_Pa
2
AR(2)
(continued)
10-13
-------
Section 10—Time Series Analysis
Table 10-2. Significant Lag and AR Model by Predictor and Site Location (cont.)
Data Period
:2011-2012
Data Period
:2012-2013
Predictor Name
(Plain Language)
Predictor Name
(Abbreviation)
Significant
Lags (days)
AR Model
Chosen
Significant
Lags(days)
AR Model
Chosen
Depth of snow on the
ground, inches
Snowdepth_daily
1
AR(1)
1
AR(1)
Soil moisture, 13 ft bis
beneath structure, cbar
Soil_H20_l n 13._cbar
1
AR(1)
1
AR(1)
Soil moisture 16.5 ft bis
beneath structure, cbar
Soil_H20_ln 16.5._cbar
1
AR(1)
0
0
Soil moisture 6 ft bis
beneath structure, cbar
Soil_H20_ln6._cbar
1
AR(1)
1
AR(1)
Soil moisture 13 ft bis
exterior, cbar
Soil_H20_Out13._cbar
1
AR(1)
Soil moisture, 3.5 ft bis
exterior, cbar
Soil_H20_Out3.5._cbar
1
AR(1)
1
AR(1)
Soil moisture 6 ft bis
exterior, cbar
Soil_H20_Out6._cbar
1
AR(1)
1
AR(1)
Soil temperature 13 ft bis
beneath structure
Soil_T_C_MW3.13
1
AR(1)
Soil temperature 16.4 ft bis
beneath structure
Soil_T_C_MW3.16.5
1
AR(1)
Soil temperature 6 ft bis
beneath structure
Soil_T_C_MW3.6
1
AR(1)
Soil temperature 9 ft bis
beneath structure
Soil_T_C_MW3.9
1
AR(1)
Soil temperature 1 ft bis
exterior
Soil_T_C_OTC.1
1
AR(1)
Soil temperature 13 ft bis
exterior
Soil_T_C_OTC. 13
1
AR(1)
Soil temperature 16.5 ft bis
exterior
S o i l_T_C_OT C. 16.5
1
AR(1)
Soil temperature 6 ft bis
exterior
Soil_T_C_OTC.6
1
AR(1)
Temperature at 420
basement north sampling
location
T_420baseN_C
1
AR(1)
Temperature at 420
basement south sampling
location
T_420baseS_C
1
AR(1)
Temperature at 420 first
floor sampling location
T_420first_C
1
AR(1)
Temperature, 422 first floor
T_422_F
1
AR(1)
1
AR(1)
Temperature 422
basement north
T_422baseN_C
1
AR(1)
Temperature 422 first floor
T_422baseS_C
1
AR(1)
Temperature on first floor of
422 side of duplex
T_422first_C
1
AR(1)
Temperature exterior
T_out_C
1
AR(1)
(continued)
10-14
-------
Section 10—Time Series Analysis
Table 10-2. Significant Lag and AR Model by Predictor and Site Location (cont.)
Data Period
:2011-2012
Data Period
:2012-2013
Predictor Name
(Plain Language)
Predictor Name
(Abbreviation)
Significant
Lags (days)
AR Model
Chosen
Significant
Lags(days)
AR Model
Chosen
Exterior temperature (°F)
T_out_F
1
AR(1)
1
AR(1)
Temperature exterior, high
during data collection period
T_out_Hi_F
1
AR(1)
1
AR(1)
Lowest exterior temperature
T_out_Lo_F
1
AR(1)
1
AR(1)
Temperature, humidity and
wind index
THW_F
1
AR(1)
1
AR(1)
Wind chill
Wnd_Chill_F
1
AR(1)
1
AR(1)
Wind direction (average)
Wind_Dir
1
AR(1)
1
AR(1)
Wind direction (of high during
measurement period)
Wind_Dir_Hi
1
AR(1)
1
AR(1)
Wind run is a function of wind
speed and duration
Wind_Run_mi
1
AR(1)
1
AR(1)
High wind speed during
measurement period
Wind_Speed_Hi_MPH
1
AR(1)
1
AR(1)
Average wind speed during
measurement period
Wind_Speed_MPH
1
AR(1)
1
AR(1)
Figure 10-9 and Figure 10-10 show the correlation matrix of the two first-difference radon concentration
variables and continuous predictors available for 2011 to 2012 and 2012 to 2013 time periods,
respectively. The color and the size of the circles describe the degree of correlation between the variables
listed. Blue color denotes positive correlation and red color denotes negative correlation. Clear or almost
white circles denote almost no correlation between pair of variables. The size of the circles is proportional
to the degree of correlation; therefore, large circles denote large correlations and small circles denote very
small correlation. The names of the first difference radon concentrations are listed as
X422baseN_AG_radon and X422office_2nd_AG_radon for Base North and Office 2nd floor,
respectively; and they are located at the end of the plots.
These figures show that the cross-correlation of the variables generally follows expected mathematical or
physical principles. The strong positive correlation among the variables that directly measure temperature
(either average or maximum) is unsurprising, as is the link to the variables calculated from temperature
such as cooling degree days, heating degree days, heat index, and wind chill. The positive correlation
between dew point and temperature also follows from the definition of dew point. The negative
correlation of temperature to air density is rational based on both the ideal gas law and the formula used
to calculate air density. The strong negative correlation between the temperature related variables and
radon concentration likely shows the influence of the stack effect in vapor intrusion—as the temperature
decreases, the stack effect increases, bringing in more radon. Similarly, the strong positive correlation
between the various rainfall-related variables is expected, and the weak positive correlation between
rainfall and humidity is expected. Also, the heat index is positively correlated to humidity by definition.
10-15
-------
Section 10—Time Series Analysis
Bar_in_Hg
Cool_Degree_Day
Dew_pt_422_F
Dew_pt_out_F
F a I l_Crk_G age_ht_ft
Heat_Degree_Day
Heat_lndex_F
Hum_422_.
Hum_out_.
I ndoor_Heat_l ndex
Rain_ln_met
RainJPH
Rain2_ln
Setra_420ss.base_Pa
Set ra_422bas e. out_Pa
Setra_422base.upst_Pa
Setra_422SGdp.ss_Pa
Setra_422ss.base_Pa
Snowdepth_daily
Soil_H20_ln13._cbar
Soil_H20_ln16.5._cbar
Soil H20_ln6._cbar
S o i l_H 20_0ut3.5. _c bar
Soil_H20_Out6._cbar
Soil_T_C_MW3.13
Soil_T_C_M W3.16.5
Soil_T_C_M W3.6
Soil_T_C_M W3.9
S o i l_T_C_OTC. 1
Soi l_T_C_OTC. 13
Soil_T_C_OTC. 16.5
Soi l_T_C_OTC. 6
T_420baseN_C
T_420baseS_C
T_420first_C
T_422_F
T_422baseN_C
T_422baseS_C
T_422first_C
T_out_C
T_out_F
T_out_Hi_F
T_out_Lo_F
THW_F
Wind_Chill_F
Wind_Dir
Wind Dir_Hi
Wind_Run_mi
Wind_Samp
Wind_Speed_Hi_MPH
X422baseN_AG_radon
X422office 2nd AG radon
0*U| i I
i ~
toT^U-
m°cVJc)l0_
C/3 0
- 4—CD co a> . I
3 a 4~-nj-Q c/3 q_cl>
CM-^rCM «5"0
s ( xh—'
—CM"vJ-CnI
IN1" U fJJ
ro c= (p^-°CD
~~~~
r\ CT3 CO
rCNv-v_0 >-°-Q
rs|CO Oo
•¦P=«S^lSg 1 .L«
.te
' co
--oO jgco I
~~~~¦nnu
¦rTrnnTW^
¦nnnnnnnii
ncnnnrm!
¦nnnninni
«1«!1SS1
oMillll
W" oi°~
wife.
CO
OLL ^ „
CM
.^CM03
2111
••• i
mil _ fN i
e i —1
INNHI99«»-^f-b<
its
_kv |
^iD
~~~~~~
~~~~~~
TTTTT]
[iiiii]
C~M <->
¦xTCM
-0.8
-0.6
-0.4
-0.2
0.2
0.4
0.6
0.8
Figure 10-9. Correlation between radon concentration and predictors for both sites (basement
north and office 2nd floor). Time period 2011-2012. NOTE: See Table 10-2 for "Plain
Language" key of abbreviations used in this figure.
10-16
-------
Section 10—Time Series Analysis
AirDens_422
Bar_drop_. Hg.hr
Bar_in_Hg
Cool_Degree_Day
Dew_pt_422_F
Dew_pt_out_F
Fal l_Crk_Gage_ht_ft
••
•
5
Jl
Ol
OJ
"O
£=
Heat_Degree_Day
•
•
•
•
•
•
5
I
E
Jl
to
Heat_lndex_F
n
•
•••
-
I
E
X
I
Q
Hum_422_.
~
•
•
••
•••
-
o
o
I
C
Hum_out_.
•
::
I
c
Hd
"to
_Q
O
lndoor_Heat_lndex
•
•
•
•
•
•
to
01
1
C
c.
1
JZ
CO
TO
_Q
TO
_Q
Rain In met
•
•
•
•
•
•
to
01
Ol
£=
Q.
a>
c.
~\
O
1
1
*_
Rain IPH
•
•
•
::
to
01
o
o
OJ
c
CO
_Q
Rain2 In
•
•
•
•
•
•
M
::
to
_l
O
OJ
O
CD
Snowdepth_daily
•
•
LJ
o>
_L
1
O
OJ
Z3
o
1
Soil_H20_ln13._cbar
¥
•
f 1
L J
&
X
_l
o
CM
Soil_H20_ln6._cbar
•
•
•
•
•
•
•
•
•
&
X
_l
l_L
1
Soil_H20_Out3.5._cbar
•
•
•
•
•
::
$>
OJ
CM
l_L
|
LL
Soil_H20_Out6._cbar
•
•
•
•
•
•
~
•
~
O
X
l_L
|
T_422_F
•
•
•
•
::
•
o
O
_l
T_°ut_F
•
M
~
•
•
k J
O
LL
LL
_l
T_out_Hi_F
n
•
M
~
•
•
1
3
O
T_out_Lo_F
z
•
•
•
~
•
•
•••
-
1
~o
c
Q
X(
THW_F
H
•
•M
•••
~
•
•
•••
S
1
~o
1=
Q
Wind_Chill_F
n
•
•
••
•••
~
•
•
•••
s
1
~o
c
Wind_Dir
•
§
Wind_Dir_Hi
•
::
Wind_Run_mi
•
•
•
•
Wi nd_S peed_Hi_M P H
•
•
•
•
Wind_Speed_MPH
•
•
•
•
X422baseN_AG_radon
X422office_2nd_AG_radon
•
i
CL
a) r i
-0.8
-0.6
-0.4
-0.2
0.2
0.4
0.6
0.8
Figure 10-10. Correlation between radon concentration and predictors for both sites (basement
north and office 2nd floor). Time period 2012-2013. NOTE: See Table 10-2 for
"Plain Language" key of abbreviations used in this figure.
Somewhat less obvious were:
The negative correlation between barometric pressure and wind speed: This makes sense because
storms are frequently associated with a falling barometer.
The 6-ft interior soil moisture and dew point correlation: The direction of this relationship is
intuitive, but one would not necessarily have expected this to have such a rapid response.
10-17
-------
Section 10—Time Series Analysis
¦ A positive correlation between wind direction and barometric pressure drop: This may suggest
that storms consistently arrive from a common direction.
10.2 Correlation between Radon Concentration and Categorical Predictors
Note that some of our predictor variables could be described only with a limited number of values; that is,
they are categorical variables. Categorical variables included the status of the air conditioning, fan used in
fan testing, and central heating. In each case, these were either "on" or "off" On was coded as a 1 and off
as a 0. Table 10-3 displays the results of the association between categorical predictors and the four time
series. In reading this and subsequent tables an* by the coefficient means that the term is significantly
correlated with radon concentration at the 5% significance level. A** denotes significance at the more
stringent 1% significance level.
As discussed before, the first difference of the time series was used as the response for the model and
mitigation was included only in the analysis of the data set collected from 2012 to 2013.
Table 10-3. Model Parameters, Standard Errors by Model, Predictor and Time Series: All Radon
Time Periods, Categorical Variables
Model : Y(t)-Y(t-1) = Intercept
+ Predictor (t)
Time Series
Predictor Name = x(t)
Model Term
Parameter
SE
X4220F2-1
AC_o n. off_420_d a i ly
Intercept
-0.010
0.089
x(t)
0.104
0.537
AC_on.off_422_daily
Intercept
0.003
0.098
x(t)
-0.048
0.218
Fan_on.off_422_daily
Intercept
-0.024
0.090
x(t)
0.344
0.400
Heat_on.off_422_daily
Intercept
-0.037
0.140
x(t)
0.050
0.179
X422BN-1
AC_o n. off_420_d a i ly
Intercept
-0.007
0.048
x(t)
-0.036
0.265
AC_on.off_422_daily
Intercept
0.003
0.053
x(t)
-0.053
0.118
Fan_on.off_422_daily
Intercept
-0.030
0.049
x(t)
0.432*
0.218
Heat_on.off_422_daily
Intercept
-0.015
0.076
x(t)
0.011
0.097
X422BN-2
Intercept
0.146
0.294
Heat_on.off_422_daily
x(t)
0.210
0.301
Mitigation(t)
-0.534*
0.215
X4220F2-2
Heat_on.off_422_daily
Intercept
0.042
0.148
x(t)
0.070
0.151
Mitigation(t)
-0.186
0.108
'Significant at 5% level of significance; ** Significant at 1 % level of significance
Red font denotes significance at 1% or 5% level. SE = standard error
10-18
-------
Section 10—Time Series Analysis
An example of how to read the results in Table 10-3 is as follows: the first pair of rows (blue shaded) on
the table are for radon concentration at 422 basement north during the 2011 to 2012 time period. In that
model, the effect of having the air conditioning on today is that the predicted difference between today's
radon concentration and yesterday's radon concentration is = -0.010 +1*.104 = 0.094 pCi/L (a result that
is not statistically significant). Similarly the effect of having the air conditioning off today is that the
predicted difference between today's radon concentration and yesterday's radon concentration is that
radon will have decreased by 0.01 pCi/L, (a result that is not statistically significant). From this model,
we can conclude that air conditioning is not correlated with radon.
The predictor (AC_on.off_420_daily) was not significant, suggesting that it is not correlated with radon
concentration. In other words, turning the air conditioning on or off on the 420 side of the duplex did not
significantly affect the amount of radon in the 422 basement.
The only models that were shown to be statistically significant were:
¦ The model using interior fan in the first data set (2011 to 2012) for the 422 basement. When the
fan was on, it increased today's radon as compared to yesterday's by 0.43 pCi/L per day. That
makes physical sense because, as discussed in Section 12.2 of our previous report (U.S. EPA,
2012a), the fan was setup in the stairway and was being used to simulate a worst case vapor
intrusion condition.
¦ The model in which both mitigation status and heating status were included for the 2012 to 2013
basement data. In that model, the slope of the mitigation term was statistically significant. That
suggests that when the mitigation system was on today the radon concentration decreased by
-0.534 pCi/L per day, regardless of whether the heating system was on or off. The heating
system (on/off) term in that same equation was not statistically significant. That makes physical
sense because the mitigation system was designed to reduce radon when it was turned on.
10.3 Correlation between Radon Concentration Time Series for 422 Basement
South in 2011-2012 (X422BN-1) and Continuous Predictor Variables
Table 10-4 shows the results of models requiring only a lag 1 day term of the predictors in the model. In
this and subsequent tables, a red font and the symbols and "**" adjacent to the numbers in the cell of
the "Estimate" column denote a statistically significant coefficients at 5% and more stringent 1%
significance level, respectively. Statistically significant coefficients for the predictor variable in the
model, suggest that the predictor and the outcome (radon concentrations) are correlated. If the coefficient
of x(t-l) (i.e., predictor at time t—1, shown in the third column of the table) is significant, that implies that
the previous day's observation of the predictor is correlated with the outcome. The sign of the coefficient
denotes the direction of the correlation.
If significance of the coefficient of x(t-l) was detected, then there is no need to analyze the reduced
model results (which are shown in columns 5 and 6). But if the variable x(t-l) was not statistically
significant in the full model (columns 3 and 4), then a reduced model (with all parameters except x(t-l))
was fit. In the reduced model (columns 5 and 6), the interest centers in the significance of the coefficient
of x(t). This coefficients measures the strength of the association between the first difference of radon
concentration (Y(t)-Y(t-l)) and the predictor.
10-19
-------
Section 10—Time Series Analysis
Table 10-4. Model Parameters, Standard Errors by Model and Predictor for
X422baseN_AG_radon (X422BN-1): 2011-2012, Lag 1 Models
Model : Y(t)-Y(t-1) = Intercept +
Model : Y(t)-Y(t-1) = Intercept +
Predictor (t)
Predictor Name
Model
Predictor (t) +
Predictor (t-1)
Estimate
SE
Estimate
SE
Air density interior
intercept
10.54412**
3.184
6.494
3.310
x(t-1)
-989.073
133.532
x(t)
845.491
133.144
-88.565
45.078
intercept
76.174
14.505
44.6707**
13.718
Barometric pressure
x(t-1)
-3.330
0.608
NA
NA
x(t)
0.788
0.609
-1.49088**
0.458
Cooling degree day
intercept
-0.026
0.102
0.093
0.110
x(t-1)
13.280
1.370
x(t)
-13.002
1.362
-0.867
0.587
intercept
-0.9325**
0.345
-0.626
0.356
Dew point, interior
x(t-1)
0.143
0.023
x(t)
-0.124
0.023
0.012
0.007
Dew point, exterior
intercept
-0.74167**
0.265
-0.151
0.290
x(t-1)
0.129
0.012
NA
x(t)
-0.114
0.012
0.003
0.006
Height measured at Fall
Creek stream gauge
intercept
0.356
0.225
-0.046
0.235
x(t-1)
-1.133
0.136
x(t)
1.026
0.136
0.012
0.064
Heating degree days
intercept
0.124
0.101
-0.062
0.115
x(t-1)
-9.343
0.744
NA
x(t)
8.768
0.745
0.264
0.358
Exterior heating index—
calculated based on
temperature and humidity
intercept
-0.405
0.247
0.245
0.299
x(t-1)
0.171
0.011
x(t)
-0.164
0.011
-0.004
0.005
Humidity interior
intercept
-0.97832*
0.407
-0.729
0.404
x(t-1)
0.07735**
0.023
x(t)
-0.05419*
0.023
0.017
0.009
intercept
-2.28123**
0.528
-2.392
0.478
Humidity exterior
x(t-1)
-0.004
0.009
x(t)
0.03714**
0.009
0.034
0.007
Interior heating index
intercept
-1.07508*
0.495
-0.676
0.512
x(t-1)
0.186
0.028
x(t)
-0.172
0.028
0.009
0.007
Rain inches totaled
during observation period
intercept
-0.29376**
0.092
-0.164
0.090
x(t-1)
72.148
14.874
x(t)
71.053
14.872
78.302
15.144
(continued)
10-20
-------
Section 10—Time Series Analysis
Table 10-4. Model Parameters, Standard Errors by Model and Predictor for
X422baseN_AG_radon (X422BN-1): 2011-2012, Lag 1 Models (cont.)
Predictor Name
Model
Term
Model : Y(t)-Y(t-1) = Intercept +
Predictor (t) + Predictor (t-1)
Model : Y(t)-Y(t-1) = Intercept +
Predictor (t)
Estimate
SE
Estimate
SE
422 side, basement vs.
upstairs differential
pressure
intercept
-0.9325**
0.345
-0.626
0.356
x(t-1)
0.143
0.023
x(t)
-0.124
0.023
0.012
0.007
422 side, deep vs.
shallow soil gas
differential pressure
intercept
-0.74167**
0.265
-0.151
0.290
x(t-1)
0.129
0.012
NA
x(t)
-0.114
0.012
0.003
0.006
422 side, subslab vs.
basement differential
pressure
intercept
0.356
0.225
-0.046
0.235
x(t-1)
-1.133
0.136
x(t)
1.026
0.136
0.012
0.064
Depth of snow on the
ground
intercept
0.010
0.088
-0.001
0.088
x(t-1)
-0.909
0.546
x(t)
0.334
0.546
-0.193
0.446
Soil moisture, 13 ft bis
beneath structure
intercept
-0.114
0.206
-0.117
0.205
x(t-1)
-0.028
0.201
x(t)
0.054
0.201
0.028
0.047
Soil moisture 16.5 ft bis
beneath structure
intercept
-0.009
0.089
0.004
0.089
x(t-1)
6.926
4.440
x(t)
-8.474
5.045
-1.209
1.927
Soil moisture 6 ft bis
beneath structure
intercept
0.053
0.201
0.028
0.195
x(t-1)
0.043
0.084
x(t)
-0.043
0.084
0.000
0.001
Soil moisture 13 ft bis
exterior
intercept
-0.008
0.193
-0.012
0.193
x(t-1)
-0.002
0.006
x(t)
0.002
0.006
0.000
0.001
Soil moisture, 3.5 ft bis
exterior
intercept
-0.013
0.111
-0.016
0.110
x(t-1)
-0.001
0.004
x(t)
0.001
0.004
0.000
0.001
Soil moisture 6 ft bis
exterior
intercept
0.033
0.152
0.042
0.152
x(t-1)
0.007
0.005
x(t)
-0.007
0.005
0.000
0.001
Soil temperature 13 ft bis
beneath structure
intercept
-0.665
0.875
-0.298
0.897
x(t-1)
-11.290
1.919
x(t)
11.326
1.925
0.019
0.058
Soil temperature 16.4 ft
bis beneath structure
intercept
-0.717
1.097
-0.405
1.140
x(t-1)
-15.470
2.391
x(t)
15.513
2.396
0.028
0.079
(continued)
10-21
-------
Section 10—Time Series Analysis
Table 10-4. Model Parameters, Standard Errors by Model and Predictor for
X422baseN_AG_radon (X422BN-1): 2011-2012, Lag 1 Models (cont.)
Model : Y(t)-Y(t-1) = Intercept +
Predictor (t) + Predictor (t-1)
Model : Y(t)-Y(t-1) = Intercept +
Predictor (t)
Predictor Name
Model
Term
Estimate
SE
Estimate
SE
Soil temperature 6 ft bis
beneath structure
intercept
-0.221
0.541
-0.450
0.578
x(t-1)
-4.526
0.543
x(t)
4.532
0.543
0.024
0.031
Soil temperature 9 ft bis
beneath structure
intercept
-0.131
0.710
-0.177
0.724
x(t-1)
-6.422
1.253
x(t)
6.423
1.254
0.010
0.042
Soil temperature 1 ft bis
exterior
intercept
-0.167
0.218
-0.186
0.217
x(t-1)
0.3096*
0.141
x(t)
-0.29855*
0.141
0.011
0.013
Soil temperature 13 ft bis
exterior
intercept
-0.033
0.621
-0.147
0.605
x(t-1)
-0.265
1.955
NA
NA
x(t)
0.267
1.962
0.010
0.043
Soil temperature 16.5 ft
bis exterior
intercept
-0.176
1.040
-0.103
1.054
x(t-1)
4.09682**
0.986
x(t)
-4.08292**
0.987
0.007
0.075
Soil temperature 6 ft bis
exterior
intercept
-0.046
0.252
-0.054
0.260
x(t-1)
-2.922
0.478
x(t)
2.917
0.478
0.003
0.016
Temperature at 420
basement north sampling
location
intercept
-0.784
0.469
-0.614
0.484
x(t-1)
0.323
0.054
x(t)
-0.311
0.054
0.010
0.008
Temperature at 420
basement south sampling
location
intercept
-0.794
0.473
-0.604
0.491
x(t-1)
0.345
0.054
x(t)
-0.332
0.054
0.010
0.008
Temperature at 420 first
floor sampling location
intercept
-0.81561*
0.388
-0.459
0.424
x(t-1)
0.297
0.030
x(t)
-0.284
0.030
0.007
0.006
Temperature, 422 first
floor
intercept
-1.41954*
0.653
-0.808
0.684
x(t-1)
0.284
0.038
x(t)
-0.264
0.038
0.011
0.009
Temperature 422
basement north
intercept
-1.201
0.669
-0.877
0.698
x(t-1)
0.476
0.071
x(t)
-0.457
0.071
0.013
0.011
Temperature 422 first
floor
intercept
-1.333
0.814
-1.079
0.829
x(t-1)
0.41748**
0.092
NA
NA
x(t)
-0.39746**
0.092
0.016
0.012
(continued)
10-22
-------
Section 10—Time Series Analysis
Table 10-4. Model Parameters, Standard Errors by Model and Predictor for
X422baseN_AG_radon (X422BN-1): 2011-2012, Lag 1 Models (cont.)
Predictor Name
Model
Term
Model : Y(t)-Y(t-1) = Intercept +
Predictor (t) + Predictor (t-1)
Model : Y(t)-Y(t-1) = Intercept +
Predictor (t)
Estimate
SE
Estimate
SE
Temperature on first floor
of 422 side of duplex
intercept
-1.80372*
0.831
-1.027
0.879
x(t-1)
0.419
0.052
x(t)
-0.395
0.052
0.014
0.012
intercept
-0.384
0.263
0.339
0.321
Temperature exterior
x(t-1)
0.186
0.012
x(t)
-0.179
0.012
-0.006
0.005
Exterior temperature (°F)
intercept
-0.384
0.263
0.339
0.321
x(t-1)
0.186
0.012
NA
NA
x(t)
-0.179
0.012
-0.006
0.005
Temperature exterior,
high during data
collection period
intercept
-0.395
0.264
0.326
0.322
x(t-1)
0.184
0.012
x(t)
-0.178
0.012
-0.005
0.005
Lowest exterior
temperature
intercept
-0.399
0.262
0.323
0.320
x(t-1)
0.186
0.012
x(t)
-0.180
0.012
-0.005
0.005
Temperature, humidity
and wind index
intercept
-0.354
0.224
0.265
0.278
x(t-1)
0.167
0.010
x(t)
-0.161
0.010
-0.005
0.004
Wind chill
intercept
-0.332
0.237
0.350
0.295
x(t-1)
0.180
0.011
x(t)
-0.174
0.011
-0.006
0.005
intercept
-0.368
0.297
-0.335
0.248
Wind direction (average)
x(t-1)
0.000
0.001
x(t)
0.002
0.001
0.002
0.001
Wind direction (of high
during measurement
period)
intercept
-0.349
0.291
-0.318
0.245
x(t-1)
0.000
0.001
x(t)
0.002
0.001
0.002
0.001
Wind run (a function of
wind speed and duration)
intercept
-0.40685*
0.199
-0.3598*
0.177
x(t-1)
0.050
0.097
x(t)
0.169
0.097
0.19307*
0.085
High wind speed during
measurement period
intercept
-0.71054**
0.244
-0.5822**
0.213
x(t-1)
0.024
0.022
x(t)
0.04746*
0.022
0.05863**
0.020
Average wind speed
during measurement
period
intercept
-0.40685*
0.199
-0.3598*
0.177
x(t-1)
0.025
0.048
x(t)
0.084
0.048
0.09654*
0.042
Note: Lag 1 models; SE = standard error
** Significant at 1 % level of significance * Significant at 5% level of significance
10-23
-------
Section 10—Time Series Analysis
In Table 10-4, the correlation of radon concentrations with wind run, high wind speed, and average wind
speed is in each case positive, so high winds are predicted to result in high radon. This agrees with a
qualitative observation in Section 6—that peaks in wall port VOCs are associated with wind speed. A
positive relationship between increasing wind speed and increasing chloroform is shown graphically in
Section 9.2.4.
The negative association of barometric pressure with radon shown in Table 10-4 makes intuitive physical
sense since increased barometric pressure should decrease the emanation of soil gas. Barometric pressure
changes are associated with some of the unusual subslab-to-interior differential pressure events discussed
in Section 9.1. However, the visual analysis of the VOC data in Section 9.2 did not reveal a clear,
monotonic barometric pressure trend. Barometric pressure drops, especially those accompanied by
rainfall, have been associated with increased indoor radon in several studies (see review by Lewis and
Houle [2009]).
Although the relationship between indoor humidity and radon was judged to be statistically significant
(not subject to chance), note that the signs of the X(t-1) and X(t) terms are opposite, and the absolute
value of those terms is similar and relatively small. Note also that indoor humidity changes slowly from
day to day (Table 10-4). Therefore, these terms will likely nearly cancel out in many cases. The exterior
humidity analysis is thus more interesting, since, although the signs are opposite, the absolute values are
very different. The term that would be expected to be dominant would be the term relating to today's
humidity, which would be associated with increased radon. An association of greater radon emanation
with higher humidity has also been observed in laboratory experiments with soil (Hosoda, 2008).24
Similarly, although relationships with exterior soil temperature at 1 ft and 16.5 ft were judged unlikely to
be attributable to chance (Table 10-4) because the signs of the terms in the full model are opposite, soil
temperatures change relatively slowly, and the magnitude of the slopes was similar; it is unlikely that soil
temperature explains a large amount of the radon variation.
Table 10-5 displays the coefficient estimate and standard error for those predictor variables that required
a lag 2 day term in the model. The predictor 420 side, subslab vs. basement differential pressure had a
relationship unlikely to be due to chance with the 422 basement radon. However, the magnitude of this
relationship is likely to be small since the signs of the coefficients for the differential pressures on various
days are opposite. For example, in the full model, the radon increases 0.55 pCi/L per pascal of differential
pressure measured blowing into the building today. But that is likely to be essentially canceled out in
many cases by the effect of the terms for yesterday's differential pressure and the day before yesterday's
differential pressure, which are negative. Note, however, that in the first reduced model, the net effect is
likely to be positive—the statistically significant term for today's differential pressure is likely to
outweigh the term for yesterday's differential pressure.
Table 10-6 displays the models for predictor 422 side basement vs. exterior differential pressure, which
necessitated 4 lag days in the model because of the results of the analysis of autocorrelation and partial
autocorrelation. The two higher order lags were not significant, so they were dropped from the full model
one at a time. Relationships were found between this predictor and radon concentration, which are
unlikely to be due to chance. However, these relationships will often cancel and not explain much of the
variation in radon because the signs of the slopes of the x(t) and x(t-l) terms are opposite. The meaning
of this relationship can perhaps be most easily understood with an example using the model that has been
reduced to only the X(t) and x(t-l) terms. Assume a case when the basement vs. exterior pressure is
24Generation and control of radon from soil
10-24
-------
Section 10—Time Series Analysis
1 pascal today and was 0 pascal yesterday. In that simplified case, the model would predict that the radon
concentration would be decreased by 0.4 pCi/L. That result is reasonable, because with a basement such
Table 10-5. Model Parameters, Standard Errors by Predictor for X422baseN_AG_radon (X422BN-
1): 2011-1012, Lag 2 Models
Y(t)-Y(t-1) = Intercept +
Predictor (t) +
Predictor(t-1)+
Predictor(t-2)
Y(t)-Y(t-1) = Intercept
+ Predictor (t) +
Predictor(t-1)+
Y (t)—Y(t—1) =
Intercept +
Predictor (t)
Predictor Name
Model Term
Estimate
SE
Estimate
SE
Estimate
SE
420 side, subslab vs.
basement differential
pressure
intercept
0.006
0.243
-0.190
0.223
-0.364
0.203
x(t)
0.55596*
0.223
0.58588**
0.223
0.348
0.178
x(t-1)
-0.088
0.264
-0.406
0.223
x(t"2)
-0.48452*
0.223
'Significant at 5% level of significance
"Significant at 1 % level of significance
Table 10-6. Model Parameters, Standard Errors by Predictor for X422baseN_AG_radon (X422BN-
1), Lag 4 Models
Predictor
Name
Model
Term
Y(t)-Y(t-1) =
Intercept +
Predictor (t) +
Predictor(t-1)+
Predictor(t-2) +
Y(t)-Y(t-1) =
Intercept +
Predictor (t) +
Predictor(t-1)+
Predictor(t-2) +
Y(t)—Y(t—1) =
Intercept +
Predictor (t) +
Predictor(t-1)+
Predictor(t-2)
Y(t)-Y(t-1) =
Intercept +
Predictor (t) +
Predictor(t-I)
Y(t)-Y(t-1) =
Intercept +
Predictor (t)
Predictor
t-4)
Predictor(t-3)+
Estimatel
SE1
Estimate2
|sE2l
Estimate3
SE3
Estimate4
SE4
Estimate5
SE5
422 side
basement
vs. exterior
differential
pressure
intercept
0.005
0.089
-0.005
0.089
-0.003
0.089
-0.002
0.088
-0.006
0.090
x(t)
-0.36984"
0.091
-0.38245"
0.090
-0.3958"
0.088
-0.37946"
0.086
-0.094
0.060
x(t—1)
0.35053"
0.104
0.3452"
0.104
0.34485"
0.103
0.39558"
0.086
NA
NA
x(t—2)
0.109
0.104
0.111
0.103
0.085
0.087
NA
NA
NA
NA
x(t—3)
0.051
0.104
-0.053
0.088
NA
NA
NA
NA
NA
NA
x(t—4)
-0.138
0.089
NA
NA
NA
NA
NA
NA
NA
NA
"Significant at 1% level of significance
as this that is partially above ground, a positive pressure vs. exterior air means that some air is directly
leaking out of the basement above ground.
10.4 Correlation between Radon Concentration Time Series for 422 2nd Floor
Office (2011-2012) and Predictor Variables
Similarly to the basement time series, the variable "drop in barometric pressure" has a significant
correlation with radon concentration (see Table 10-7). The positive coefficients indicate that as
barometric pressure goes down, the radon concentration upstairs goes up. This is an expected result.
Table 10-8 shows the results of models including predictors requiring Lag 1 terms to model X4220F2-1
for the period 2011 to 2012.
10-25
-------
Section 10—Time Series Analysis
Table 10-7. Model Parameters, Standard Errors by Predictor for X422office_2nd_AG_radon
Concentration (X422QF2-1): 2011-2012, No Lag Terms in Model
Time Series
Model : Y(t)-Y(t-1) = Intercept + Predictor (t)
Predictor Name = x(t)
Model Term
Estimate
SE
X422baseN_AG_radon
Drop in barometric pressure
intercept
-0.009
0.044
x(t)
0.986**
0.102
Rain highest rate during
observation period
intercept
-0.035
0.049
x(t)
1.622
0.855
"Significant at 1% level of significance
Table 10-8. Model Parameters, Standard Errors by Model and Predictor for Time Series Analysis
of Radon in 422 Office: 2011-2012, Lag 1 Models
Predictor Name
Model
Model : Y(t)-Y(t-1) = Intercept +
Predictor (t) + Predictor (t-1)
Model : Y(t)-Y(t-1) = Intercept +
Predictor (t)
Estimate
SE
Estimate
SE
Air density interior
Intercept
3.98887*
1.777
2.252
1.810
x(t)
377.08738**
74.255
-30.798
24.659
x(t-1)
-431.49046**
74.388
Intercept
31.96701**
7.873
10.989
7.550
Barometric pressure
x(t)
1.14704**
0.330
-0.367
0.252
x(t-1)
-2.21408**
0.330
Cooling degree day
Intercept
-0.023
0.058
0.026
0.060
x(t)
-5.26306**
0.776
-0.292
0.320
x(t-1)
5.43776**
0.781
Intercept
-0.44352*
0.193
-0.336
0.194
Dew point, interior
x(t)
-0.04195**
0.013
0.007
0.004
x(t-1)
0.05078**
0.013
Dew point, exterior
Intercept
-0.37804**
0.146
-0.065
0.158
x(t)
-0.06071**
0.006
0.001
0.003
x(t-1)
0.06835**
0.006
Height measured at Fall
Creek stream gauge
Intercept
0.101
0.126
-0.060
0.128
x(t)
0.43301**
0.076
0.016
0.035
x(t-1)
-0.46503**
0.076
Heating degree days
Intercept
0.055
0.054
-0.047
0.062
x(t)
4.91909**
0.405
0.188
0.195
x(t-1)
-5.19542**
0.404
(continued)
10-26
-------
Section 10—Time Series Analysis
Table 10-8. Model Parameters, Standard Errors by Model and Predictor for Time Series Analysis
of Radon in 422 Office: 2011-2012, Lag 1 Models (cont.)
Predictor Name
Model
Model : Y(t)-Y(t-1) = Intercept +
Predictor (t) + Predictor (t-1)
Model : Y(t)-Y(t-1) = Intercept +
Predictor (t)
Estimate
SE
Estimate
SE
Exterior heating index—
calculated based on
temperature and humidity
Intercept
-0.204
0.139
0.127
0.163
x(t)
-0.08386**
0.006
-0.002
0.003
x(t-1)
0.08723**
0.006
Humidity interior
Intercept
-0.61177**
0.220
-0.47041*
0.218
x(t)
-0.02534*
0.012
0.011*
0.005
x(t-1)
0.03973**
0.012
Intercept
-1.17568**
0.290
-1.18449**
0.262
Humidity exterior
x(t)
0.0172**
0.005
0.01698**
0.004
x(t-1)
0.000
0.005
Interior heating index
Intercept
-0.390
0.279
-0.272
0.280
x(t)
-0.05113**
0.016
0.004
0.004
x(t-1)
0.05631**
0.016
Rain inches totaled during
observation period
Intercept
-0.12091*
0.051
-0.056
0.050
x(t)
20.3433*
8.317
23.93887**
8.427
x(t-1)
35.95066**
8.317
422 side, basement vs.
upstairs differential
pressure
Intercept
0.000
0.047
0.008
0.049
x(t)
-0.68146**
0.100
-0.10379*
0.050
x(t-1)
0.65764**
0.100
422 side, deep vs. shallow
soil gas differential
pressure
Intercept
0.058
0.056
0.010
0.054
x(t)
0.003
0.015
-0.009
0.015
x(t-1)
-0.04185**
0.015
422 side, subslab vs.
basement differential
pressure
Intercept
-0.041
0.135
-0.068
0.120
x(t)
0.039
0.050
0.024
0.042
x(t-1)
-0.024
0.050
Depth of snow on the
ground
intercept
0.001
0.048
-0.006
0.048
x(t)
0.256
0.298
-0.076
0.244
x(t-1)
-0.573
0.298
Soil moisture, 13 ft bis
beneath structure
intercept
-0.039
0.111
-0.036
0.111
x(t)
-0.024
0.110
0.007
0.025
x(t-1)
0.032
0.110
Soil moisture 16.5 ft bis
beneath structure
intercept
-0.009
0.048
-0.002
0.048
x(t)
-8.04869**
2.752
-1.482
1.053
x(t-1)
6.25195*
2.422
(continued)
10-27
-------
Section 10—Time Series Analysis
Table 10-8. Model Parameters, Standard Errors by Model and Predictor for Time Series Analysis
of Radon in 422 Office: 2011-2012, Lag 1 Models (cont.)
Predictor Name
Model
Model : Y(t)-Y(t-1) = Intercept +
Predictor (t) + Predictor (t-1)
Model : Y(t)-Y(t-1) = Intercept +
Predictor (t)
Estimate
SE
Estimate
SE
Soil moisture 6 ft bis
beneath structure
intercept
0.026
0.109
-0.009
0.106
x(t)
-0.058
0.046
0.000
0.001
x(t-1)
0.058
0.046
Soil moisture 13 ft bis
exterior
intercept
-0.042
0.104
-0.042
0.104
x(t)
0.001
0.003
0.000
0.000
x(t-1)
-0.001
0.003
Soil moisture, 3.5 ft bis
exterior
intercept
-0.015
0.060
-0.013
0.060
x(t)
0.000
0.002
0.000
0.001
x(t-1)
0.001
0.002
Soil moisture 6 ft bis
exterior
intercept
-0.029
0.082
-0.026
0.082
x(t)
-0.002
0.003
0.000
0.000
x(t-1)
0.002
0.003
Soil temperature 13 ft bis
beneath structure
intercept
-0.280
0.489
-0.190
0.487
x(t)
3.25428**
1.075
0.012
0.032
x(t-1)
-3.23827**
1.072
Soil temperature 16.4 ft bis
beneath structure
intercept
-0.330
0.616
-0.245
0.618
x(t)
4.58871**
1.342
0.017
0.043
x(t-1)
-4.56774**
1.338
Soil temperature 6 ft bis
beneath structure
intercept
-0.064
0.304
-0.163
0.314
x(t)
1.89452**
0.304
0.009
0.017
x(t-1)
-1.89359**
0.305
Soil temperature 9 ft bis
beneath structure
intercept
-0.057
0.395
-0.091
0.393
x(t)
1.70464*
0.696
0.005
0.023
x(t-1)
-1.70318*
0.696
Soil temperature 1 ft bis
exterior
intercept
-0.071
0.118
-0.081
0.118
x(t)
-0.15712*
0.077
0.005
0.007
x(t-1)
0.16196*
0.077
Soil temperature 13 ft bis
exterior
intercept
0.019
0.337
-0.105
0.329
x(t)
-1.180
1.063
0.007
0.024
x(t-1)
1.180
1.059
Soil temperature 16.5 ft bis
exterior
intercept
-0.248
0.571
-0.228
0.574
x(t)
-1.62806**
0.540
0.016
0.041
x(t-1)
1.64586**
0.539
(continued)
10-28
-------
Section 10—Time Series Analysis
Table 10-8. Model Parameters, Standard Errors by Model and Predictor for Time Series Analysis
of Radon in 422 Office: 2011-2012, Lag 1 Models (cont.)
Predictor Name
Model
Model : Y(t)-Y(t-1) = Intercept +
Predictor (t) + Predictor (t-1)
Model : Y(t)-Y(t-1) = Intercept +
Predictor (t)
Estimate
SE
Estimate
SE
Soil temperature 6 ft bis
exterior
intercept
-0.020
0.140
-0.030
0.141
x(t)
1.00762**
0.266
0.002
0.009
x(t-1)
-1.00914**
0.267
Temperature at 420
basement north sampling
location
intercept
-0.404
0.256
-0.312
0.264
x(t)
-0.1692**
0.030
0.005
0.004
x(t-1)
0.17576**
0.030
Temperature at 420
basement south sampling
location
intercept
-0.410
0.258
-0.309
0.268
x(t)
-0.17917**
0.030
0.005
0.004
x(t-1)
0.18575**
0.030
Temperature at 420 first
floor sampling location
intercept
-0.400
0.215
-0.221
0.232
x(t)
-0.14556**
0.017
0.003
0.003
x(t-1)
0.15167**
0.017
intercept
-0.457
0.373
-0.312
0.374
Temperature, 422 first floor
x(t)
-0.06239**
0.021
0.004
0.005
x(t-1)
0.06843**
0.021
Temperature 422
basement north
intercept
-0.563
0.371
-0.419
0.380
x(t)
-0.20225**
0.039
0.006
0.006
x(t-1)
0.21089**
0.039
intercept
-0.595
0.452
-0.546
0.452
Temperature 422 first floor
x(t)
-0.071
0.051
0.008
0.007
x(t-1)
0.080
0.051
Temperature on first floor
of 422 side of duplex
intercept
-0.647
0.475
-0.421
0.480
x(t)
-0.11474**
0.030
0.006
0.007
x(t-1)
0.1235**
0.030
intercept
-0.200
0.148
0.167
0.175
Temperature exterior
x(t)
-0.09139**
0.007
-0.003
0.003
x(t-1)
0.09471**
0.007
Exterior temperature (°F)
intercept
-0.200
0.148
0.167
0.175
x(t)
-0.09139**
0.007
-0.003
0.003
x(t-1)
0.09471**
0.007
Temperature exterior, high
during data collection
period
intercept
-0.206
0.149
0.159
0.175
x(t)
-0.09043**
0.007
-0.003
0.003
x(t-1)
0.09383**
0.007
(continued)
10-29
-------
Section 10—Time Series Analysis
Table 10-8. Model Parameters, Standard Errors by Model and Predictor for Time Series Analysis
of Radon in 422 Office: 2011-2012, Lag 1 Models (cont.)
Predictor Name
Model
Model : Y(t)-Y(t-1) = Intercept +
Predictor (t) + Predictor (t-1)
Model : Y(t)-Y(t-1) = Intercept +
Predictor (t)
Estimate
SE
Estimate
SE
Lowest exterior
temperature
intercept
-0.208
0.148
0.158
0.175
x(t)
-0.09138**
0.007
-0.003
0.003
x(t-1)
0.09487**
0.007
Temperature, humidity and
wind index
intercept
-0.178
0.129
0.128
0.152
x(t)
-0.07969**
0.006
-0.002
0.002
x(t-1)
0.08267**
0.006
Wind chill
intercept
-0.172
0.137
0.163
0.161
x(t)
-0.08563**
0.006
-0.003
0.003
x(t-1)
0.08853**
0.006
intercept
-0.141
0.160
-0.30692*
0.134
Wind direction (average)
x(t)
0.00203**
0.001
0.0016*
0.001
x(t-1)
-0.001
0.001
Wind direction (of high
during measurement
period)
intercept
-0.159
0.157
-0.30257*
0.133
x(t)
0.00204**
0.001
0.0016*
0.001
x(t-1)
-0.001
0.001
Wind run is a function of
wind speed and duration
intercept
-0.128
0.106
0.082
0.097
x(t)
-0.16169**
0.052
-0.050
0.046
x(t-1)
0.22816**
0.052
High wind speed during
measurement period
intercept
-0.236
0.131
0.042
0.117
x(t)
-0.02996*
0.012
-0.005
0.011
x(t-1)
0.05337**
0.012
Average wind speed
during measurement
period
intercept
-0.128
0.106
0.082
0.097
x(t)
-0.08085**
0.026
-0.025
0.023
x(t-1)
0.11408**
0.026
"Significant at 1% level of significance
'Significant at 5% level of significance
The magnitude of effect is large for rain (Table 10-8), which increased indoor radon substantially
whether it fell on the day in question or the previous day. This effect has been documented by others in
other structures. Lewis and Houle (2009), for example, review Nazaroff s work and state:
In a house with a crawl space, a modest drop in barometric pressure and a period of heavy rain caused
the indoor radon and crawl space radon to rise to its highest level during a 5-week measurement period.
The rain may be acting in one of two ways; it could act by tunneling the radon from the soil into the crawl
space: with heavy rain, the permeability of the soil surrounding the house is greatly reduced while the
permeability of the soil beneath the house remains unchanged; as the barometric pressure falls, soil gas
then flows into the crawl space at a higher rate than it does out of the soil surrounding the house. The
10-30
-------
Section 10—Time Series Analysis
alternative explanation is that the downward movement of water through the soil may act like a piston and
displace the radon, which then flows into the crawl space (Lewis & Houle, 1985).
Numerous predictor variables showed the same pattern in the results:
¦ The results for x(t) and x(t-l) were both judged to be unlikely to be due to chance.
¦ The results for x(t) and x(t-l) slopes were similar in magnitude but opposite in sign.
¦ The reduced model is not significant.
Results of this type suggest that the change from day to day in the predictor variable may be associated
with a change in radon concentration but that the absolute value of the predictor variable is less important
(Nau, 2005b). In at least some cases, this result can make physical sense. For example, let us use the
example of exterior temperature. For a day when the temperature is 40°F and yesterday's temperature was
60°F, the model predicts that the radon concentration will have increased 1.9 pCi/L—falling temperatures
increase radon, which is consistent with the stack effect. In the opposite case where the temperature
yesterday was 40°F and today 60°F, a decrease in radon is predicted.
Another example where the modeled behavior appears reasonable is the height of the Fall Creek stream
gauge, which we have shown to be tightly linked to the on-site water table (Chapter 11, U.S. EPA,
2012a). In an example where the stream gauge is 5 ft today and 2 ft yesterday, the model predicts an
increase in radon of 1.3 pCi/L. This is what would be expected—the rising stream and; thus, water table
would tend to "squeeze" radon containing soil gas up into the structure. It could also be that the rise in the
stream gauge is a surrogate parameter for on-site rainfall, which could be expected to have an effect on
the vadose zone, increasing radon indoors.
Since the change in so many parameters appears to be related to radon concentrations, it may be more
fruitful to examine those that stand out from the pattern. Surprisingly, the parameters that were not
significantly associated with radon in this data set included 422 subslab vs. basement differential pressure
and the depth of snow on the ground.
The few parameters that were significant even in the reduced model of this data set include rainfall
(discussed above as positively correlated), humidity (positively correlated, see discussion in Section 10.3)
and wind direction (positively correlated).
Table 10-9 shows the result for models exploring the correlation between predictor 420 side, subslab vs.
basement differential pressure, and radon concentration (X4220F2-1) for time period 2011 to 2012. The
non-significant results suggest that subslab vs. basement differential pressure on the 420 side of the
duplex is not correlated with radon concentrations in the office on the second floor of the 422 side. The
range of variation of this differential pressure is shown in our previous report (Figure 10-5 of U.S. EPA
2012a).
Table 10-10 shows the results of the analysis to determine the correlation between 422 side basement vs.
exterior differential pressure and radon concentration (X4220F2-1) for time period 2011 to 2012. Current
and past observations of the 422 side basement vs. exterior differential pressure are highly correlated with
radon concentrations. The current observation has negative correlation and the past observation has a
positive correlation. The difference in signs indicates that the change in the predictor variable (differential
pressure) is likely more important than its absolute magnitude.
10-31
-------
Section 10—Time Series Analysis
Table 10-9. Model Parameters, Standard Errors by Predictor for X422office_2nd_AG_radon
Concentration (X422QF2-1): 2011-2012. Lag 2 Models
Y(t)-Y(t-1) = Intercept +
Predictor (t) +
Predictor(t-1)+
Predictor(t-2)
Y(t)-Y(t-1) = Intercept
+ Predictor (t) +
Predictor(t-1)+
Y (t)—Y (t—1) =
Intercept +
Predictor (t)
Predictor Name
Model
Term
Estimate
SE
Estimate
SE
Estimate
SE
420 side, subslab vs.
basement differential
pressure
intercept
-0.016
0.131
-0.086
0.120
-0.106
0.109
x(t)
0.101
0.120
0.116
0.120
0.101
0.096
x(t-1)
0.109
0.142
-0.034
0.120
x(t"2)
-0.203
0.120
Table 10-10. Model Parameters, Standard Errors by Predictor for X422baseN_AG_radon
(X4220F2-1): 2011-2012. Lag4 Models
Y (t)—Y (t—1) =
Intercept +
Predictor (t) +
Predictor(t-1)+
Predictor(t—2) +
Predictor(t-3)+
Predictor(t-4)
Y (t)—Y (t—1) =
Intercept +
Predictor (t) +
Predictor(t-1)+
Predictor(t—2) +
Predictor(t-3)+
Y(t)—Y(t—1) =
Intercept +
Predictor (t) +
Predictor(t-1)+
Y (t) —Y(t—1) =
Intercept +
Predictor (t) +
Predictor(t—1)
Y(t)—Y (t—1) =
Intercept +
Predictor (t)
Predictor
Name
Model
Term
Estimate
I SE
Estimate
I SE
Estimate
I SE
Estimate
I SEI
Estimate
lSEl
422 side
basement
vs. exterior
differential
pressure
intercept
-0.001
0.047
-0.011
0.047
-0.008
0.047
-0.007
0.047
-0.008
0.049
x(t)
-0.25521**
0.048
-0.25277**
0.048
-0.25845**
0.047
-0.2452**
0.046
-0.052
0.032
x(M)
0.21061**
0.055
0.21655**
0.055
0.21495**
0.055
0.2674**
0.046
x(t-2)
0.065
0.055
0.079
0.055
0.082
0.046
x(t-3)
0.017
0.055
-0.008
0.047
x(t-4)
0.003
0.047
"Significant at 1% level of significance
'Significant at 5% level of significance
10.5 Correlation between Radon Concentration Time Series for 422 Basement
South (2012-2013) and Predictor Variables
For time period 2012 to 2013, the variable mitigation was incorporated in the model as a controlling
variable to account for the testing of subslab depressurization. Thus, all of the model equations have a
mitigation term incorporated. Table 10-11 displays the results for models that include predictors not
needing lag terms. Except for soil moisture 16.5 ft bis beneath structure, all three variables in Table 10-11
have significant positive association with radon concentration after accounting for the effect of mitigation
being on during the time period. This is a reasonable result because, as discussed in previous sections, we
expect from the literature that drops in barometric pressure and rainfall increase indoor radon. In all
models, mitigation was significant and the coefficients were negative, suggesting the expected negative
correlation that falling barometric pressure increases radon concentration.
10-32
-------
Section 10—Time Series Analysis
Table 10-11. Model Parameters, Standard Errors by Predictor for 422 Basement Radon: 2012-
2013. No Lag Terms in Model
Time Series
Model: Y(t)-Y(t-1) =
Intercept + Mitigation(t) + Predictor (t
Predictor Name = x(t)
Model Term
Estimate
SE
X422baseS_AG_radon
intercept
0.299
0.157
Drop in barometric pressure
x(t)
0.798**
0.193
mitigation(t)
-0.518*
0.216
Rain inches totaled during
observation period
intercept
0.175
0.166
x(t)
65.318**
24.607
mitigation(t)
-0.474*
0.221
Rain highest rate during
observation period
intercept
0.207
0.167
x(t)
5.1572*
2.561
mitigation(t)
-0.4584*
0.223
Soil moisture 16.5 ft bis beneath
structure
intercept
0.32675*
0.158
x(t)
-0.336
0.690
mitigation(t)
-0.532*
0.216
"Significant at 1% level of significance
'Significant at 5% level of significance
Table 10-12 shows the results of models including predictors Lag 1 terms for radon in the 422 basement
for the period 2012 to 2013. In a large number of cases, the patterns seen in the 422 office data set in 2011
to 2012 (Table 10-8) are repeated in this basement 2012 to 2013 data set. The sign and magnitude of the
coefficients for external temperature, barometric pressure, and height of Fall Creek rain gauge are quite
similar for example. Thus, the interpretations provided in Section 10.4 also apply here and will not be
repeated for brevity.
Note that in Table 10-12 the mitigation effect is calculated in each model, but the calculated values are
similar to each other although not identical.
The Table 10-12 data set is also similar to Table 10-8 in that the depth of snow on the ground is not
significant in either data set. As discussed in Section 9, some literature suggests that radon moves easily
through snow packs, so this result is reasonable.
Table 10-12. Model Parameters, Standard Errors by Model and Predictor for 422 Basement Radon:
2012-2013 Lag 1 Models
Model : Y(t)-Y(t-1) =
Intercept + Mitigation(t) +
Predictor (t) + Predictor (t-1)
Model : Y(t)-Y(t-1) =
Intercept + Mitigation(t) +
Predictor (t)
Predictor Name
Model Term
Estimate
SE
Estimate
SE
Air density interior
Intercept
34.46805**
9.136
20.88483*
8.837
mitigation(t)
-0.51691*
0.214
-0.50571*
0.221
x(t)
354.740
187.911
-276.63943*
118.699
x(t-1)
-813.64637**
191.384
(continued)
10-33
-------
Section 10—Time Series Analysis
Table 10-12. Model Parameters, Standard Errors by Model and Predictor for422 Basement Radon:
2012-2013 Lag 1 Models (cont.)
Model : Y(t)-Y(t-1) =
Intercept + Mitigation(t) +
Predictor (t) + Predictor (t-1)
Model : Y(t)-Y(t-1) =
Intercept + Mitigation(t) +
Predictor (t)
Predictor Name
Model Term
Estimate
SE
Estimate
SE
Intercept
40.47232*
18.416
10.929
16.622
Barometric pressure
mitigation(t)
-0.6509**
0.228
-0.54489*
0.230
x(t)
0.945
0.658
-0.353
0.552
x(t-1)
-2.27969**
0.660
Cooling degree day
Intercept
0.154
0.159
0.293
0.172
mitigation(t)
-0.43284*
0.208
-0.51052*
0.226
x(t)
-18.75498**
3.894
0.023
2.971
x(t-1)
26.55995**
3.950
Intercept
-0.781
0.469
-0.487
0.465
Dew point, interior
mitigation(t)
-0.389
0.225
-0.422
0.228
x(t)
-0.049
0.026
0.020
0.011
x(t-1)
0.07627**
0.026
Dew point, exterior
intercept
-0.474
0.335
0.083
0.350
mitigation(t)
-0.413
0.210
-0.48707*
0.226
x(t)
-0.067**
0.014
0.006
0.008
x(t-1)
0.08762**
0.014
intercept
0.438
0.350
0.217
0.341
Height measured at Fall
Creek stream gauge
mitigation(t)
-0.55895*
0.221
-0.53849*
0.221
x(t)
0.411*
0.166
0.030
0.081
x(t-1)
-0.43615**
0.166
Heating degree days
intercept
0.57716*
0.233
0.296
0.244
mitigation(t)
-0.46499*
0.209
-0.51029*
0.225
x(t)
3.89451**
0.742
-0.008
0.419
x(t-1)
-4.56292**
0.739
intercept
-0.392
0.367
0.256
0.395
Exterior heating index—
calculated based on
temperature and humidity
mitigation(t)
-0.44966*
0.204
-0.50762*
0.226
x(t)
-0.08667**
0.014
0.001
0.008
x(t-1)
0.10184**
0.014
Humidity interior
intercept
-0.122
0.353
0.059
0.356
mitigation(t)
-0.44899*
0.225
-0.47586*
0.229
x(t)
-0.06728**
0.025
0.007
0.010
x(t-1)
0.07987**
0.024
intercept
-1.370
0.840
-1.249
0.749
Humidity exterior
mitigation(t)
-0.4653*
0.224
-0.47371*
0.222
x(t)
0.019
0.012
0.02065*
0.010
x(t-1)
0.004
0.012
(continued)
10-34
-------
Section 10—Time Series Analysis
Table 10-12. Model Parameters, Standard Errors by Model and Predictor for422 Basement Radon:
2012-2013 Lag 1 Models (cont.)
Model : Y(t)-Y(t-1) =
Intercept + Mitigation(t) +
Predictor (t) + Predictor (t-1)
Model : Y(t)-Y(t-1) =
Intercept + Mitigation(t) +
Predictor (t)
Predictor Name
Model Term
Estimate
SE
Estimate
SE
Interior Heating Index
intercept
-5.03405**
1.720
-3.92999*
1.647
mitigation(t)
-0.386
0.223
-0.411
0.224
x(t)
-0.031
0.049
0.06073*
0.024
x(t-1)
0.10743*
0.051
intercept
0.270
0.162
0.275
0.161
Depth of snow on the
ground
mitigation(t)
-0.50981*
0.216
-0.51539*
0.215
x(t)
0.069
0.158
0.104
0.085
x(t-1)
0.042
0.159
Soil moisture, 13 ft bis
beneath structure
intercept
0.413
0.263
0.406
0.259
mitigation(t)
-0.54499*
0.221
-0.54187*
0.219
x(t)
0.017
0.229
-0.018
0.042
x(t-1)
-0.036
0.230
intercept
-0.802
0.804
-0.555
0.799
Soil moisture 6 ft bis
mitigation(t)
-0.427
0.219
-0.5213*
0.215
beneath structure
x(t)
-0.127
0.067
0.005
0.005
x(t-1)
0.133
0.068
Soil moisture, 3.5 ft bis
exterior
intercept
0.347
0.290
0.464
0.283
mitigation(t)
-0.4946*
0.233
-0.46782*
0.234
x(t)
-0.005
0.002
-0.001
0.001
x(t-1)
0.004
0.002
intercept
-0.596
0.452
-0.309
0.447
Soil moisture 6 ft bis
exterior
mitigation(t)
-0.60875**
0.215
-0.58545**
0.218
x(t)
-0.005
0.003
0.003
0.002
x(t-1)
0.00914**
0.003
Temperature, 422 first floor
intercept
-5.207*
2.017
-4.3276*
1.933
mitigation(t)
-0.435
0.223
-0.44668*
0.223
x(t)
-0.013
0.057
0.06414*
0.027
x(t-1)
0.090
0.059
intercept
-0.358
0.373
0.284
0.399
Exterior temperature (°F)
mitigation(t)
-0.45692*
0.205
-0.51003*
0.226
x(t)
-0.08621**
0.014
0.000
0.008
x(t-1)
0.10054**
0.014
Temperature exterior, high
during data collection period
intercept
-0.374
0.375
0.260
0.401
mitigation(t)
-0.4561*
0.205
-0.50818*
0.226
x(t)
-0.08492**
0.014
0.001
0.008
x(t-1)
0.09949**
0.014
(continued)
10-35
-------
Section 10—Time Series Analysis
Table 10-12. Model Parameters, Standard Errors by Model and Predictor for422 Basement Radon:
2012-2013 Lag 1 Models (cont.)
Model : Y(t)-Y(t-1) =
Intercept + Mitigation(t) +
Predictor (t) + Predictor (t-1)
Model : Y(t)-Y(t-1) =
Intercept + Mitigation(t) +
Predictor (t)
Predictor Name
Model Term
Estimate
SE
Estimate
SE
intercept
-0.370
0.371
0.266
0.397
Lowest exterior temperature
mitigation(t)
-0.45514*
0.205
-0.50864*
0.226
x(t)
-0.08599**
0.014
0.001
0.008
x(t-1)
0.10073**
0.014
Temperature, humidity and
wind index
intercept
-0.224
0.322
0.309
0.345
mitigation(t)
-0.45807*
0.205
-0.51239*
0.226
x(t)
-0.07583**
0.013
0.000
0.007
x(t-1)
0.08785**
0.012
intercept
-0.197
0.326
0.331
0.348
Wind chill
mitigation(t)
-0.46436*
0.206
-0.51439*
0.226
x(t)
-0.07522**
0.013
-0.001
0.007
x(t-1)
0.08654**
0.013
Wind direction (average)
intercept
0.066
0.405
-0.022
0.346
mitigation(t)
-0.50485*
0.225
-0.50827*
0.223
x(t)
0.002
0.002
0.002
0.002
x(t-1)
-0.001
0.002
intercept
0.119
0.405
0.045
0.348
Wind direction (of high
during measurement
period)
mitigation(t)
-0.50456*
0.225
-0.50736*
0.224
x(t)
0.002
0.002
0.001
0.002
x(t-1)
-0.001
0.002
Wind run is a function of
wind speed and duration
intercept
0.092
0.256
0.244
0.236
mitigation(t)
-0.55184*
0.229
-0.52169*
0.227
x(t)
-0.064
0.113
0.027
0.096
x(t-1)
0.175
0.113
intercept
0.009
0.289
0.181
0.265
High wind speed during
measurement period
mitigation(t)
-0.56257*
0.229
-0.5332*
0.228
x(t)
-0.009
0.027
0.012
0.023
x(t-1)
0.040
0.027
Average wind speed during
measurement period
intercept
0.092
0.256
0.244
0.236
mitigation(t)
-0.55184*
0.229
-0.52169*
0.227
x(t)
-0.032
0.056
0.014
0.048
x(t-1)
0.088
0.056
"Significant at 1% level of significance
'Significant at 5% level of significance
10-36
-------
Section 10—Time Series Analysis
10.6 Correlation between Radon Concentration Time Series for 422 Office on
2nd Floor and Predictor Variables
For time period 2012 to 2013, the variable mitigation was incorporated in the model for the office as well
as a controlling variable to account for the testing of subslab depressurization. Table 10-13 displays the
results for models that include predictors not needing lag terms in the model. Drops in barometric
pressure and rainfall have significant positive association with radon concentration after accounting for
the effect of mitigation being on during the time period. These results are expected from the literature as
discussed earlier in this section.
Table 10-13. Model Parameters, Standard Errors by Predictor for X422office_2nd_AG_radon
Concentration (X422QF2-2): 2012-2013. No Lag Terms in Model
Time Series
Model : Y(t)-Y(t-1)
= Intercept + Predictor (t)
Predictor Name = x(t)
Model Term
Estimate
SE
X422office_2nd_AG_radon
intercept
0.084
0.076
Drop in Barometric Pressure
x(t)
0.561**
0.093
mitigation(t)
-0.174
0.104
Rain inches totaled during
observation period
intercept
0.019
0.083
x(t)
34.104**
12.313
mitigation(t)
-0.149
0.111
Rain highest rate during
observation period
intercept
0.046
0.084
x(t)
2.063
1.288
mitigation(t)
-0.147
0.112
Soil moisture 16.5 ft bis
beneath structure
intercept
0.099
0.079
x(t)
0.055
0.347
mitigation(t)
-0.182
0.108
"Significant at 1% level of significance
Table 10-14 shows the results of models including predictors requiring Lag 1 terms to model for the 422
Office for the period 2012 to 2013. The patterns of sign of the x(t) and x(t-l) terms for some of the most
interesting parameters, such as exterior temperature and Fall Creek gauge height, parallel those seen in
the previous year's data set in Table 10-8. The coefficients were somewhat different than for the previous
year's data set. Therefore, the qualitative interpretations provided in Section 10.4 for those parameters
apply here and will not be repeated for brevity.
There were some modest differences with this 2012-2013 data set as compared to the 2011-2012 data set.
For example, the 2012-2013 office data set does not show a statistically significant humidity effect in the
full model (Table 10-14), although the effect in the reduced model is significant and similar to the result
from the 2011-2012 data set (Table 10-8).
Mitigation provides a benefit in almost all cases, just as was seen in the basement data set. The magnitude
of the mitigation benefit in pCi/L is lower upstairs than in the basement, which is expected since the
unmitigated concentrations upstairs are lower than the unmitigated concentrations in the basement.
10-37
-------
Section 10—Time Series Analysis
Table 10-14. Model Parameters, Standard Errors by Model and Predictor for
X422office_2nd_AG_radon Concentration (X422QF2-2): 2012-2013 Lag 1 Models
Model : Y(t)-Y(t-1) = Intercept
+ Predictor (t) + Predictor (t-1)
Model : Y(t)-Y(t-1) =
Intercept + Predictor (t)
Predictor Name
Model Term
Estimate
SE
Estimate
SE
Air density interior
intercept
12.06096**
4.61969
4.97669
4.47063
mitigation(t)
-0.17894
0.10823
-0.16696
0.11191
x(t)
244.35736*
95.02056
-65.7802
60.05041
x(t-1)
-405.1163**
96.77658
intercept
22.75505*
9.1404
5.43045
8.32925
Barometric pressure
mitigation(t)
-0.2547*
0.11311
-0.18533
0.11525
x(t)
0.54341
0.32666
-0.17777
0.27675
x(t-1)
-1.29649**
0.32735
Cooling degree day
intercept
0.09455
0.08373
0.13298
0.0856
mitigation(t)
-0.18022
0.10915
-0.19906
0.11248
x(t)
-8.76847**
2.04811
-2.8046
1.47652
x(t-1)
8.48575**
2.07788
intercept
-0.3166
0.23253
-0.11388
0.23414
Dew point, interior
mitigation(t)
-0.12924
0.11153
-0.14608
0.11476
x(t)
-0.03995**
0.01282
0.005
0.00565
x(t-1)
0.05038**
0.01287
Dew point, exterior
intercept
-0.15241
0.16129
0.18395
0.17534
mitigation(t)
-0.14079
0.10092
-0.17985
0.11346
x(t)
-0.04584**
0.0067
-0.00281
0.00421
x(t-1)
0.05222**
0.0067
intercept
0.04278
0.17284
-0.08965
0.16992
Height measured at Fall
Creek stream gauge
mitigation(t)
-0.17441
0.10914
-0.15852
0.10981
x(t)
0.26485**
0.08217
0.04413
0.04048
x(t-1)
-0.25319**
0.08213
Heating degree days
intercept
0.15756
0.11004
-0.03351
0.12186
mitigation(t)
-0.16145
0.09882
-0.18448
0.11256
x(t)
2.6996**
0.35027
0.2624
0.20917
x(t-1)
-2.87089**
0.34897
intercept
-0.0601
0.1755
0.32065
0.19695
Exterior heating index—
calculated based on
temperature and humidity
mitigation(t)
-0.1603
0.09756
-0.1885
0.11278
x(t)
-0.0554**
0.00672
-0.00536
0.004
x(t-1)
0.05861**
0.00671
Humidity interior
intercept
-0.11104
0.17324
0.01439
0.17854
mitigation(t)
-0.14609
0.11011
-0.15832
0.11465
x(t)
-0.04852**
0.0121
0.00204
0.0049
x(t-1)
0.05451**
0.01196
(continued)
10-38
-------
Section 10—Time Series Analysis
Table 10-14. Model Parameters, Standard Errors by Model and Predictor for
X422office_2nd_AG_radon Concentration (X4220F2-2): 2012-2013 Lag 1 Models
(cont.)
Model : Y(t)-Y(t-1) = Intercept +
Predictor (t) + Predictor (t-1)
Model : Y(t)-Y(t-1) =
Intercept + Predictor (t)
Predictor Name
Model Term
Estimate
SE
Estimate
SE
intercept
-0.85948*
0.41918
-0.77538*
0.37462
Humidity exterior
mitigation(t)
-0.14924
0.11204
-0.1476
0.11118
x(t)
0.00927
0.00617
0.01146*
0.0049
x(t-1)
0.00338
0.00618
Interior heating index
intercept
-1.18663
0.87626
-0.75597
0.83561
mitigation(t)
-0.14504
0.11381
-0.14835
0.11368
x(t)
-0.01957
0.02497
0.01203
0.01196
x(t-1)
0.03793
0.02575
intercept
0.07105
0.08145
0.07316
0.0806
Depth of snow on the
ground
mitigation(t)
-0.17566
0.10847
-0.17781
0.10772
x(t)
0.04996
0.0795
0.06345
0.04261
x(t-1)
0.01604
0.07972
Soil moisture, 13 ft bis
beneath structure
intercept
0.08809
0.1319
0.07184
0.13014
mitigation(t)
-0.18544
0.11073
-0.17751
0.11017
x(t)
0.09404
0.11492
0.00574
0.02099
x(t-1)
-0.09041
0.11569
intercept
-0.10557
0.40645
-0.01077
0.4025
Soil moisture 6 ft bis
mitigation(t)
-0.14658
0.11059
-0.18287
0.10816
beneath structure
x(t)
-0.0499
0.03399
0.00064
0.00228
x(t-1)
0.05101
0.03423
Soil moisture, 3.5 ft bis
exterior
intercept
0.07703
0.14646
0.10444
0.14243
mitigation(t)
-0.18782
0.11796
-0.18157
0.11762
x(t)
-0.00089
0.00126
-2.00E-05
0.00067
x(t-1)
0.00101
0.00124
intercept
-0.16799
0.2302
-0.07969
0.22524
Soil moisture 6 ft bis
exterior
mitigation(t)
-0.20792
0.10957
-0.20076
0.10994
x(t)
-0.0016
0.00169
8.00E-04
0.00094
x(t-1)
0.00281
0.00165
Temperature, 422 first floor
intercept
-1.03116
1.0247
-0.7593
0.97944
mitigation(t)
-0.15938
0.11332
-0.15652
0.11284
x(t)
-0.00889
0.02903
0.01166
0.01355
x(t-1)
0.02443
0.02976
intercept
-0.04636
0.17824
0.33336
0.19908
Exterior Temperature (°F)
mitigation(t)
-0.16269
0.09789
-0.18821
0.11264
x(t)
-0.0553**
0.0068
-0.00561
0.00403
x(t-1)
0.0582**
0.00679
(continued)
10-39
-------
Section 10—Time Series Analysis
Table 10-14. Model Parameters, Standard Errors by Model and Predictor for
X422office_2nd_AG_radon Concentration (X4220F2-2): 2012-2013 Lag 1 Models
(cont.)
Model : Y(t)-Y(t-1) =
Intercept + Predictor (t) +
Predictor (t-1)
Model : Y(t)-Y(t-1) =
Predictor Name
Model Term
Intercept + Predictor (t)
Estimate
SE
Estimate
SE
Temperature exterior, high
during data collection
period
intercept
-0.05542
0.18007
0.3181
0.19999
mitigation(t)
-0.16208
0.09851
-0.18684
0.1127
x(t)
-0.0542**
0.00685
-0.00522
0.00401
x(t-1)
0.05727**
0.00684
intercept
-0.05472
0.17769
0.32169
0.19822
Lowest exterior
temperature
mitigation(t)
-0.16152
0.09804
-0.18725
0.11267
x(t)
-0.05517**
0.00684
-0.0054
0.00404
x(t-1)
0.05828**
0.00683
Temperature, humidity and
wind index
intercept
-0.00968
0.1538
0.30327
0.1718
mitigation(t)
-0.1647
0.0982
-0.19085
0.11272
x(t)
-0.04851**
0.00598
-0.00521
0.00354
x(t-1)
0.05072**
0.00597
Wind chill
intercept
0.00109
0.1559
0.31297
0.17335
mitigation(t)
-0.16674
0.09852
-0.19049
0.11259
x(t)
-0.04828**
0.00604
-0.0054
0.00356
x(t-1)
0.05023**
0.00603
Wind direction (average)
intercept
-0.09544
0.2019
-0.16009
0.17276
mitigation(t)
-0.17169
0.11218
-0.16626
0.11159
x(t)
0.00137
0.00083
0.00121
0.00077
x(t-1)
-0.00046
0.00083
intercept
-0.07846
0.20227
-0.1357
0.17383
Wind direction (of high
during measurement
period)
mitigation(t)
-0.171
0.11235
-0.1652
0.11173
x(t)
0.00123
0.00085
0.00109
0.00078
x(t-1)
-4.00E-04
0.00085
Wind run is a function of
wind speed and duration
intercept
-0.0663
0.12726
0.03087
0.11839
mitigation(t)
-0.20776
0.11376
-0.17926
0.11377
x(t)
-0.03347
0.0561
0.02779
0.04824
x(t-1)
0.11895*
0.05615
intercept
-0.12455
0.14387
-0.01764
0.13236
High wind speed during
measurement period
mitigation(t)
-0.21512
0.11412
-0.18769
0.1139
x(t)
-0.00303
0.01346
0.01074
0.01144
x(t-1)
0.02608
0.01346
Average wind speed during
measurement period
intercept
-0.0663
0.12726
0.03087
0.11839
mitigation(t)
-0.20776
0.11376
-0.17926
0.11377
x(t)
-0.01674
0.02805
0.0139
0.02412
x(t-1)
0.05948*
0.02808
"Significant at 1% level of significance; * Significant at 5% level of significance
10-40
-------
Section 10—Time Series Analysis
10.7 Correlation between VOC (Radiello) Time Series and Predictor Variables in
422 Basement South
The association between four Radiello time series and a series of predictors were investigated in this
section. Details of the periodicity, time period and location are displayed in Table 10-15. Please refer to
the discussion of the terminology, and methods of time series analysis presented earlier in this section in
the context of a discussion of the radon time series. Examples of how to read the output tables were also
provided earlier in this section.
Table 10-15. Name, Periodicity, Time Period, and Location of Time Series (Outcome) Considered
Time Series Name
Variable Name in Data Set
Periodicity
Time Period
Location
X422BaseS Radiello
X422BaseS_Radiello_Weekly_CHCI3
Weekly
Jan 5, 2011-Feb 15,
Basement
CHCI3-1
2012
south
X422BaseS Radiello
X422BaseS_Radiello_Weekly_PCE
Weekly
Jan 5, 2011-Feb 15,
Basement
PCE-1
2012
south
X422BaseS Radiello
X422BaseS_Radiello_Weekly_CHCI3
Weekly
Sept 26, 2012-
Basement
CHCI3-2
April 10, 2013
south
X422BaseS Radiello
X422BaseS_Radiello_Weekly_PCE
Weekly
Sept 26, 2012-
Basement
PCE-2
April 10, 2013
south
10.7.1 Stationarity and Serial Correlation Analysis
As a first step in the time series data analysis, we explored the stationarity and serial correlation of the
time series. The p-values of the stationarity tests suggested the chloroform time series, period Jan 5,
2011-Feb 15, 2012, is non-stationary (Figure 10-11). A first difference of that time series was calculated
(i.e., y(t)-y(t-l) where y(t) denotes the measurement of 422 Base South Chloroform at time t), and
corresponding plots are displayed in Figure 10-12. The stationarity tests results suggested the first
differences constitute a stationary time series; therefore, no further transformations were needed. None of
the partial autocorrelation functions (spikes) in the PACF plot were statistically significant (none of the
spikes are crossing blue lines), suggesting no significant serial correlation remained in the first differences
time series. As a result, a model for this time series did not include any lag-terms. Note that while
autocorrelation was found in the radon analyses, those analyses used daily data sets. Significant
autocorrelation in indoor concentrations between one week and the next is not necessarily physically
expected.
The plot of the time series 422BaseSouth PCE for the, period Jan 5, 2011-Feb 15, 2012, suggested some
non-stationarity in the mean (non-constant means) in the first weeks, and a more stationarity mean and
variance (more stable plot) in the later weeks (Figure 10-13). For example, the average of the first 4
observations is not the same as the average of the last observations. The test for non-stationarity suggests
that the deviations for non-stationarity were not significant (both p-values < 0.01). Therefore, no
transformation was needed of this data set. The PACF plot suggested serial autocorrelation at lag 2 weeks,
this indicates that three weekly consecutive measurements are correlated. We accounted for the serial
10-41
-------
I I I I I I
01/05 02/09 03/16 04/20 05/25 06/29 08/03 09/07 1 0/12 11/16 12/21 01/25
Date: Jan 2011-Feb 2012
Weekly Time Series of Radiello CHCI3. Location: X422 Basement South
ADF, PACF and stationary tests pvalues = ( 0.75 , 0.37 )
-1—i—i—i—i—i—i—i—i—r-
01 23456789
Lag
~I 1 1 1 1 1 1 r~
10 11 12 13 14 15 16 17
-1 1 1 1 I-
1 2 3 4 5
10 11 12 13 14 15 16 17
Lag
Figure 10-11. Time series plot, ACF and PACF for weekly Radiello chloroform. Location X422 basement south. Time period: Jan 5,
2011-Feb 15, 2012.
-------
I I I I I I I
01/05 02/09 03/16 04/20 05/25 06/29 08/03 09/07 1 0/12 11/16 12/21 01/25
Date: Jan 2011-Feb 2012
First Difference of Weekly Radiello CHCI3. Location: X422 Basement South
ADF, PACF and stationarityteste pvalues = (0.01 , 0.01 )
I I I
~l 1 1 1 1 I-
0 1 2 3 4 5
-1 1 1 1 1 1 1 1 1 1 T~
7 8 9 10 11 12 13 14 15 16 17
1 1 1 1 I-
1 2 3 4 5
-1 1 1 1 1 1 1 1—
10 11 12 13 14 15 16 17
Figure 10-12. Time series plot, ACF and PACF for first difference of weekly Radiello. Chloroform. Location X422 basement south. Time
period: Jan 5, 2011-Feb 15, 2012.
-------
Weekly Time Series of Radiello PCE. Location: X422 Basement South
ADF, PACF and stationarity tests pvalues = (0.01 , 0.01 )
"H"
01/05 02/09 03/16 04/20 05/25 06/29 08/03 09/07 1 0/12 11/16
Date: Jan 2011-Feb 2012
10 11 12 13 14 15 16 17
10 11 12 13 14 15 16 17
Figure 10-13. Time series plot, ACF and PACF for weekly Radiello PCE. Location X422 basement south. Time period: Jan 5, 2011—
Feb 15, 2012.
-------
Section 10—Time Series Analysis
autocorrelation in the models by modeling the error term25 instead of adding lag terms of the response as
predictors (Cochrane and Orcutt, 1949).
Figure 10-14 display the time series and corresponding ACF and PACF plots for 422 Base South
chloroform data set for the period Sept 26, 2012-April 10, 2013. The stationarity test results were
conflicting (one concluding non-stationarity and the other suggesting stationarity). We investigated the
non-stationarity of the first differences time series (Figure 10-15), and both tests concluded stationarity at
5% and 1% significance level. None of the partial autocorrelation functions (spikes) in the PACF plot
were statistically significant (none of the spikes are crossing blue lines), suggesting no significant serial
correlation remained in the first differences time series. A model for X422BaseS Radiello CHCI3-2 thus
did not include any lag-term of the outcome.
Similarly, Figures 10-16 and 10-17 display the time series, and corresponding ACF and PACF plots for
the time series and first differences of the 422 Base South PCE data set for the, period Sept 26, 2012-
April 10, 2013. The original time series did not pass both stationarity tests. We proceeded to calculate the
first differences of the time series, which passed both stationarity tests (Figure 10-17). No indication of
significant serial correlation was detected from the PACF plot (Figure 10-17).
Table 10-16 summarizes the transformation (first difference or no difference needed) and serial
correlation determined for each of the four time series data sets. The needed transformations were then
applied before modeling the data. The serial correlations identified were used to determine how many lag
terms were needed in the equation fit to the data.
Table 10-16. Transformation and Terms Required by Time Series
Time Series Name
X422BaseS Radiello CHCI3-1
X422BaseS Radiello PCE-1
X422BaseS Radiello CHCI3-2
X422BaseS Radiello PCE-2
Transformation
First difference
No difference needed
First difference
First difference
Serial Correlation
None
Second order
None
None
25The error term is the difference between the actual value of the independent variable and the value predicted by the regression equation. In the
presence of non-significant serial correlation, the outcome is approximately independent, which results in error terms that are approximately
independent. The serial correlation also affects the constant variance assumption that is inherent in regression analysis. Having non-constant
variance will affect the results of the test-statistics used to evaluate the significance of the parameters in the regression equation. This is
addressed at the model level by incorporating a model for the errors that accounts for the serial correlation. Using an appropriate model for the
errors results in correct results for the testing of the significance of the parameters (Judge et al., 1985).
10-45
-------
Weekly Time Series of Radiello CHCI3. Location: X422 Basement South
ADF, PACF and stationary tests pvalues = (0.02 , 0.14 )
Sep 2012-April 2013
-1 1 1 1 1 I-
0 1 2 3 4 5
~l 1 1 1 1 1—
9 10 11 12 13 14
~l 1 1 1 1 1 1 1 1 1 1 1 1 1—
1 2 3 4 5 6 7 8 9 10 11 12 13 14
Figure 10-14. Time series plot, ACF and PACF for weekly chloroform. Location X422 basement south. Time period: Sept 26, 2012-
April 10, 2013.
£
>3
-------
-1^
-»J
First Difference of Weekly Radiello CHCI3. Location: X422 Basement South
ADF, PACF and stationarity tests pvalues = (0.02 , 0.01 )
-1 1 1 1 1 r~
0 1 2 3 4 5
-i 1 1 1 1 1—
9 10 11 12 13 14
i 1 1 1 1 1 1 1 1 1 1 1 1 r~
1 2 3 4 5
10 11 12 13 14
Sep 2012-April 2013
Lag
Lag
Figure 10-15. Time series plot, ACF and PACF for first difference weekly Radiello CHCI3-2 (X422BaseS_Radiello_Weekly_CHCI3).
Location X422 basement south. Time period: Sept 26, 2012-April 10, 2013.
-------
09/26 10/31
Weekly Time Series of Radiello PCE. Location: X422 Basement South
ADF, PACF and stationary tests pvalues = (0.02 , 0.33)
-1 1 1 1 1 r~
0 1 2 3 4 5
-i 1 1 1 1 i—
9 10 11 12 13 14
n 1 1 1 1 1 1 1 1 1 1 1 1 r~
1 2 3 4 5
10 11 12 13 14
Sep 2012-April 2013
Lag
Lag
Figure 10-16. Time series plot, ACF and PACF for weekly Radiello PCE-2 (X422BaseS_Radiello_Weekly_PCE). Location X422 basement
south. Time period: Sept 26, 2012-April 10, 2013.
-------
09/26 10/31
First Difference of Weekly Radiello PCE. Location: X422 Basement South
ADF, PACF and stationary tests pvalues = (0.05 , 0.01 )
y 0.3
-1 1 1 1 1 r~
0 1 2 3 4 5
"I 1 1 1 1 1—
9 10 11 12 13 14
n 1 1 1 1 1 1 1 1 1 1 1 1 r~
1 2 3 4 5
10 11 12 13 14
Sep 2012-April 2013
Figure 10-17. Time series plot, ACF and PACF for first difference weekly Radiello PCE-2 (X422BaseS_Radiello_Weekly_PCE). Location
X422 basement south. Time period: Sept 26, 2012-April 10, 2013.
-------
Section 10—Time Series Analysis
10.7.2 Predictor Variables Modeled and Their Potential for Autocorrelation
Table 10-17 lists a set of variables considered potential predictors for the VOC concentrations. We
investigated their serial correlation (significant lags) and determined the transformation (e.g., no
transformation, include a lag variable (e.g., lag 1 or past observation= x(t-l)) needed for the inclusion of
each predictor in the model. Variables with 0 significant lags did not show significant serial correlation
among consecutive measurements and did not need any lag transformation. Variables with 1 significant
lag necessitated two terms in the model, the variable (x(t)) and the previous observation (x(t-l)).
Variables with 2 significant lags required a model with three terms, the variable (x(t)), the previous
observation (x(t-l)) and the second previous observation (x(t-2)).
A set of categorical variables were also considered. For the period Jan 5, 2011-Feb 15, 2012, the
following variables were considered: AC (on or off), fan (used for fan testing, on or off), and heating
system (on or off). In each case, the on was modeled as 1 and the off as zero. The variable for mitigation
status was set to OFF during the Jan 2011 to Feb 2012 period (because the mitigation system had not yet
been installed). For the period Sept 26, 2012-April 10, 2013, most of the categorical variables were
consistently OFF except for heat and mitigation status, which was ON and OFF during the time period.
As with the radon analysis, mitigation status was considered as a control variable and was included in all
models for variables listed for the period Sept 26, 2012-April 10, 2013. In other words, as each of the
predictor variables was individually modeled, a mitigation (on/off) term was always used as part of the
equation.
Table 10-17. Continuous Covariates by Time Period
Sept 26, 2012-April 10,
2013
Plain Language
Ijan 5, 2011 —
Feb 15, 20121
Variable Name
Variable Name
Significant
Significant
I AR Model I
Lags
AR Model
Lags
Barometric rate of change in
inches of mercury pressure
Bar_drop_Hg.hr
0
0
per hour
Barometric pressure in
inches of mercury
Bar_in_Hg
1
AR(1)
0
Net pressure change over
measurement period in
BP_Net_Change
0
0
inches of mercury
Standard deviation of
pressure change over
measurement period in
inches of mercury
BP_Pump_Speed
1
AR(1)
0
Largest pressure change
over measurement period
("stroke length" of
BP_Stroke_Length
1
AR(1)
0
barometric pumping) in
inches of mercury
Cooling degree day
Cool_Degree_Day
1
AR(1)
0
Dew point, interior, 422
Fahrenheit
Dew_pt_422_F
1
AR(1)
1
AR(1)
Height measured at Fall
Creek stream gauge in feet
Fall_Crk_Gage_ht_ft
1
AR(1)
0
(continued)
10-50
-------
Section 10—Time Series Analysis
Table 10-17. Continuous Covariates by Time Period (cont.)
Sept 26, 2012-April 10,
2013
Plain Language
Jan 5, 2011-
-Feb 15, 2012
Variable Name
Variable Name
Significant
Lags
Significant
Lags
I AR Model 1
AR Model
Heating degree days
Heat_Degree_Day
2
AR(2)
1
AR(1)
% Humidity in interior of
422
Hum_422_.
1
AR(1)
1
AR(1)
% Humidity in exterior
Hum_out_.
1
AR(1)
0
Rain (inches) totaled
during observation period
Rain_ln_met
1
AR(1)
0
Rain highest rate during
observation period in
RainJPH
1
AR(1)
0
inches/hour
420 side, subslab vs.
basement differential
pressure
Setra_420ss. base_Pa
1
AR(1)
1
AR(1)
422 side basement vs
exterior differential
Setra_422base.out_Pa
1
AR(1)
1
AR(1)
pressure, Pascals
422 side, basement vs.
upstairs differential
pressure, Pascals
Setra_422base.upst_Pa
1
AR(1)
1
AR(1)
422 side, deep vs. shallow
soil gas differential
Setra_422SGdp.ss_Pa
1
AR(1)
1
AR(1)
pressure, Pascals
422 side, subslab vs.
basement differential
pressure, Pascals
Setra_422ss. base_Pa
1
AR(1)
1
AR(1)
Depth of snow on the
ground, inches
Snowdepth_daily
1
AR(1)
1
AR(1)
Soil moisture 6 ft bis
beneath structure, cbar
Soil_H20_ln6._cbar
1
AR(1)
1
AR(1)
Soil moisture, 3.5 ft bis
exterior, cbar
Soil_H20_Out3.5._cbar
1
AR(1)
1
AR(1)
Soil moisture 6 ft bis
exterior, cbar
Soil_H20_Out6._cbar
1
AR(1)
1
AR(1)
Soil temperature 9 ft bis
beneath structure, C
Soil_T_C_MW3.9
1
AR(1)
1
AR(1)
Soil temperature 1 ft bis
exterior, C
Soil_T_C_OTC.1
1
AR(1)
1
AR(1)
Soil temperature 6 ft bis
exterior, C
Soil_T_C_OTC.6
1
AR(1)
1
AR(1)
Temperature, 422 first
floor from weather station
T_422_F
1
AR(1)
1
AR(1)
(continued)
10-51
-------
Section 10—Time Series Analysis
Table 10-17. Continuous Covariates by Time Period (cont.)
Sept 26, 2012-April 10,
Plain Language Variable
Variable Name
Jan 5, 2011-
¦Feb 15, 2012
2013
Name
Significant
Lags
Significant
Lags
AR Model
AR Model
Temperature 422
basement north from
T_422baseN_C
1
AR(1)
1
AR(1)
HOBO
Temperature 422 first floor
from HOBO
T_422baseS_C
1
AR(1)
1
AR(1)
Temperature on first floor
of 422 side of duplex from
T_422first_C
1
AR(1)
0
HOBO
Temperature exterior
T_out_F
1
AR(1)
1
AR(1)
Temperature exterior, high
during data collection
T_out_Hi_F
1
AR(1)
1
AR(1)
period
Lowest exterior
temperature in Fahrenheit
T_out_Lo_F
2
AR(2)
1
AR(1)
Wind chill
Wind_Chill_F
1
AR(1)
1
AR(1)
Average wind direction in
degrees
Wind_Dir
0
0
Wind direction of high
speed during
measurement period in
Degrees
Wind_Dir_Hi
0
0
Wind run ( a function of
wind speed and duration)
Wind_Run_mi
0
2
AR(2)
High wind speed during
measurement period
Wind_Speed_Hi_MPH
0
0
Average wind speed
during measurement
period
Wind_Speed_MPH
2
AR(2)
0
422 Base North radon
measured by electret
X422baseN_Wkly_Elect
radon
1
AR(1)
1
AR(1)
422 Base south radon
measured by electret
X422baseS_Wkly_Elect
_radon
0
1
AR(1)
422 first floor radon
measured by electret
X422fi rst_Wkly_E lect_r
adon
1
AR(1)
1
AR(1)
Soil temperature from
exterior thermocouple at
3.5 ft bis, C
S o i l_T_C_OT C. 3.5
1
AR(1)
422 second floor office
radon measured by
electret
X422office2nd_Wkly_EI
ect radon
1
AR(1)
10-52
-------
Section 10—Time Series Analysis
10.7.3 Time Series Analysis Results for 2011-2012 Chloroform Data Set
Table 10-18 lists the results for the analysis of association between the outcome (chloroform
concentration in 422 base south) and predictors not needing a lag term in the model. None of these
predictors were found to be statistically significant. Note that the finding that indoor radon and
chloroform were not significantly correlated in this data set suggests that the mechanisms controlling
vapor intrusion for these two pollutants must have differed.
Table 10-18. Analysis for Outcome First Difference of X422BaseS_Radiello_Weekly_CHCl3.
Variables That Did Not Need Lag Terms. Period Jan 5, 2011-Feb 15, 2012
Predictor Name
Model Term
Estimate
SE
Wind_Speed_Hi_MPH
intercept
-0.020
0.130
x(t)
0.000
0.000
RainJPH
intercept
-0.030
0.040
x(t)
0.010
0.020
Soil_T_C_OTC.6
intercept
-0.070
0.070
x(t)
0.000
0.000
X422baseN_Wkly_Elect_radon
intercept
-0.080
0.100
x(t)
0.010
0.010
X422baseS_Wkly_Elect_radon
intercept
-0.040
0.070
x(t)
0.000
0.010
X422first_Wkly_Elect_radon
intercept
-0.070
0.070
x(t)
0.020
0.020
BP_Pump_Speed
intercept
0.100
0.080
x(t)
-0.770
0.440
Table 10-19 shows the results of the regression analysis of predictor variables that needed a lag-1 week
term in the model. Only heating degree days, snow depth, outside soil temperature at one foot bis, exterior
temperature, exterior high temperature, and wind chill showed a significant association with the outcome.
As discussed previously heating degree days, exterior temperature and daily high exterior temperature are
closely linked variables that are all expected to be measures of the strength of the stack effect. The finding
that snow depth had a significant correlation but rain did not is valuable. Current sampling guidance often
imposes logistically difficult requirements for vapor intrusion related sampling based on rain. For
example, California requires that soil gas sampling be delayed for 5 days after any rain of more than Vi
inch (California EPA, 2012). The California guidance refers specifically to rain and the term snow does
not appear in the guidance. NJ DEP (2012), in contrast, requires documentation of "significant
precipitation" but does not further define the types or quantities of precipitation which may be significant.
As discussed earlier in this section, when the slopes of the x(t) and x(t-l) terms differ in sign but are
similar in magnitude, that can be interpreted as an indication that the change in the predictor variable from
week to week is what is correlated with the change in chloroform concentration. To understand the
implications of the model predictions, it is useful to plug in some realistic trial values and compute the
outcome.
Example A: If the exterior temperature this week averages 20°F and the temperature last week averaged
0° F, the model predicts this week's chloroform will be 0.258 |Jg/m3 lower than last week's chloroform.
This is consistent with the expected weaker stack effect with warmer temperatures.
10-53
-------
Section 10—Time Series Analysis
Table 10-19. Analysis for Outcome X422BaseS_Radiello_Weekly_CHCl3. Variables that Needed a
lag-1 Term. Period Jan 5, 2011-Feb 15, 2012
Model Y(t)-Y(t-1) =
Model Y(t) - Y(t-1)
=
intercept +
Predictor Name (x(t))
Model Term
intercept + x(t)
h x(t—1)
x(t)
Estimate
SE
Estimate
SE
intercept
-0.018
0.029
-0.019
0.029
Bar_drop_Hg.hr
x(t)
17.416
16.828
28.654
14.306
x(t-1)
-21.010
16.782
NA
NA
intercept
-1.352
7.760
-0.518
6.669
Bar_in_Hg
x(t)
-0.012
0.259
0.017
0.222
x(t-1)
0.056
0.260
NA
NA
intercept
-0.018
0.029
-0.019
0.029
BP_Net_Change
x(t)
-0.129
0.100
-0.18906*
0.084
x(t-1)
0.108
0.100
NA
NA
intercept
0.117
0.088
0.101
0.079
B P_Stro ke_Le n gt h
x(t)
-0.156
0.145
-0.192
0.117
x(t-1)
-0.062
0.144
NA
NA
intercept
-0.035
0.036
-0.032
0.036
Cool_Degree_Day
x(t)
-0.001
0.002
0.000
0.001
x(t-1)
0.001
0.002
NA
NA
intercept
-0.065
0.105
-0.041
0.105
Dew_pt_422_F
x(t)
-0.008
0.006
0.000
0.002
x(t-1)
0.009
0.006
NA
NA
intercept
0.050
0.093
0.046
0.084
Fall_Crk_Gage_ht_ft
x(t)
-0.018
0.030
-0.018
0.021
x(t-1)
-0.001
0.030
NA
NA
intercept
0.001
0.044
-0.021
0.046
Heat_Degree_Day
x(t)
0.00143*
0.001
0.000
0.000
x(t-1)
-0.00161**
0.001
NA
NA
intercept
-0.074
0.119
-0.052
0.118
Hum_422_.
x(t)
-0.005
0.006
0.001
0.003
x(t-1)
0.006
0.006
NA
NA
intercept
0.471
0.380
0.258
0.324
Hum_out_.
x(t)
-0.001
0.005
-0.004
0.004
x(t-1)
-0.005
0.005
NA
NA
intercept
0.029
0.045
0.010
0.041
Rain_ln_met
x(t)
-0.027
0.037
-0.038
0.035
x(t-1)
-0.035
0.037
NA
NA
(continued)
10-54
-------
Section 10—Time Series Analysis
Table 10-19. Analysis for Outcome X422BaseS_Radiello_Weekly_CHCl3. Variables That Needed
a lag-1 Term. Period Jan 5, 2011-Feb 15, 2012 (cont.)
Model Y(t) -
-<
r
II
Model Y(t) -
-<
r
II
Predictor Name (x(t))
Model Term
intercept + x(t
h x(t—1)
intercept
+ x(t)
Estimate
SE
Estimate
SE
Setra_420ss.base_Pa
intercept
0.001
0.073
-0.054
0.063
x(t)
0.057
0.056
0.032
0.053
x(t-1)
-0.079
0.056
NA
NA
intercept
-0.018
0.030
-0.021
0.030
Setra_422base.out_Pa
x(t)
-0.019
0.027
0.001
0.021
x(t-1)
0.030
0.027
NA
NA
Setra_422base.upst_Pa
intercept
-0.020
0.029
-0.022
0.029
x(t)
-0.051
0.031
-0.041
0.030
x(t-1)
0.038
0.031
NA
NA
intercept
0.015
0.049
-0.020
0.045
Setra_422SGdp.ss_Pa
x(t)
0.024
0.030
-0.001
0.022
x(t-1)
-0.047
0.035
NA
NA
Setra_422ss. base_Pa
intercept
-0.110
0.112
-0.128
0.097
x(t)
0.051
0.044
0.044
0.038
x(t-1)
-0.014
0.044
NA
NA
intercept
-0.019
0.030
-0.006
0.031
Snowdepth_daily
x(t)
-0.19513**
0.063
-0.057
0.042
x(t-1)
0.17307**
0.061
NA
NA
Soil_H20_ln6._cbar
intercept
-0.068
0.053
-0.072
0.051
x(t)
-0.001
0.006
0.001
0.001
x(t-1)
0.002
0.006
NA
NA
intercept
-0.041
0.037
-0.041
0.036
Soil_H20_Out3.5._cbar
x(t)
0.000
0.001
0.001
0.001
x(t-1)
0.000
0.001
NA
NA
Soil_H20_Out6._cbar
intercept
-0.037
0.040
-0.038
0.039
x(t)
0.000
0.002
0.000
0.000
x(t-1)
0.000
0.002
NA
NA
intercept
-0.061
0.050
-0.002
0.062
Soil_T_C_OTC.1
x(t)
-0.05861**
0.015
-0.001
0.004
x(t-1)
0.06156**
0.015
NA
NA
T_422_F
intercept
-0.103
0.270
-0.009
0.259
x(t)
-0.006
0.006
0.000
0.004
x(t-1)
0.007
0.006
NA
NA
(continued)
10-55
-------
Section 10—Time Series Analysis
Table 10-19. Analysis for Outcome X422BaseS_Radiello_Weekly_CHCl3. Variables That Needed a
lag-1 Term. Period Jan 5, 2011-Feb 15, 2012 (cont.)
Model Y(t) -
-<
r
II
Model Y(t) -
Y (t—1) =
Predictor Name (x(t))
Model Term
intercept + x(t
h x(t—1)
intercept
+ x(t)
Estimate
SE
Estimate
SE
intercept
-0.208
0.203
-0.188
0.205
T_422baseN_C
x(t)
-0.016
0.012
0.003
0.003
x(t-1)
0.019
0.012
NA
NA
T_422baseS_C
intercept
-0.293
0.243
-0.275
0.241
x(t)
-0.006
0.013
0.004
0.004
x(t-1)
0.010
0.012
NA
NA
intercept
-0.327
0.305
-0.243
0.299
T_422first_C
x(t)
-0.007
0.009
0.003
0.004
x(t-1)
0.011
0.009
NA
NA
T_out_F
intercept
-0.067
0.091
-0.035
0.095
x(t)
-0.00957*
0.004
0.000
0.002
x(t-1)
0.01048*
0.004
NA
NA
intercept
-0.117
0.124
-0.076
0.127
T_out_Hi_F
x(t)
-0.006
0.004
0.001
0.002
x(t-1)
0.00757*
0.003
NA
NA
T_out_Lo_F
intercept
-0.043
0.069
-0.043
0.067
x(t)
0.001
0.004
0.001
0.002
x(t-1)
0.000
0.004
NA
NA
intercept
-0.060
0.081
-0.030
0.086
Wind_Chill_F
x(t)
-0.00912*
0.004
0.000
0.002
x(t-1)
0.00994**
0.004
NA
NA
Wind_Run_mi
intercept
0.099
0.103
-0.001
0.085
x(t)
0.000
0.000
0.000
0.000
x(t-1)
0.000
0.000
NA
NA
intercept
0.073
0.104
0.000
0.092
Wind_Speed_MPH
x(t)
0.012
0.025
-0.005
0.023
x(t-1)
-0.036
0.025
NA
NA
"Significant at 1% level of significance
'Significant at 5% level of significance
NA = Not Applicable
10-56
-------
Section 10—Time Series Analysis
Example B: If the exterior temperature this week averages 20° F and the temperature last week averaged
40°F, the model predicts this week's chloroform will be 0.162 |lg/m3 greater than last week's chloroform.
This is consistent with the expected stronger stack effect with cooler temperatures.
Example C: If snow depth this week = 2 in and snow depth last week = 0 in, the model predicts that the
concentration of chloroform will be 0.41 |ig/m3 LOWER this week than last week. This appears to be the
opposite of the effect we might have predicted from our analysis in Section 9.
Only the current observation of the high temperature is significantly associated with the chloroform
concentration, not the previous week's concentration.
Table 10-20 displays the results for the analysis of the association between variables needing a Lag 1 and
Lag 2 terms in the model. The recent past and current measurement of soil temperature at 9 ft bis (3ft
below the basement floor) was associated with the chloroform concentration. Again an example may be
needed to interpret the results. Assume the soil temperature at that location was gradually warming so that
it averaged 18°C this week, 17°C the week before, and 16°C two weeks ago. The model would predict
that the chloroform concentration in indoor air would have increased 0.51 |_ig/m\ While the strength of
this effect is surprising, the direction is physically reasonable from a consideration of volatility.
10.7.4 Time Series Analysis Results for 2011-2012 PCE Data Set
We encountered convergence issues when fitting the models for PCE during the period Jan 2011 to Feb
2012. These issues were likely the results of the non-constant mean observed at the beginning of the time
series (see Figure 10-3). We decided to take the first difference to get a more stationary time series and to
ensure convergence. The first difference improved the stationarity of the mean but it did not eliminate the
serial correlation observed between the consecutive time terms in the time series. To address this issue,
we investigated the distribution (specifically the covariance structure) of the error term in the model. We
fitted several models using several predictors and explored the corresponding PACF of the error term and
determined that a lag-1 error term was required. To incorporate a model term to the error, we specify in
the model equation a distribution for the variance of the error terms, in this case an AR(1) structure. A
more complex model (AR(2)) was also considered, but convergence issues were encountered likely the
result of small variability in some predictors and the small data set.
Table 10-21 shows the results of the analysis of the association between various predictor variables and
the PCE concentration in the 422 basement for the period Jan 2011 to Feb 2012. Only the variables soil
temperature at 6ft and barometric pressure "pump speed" were significantly associated with the PCE
concentration. As shown in Section 3, we defined the variable "barometric pressure pump speed" to be
the standard deviation of pressure change over measurement period in inches of mercury. While a
relationship between barometric pressure change and vapor intrusion is expected, the negative coefficient
is at first counterintuitive. We would have expected that the greater the amount of variability in
barometric pressure the more air that would be pumped in and out of the building. However, the negative
coefficient of this significant correlation may point to a more subtle interpretation—a pump such as the
human heart can be less effective when it has an irregular rapid pumping, described in the human as atrial
fibrillation or atrial flutter.26Further investigation in the literature of barometric pumping, building science
and fluid dynamics may be needed to better understand how to correlate the variability of barometric
pressure to vapor intrusion.
26http ://www.nlm.nih. gov/medlineplus/ency/article/000184.htm
10-57
-------
Table 10-20. Analysis for Outcome X422BaseS_Radiello_Weekly_CHCl3. Variables that Needed Lag-1 and Lag-2 Terms. Period Jan 5,
2011-Feb 15, 2012
Predictor Name
Soil T C MW3.9
Wind Dir
Wind Dir Hi
Model Term
x(t) +
x(t-1)+x(t-2)
x(t) +
x(t-1)
x(t)
Estimate
SE
Estimate
SE
Estimate
SE
intercept
-0.247180
0.186580
-0.333260
0.193930
-0.267510
0.190440
x(t)
0.25581*
0.115020
0.089320
0.084300
0.015130
0.011550
x(t-1)
-0.49172*
0.210950
-0.070670
0.083090
NA
NA
x(t-2)
0.25033*
0.112990
NA
NA
NA
NA
intercept
-0.024830
0.102390
0.128880
0.095890
0.066990
0.068320
x(t)
-0.000400
0.000260
-0.000440
0.000290
-0.000420
0.000290
x(t-1)
-0.000180
0.000260
-0.000270
0.000300
NA
NA
x(t-2)
0.00065*
0.000260
NA
NA
NA
NA
intercept
0.022160
0.135680
0.051860
0.122550
-0.066890
0.094640
x(t)
0.000320
0.000350
0.000260
0.000380
0.000200
0.000380
x(t-1)
-0.000520
0.000350
-0.000570
0.000380
NA
NA
x(t-2)
0.000070
0.000350
NA
NA
NA
NA
'Significant at 5% level of significance
-------
Section 10—Time Series Analysis
Table 10-21. Time Series Analysis for Outcome First Difference of 422 Basement South PCE
Concentration Variables that Did Not Need Lag Terms. Period Jan 2011 to Feb 2012
Predictor Name
Model Term
Estimate
SE
Intercept
-0.419
0.720
Wind_Speed_Hi_MPH
x(t)
0.006
0.024
RainJPH
Intercept
-0.352
0.202
x(t)
0.067
0.085
Intercept
-1.048**
0.271
Soil_T_C_OTC.6
x(t)
0.058**
0.018
X422baseN_Wkly_Elect_radon
Intercept
0.289
0.498
x(t)
-0.074
0.067
Intercept
-0.041
0.366
X422baseS_Wkly_Elect_radon
x(t)
-0.023
0.040
X422f i rst_Wkly_E 1 ect_ra d o n
Intercept
0.180
0.371
x(t)
-0.123
0.103
Intercept
0.593
0.405
BP_Pump_Speed
x(t)
-5.279*
2.395
"Significant at 1% level of significance
'Significant at 5% level of significance
The positive correlation to soil temperature is a physically reasonable result. It is likely that this reflects
the enhanced volatility of PCE at higher temperatures.
Table 10-22 displays the results of an analysis of the association between PCE concentration in the
basement of 422 for the period Jan 2011 to Feb 2012, and predictor variables needing a lag-1 week term
in their models. Relatively few variables were found to be significantly correlated:
¦ the previous week's dew point and humidity.
¦ Several exterior temperature related variables were significant, although the signs of the
coefficients were in many cases counterintuitive (and opposite of those observed for radon):
exterior temperature, low exterior temperature for the week, heating degree days, and wind chill.
¦ The positive correlation with the interior basement temperatures is expected: warmer interior
temperatures should lead to a stronger stack effect.
¦ The highly significant correlation to snow depth was expected from a visual examination of the
data set but again the negative coefficient was not expected. As shown in Figure 10-18, a visual
inspection of the XY plot of the weekly data would have suggested a positive correlation.
However as shown in Figures 10-19 and 10-20, the first difference of snow depth (difference
between last week's snow depth and this week's) and first difference in PCE concentration
shows a more complex behavior. After examining the data set, we believe that there are some
confounding factors that may have affected the results observed in this part of the time series
analysis:
1. There are relatively few weeks with a non-zero amount of snow in the data set.
2. The project began at a time with snow already on the ground.
3. The weekly time resolution was probably too coarse.
10-59
-------
Section 10—Time Series Analysis
Table 10-22. Analysis for PCE Concentration at 42 Base South, Variables that Needed a Lag-1
Term. Period Jan 2011 to Feb 2012
Model Y(t)-Y(t-1) =
Model Y(t)-Y(t-1) =
Predictor Name (x(t))
Model Term
intercept + x(t]
+ x(t-1)
intercept +
x(t)
Estimate
SE
Estimate
SE
intercept
-0.223
0.137
-0.232
0.134
Bar_drop_Hg.hr
x(t)
67.183
122.050
96.997
122.437
x(t-1)
-205.252
124.905
Bar_in_Hg
intercept
50.044
36.234
56.164
33.303
x(t)
-2.620
1.762
-1.880
1.110
x(t-1)
0.944
1.764
intercept
-0.222
0.136
-0.232
0.134
BP_Net_Change
x(t)
-0.574
0.729
-0.819
0.724
x(t-1)
1.231
0.755
B P_Stro ke_Le ngth
intercept
0.574
0.388
0.386
0.404
x(t)
0.498
0.912
-0.988
0.603
x(t-1)
-1.763
0.896
intercept
-0.358*
0.167
-0.35023*
0.162
Cool_Degree_Day
x(t)
0.001
0.012
0.004
0.003
x(t-1)
0.004
0.012
Dew_pt_422_F
intercept
-1.235**
0.460
-1.201*
0.454
x(t)
-0.012
0.039
0.0213*
0.010
x(t-1)
0.034
0.039
intercept
-0.151
0.417
-0.008
0.411
Fall_Crk_Gage_ht_ft
x(t)
-0.250
0.170
-0.062
0.104
x(t-1)
0.229
0.169
Heat_Degree_Day
intercept
0.121
0.192
0.091
0.196
x(t)
0.005
0.005
-0.003*
0.001
x(t-1)
-0.008
0.005
intercept
-1.494**
0.498
-1.473**
0.512
Hum_422_.
x(t)
-0.028
0.040
0.030*
0.012
x(t-1)
0.059
0.038
Hum_out_.
intercept
2.069
1.926
1.876
1.727
x(t)
-0.025
0.032
-0.029
0.024
x(t-1)
-0.007
0.031
intercept
-0.273
0.230
-0.197
0.209
Rain_ln_met
x(t)
-0.150
0.234
-0.047
0.194
x(t-1)
0.193
0.233
(continued)
10-60
-------
Section 10—Time Series Analysis
Table 10-22. Analysis for PCE Concentration at 42 Base South, Variables that Needed A Lag-1
Term. Period Jan 2011 to Feb 2012 (cont.)
Model Y(t)-Y(t-1) =
Model Y(t)-Y(t-1) =
Predictor Name (x(t))
Model Term
intercept + x(t]
+ x(t-1)
intercept +
x(t)
Estimate
SE
Estimate
SE
Setra_420ss. base_Pa
intercept
-0.208
0.377
-0.248
0.335
x(t)
0.046
0.364
0.004
0.295
x(t-1)
-0.081
0.364
intercept
-0.236
0.139
-0.244
0.136
Setra_422base.out_Pa
x(t)
-0.248
0.179
-0.055
0.102
x(t-1)
0.234
0.176
Setra_422base.upst_Pa
intercept
-0.226
0.137
-0.231
0.136
x(t)
0.040
0.207
0.127
0.170
x(t-1)
0.157
0.207
intercept
-0.079
0.210
-0.268
0.199
Setra_422SGdp.ss_Pa
x(t)
0.331
0.183
0.044
0.103
x(t-1)
-0.432
0.234
Setra_422ss. base_Pa
intercept
0.249
0.542
-0.154
0.506
x(t)
0.347
0.286
-0.034
0.202
x(t-1)
-0.552
0.299
intercept
-0.059
0.123
-0.057
0.118
Snowdepth_daily
x(t)
-0.746
0.406
-0.692**
0.170
x(t-1)
0.050
0.387
Soil_H20_ln6._cbar
intercept
-0.705**
0.225
-0.648**
0.211
x(t)
0.027
0.027
0.005*
0.002
x(t-1)
-0.021
0.027
intercept
-0.342
0.173
-0.334*
0.166
Soil_H20_Out3.5._cbar
x(t)
0.002
0.005
0.002
0.002
x(t-1)
0.001
0.005
Soil_H20_Out6._cbar
intercept
-0.41*
0.180
-0.397*
0.175
x(t)
0.007
0.009
0.002
0.001
x(t-1)
-0.006
0.009
intercept
-0.205
0.115
-0.434
0.332
Soil_T_C_OTC.1
x(t)
-0.076*
0.031
0.017
0.021
x(t-1)
0.086**
0.031
T_422_F
intercept
-1.680
1.246
-1.667
1.185
x(t)
0.019
0.041
0.020
0.017
x(t-1)
0.001
0.041
(continued)
10-61
-------
Section 10—Time Series Analysis
Table 10-22. Analysis for PCE Concentration at 42 Base South, Variables that Needed A Lag-1
Term. Period Jan 2011 to Feb 2012 (cont.)
Model Y(t)-Y(t-1) =
Model Y(t)-Y(t-1) =
Predictor Name (x(t))
Model Term
intercept + x(t]
+ x(t-1)
intercept +
x(t)
Estimate
SE
Estimate
SE
intercept
-2.298*
0.894
-2.274*
0.871
T_422baseN_C
x(t)
0.009
0.074
0.033*
0.014
x(t-1)
0.024
0.073
T_422baseS_C
intercept
-2.677*
1.052
-2.66*
1.027
x(t)
0.047
0.074
0.078*
0.016
x(t-1)
-0.009
0.073
intercept
-2.631
1.340
-2.720*
1.309
T_422first_C
x(t)
0.061
0.056
0.035
0.018
x(t-1)
-0.027
0.055
T_out_F
intercept
-1.077*
0.407
-1.063*
0.416
x(t)
-0.026
0.030
0.015*
0.007
x(t-1)
0.041
0.029
intercept
-1.380*
0.536
-1.340*
0.573
T_out_Hi_F
x(t)
-0.035
0.023
0.015
0.008
x(t-1)
0.051*
0.022
T_out_Lo_F
intercept
-0.826**
0.295
-0.809**
0.287
x(t)
0.003
0.026
0.016*
0.007
x(t-1)
0.014
0.026
intercept
-0.980**
0.360
-0.973*
0.375
Wind_Chill_F
x(t)
-0.031
0.027
0.014*
0.007
x(t-1)
0.045
0.026
Wind_Run_mi
intercept
0.222
0.496
-0.091
0.465
x(t)
0.001
0.001
0.000
0.001
x(t-1)
-0.001
0.001
intercept
0.318
0.479
0.034
0.471
Wind_Speed_MPH
x(t)
0.196
0.181
-0.070
0.117
x(t-1)
-0.339
0.181
"Significant at 1% level of significance
'Significant at 5% level of significance
10-62
-------
Section 10—Time Series Analysis
Weekly Average Snow Depth vs. PCE Concentration (Radiello)
422 Base South; Mitigation Off or Not Installed Data Only
18
16
14
f
~6D 12
A.
e
e
V
r°
u
o
s
©
H 8
a
6
4
J
~
~
~
~ ~
~
0.5 1 1.5 2 2.5 3 3.5
Snow Depth on Ground (inches)
Figure 10-18. XY plot of weekly average snow depth vs. PCE concentration.
Change in Weekly Average Snow Depth vs. PCE Concentration (Radiello)
422 Base South; Mitigation Off or Not Installed Data Only
26—
~
~
~
~
~
~
4 ~
*
-2.5 -2 -1.5 -1 -0.5 0 0.5 1 1.5 2
Snow Depth on Ground change from previous week
Figure 10-19. XY Plot of change in weekly average snow depth vs. PCE concentration.
10-63
-------
Section 10—Time Series Analysis
Weekly Average Snow Depth Change vs. PCE Concentration (Radiello)
422 Base South; Mitigation Off or Not Installed Data Only
~ ~
-10
Snow Depth on Ground {inches)
Figure 10-20. XY Plot of weekly average snow depth vs. change in PCE.
As shown in Table 10-23, only one significant association was found between predictors needing two lag
terms and the PCE concentration. The one case where significance was shown was a correlation between
the soil temperature at 9 ft (3 ft directly below the basement floor) in the reduced model. This positive
coefficient is physically realistic, most likely suggesting that warmer soils enhanced PCE volatility.
Table 10-23. Analysis for PCE Concentration at 422 Base South Variables that Needed Both
Lag-1 Week And Lag-2 Week Terms. Period Jan 2011 to Feb 2012
Model Y(t)-Y(t-1) =
intercept + x(t) +
x(t-1)+x(t-2)
Model
Term
Model Y(t)-Y(t-1) =
intercept + x(t) + x(t—1)
Model Y(t)-Y(t-1) =
intercept + x(t)
Predictor Name
Estimate
SE
Estimate
SE
Estimate
SE
Intercept
-1.419*
0.684
-2.546
20,354.593
-2.677**
0.746
Soil_T_C_MW3.9
x(t)
0.326
0.255
0.543
0.683
0.14655**
0.045
x(t-1)
-0.598
0.419
-0.775
0.681
x(t"2)
0.350
0.251
(continued)
10-64
-------
Section 10—Time Series Analysis
Table 10-23. Analysis for PCE Concentration at 422 Base South Variables that Needed Both
Lag-1 Week And Lag-2 Week Terms. Period Jan 2011 to Feb 2012 (cont.)
Model Y(t)-Y(t-1) =
intercept + x(t) +
x(t-1)+x(t-2)
Model
Term
Model Y(t)-Y(t-1) =
intercept + x(t) + x(t—1)
Model Y(t)-Y(t-1) =
intercept + x(t)
Predictor Name
Estimate
SE
Estimate
SE
Estimate
SE
Intercept
-0.322
0.501
-0.172
0.506
-0.239
0.423
Wind_Dir
x(t)
0.000
0.001
0.000
0.002
0.000
0.002
x(t-1)
0.000
0.001
-0.001
0.002
x(t"2)
0.000
0.001
Intercept
0.175
0.641
0.245
0.632
0.079
0.544
Wind_Dir_Hi
x(t)
0.000
0.001
-0.001
0.003
-0.001
0.002
x(t-1)
0.000
0.001
-0.001
0.003
x(t"2)
-0.001
0.001
"Significant at 1% level of significance
'Significant at 5% level of significance
10.7.5 Time Series Analysis of 422 Basement South Chloroform Data Set from the
period Sept 2012-Apr 2013
In this section, we discussed the evaluation of the association between chloroform concentrations at 422
base south and a list of predictor variables measured weekly during between Sept 2012 and April 2013.
All models in this section (period Sept 2013 to April 2013) include mitigation as a control variable.
Mitigation was coded as 1 (on both sides of duplex,) and 0 (OFF, passive or not yet installed).
Table 10-24 display the analysis results for the chloroform time series, period Sept 2012 to April 2013 at
422 base south for variables not requiring lag terms in these models. Only barometric pressure was
associated with X422BaseS_Radiello_CHC13-2 after controlling for the effect of mitigation. Note that the
mitigation effect was not statistically significant, although it almost always had a consistent coefficient of
the expected sign.
Table 10-25 displays the results of the analysis to of the association between basement 422 chloroform
for the period Sept 2012 to April 2013, with predictors needing a lag-1 week term in the model. Both the
past and current measurements of shallow soil temperature at 1 ft bis and exterior temperature were
correlated to the chloroform concentration. The current measurements of predictor heating degree days,
and soil moisture at 6 ft were correlated with the chloroform concentrations. Only the previous weeks
measurements of the basement temperatures were correlated to chloroform, not the current measurement.
The correlation with heating degree days was in the expected direction and was large in magnitude. An
example is helpful to understand the implications of the model of shallow soil temperature. Assume
mitigation is off and that the shallow soil temperature is 15°C this week and 10°C last week. The model
predicts that the indoor chloroform will be 0.42 (ig/m3 higher this week than last week.
As with Table 10-24, it is notable that the effect of mitigation being on is almost always to decrease
chloroform and usually by a similar amount (less than 0.250 (ig/m3) but that this never rises to statistical
significance.
10-65
-------
Section 10—Time Series Analysis
Table 10-24. Analysis for First Difference of Chloroform Concentration at 422 Basement South.
Variables that Did Not Need Lag Terms. Period Sept 2012 to April 2013
Model: Y(t) - Y(t-1) = intercept + mitigation (t) + x(t)
Predictor Name
Model Term
Estimate
SE
Intercept
0.086
0.143
Bar_drop_Hg.hr
Mitigation
-0.117
0.158
x(t)
9.488
27.176
Intercept
-35.363*
13.862
Bar_in_Hg
Mitigation
-0.050
0.143
x(t)
1.177*
0.460
Intercept
0.086
0.143
BP_Net_Change
Mitigation
-0.117
0.158
x(t)
-0.060
0.167
Intercept
0.384
0.252
BP_Pump_Speed
Mitigation
-0.146
0.154
x(t)
-1.486
1.048
Intercept
0.333
0.275
BP_Stroke_Length
Mitigation
-0.163
0.162
x(t)
-0.292
0.277
Intercept
0.106
0.144
Cool_Degree_Day
Mitigation
-0.115
0.156
x(t)
-0.011
0.014
Intercept
0.290
0.328
Fall_Crk_Gage_ht_ft
Mitigation
-0.105
0.157
x(t)
-0.006
0.009
Intercept
0.025
0.283
Hum_out_.
Mitigation
-0.111
0.158
x(t)
0.016
0.066
Intercept
0.048
0.193
Rain_ln_met
Mitigation
-0.110
0.158
x(t)
0.000
0.001
Intercept
0.244
0.247
RainJPH
Mitigation
-0.136
0.159
x(t)
-0.005
0.006
Intercept
0.356
0.633
T_422first_C
Mitigation
-0.107
0.158
x(t)
-0.004
0.008
(continued)
10-66
-------
Section 10—Time Series Analysis
Table 10-24. Analysis for First Difference of Chloroform Concentration at 422 Basement South.
Variables that Did Not Need Lag Terms. Period Sept 2012 to April 2013 (cont.)
Model: Y(t) - Y(t-1) = intercept + mitigation (t) + x(t)
Predictor Name
Model Term
Estimate
SE
Intercept
0.102
0.151
Wind_Dir
Mitigation
-0.116
0.158
x(t)
-0.035
0.096
Intercept
0.157
0.176
Wind_Dir_Hi
Mitigation
-0.164
0.173
x(t)
-0.041
0.059
Intercept
0.062
0.178
Wind_Speed_Hi_MPH
Mitigation
0.132
0.225
x(t)
0.020
0.011
Intercept
0.028
0.162
Wind_Speed_MPH
Mitigation
-0.083
0.172
x(t)
0.005
0.004
'Significant at 5% level of significance
Table 10-25. Analysis Chloroform Concentration at 422 Base South. Variables that Needed A Lag-1
One Week Term. Period Sept 2012 to April 2013
Model: Y(t)-Y(t-1) =
intercept + mitigation (t)
+ x(t)+ x(t-1)
Model: Y(t)-Y(t-1) =
intercept + mitigation (t) +
x(t)
Predictor Name
Model Term
Estimate
SE
Estimate
SE
intercept
0.096
0.150
0.086
0.143
Dew_pt_422_F
mitigation(t)
-0.132
0.171
-0.117
0.158
x(t)
13.751
32.226
9.488
27.176
x(t-1)
8.819
34.080
intercept
-28.500
17.429
-35.36332*
13.862
mitigation(t)
-0.059
0.145
-0.050
0.143
Heat_Degree_Day
x(t)
1.290*
0.496
1.17744*
0.460
x(t-1)
-0.341
0.514
intercept
0.099
0.151
0.086
0.143
Hum_422_.
mitigation(t)
-0.136
0.170
-0.117
0.158
x(t)
-0.093
0.197
-0.060
0.167
x(t-1)
-0.069
0.207
(continued)
10-67
-------
Section 10—Time Series Analysis
Table 10-25. Analysis Chloroform Concentration at 422 Base South. Variables that Needed A Lag-
1 One Week Term. Period Sept 2012 to April 2013 (cont.)
Model: Y(t)- Y(t-1) =
intercept + mitigation (t)
+ x(t)+ x(t-1)
Model: Y(t)- Y(t-1) =
intercept + mitigation (t) +
x(t)
Predictor Name
Model Term
Estimate
SE
Estimate
SE
Setra_420ss. base_Pa
intercept
0.357
0.328
0.384
0.252
mitigation(t)
-0.145
0.157
-0.146
0.154
x(t)
-1.483
1.069
-1.486
1.048
x(t-1)
0.137
1.045
Setra_422base.out_Pa
intercept
0.378
0.314
0.333
0.275
mitigation(t)
-0.164
0.165
-0.163
0.162
x(t)
-0.270
0.291
-0.292
0.277
x(t-1)
-0.086
0.273
Setra_422base.upst_Pa
intercept
0.092
0.149
0.106
0.144
mitigation(t)
-0.112
0.158
-0.115
0.156
x(t)
-0.012
0.014
-0.011
0.014
x(t-1)
0.007
0.014
Setra_422SGdp.ss_Pa
intercept
0.084
0.346
0.290
0.328
mitigation(t)
-0.032
0.160
-0.105
0.157
x(t)
-0.020
0.012
-0.006
0.009
x(t-1)
0.018
0.012
Setra_422ss. base_Pa
intercept
0.094
0.411
0.025
0.283
mitigation(t)
-0.124
0.170
-0.111
0.158
x(t)
0.016
0.068
0.016
0.066
x(t-1)
-0.017
0.071
Snowdepth_daily
intercept
0.086
0.198
0.048
0.193
mitigation(t)
-0.091
0.160
-0.110
0.158
x(t)
0.001
0.001
0.000
0.001
x(t-1)
-0.001
0.001
Soil_H20_ln6._cbar
intercept
0.138
0.241
0.244
0.247
mitigation(t)
-0.082
0.153
-0.136
0.159
x(t)
-0.023*
0.011
-0.005
0.006
x(t-1)
0.020
0.010
Soil_H20_Out3.5._cbar
intercept
0.370
0.803
0.356
0.633
mitigation(t)
-0.107
0.163
-0.107
0.158
x(t)
-0.004
0.009
-0.004
0.008
x(t-1)
0.000
0.009
(continued)
10-68
-------
Section 10—Time Series Analysis
Table 10-25. Analysis Chloroform Concentration at 422 Base South. Variables that Needed A Lag-
1 One Week Term. Period Sept 2012 to April 2013 (cont.)
Model: Y(t)- Y(t-1) =
intercept + mitigation (t)
+ x(t)+ x(t-1)
Model: Y(t)- Y(t-1) =
intercept + mitigation (t) +
x(t)
Predictor Name
Model Term
Estimate
SE
Estimate
SE
Soil_H20_Out6._cbar
intercept
-0.235
0.272
0.102
0.151
mitigation(t)
0.145
0.235
-0.116
0.158
x(t)
0.019
0.101
-0.035
0.096
x(t-1)
0.267
0.182
Soil_T_C_MW3.9
intercept
0.247
0.249
0.157
0.176
mitigation(t)
-0.232
0.219
-0.164
0.173
x(t)
-0.051
0.063
-0.041
0.059
x(t-1)
-0.040
0.077
Soil_T_C_OTC.1
intercept
0.075
0.117
0.062
0.178
mitigation(t)
0.033
0.161
0.132
0.225
x(t)
0.054**
0.010
0.020
0.011
x(t-1)
-0.046**
0.009
So i l_T_C_OT C. 3.5
intercept
-0.231
1.657
-0.461
1.387
mitigation(t)
-0.141
0.175
-0.139
0.171
x(t)
0.011
0.023
0.008
0.020
x(t-1)
-0.006
0.022
Soil_T_C_OTC.6
intercept
0.098
0.189
0.028
0.162
mitigation(t)
-0.155
0.204
-0.083
0.172
x(t)
0.011
0.012
0.005
0.004
x(t-1)
-0.005
0.011
T_422_F
intercept
-0.054
0.431
-0.236
0.372
mitigation(t)
0.025
0.225
-0.012
0.204
x(t)
1.721*
0.629
0.464
0.514
x(t-1)
-1.706**
0.532
T_422baseN_C
intercept
-0.053
0.181
0.054
0.165
mitigation(t)
-0.130
0.198
-0.121
0.198
x(t)
-0.005
0.006
0.002
0.006
x(t-1)
0.014*
0.005
T_422baseS_C
intercept
0.123
0.169
0.062
0.163
mitigation(t)
-0.275
0.228
-0.062
0.211
x(t)
0.023
0.012
0.003
0.011
x(t-1)
-0.036**
0.012
(continued)
10-69
-------
Section 10—Time Series Analysis
Table 10-25. Analysis Chloroform Concentration at 422 Base South. Variables that Needed A
Lag-1 One Week Term. Period Sept 2012 to April 2013 (cont.)
Model: Y(t)- Y(t-1) =
intercept + mitigation (t)
+ x(t)+ x(t-1)
Model: Y(t)- Y(t-1) =
intercept + mitigation (t) +
x(t)
Predictor Name
Model Term
Estimate
SE
Estimate
SE
T_out_F
intercept
0.080
0.143
0.081
0.142
mitigation(t)
-0.139
0.161
-0.126
0.158
x(t)
0.016
0.053
0.030
0.049
x(t-1)
0.039
0.053
T_out_Hi_F
intercept
0.245
0.811
0.233
0.661
mitigation(t)
-0.112
0.171
-0.111
0.158
x(t)
-0.001
0.013
-0.001
0.004
x(t-1)
0.000
0.015
T_out_Lo_F
intercept
0.130
0.231
0.133
0.232
mitigation(t)
-0.086
0.184
-0.087
0.185
x(t)
-0.002
0.002
0.000
0.001
x(t-1)
0.002
0.001
Wind_Chill_F
intercept
-0.154
0.387
-0.121
0.367
mitigation(t)
-0.174
0.187
-0.170
0.183
x(t)
0.001
0.002
0.001
0.002
x(t-1)
0.000
0.001
X422baseN_Wkly_Elect_radon
intercept
-0.852
0.776
-0.270
0.682
mitigation(t)
-0.084
0.204
-0.106
0.203
x(t)
0.184
0.441
0.023
0.043
x(t-1)
-0.122
0.420
X422baseS_Wkly_Elect_radon
intercept
0.019
0.197
0.086
0.194
mitigation(t)
-0.094
0.219
-0.089
0.218
x(t)
-0.024
0.043
-0.001
0.020
x(t-1)
0.041
0.039
X422first_Wkly_Elect_radon
intercept
-0.070
0.218
0.020
0.213
mitigation(t)
-0.139
0.213
-0.131
0.209
x(t)
0.030
0.084
0.013
0.020
x(t-1)
0.000
0.075
X422office2n d_Wkly_E 1 ect_rad o n
intercept
0.422
1.640
0.466
1.648
mitigation(t)
-0.029
0.206
-0.111
0.193
x(t)
-0.060
0.057
-0.007
0.029
x(t-1)
0.052
0.048
"Significant at 1% level of significance; * Significant at 5% level of significance
10-70
-------
Section 10—Time Series Analysis
The only variable that required two lag terms, Wind Run was not correlated with chloroform (see Table
10-26)."
Table 10-26. Analysis for Chloroform Concentration at 422 Base South. Variables Needing Lag-1
And Lag-2 Week Terms. Period Sept 2012 to April 2013
Model Y(t)-Y(t-1) = intercept
+ x(t) + x(t-1)+x(t-2)
Model Y(t)-Y(t-1) =
intercept + x(t) + x(t-1)
Model Y(t)-Y(t-1) =
intercept + x(t)
Predictor
Name (x(t))
Model
Term
Estimate
SE
Estimate
SE
Estimate
SE
intercept
0.130
0.170
0.096
0.150
0.086
0.143
Mitigation
-0.199
0.193
-0.132
0.171
-0.117
0.158
Wind_Run_mi
x(t)
16.281
32.314
13.751
32.226
9.488
27.176
x(t—1)
37.376
37.720
8.819
34.080
x(t—2)
60.663
33.665
10.7.6 Time Series Analysis Results of 422 Basement South PCE Data from Sept 2012
through April 2013
In this section, we discussed the evaluation of the association between PCE a concentrations at 422 base
south and a list of predictor variables measured weekly during between Sept 2012 and April 2013. All
models in this section (period Sept 2013 to April 2013) include mitigation as a control variable.
Mitigation was coded as 1 (on both sides of duplex,) and 0 (OFF, passive or not yet installed).
Table 10-27 displays the analysis results for variables that did not need a lag term. The variable
mitigation was included as a control variable but was not statistically significant. Only the barometric
pressure predictor variable was significantly associated with PCE.
Table 10-27. Analysis for First Difference of422 Base South PCE Concentration. Variables that
Did Not Need Lag Terms. Period Sept 2012 to April 2013
Model: Y(t) - Y(t-1) = intercept + mitigation (t) + x(t)
Predictor Name
Model Term
Estimate
SE
intercept
0.338
0.587
Bar_drop_Hg.hr
mitigation
-0.432
0.649
x(t)
-23.020
111.636
intercept
-131.243*
58.183
BarJnJHg
mitigation
-0.211
0.599
x(t)
4.371*
1.933
intercept
0.338
0.587
BP_Net_Change
mitigation
-0.435
0.649
x(t)
0.110
0.686
intercept
0.877
1.068
BP_Pump_Speed
mitigation
-0.504
0.651
x(t)
-2.652
4.436
(continued)
10-71
-------
Section 10—Time Series Analysis
Table 10-27. Analysis for First Difference of422 Base South PCE Concentration. Variables that
Did Not Need Lag Terms. Period Sept 2012 to April 2013 (cont.)
Model: Y(t) - Y(t-1) = intercept + mitigation (t) + x(t]
Predictor Name
Model Term
Estimate
SE
intercept
1.097
1.137
B P_St ro ke_Le n gth
mitigation
-0.598
0.670
x(t)
-0.886
1.148
intercept
0.433
0.590
Cool_Degree_Day
mitigation
-0.454
0.639
x(t)
-0.045
0.055
intercept
1.178
1.346
Fa 1 l_C rk_Ga g e_ht_ft
mitigation
-0.412
0.643
x(t)
-0.025
0.036
intercept
1.087
1.148
Hum_out_.
mitigation
-0.462
0.641
x(t)
-0.203
0.269
intercept
0.221
0.792
Rain_ln_met
mitigation
-0.437
0.648
x(t)
0.001
0.003
intercept
1.012
1.014
RainJPH
mitigation
-0.541
0.651
x(t)
-0.021
0.026
intercept
1.860
2.587
T_422first_C
mitigation
-0.411
0.645
x(t)
-0.021
0.035
intercept
0.371
0.619
Wind_Dir
mitigation
-0.450
0.648
x(t)
-0.058
0.395
intercept
0.076
0.725
Wind_Dir_Hi
mitigation
-0.256
0.711
x(t)
0.149
0.241
intercept
0.457
0.745
Wind_Speed_Hi_MPH
mitigation
-0.258
0.945
x(t)
0.039
0.045
intercept
0.328
0.671
Wind_Speed_MPH
mitigation
-0.513
0.714
x(t)
0.014
0.017
'Significant at 5% level of significance
10-72
-------
Section 10—Time Series Analysis
Among the predictors needing a lag-1 week term in the model (Table 10-28), the following predictor
variables were statistically significantly related to PCE:
¦ the current week's total of heating degree days (in the reduced model),
¦ the past and current weeks" measurement of soil temperature at 1 ft bis, and
¦ the past week's measurements of temperatures in the 422 basement.
The relationship with heating degree days was in the expected direction: colder weather would result in
more heating degree days in the current week and higher PCE in indoor air. Once again since the
coefficients for the soil temperature are opposite, a worked example will be beneficial. Assume that the
shallow soil is 15°C this week and was 10°C last week. Assume mitigation to be turned off. Under those
conditions, the model predicts PCE will increase by 1.4 (ig/m3.
Table 10-28. Analysis for 422 Base South PCE Concentration. Variables that Needed A Lag-1
Week Term. Period Sept 2012 to April 2013
Model: Y(t)- Y(t-1) =
Model: Y(t)- Y(t-1) =
Predictor Name
Model Term
x(t)+ x(
t-1)
(t) +
x(t)
Estimate
SE
Estimate
SE
Dew_pt_422_F
intercept
0.352
0.619
0.338
0.587
mitigation(t)
-0.454
0.702
-0.432
0.649
x(t)
-16.713
132.540
-23.020
111.636
x(t-1)
13.050
140.164
Heat_Degree_Day
intercept
-190.689*
70.939
-131.243*
58.183
mitigation(t)
-0.136
0.590
-0.211
0.599
x(t)
3.392
2.018
4.37081*
1.933
x(t-1)
2.954
2.094
Hum_422_.
intercept
0.361
0.621
0.338
0.587
mitigation(t)
-0.467
0.702
-0.435
0.649
x(t)
0.053
0.813
0.110
0.686
x(t-1)
-0.117
0.854
Setra_420ss. base_Pa
intercept
1.685
1.362
0.877
1.068
mitigation(t)
-0.545
0.653
-0.504
0.651
x(t)
-2.724
4.444
-2.652
4.436
x(t-1)
-4.167
4.344
Setra_422base.out_Pa
intercept
1.720
1.274
1.097
1.137
mitigation(t)
-0.610
0.669
-0.598
0.670
x(t)
-0.578
1.180
-0.886
1.148
x(t-1)
-1.186
1.107
(continued)
10-73
-------
Section 10—Time Series Analysis
Table 10-28. Analysis for 422 Base South PCE Concentration. Variables that Needed A Lag-1
Week Term. Period Sept 2012 to April 2013 (cont.)
Model: Y(t)- Y(t-1) =
intercept + mitigation (t) +
x(t)+ x(t-1)
Model: Y(t)- Y(t-1) =
Predictor Name
Model Term
(t) +
x(t)
Estimate
SE
Estimate
SE
Setra_422base.upst_Pa
intercept
0.415
0.612
0.433
0.590
mitigation(t)
-0.450
0.653
-0.454
0.639
x(t)
-0.047
0.058
-0.045
0.055
x(t-1)
0.009
0.057
Setra_422SGdp.ss_Pa
intercept
0.964
1.483
1.178
1.346
mitigation(t)
-0.336
0.685
-0.412
0.643
x(t)
-0.040
0.054
-0.025
0.036
x(t-1)
0.019
0.050
Setra_422ss. base_Pa
intercept
0.548
1.665
1.087
1.148
mitigation(t)
-0.362
0.687
-0.462
0.641
x(t)
-0.204
0.274
-0.203
0.269
x(t-1)
0.131
0.288
Snowdepth_daily
intercept
0.223
0.825
0.221
0.792
mitigation(t)
-0.435
0.667
-0.437
0.648
x(t)
0.001
0.004
0.001
0.003
x(t-1)
0.000
0.004
Soil_H20_ln6._cbar
intercept
0.737
1.034
1.012
1.014
mitigation(t)
-0.402
0.657
-0.541
0.651
x(t)
-0.067
0.047
-0.021
0.026
x(t-1)
0.051
0.043
Soil_H20_Out3.5._cbar
intercept
3.641
3.224
1.860
2.587
mitigation(t)
-0.500
0.654
-0.411
0.645
x(t)
-0.011
0.036
-0.021
0.035
x(t-1)
-0.033
0.036
Soil_H20_Out6._cbar
intercept
1.076
1.154
0.371
0.619
mitigation(t)
-0.997
0.998
-0.450
0.648
x(t)
-0.171
0.428
-0.058
0.395
x(t-1)
-0.560
0.771
S o i l_T_C_M W3.9
intercept
1.023
0.991
0.076
0.725
mitigation(t)
-0.973
0.872
-0.256
0.711
x(t)
0.040
0.250
0.149
0.241
x(t-1)
-0.420
0.306
(continued)
10-74
-------
Section 10—Time Series Analysis
Table 10-28. Analysis for 422 Base South PCE Concentration. Variables that Needed A Lag-1
Week Term. Period Sept 2012 to April 2013 (cont.)
Model: Y(t)- Y(t-1) =
intercept + mitigation (t) +
x(t)+ x(t-1)
Model: Y(t)- Y(t-1) =
Predictor Name
Model Term
(t) +
x(t)
Estimate
SE
Estimate
SE
Soil_T_C_OTC.1
intercept
0.480
0.677
0.457
0.745
mitigation(t)
-0.372
0.929
-0.258
0.945
x(t)
0.138*
0.056
0.039
0.045
x(t-1)
-0.118*
0.054
S o i l_T_C_OT C. 3.5
intercept
0.835
6.581
-3.259
5.660
mitigation(t)
-0.648
0.693
-0.619
0.698
x(t)
0.098
0.090
0.053
0.083
x(t-1)
-0.103
0.086
Soil_T_C_OTC.6
intercept
0.515
0.730
0.328
0.671
mitigation(t)
-0.862
0.791
-0.513
0.714
x(t)
-0.067
0.048
0.014
0.017
x(t-1)
0.086
0.044
T_422_F
intercept
0.583
2.111
-0.008
1.535
mitigation(t)
-0.381
1.102
-0.355
0.843
x(t)
3.132
3.077
0.689
2.119
x(t-1)
-3.598
2.604
T_422baseN_C
intercept
0.072
0.734
0.508
0.670
mitigation(t)
-0.326
0.803
-0.283
0.805
x(t)
-0.043
0.023
-0.016
0.023
x(t-1)
0.0563*
0.022
T_422baseS_C
intercept
0.599
0.744
0.429
0.666
mitigation(t)
-0.994
1.003
-0.349
0.862
x(t)
0.082
0.054
0.018
0.045
x(t-1)
-0.11453*
0.054
T_out_F
intercept
0.330
0.595
0.331
0.583
mitigation(t)
-0.507
0.667
-0.495
0.650
x(t)
0.101
0.221
0.114
0.202
x(t-1)
0.037
0.222
T_out_Hi_F
intercept
2.232
3.294
2.322
2.682
mitigation(t)
-0.406
0.693
-0.418
0.641
x(t)
-0.015
0.054
-0.012
0.016
x(t-1)
0.003
0.062
(continued)
10-75
-------
Section 10—Time Series Analysis
Table 10-28. Analysis for 422 Base South PCE Concentration. Variables that Needed A Lag-1
Week Term. Period Sept 2012 to April 2013 (cont.)
Model: Y(t)- Y(t-1) =
intercept + mitigation (t) +
x(t)+ x(t-1)
Model: Y(t)- Y(t-1) =
Predictor Name
Model Term
(t) +
x(t)
Estimate
SE
Estimate
SE
T_out_Lo_F
intercept
0.404
0.944
0.421
0.954
mitigation(t)
-0.401
0.750
-0.402
0.759
x(t)
-0.007
0.007
0.000
0.005
x(t-1)
0.007
0.006
Wind_Chill_F
intercept
-1.195
1.564
-1.038
1.484
mitigation(t)
-0.848
0.755
-0.829
0.741
x(t)
0.006
0.008
0.007
0.007
x(t-1)
0.002
0.006
X422baseN_Wkly_Elect_radon
intercept
-0.520
3.133
0.086
2.722
mitigation(t)
-0.599
0.822
-0.715
0.810
x(t)
-1.492
1.781
0.027
0.170
x(t-1)
1.539
1.694
X422baseS_Wkly_Elect_radon
intercept
0.302
0.787
0.548
0.766
mitigation(t)
-0.515
0.877
-0.591
0.863
x(t)
-0.174
0.171
-0.028
0.079
x(t-1)
0.194
0.154
X422first_Wkly_Elect_radon
intercept
0.137
0.896
0.370
0.851
mitigation(t)
-0.655
0.874
-0.775
0.834
x(t)
-0.125
0.343
0.026
0.079
x(t-1)
0.178
0.307
X422office2nd_Wkly_Elect_radon
intercept
1.448
6.816
1.410
6.632
mitigation(t)
-0.487
0.858
-0.417
0.779
x(t)
0.026
0.235
-0.019
0.116
x(t-1)
-0.044
0.199
'Significant at 5% level of significance
The variable wind run which required two lag terms was not correlated with PCE (Table 10-29).
10-76
-------
Section 10—Time Series Analysis
Table 10-29. Analysis for PCE Concentration at 422 Base South. Variables Needing Both Lag-1
And Lag-2 Week Terms. Period Sept 2012 to April 2013
Model Y(t)-Y(t-1) = intercept
+ x(t) + x(t-1)+x(t-2)
Model Y(t)-Y(t-1) =
intercept + x(t) + x(t—1)
Model Y(t)-Y(t-1) =
intercept + x(t)
Predictor
Name (x(t))
Model
Term
Estimate
SE
Estimate
SE
Estimate
SE
intercept
0.681
0.693
0.352
0.619
0.338
0.587
Mitigation
-0.935
0.787
-0.454
0.702
-0.432
0.649
Wind_Run_mi
x(t)
10.321
131.649
-16.713
132.540
-23.020
111.636
x(t-1)
151.413
153.677
13.050
140.164
x(t-2)
262.750
137.155
10-77
-------
Section 10—Time Series Analysis
Addendum
To capture the existing serial correlation between consecutive observations, a time series regression
model requires incorporating a model for the errors. For example, assume radon concentrations can be
modeled as a function of observed predictors and a random error. The random error includes
measurement error and any variability not explained by the current variables in the model. A model like
n
that is formulated as ^ /3jxjt + £t where xjt, i = 1,...,n denote n time-dependent predictors observed
i=1
at time t, y, is the outcome of interest (radon emissions) at time t, and £, is the associated error at time t.
A standard model for the error term is : £t = p£t_x + Vt, where —\, can be rewritten as: = + Pet-1 + vt and since
7=1
n n
yt-1 = 'YjP1xu-i + £t-\ then £t_x = yt_x - . which result in
Z=1
n f n
yt=+ p yt~i -
i=1 i=1
A
. . +K
;= i V /= i y
The above model is termed an autoregressive (because it includes lag values of yt) distributed lag
(because it includes lag values of the predictors: xit,i =
10-78
-------
Section 11—Results and Discussion: Do Groundwater Concentrations Control Soil Gas
Concentrations at this Site? And Thus Indoor Air Concentrations?
Table of Contents
11.0 Results and Discussion: Do Groundwater Concentrations Control Soil Gas Concentrations
at this Site? And Thus Indoor Air Concentrations? 11-1
11.1 Groundwater Level Changes 11-1
11.2 Groundwater Concentration Trends 11-2
11.2.1 Is the Groundwater Concentration Trend Correlated to Well or Water Table
Depth? 11-5
11.3 Revision to the Conceptual Model—Is the Groundwater Concentration Related to
Soil Gas and Indoor Air Concentrations? 11-6
List of Figures
11-1. Stacked hydrological graph with rainfall in inches (top—green line), depth to water in
feet (middle—red circles), and discharge at Fall Creek in ft3/s (bottom—blue line) 11-1
11-2. Actual groundwater levels along with the daily time series predicted from Fall Creek
gage height data 11-2
11-3. Groundwater concentrations over time for Indianapolis duplex 11-4
11-4. PCE groundwater concentrations over time, showing concentrations by individual well
and soil gas ports 11-4
11-5. Plot of groundwater PCE and chloroform concentrations against well screen (or soil gas
port) depth (well depth measured to top of screen) 11-5
11-6. Plot of groundwater PCE and chloroform concentrations against groundwater depth 11-6
List of Tables
11-1. Groundwater Monitoring Locations 11-3
11-i
-------
Section 11—Results and Discussion: Do Groundwater Concentrations Control Soil Gas
Concentrations at this Site? And Thus Indoor Air Concentrations?
11-ii
-------
Section 11—Results and Discussion: Do Groundwater Concentrations Control Soil Gas
Concentrations at this Site? And Thus Indoor Air Concentrations?
11.0 Results and Discussion: Do Groundwater Concentrations
Control Soil Gas Concentrations at this Site? And Thus Indoor
Air Concentrations?
11.1 Groundwater Level Changes
Figure 11-1 presents the relationship between depth to water readings taken at the 422/420 house, Fall
Creek discharge, and the rainfall taken at the house for the duration of the project. Fall Creek discharge
was obtained from the online stream gage data collected by the USGS at gaging station number
03352500.
Stream Discharge, Average Depth to Water, and Rainfall
re 4
H-
C
*re
cc i
6000
m 4000
tu>
J 2000
.a o:
A
1 1
Li
j j . r j
*
•
•
%
...
•
•
. •
•
•
•
-
•
*
\
*5i
_ 1 *v>
•w.
¦ —
JLi
M-
jjJ
3
oJ
r-
Ci
rM
rsi
lO
O
rM
O
fM
ft
vD
O
fN
3
CO
o
N
O
m
m
o
2D
>
<
IB
CD
•n
01
16
on*
fD
li
D
12
¦ ScrMim Dwefurg*
• Avenue DTW
- Rainfall
Date
Figure 11-1. Stacked hydrological graph with rainfall in inches (top—green line), depth to water
in feet (middle—red circles), and discharge at Fall Creek in ft3/s (bottom—blue
line).
The most striking feature is during the heavy rain events that occurred in early March 2011. This rain was
immediately reflected in the discharge at Fall Creek and in the rapid decrease in depth to groundwater as
the water level responded to the increasing height of the creek. The hotter summer 2011 months that
followed were drier, with deepening groundwater levels down to the lowest water table depth recorded
during this project (about 20 ft bis). Spikes in Fall Creek discharge in early 2013 also were reflected in a
higher water table.
As described in the previous report (U.S. EPA, 2012a), the water table had been expected to be at about
17 ft, just under the deepest of the soil gas ports (16.5 ft), but these deep ports were flooded in 2011. In
27-
http://waterdata.usgs.gov/in/nwis/uv?site no=03352500
11-1
-------
Section 11—Results and Discussion: Do Groundwater Concentrations Control Soil Gas
Concentrations at this Site? And Thus Indoor Air Concentrations?
response, groundwater samples were taken from the deeper soil gas ports (16.5 ft and sometimes 13 ft)
when they flooded, analyzed along with the more conventional well samples and included in the
groundwater sampling data discussed in this report.
Starting on November 9, 2012, a Solinst water Levelogger Model 3001 was used to obtain higher time
resolution, with data recorded each half hour. The device was installed in the deepest well (MW1A). The
higher resolution of this instrument enabled us to confirm and model the strong and rapid connection
between surface water levels in Fall Creek and groundwater levels at the duplex. The strong correlation
between groundwater levels and Fall Creek gauge heights enabled the use of the historic Fall Creek
USGS gauge height data to accurately hind cast groundwater levels for the earlier part of the project at a
much finer (daily) resolution than was available before. This empirical model was used to create a
detailed time series of water table elevations that was used as an independent variable in the statistical
analyses discussed in Sections 9 and 10 of this report and compared with groundwater VOC concentration
trends in this section. Figure 11-2 shows the daily time series of actual and predicted groundwater levels
from the model.
•
1 V
* •
• * •
t ~
* ***
• •
* V
*£**
** *
X
:
*•
* *
•
••
A,
•• v *j •
. u V:
v> V:
W '
: < l?'1'
V * . i
U : i
;
•
Actual
Predicted
&
&
&
Figure 11-2. Actual groundwater levels along with the daily time series predicted from Fall
Creek gage height data.
11.2 Groundwater Concentration Trends
During initial screening conducted in late spring 2010 at the site (U.S. EPA, 2012a), groundwater had
detectable but low concentrations of PCE and chloroform, with PCE concentrations ranging from 0.46 to
0.61 ng/L and chloroform levels ranging from 1.9 to 3.0 ng/L. Although these data were subject to
qualifiers related to the very low levels of analytes in the samples, the analyst believed that the analytes
were present. In addition, a June 2005 groundwater sampling event associated with the nearby Mapleton-
Fall Creek brownfields site found detectable chloroform (8.9 to 22.1 j^ig/L) in groundwater. However,
initial groundwater sampling at the duplex did not find comparable chloroform levels in groundwater,
with only nondetects being seen in the earlier phase of this project (U.S. EPA, 2012a). The detectable
chloroform concentrations measured in late 2012 and early 2013 were also lower than those observed
11-2
-------
Section 11—Results and Discussion: Do Groundwater Concentrations Control Soil Gas
Concentrations at this Site? And Thus Indoor Air Concentrations?
during screening. However, PCE levels have remained fairly constant from spring 2010 screening through
the spring 2013 measurements presented in this report.
As described in Section 3, groundwater was sampled for VOC analysis approximately monthly during the
active project. One hundred eighteen groundwater samples were collected and analyzed over the two
project phases from six monitoring wells (two 3-well clusters) and from the flooded soil gas ports. Table
11-1 shows the sampling locations, screened interval, and number of samples collected from each well
and from the soil gas ports.
Table 11-1. Groundwater Monitoring Locations
Screened Interval
No
1 Well ID
Depth (ft bis)
Measurements
Location
MW1A
24-26
9
Exterior South
MW1B
21-24
14
Exterior South
MW1C
16-21
15
Exterior South
MW2A
24-26
14
Exterior North
MW2B
21-24
14
Exterior North
MW2C
16-21
13
Exterior North
MW3
19.5-24.5
12
422 Basement
SGPGW
points1
13-16.5
27
Various
1flooded soil gas ports
Figure 11-3 shows the groundwater concentrations of PCE and chloroform over both phases of the
project. Nondetect samples were not plotted for the first phase of this project because the detection limits
for PCE (2.8 ng/L) and chloroform (2.0 ng/L) were considerably higher than the detectable
concentrations in the figure. As a result of efforts to improve the detection limits and instrument
sensitivity at EPA NERL, chloroform was detected beginning in late 2012.
Chloroform concentrations were not detectable until late March through mid-April 2013, when two
samples showed detectable concentrations around 0.6 ng/L. Detectable mid-April chloroform
concentrations otherwise ranged from 0.07 to 0.37 ng/L. The reason for the difference between the two
high samples and the rest of the chloroform detections is not clear at this point in time. Overall, the
chloroform groundwater data showed 99 nondetects, or 84% of the 118 groundwater measurements.
There were 19 detections and 21 nondetects in the 2013 data for a 48% detection rate.
For PCE, groundwater concentrations were consistent with screening and fairly stable over the course of
the project, with almost all detectable concentrations ranging between 0.2 and 0.8 ng/L. A single higher
concentration event occurred in September-October 2011 when PCE concentrations in some wells ranged
up to 1.3 ng/L. Figure 11-4 plots the PCE concentrations by well so that one can see the fairly consistent
relationship between the wells during a particular sampling event. Note that the groundwater
concentrations taken from the soil gas ports generally plot well within the range of the more conventional
monitoring well samples. Overall, there were 50 nondetects for PCE for a 42% detection rate. In 2013,
PCE was detected in 13 of 40 samples, for a 33% detection rate.
11-3
-------
Section 11—Results and Discussion: Do Groundwater Concentrations Control Soil Gas
Concentrations at this Site? And Thus Indoor Air Concentrations?
~ PCE ¦Chloroform
Figure 11 -3. Groundwater concentrations over time for Indianapolis duplex.
to
E
o
~
A
~
A
~
~~
o
~ o
~
A
t> •
81
A
A
~
m
~
oo
A
~
4
~ A
£ ~
§
A A
A
~
A
n
o
~
A
O
'L
o
~
O
o
0 ~
o o
CO
o
"O
¦ MW1A
OMW2A nd
~ MW1AND
-MW2B
AMW1B
O MW2B nd
% ^
V
AMWIBnd
~ MW2C
~ MW1C
MW2C rid
U Vb
% -
~
OMWICnd
AMW3
OMW2A
OSGP GW samp
Figure 11 -4. PCE groundwater concentrations over time, showing concentrations by individual
well and soil gas ports.
11-4
-------
Section 11—Results and Discussion: Do Groundwater Concentrations Control Soil Gas
Concentrations at this Site? And Thus Indoor Air Concentrations?
11.2.1 Is the Groundwater Concentration Trend Correlated to Well or Water Table
Depth?
Figure 11-5 shows the relationship between monitoring well screen depth (or soil gas port depth) and
PCE and chloroform groundwater concentration. As the figure shows, there does not appear to be a strong
correlation, although the higher concentrations are associated with the 16-ft, 16.5-ft, and 19-ft well screen
depths. (Note that the well screen depths discussed here are at the top of the screen.) This supports the
hypotheses that this site lacks a "clean water lens" and that the shallowest groundwater contains VOCs
available to partition into the soil gas.
1.4
1.2
tXQ
C
o
0.8
C
QJ
u
c
O
u
u
CL 0.6
i_
0>
4—1
ro
5
-o
§ 0.4
15
~F
1
0.2
H
il
~
I
t
~
~
*!
10
12
14 16 18 20
Well Depth (feet below landsurface)
22
24
~ PCE ¦ Chloroform
Figure 11 -5. Plot of groundwater PCE and chloroform concentrations against well screen (or
soil gas port) depth (well depth measured to top of screen).
Figure 11-6 plots the groundwater PCE and chloroform concentration against the water level measured or
estimated for the day the groundwater sample was taken. Again, a strong correlation is not visually
apparent, but the PCE high concentrations do correspond with the low groundwater levels observed in
September-October 2011. This could be related to the extremely rapid groundwater level decrease shown
for this time period in Figure 4-2. Possible conceptual models consistent with this finding include:
• When freshwater comes into the aquifer, it is generally low concentration. Over time, the
freshwater mixes with existing water and picks up VOCs at a rate faster than it can give them off
through volatilization. Thus, the maximum concentration is seen when little freshwater has come
in recently. OR
11-5
-------
Section 11—Results and Discussion: Do Groundwater Concentrations Control Soil Gas
Concentrations at this Site? And Thus Indoor Air Concentrations?
Q. 0.6
12 13 14 15 16
Groundwater Level (feet below landsurface, ft bgl)
~ PCE Mean PCE ¦ Chloroform Mean Chloroform
Figure 11-6. Plot of groundwater PCE and chloroform concentrations against ground water
depth.
• Figure 11-6. Plot of groundwater PCE and chloroform concentrations against groundwater
depth.The well screens are composed as most screens are by materials of varying conductivity. At
high water levels in the well, the screens are mostly or completely submerged. The water in the
PDB then preferentially comes from the high conductivity parts of the formation within the well
screen. At low water levels in the well, it may be that all the water in the well is coming from a
lower flow part of the formation. Formation layers with higher hydraulic conductivities can be
lower in concentration than the layers with lower hydraulic conductivities because finer grained
materials (e.g., clays, organic matter) tend to have higher sorption affinities for VOCs than
coarser sand and gravels.
11.3 Revision to the Conceptual Model—Is the Groundwater Concentration
Related to Soil Gas and Indoor Air Concentrations?
The monthly or longer sampling intervals make it very difficult to quantitatively assess the correlation of
groundwater PCE against soil gas or indoor air concentrations, although the narrow range of variability in
PCE concentrations (below an order of magnitude) and stability of this variability overtime (see Figures
11-3 and 11-4) make it unlikely that changes in groundwater concentrations are strongly related to the
changes in soil gas or indoor air concentrations over the time scales observed in this study. Chloroform's
limited length of record for detectable concentrations in groundwater severely limits what can be assessed
in that regard.
11-6
-------
Section 11—Results and Discussion: Do Groundwater Concentrations Control Soil Gas
Concentrations at this Site? And Thus Indoor Air Concentrations?
However, given the groundwater concentrations are relatively stable under the site, there are sufficient
data to evaluate the potential for groundwater to be the source of indoor air concentrations in the 420/422
duplex on the basis of Henry's law calculations. For PCE, the prevalent range of groundwater
concentrations, from 0.2 to 0.8 ng/L, and mean of 0.4 ng/L, corresponds to vapor concentrations of 175,
579, and 275 ng/m3, respectively, with the maximum groundwater concentration, 1.3 ng/L, corresponding
to 941 ng/m3. These concentrations are probably sufficient to produce the soil gas and indoor air
concentrations observed in this study, especially considering the coarse-grained nature of the subsurface
that allow ready diffusion and flow of contaminants from the water table to the building. Thus, the
available groundwater data indicate that groundwater is a likely source of the PCE vapors observed in the
subsurface and indoor air during this study, although additional vadose zone PCE sources cannot be ruled
out based on the groundwater evidence.
As mentioned above, although the groundwater PCE concentrations are sufficient to be the primary
source of the PCE measured in indoor air in this study, their observed variability does not explain the
variability in the indoor air PCE measurements. As has been shown in this and other studies, the
variability in indoor air PCE concentrations is also influenced by subsurface, building-related, and
meteorological variables that affect the concentration of PCE as it migrates from the water table, enters
the building, and mixes with indoor air. In addition, other sources of PCE that may also exist in the
vadose zone or sewer lines cannot be ruled out at this point, and variability in those sources could also
influence the PCE concentrations in indoor air. We will continue to test this assertion (that groundwater is
the primary source of PCE vapors) as additional data are collected and analyzed at this site.
The same VOC source situation may not hold true for chloroform. Most of the measured concentrations
of chloroform in groundwater are nondetects, with the maximum groundwater concentration (0.64 ng/L)
corresponding to a vapor concentration of 63 ng/m3, and the mean (0.12 ng/L) corresponding to 12 ng/m3.
These groundwater concentrations are not sufficient to drive the soil gas and indoor air concentrations
measured in this study, indicating that other sources, such as vadose zone sources from nearby former
businesses using chloroform or disinfection by-products from leaky water mains, may be responsible for
the observed peak soil gas and indoor air chloroform concentrations.
11-7
-------
Section 11—Results and Discussion: Do Groundwater Concentrations Control Soil Gas
Concentrations at this Site? And Thus Indoor Air Concentrations?
11-8
-------
Section 12—Results and Discussion: Special Studies
Table of Contents
12.0 Results and Discussion: Special Studies 12-1
12.1 Summary of Geophysics Study 12-1
12.2 Summary of Tracer Testing 12-3
12.2.1 Introduction to Tracer Testing 12-3
12.2.2 Tracer Testing Objective 12-3
12.2.3 Tracer Test Experimental Methods 12-3
12.2.4 Tracer Test Results and Discussion 12-4
12.2.5 Summary of Tracer Test Conclusions 12-15
12.3 Testing Utility of Consumer Grade Radon Device (Safety Siren Pro) 12-15
12.3.1 Introduction to the Use of Consumer Grade Radon Monitoring Equipment
in Vapor Intrusion 12-15
12.3.2 Objective of Consumer-Grade Radon Device Testing 12-15
12.3.3 Methods of Consumer-Grade Radon Device Testing 12-15
12.3.4 Consumer Grade Radon Detector Results and Discussion 12-16
List of Figures
12-1. Response at all locations to first helium injection in front yard 12-5
12-2. Response at all locations to second helium injection in front yard 12-6
12-3. Response at all locations to third helium injection in backyard 12-6
12-4. Response at all locations to fourth helium injection in backyard 12-7
12-5. Helium response at SGP2 cluster (south of duplex) directly above injection point after
injections 1 (top graph, mitigation off) and 2 (bottom graph, mitigation on) 12-8
12-6. Helium response at SGP6 cluster at SGP6 cluster (north of duplex) directly above
injection point after injections 3 (top graph, mitigation off) and 4 (bottom graph,
mitigation on) 12-9
12-7. Helium response at SGP1 cluster (south of duplex, approximately 6 ft closer to duplex
than injection point) after first Injection (top graph, mitigation off) and second injection
(bottom graph, mitigation on) 12-10
12-8. Helium response at SGP5 cluster (north of duplex, approximately 6 ft closer to duplex
than injection point) after third injection (top graph, mitigation off) and fourth injection
(bottom graph, mitigation on) 12-11
12-9. Helium response to first and second helium injections at SGP9 (interior) 12-12
12-10. Helium response to third and fourth helium injections at SGP10 (interior) 12-12
12-11. Cross-section view of helium tracer arrival at 6-ft depth intervals 12-14
12-12. Comparison of electret and Safety Siren results for first phase of the project (top graph)
and second project phase (bottom graph) 12-17
12-13. Comparison between the 422 office Safety Siren and electret 12-22
12-14. Comparison between the 422 basement N Safety Siren and electret 12-22
12-i
-------
Section 12—Results and Discussion: Special Studies
List of Tables
12-1. Data Quality Objectives and Performance/Acceptance Criteria for Special Studies 12-1
12-2. Comparison of Safety Siren, AlphaGUARD, and Electret Radon Data 12-18
12-3. Comparison of Safety Siren and Electret Radon Data 12-20
12-ii
-------
Section 12—Results and Discussion: Special Studies
12.0 Results and Discussion: Special Studies
We conducted five special studies in the Indianapolis duplex. These studies either go beyond the
objectives articulated in Section 2 to address specific EPA research needs or address the primary project
objectives using different tools. Two of these special studies were reported completely in our previous
report and are not discussed in detail here:
¦ VOC measurement efficiency of temporary subslab ports as compared with permanent subslab
port constructions (Section 12.1 and Appendix B of U.S. EPA [2012a])
¦ Use of box fans to induce flow into the structure in an attempt to create "worst case" conditions
for vapor intrusion (Section 12.2 of U.S. EPA [2012a])
In this section, we present methods, results, and discussion for three special studies:
¦ An EPA ORD study of this duplex and its subsurface environs using geophysical methods
(Section 12.1)
¦ Helium tracer testing in which helium was injected into deep soil gas ports and its migration
observed. (Section 12.2)
¦ Continued testing of a consumer-grade radon detector as an indicator of vapor intrusion,
providing additional data collected since U.S. EPA (2012a) (Section 12.3)
Data quality objectives and criteria for these studies are described in Table 12-1.
12.1 Summary of Geophysics Study
In August 2012, the EPA National Exposure Research Laboratory (NERL) in Las Vegas conducted
magnetic, electromagnetic induction, and ground penetrating radar (GPR) surveys within the basement
and surrounding exterior of the Indianapolis study duplex (U.S. EPA, in press). The exterior surveys
included the house property, East 28th Street to the south, the adjacent alley and property to the west,
and the property to the north and northeast. The interior surveys were performed in the basement of both
the 420 and 422 residences. The objectives of these surveys were to locate anthropogenic objects and
identify subsurface conditions that may influence subsurface vapor flow to aid in the conceptual site
model refinement.
The exterior magnetic and EMI survey identified known surface objects and known utilities as well as
likely utilities. The exterior 200 MHz and 500 MHz GPR results suggested an electrically conductive
unsaturated zone with irregular hummocky layers overlaying a more regular horizontally stratified zone
at approximately 7.5 ft deep, which likely corresponds to the facies contrast between the underlying
sand/gravel layer and the shallower silt/clay layer. This contact may also form local perched water
conditions that may also cause such a geophysical response. These stratigraphic features may influence
subsurface fluid fate and transport. A large hyperbolic reflection was observed in the middle of the
backyard and about 10 ft north of the house. The source of this object is unknown and should be
investigated. A sloping feature near the back wall of the house may be a backfill horizon or perhaps old
brick stairs formerly used for external access to the basements. Additional observed features are
described in U.S. EPA (in press).
12-1
-------
Section 12—Results and Discussion: Special Studies
Table 12-1. Data Quality Objectives and Performance/Acceptance Criteria for Special Studies
Measurement Performance or
Acceptance Criteria
Task Order Objectives
Study Questions
Measurements Used
Better define the
subsurface conditions
that influence the
movement of VOCs and
radon into the study
duplex.
What is the nature of
the subsurface
environment? What
stratigraphic and
human-made features
influence contaminant
transport?
Geophysical techniques (i.e.,
GPR, magnetic, and
electromagnetic induction),
Helium tracer testing.
Acquire geophysical data from
multiple transects within each
third of the basement.
Perform at least two helium
tracer injections with monitoring
on 1-hour frequency at five or
more locations.
Evaluate the ability of a
low-cost consumer-grade
radon detector (Safety
Siren Pro Series 3,
Family Safety Products
Inc.) to provide
continuous indication of
soil gas entry into the
structure.
Does the measurement
of radon concentration
using this consumer-
based analyzer agree
within +/- 30% to the
readings from the
electret and
AlphaGUARD methods
>90% of the time?
The Safety Siren has two
displays—the "short term" is an
average over the previous 7
days, and the "long term" is the
average from time of last reset
(up to 5 years). Readings are
available after a minimum of 48
hours of operation. Record the
short-term reading at each of six
indoor stations weekly and
compare with ongoing electret
and AlphaGUARD
measurements.
Conduct tests at six stations:
422 south basement
(AlphaGUARD), 422 north
basement, 420 south basement,
422 first floor (AlphaGUARD),
422 second floor, and 420 first
floor. Add an electret
measurement during the Safety
Siren test period at 422 second
floor for additional comparison.
At six stations, compare Safety
Siren results with weekly
electrets. Collect at least 4
weeks of comparative data = 28
pairs of data points.
At two stations, compare the
Safety Siren against weekly
averages of hourly
AlphaGUARD data, providing at
least 8 additional pairs of data
points.
Does the consumer-
grade radon detector
provide a useful
indication of the weekly
average infiltration of
VOC containing soil
gas?
Month-long correlation test
between consumer-grade radon
detector and other radon
detectors as discussed above.
Year-long data set on
radon/VOC correlation in this
house.
Safety Siren adequately
correlates with the electrets (see
above) and radon correlates to
VOCs in the main study data set.
The interior GPR results suggest the concrete slab varies from 0.5 to 0.7 ft in thickness with an irregular
undulating contact with the underlying material. This underlying material is electrically conductive and
contains many discontinuities and non\horizontal interfaces, which may govern subsurface fluid fate and
transport. This suggests the concrete floor was not poured on well-flattened fill material and may provide
pockets into which soil gas may pool or move preferentially. However, the presence of fill material of
some sort is evident at many locations beneath the slab. One prominent area of fill lies to the west of the
cistern in the 420 basement. This may be consistent with a construction practice in which cuts and fills
were incompletely done prior to pouring the slab. The GPR data suggest there is a thicker concrete
footing below the doorways that separate the rooms in the basement. The GPR response to the cistern and
the covered sewer line is evident as are GPR reflections likely due to fill material or other features related
to the cistern. Other anomalous GPR reflections are detailed in U.S. EPA (in press).
12-2
-------
Section 12—Results and Discussion: Special Studies
12.2 Summary of Tracer Testing
12.2.1 Introduction to Tracer Testing
In addition to geophysical imaging and stratigraphic visualization based on soil logs, an additional line of
evidence for understanding subsurface conditions is tracer testing. Tracer testing uses injected gases, like
helium, to help determine subsurface air permeability, possible vapor entry pathways into a structure, and
potential preferential pathways for vapor migration.
12.2.2 Tracer Testing Objective
During the first phase of this project, three-dimensional characterization and visualization was done on
the 422/420 house based on survey maps, soil data, VOC, and radon data. In autumn 2012, geophysical
work was performed on the 422/420 house as an additional line of evidence of subsurface conditions in
the near- and sub-building environments. A helium tracer test can add to the already acquired evidence by
suggesting flow pathways and points of ingress into the 422/420 house. Additionally, tracer testing, by
sampling proximally related points, can describe how subsurface conditions change from north to south in
the house environment. The following is a quote from the study objectives in U.S. EPA (2012a):
A helium tracer test will also be used in order to gather more information about the
conditions beneath the 422/420 house and its immediate surroundings, as well as sewer
outflows, utility corridors, and public utility lines (ifpossible). The sample plan will also
incorporate a way of determining the difference in the dynamics of collection and motion
of vapors and gases between the 422 and 420 sides of the house and the changes between
the north (backyard), central (house basement), and south (front yard) sections of the
house.
Helium was selected as a tracer because it was easily obtained and measured and would meet the above
objectives. And although helium diffuses much faster than chlorinated VOCs, the coarse (sand and
gravel) subsurface beneath the home makes this less of an issue because the system is probably not
diffusion driven.
12.2.3 Tracer Test Experimental Methods
Although tracer testing occurred on four separate occasions, the basic structure of the tests remained
essentially the same: a measureable volume of pure industrial helium was injected into a 13-ft deep soil
gas location and then a handheld helium detector was used to make observations at a preselected series of
subslab and soil-gas ports. The 13-ft depth was selected for injection because it is the deepest depth in the
nested soil gas points that is normally above the water table. Readings were taken over a period of time,
frequently at first, and then with greater temporal spacing as helium concentrations fell to 0 ppm.
Injections were done by a Cole-Parmer rotameter connected to a helium tank and consisted of 18.6 L of
helium for injections 1 and 2, 23.25 L for injection 3, and approximately 26 L for injection 4. The first
two tests focused on the nested soil gas ports (SGP1, SGP2, and SGP9), WP-3, and SSP-1 and -4; soil gas
ports located in the front yard and several soil gas ports, sub-slab ports, and wall ports located within the
422/420 house proximal to the front yard ports. SGP2-13 was used as the injection port.
Tests 3 and 4 switched to the backyard locations, using a series of exterior soil gas ports and proximally
related interior soil gas ports, wall ports, and sub-slab ports. SGP6-13 was the injection point in tests 3
and 4. These tests were monitored at SGP5; SGP6; SGP10; SGP8-6; SSP-1, -2, and -6; and WP-1. The
sample locations are shown in plan view in Section 3 and in cross section in Figure 12-11.
12-3
-------
Section 12—Results and Discussion: Special Studies
Readings were taken during all four tests with a rented Dielectric MGD-2002 helium detector. However,
during the first test, the first MGD-2002 rented suffered from battery and filter issues, so it was returned.
Another MGD-2002 was rented immediately from a different rental firm, and all subsequent rentals of the
MGD-2002 were from the second firm. After injecting the helium, readings were taken rapidly in order to
record the initial stages of helium dispersal. Over the following day or two, the readings would be taken
further apart until measurements dropped to 0 ppm.
Test 1 occurred on October 13, 2012, to October 17, 2012. Test 2 occurred on October 17, 2012, to
October 22, 2012. Test 3 occurred on November 19, 2012, to November 21, 2012. Test 4 occurred on
April 10, 2013, to April 12, 2013. The mitigation system was on during tests 2 and 4, and in passive mode
during test 3, and off except for the last two readings during test 1 (the full mitigation system was not
running until after all of test 1 was complete).
Given that the geologic material at the 13-ft depth is relatively coarse, it was expected that these
injections would generate a brief pulse of advective pressure-driven migration that would then dissipate.
Helium would be expected to migrate upward based on its lower gas density than air and due to the
driving forces for VOC vapor intrusion such as the stack effect.
12.2.4 Tracer Test Results and Discussion
Figures 12-1 through 12-4 show the combined results of each of the helium injections plotted versus the
time from injection in days. Tests that occurred during mitigation system on cycles (tests 2 and 4) had
peak concentrations at monitoring points vertically separated from the injection point, approximately
1,800 ppm to 2,500 ppm; however, those that occurred during mitigation system off or passive cycles had
concentrations ranging from approximately 3,500 ppm to 4,000 ppm. Recall that our differential pressure
measurements show that when the mitigation system is on the deep to shallow soil gas differential
pressure is enhanced, pulling soil gas toward the building (Section 5). Thus, it is likely that the mitigation
system pulled the injected gasses away from the point of injection (at 13 ft outside the footprint of the
house) and helped dilute the tracer in the immediate vicinity of the injection point.
The paired tests have similar patterns for the same injection location (with and without mitigation) when
comparing the exterior soil gas ports immediately above the injection point (SGP2 and SGP6 clusters,
respectively; see Figures 12-5 and 12-6). There is considerable similarity in paired tests at the exterior
soil gas clusters that are approximately 6 ft closer to the house than the injection points (SGP1 and SGP5
clusters, respectively; see Figures 12-7 and 12-8).
The tests with the injection point in the front yard can be seen in Figures 12-1 and 12-2 (tests 1 and 2,
respectively). Test 1 had its high concentrations of helium at the monitoring points by the end of 24
hours, but the helium was nearly gone after -48 hours. Test 2 had its highest concentration by the end of
24 hours and showed very low helium concentrations 48 hours later.
Test 3 had its high helium shortly after injection, was low but steady by the end of 24 hours, and roughly
steady at 48 hours (Figure 12-3). Test 4 showed roughly the same pattern as test 3 (Figure 12-4).
In general, the comparisons of these paired tests at the exterior monitoring ports suggest that the
mitigation system had a relatively limited effect on the speed of tracer migration. Most often, among all
the tests, the ports that showed the highest helium concentrations were the deepest, such as the 9-ft and
13-ft depths. This suggests attenuation during vertical migration by dispersion or dilution. Attenuation
was especially pronounced in the 3.5 ft bis soil gas ports, even directly above the injection point. The
similarity between paired tests performed in common locations suggests that tracer migration from deep
soil gas zones is controlled primarily by localized differences in soil types and/or building structure more
than by the mitigation system status.
12-4
-------
Section 12—Results and Discussion: Special Studies
Figures 12-9 and 12-10 focus on two of the sets of nested soil gas ports located under the building and
tell a slightly different story from the exterior soil gas points. Figure 12-9 shows data from SGP9 (a
cluster under the southern portion of the 422 side of the duplex) from the first two injections. Based on its
location, it would be expected to be the first soil gas cluster to detect migration from the front yard deep
injection point beneath the structure. The data from the first injection (mitigation system off) is on the
bottom half of the graph, and the top half plots the second injection data (mitigation system on). The first
injection data for SGP9 shows a peak nearly at the time of the injection, which almost immediately drops
to 0 ppm. This finding would suggest a very rapid migration propelled by the advective force of the gas
injection. However, the second injection shows a peak after 24 hours, with only other minor fluctuations.
Also note the helium detection that occurs right before the second injection in SGP9. The explanation for
this detection is not clear at this time, but its timing could represent a delayed migration of helium from
the first injection to SGP9 or carryover from another nearby injection point.
SGP10 is located under the north side of the 422 duplex and is expected to be the first soil gas cluster to
detect migration from the exterior injection point in the backyard beneath the structure. Figure 12-10
plots SGP10 during the third and fourth injections, the bottom half of the graph showing the third
injection (mitigation system off) and the top showing the fourth (mitigation system on). There is
considerable fluctuation for the third injection, with helium concentrations at some points rising and
falling (SGP10-9 and -13). During test 4, there are initial peaks roughly at the time of injection and then
another after -24 hours. This could indicate a difference between initial injection pressure-driven flow
and subsequent buoyant or stack effect flow. The arrival time at the SGP10 cluster appears to be
somewhat delayed with the mitigation off vs. mitigation on. The mitigation on results show much less
observable concentration at the shallow depths (6 and 9 ft) in this case; this result differs from the results
discussed in the previous paragraph from the injection at SGP9. However, SGP10 is similar to SGP9 in
that there were elevated helium concentrations prior to the injection time.
First Helium Injection: All Locations
jouo
E
a.
a ?ox)
-------
Section 12—Results and Discussion: Special Studies
Second Helium Injection: All Locations
2500
20Q0
?
Q»
CL I WO
Q}
X
1000
SCO
O
©
> $
G 1
w U
J *
S §3 B B
-1.00
0 00
4 CO
SGW-33
5GP2 6
SGPl 6
sai>i 13
WP-3
£SP 1
SGP9fc
5GP9 9
n
Ambient
422 e^s
422 H jw N
¦ Inprilwm Po*i|
s.oo
Time from injection (Days)
Figure 12-2. Response at all locations to second helium injection in front yard.
Third Helium Injection: All Locations
MOO
WOO
JSOO
E
a
CL MOO
01
I Wo
1030
MM
o
0
o
o
% e
• O 0 § 3 a
So : 8 e 8 Is
r> «
o.se
aao
aw
1.00
I 50
?.a>
0
5GP6-9
•
SGP6 6
•
SGP6-1.5
0
SGP5 13
•
5&PS9
•
•
SGP10 6
0
HPIO-Q
o
SGPH013
&
SGfPB'G
0
5SP-J
•
$5P 1
o
SSP-6
WP 1
Ambient
422 fljse N
4« BJMfC
*
422 8.35*5
Wear WP-l
'lfiredkNi> Pari?
Time from Injection (Days)
Figure 12-3. Response at all locations to third helium injection in backyard.
12-6
-------
Section 12—Results and Discussion: Special Studies
Fourth Helium injection: All Locations
!MO
I MO
E 1030
CL
CL
800
X
4-CO
2" CO
SGP6-13
t
0
o
0
a i
*. o » P o I?.
too
4)S0
000
0 50
1.00
1-50
2.00
0
S6P6-1J
O
SGP6-9
•
•
5GP6-3.S
&
5GP5 13
•
SGP5-9
•
•
SGP10 6
•
5&P109
0
Q
SGP9--&
O
SSP-2
•
56P1
0
0
WP 1
Ambient
412 B-ase N
*22 Base C
422 Base S
Nrar WP-I
lajctlMm Po^il
Time from Injection (Days)
Figure 12-4. Response at all locations to fourth helium injection in backyard.
12-7
-------
Section 12—Results and Discussion: Special Studies
First Helium injection (mitigation system off): 5GP2
KO0
JSOO
£
Q.
Cl him
W&
0.00 0-50
~
.00
-J .00 -0.50
Time from Injection (Days)
130
d 5GPJ -15
A SGP?-6
O SGPi«J
injection Pwm
A
OS
}.00
js)
Second Hetium Injection (mitigation system on): SGP2
3000
?KK»
2000
E
OL
&
X
o
A
o
i 4
J
iQfi \
| EE a 1
O S6P2S-S
it SGP1-6
o SGP2-9
— Infection Poifrt
Time from Injection {Days)
Figure 12-5. Helium response at SGP2 cluster (south of duplex) directly above injection point
after injections 1 (top graph, mitigation off) and 2 (bottom graph, mitigation on).
12-8
-------
Section 12—Results and Discussion: Special Studies
4. MO
3500
mo
mo
E
&
Q_ 1. roo
-------
Section 12—Results and Discussion: Special Studies
103
!»
10)
E
a
£L go
.50
Time from Injection (Days)
1-50
2dI
~ 5GPI-6
A WW
o sen-u
— mtcctlM p»nd
130
Second Helium Injection (mitigation system on): SGP1
TOO
bOO
soo
£ 409
a
a.
-------
Section 12—Results and Discussion: Special Studies
Third Helium Injection {mitigation system off): SGPS
*50
<1 DO
350
"e ^
a
a 2w
X *00
ISO
100
-050
O O
~
~
~ B
0-00 O.SO 1.00
Time from Injection (Days)
1.50
A
o ~
~ S6PS-6
4 SGP5-9
© 5GPS-13
ifijearofl Pcmwi
100
Fourth Helium Injection: SGPS
JOO
J50
?
a
EL TO
<3>
X
150
100
so
9
/ °
hd • • •
o
»-5&P5-6
»- SWS9
K&GPS.tl
—ftijrtf km Pt»i nt
-0J50
000
0.^0
1.00
1.50
2.00
Time from Injection (Days)
Figure 12-8. Helium response at SGPS cluster (north of duplex, approximately 6 ft closer to
duplex than injection point) after third injection (top graph, mitigation off) and
fourth injection (bottom graph, mitigation on).
12-11
-------
Section 12—Results and Discussion: Special Studies
E
Q.
a
0>
o
First (mitigation system off) and Second (mitigation system on)
Helium Injections: SGP9
'
aoo
300
a
.2. loo
B
-2.00
£Z3
~
A
-1.00
n
I. DC
m
J.00
Time from Injection (days)
fiT
N>
~
Ol
rt>
n
rt
3
X
w
13
£
Q S6P9-6LH
inaction
A 9 1*3
In jeer to*
o sara-uiit
Injection
Q SGP9£Jnd
Injvrtfetfi
A SGP9-9 2nd
Iflpetttotfi
0 WP^I 3 2nd
inject lw
— irtJtKKJofl Poll*
Figure 12-9. Helium response to first and second helium injections at SGP9 (interior).
Third (mitigation system off) and Fourth (mitigation system on}
E
Q.
CL
&
4th hjncton
$GP10*
4th Injoctiqfi
SGP10-H
= Mrtbfi MW
1 SO
3 0(J
Time from Injection (days)
Figure 12-10. Helium response to third and fourth helium injections at SGP10 (interior).
12-12
-------
Section 12—Results and Discussion: Special Studies
These differences in the time profiles observed at SGP9 and SGP10 within the test groups performed in
common locations must be controlled by a factor other than just soil differences or building structure. The
first two tests were performed in close sequence, but a considerable amount of time passed between the
third and fourth tests, so for the third and fourth tests, differences in soil moisture cannot be ruled out as
an explanation for differences in tracer behavior.
The first three tests showed fleeting but moderately strong detections of helium tracer at the wall ports
located closest to the injection locations (WP-3 and WP-1) within the first day after injection. These wall
ports are approximately 10 ft above and 10 ft laterally from the injection location. This suggests rapid
migration to the building envelope is possible from near the water table if the buoyant and advective
driving forces are sufficiently strong.
Observation of tracer arrival at the points immediately below the slab—either conventional subslab ports
or the shallow 6-ft intervals of the nested soil gas probes was highly erratic (Figure 12-11). Note, for
example, in tests 1 and 2 how some tracer was seen at both SGP9-6 and SSP-1 but not at SSP-4, which
lies closer to the point of injection than SSP-1 and quite near SGP9-6. Similarly, in test 1, no tracer was
observed at SGP1-6 and SGP1-9.
12-13
-------
Section 12—Results and Discussion: Special Studies
Helium Tests House Cross-section View
Test 1 (mitigation system off)
N
w
WP-3VV
j>?p^ND
5
ND
ND
S
ND
W
SGP2
S6P1 yy
SGP9W
House Basement
in All Four Diagrams
Test 2 (mitigation system on)
W
WP-lVV
jjfMND jjgp-iVV
s
ND
W
s
W
ND
SCP2
SGP1 S
SGP9 §
Test 3 (mitigation system off)
ND
fcjpi SSPG (420 side)
roUspp-1
V
SGPB-6 W
W
w
w
W wp-i
JSP-2 W
¦S6P10 SOPS-
w
w
w
Test 4 (mitigation system on)
W
W
S
SGP6
Key
S=>500ppm
W=<500ppm
ND=non detect
Kin ND
N D sspe (4Zflside)
\
W e
JjSP-2
WP-1
ND
w
SSPB-eND fsID
W
W
w
W
s
W
-5GP10 SGP5-
W
SGP6
Note; soil gas port depths are 3.5', 6', 9", and 13" bgs, except the indoor locations, which
omit the 3.5* depths, SGP8-6 is truncated in this figure, since only its 6' depth was sampled
during the helium tests.
Figure 12-11. Cross-section view of helium tracer arrival at 6-ft depth intervals.
12-14
-------
Section 12—Results and Discussion: Special Studies
12.2.5 Summary of Tracer Test Conclusions
The four tracer tests provide more information about the subsurface conditions surrounding the 422/420
house. The tests performed with common injection locations yielded similar overall patterns of tracer
distribution with and without mitigation in the exterior soil gas clusters. The variability between paired
tests (mitigation on and mitigation off) was more pronounced beneath the building where the mitigation
system would be expected to have the most significant influence on airflow. The highest concentrations
observed were usually those close to the 13-ft deep injection depth, suggesting attenuation as the tracer
migrated upward. The similar patterns between tests performed in different subsurface areas (different
injection wells) suggest control by common features of soil stratigraphy or the building envelope because
the tests would have these factors in common. The high concentrations observed at the 9-ft and 13-ft
depths for all the injections, despite the buoyancy of the helium-enriched soil gas, suggest that
transference of helium is easy horizontally toward the building over distances of up to 20 ft typically
within 2 days. Vertical migration from 13 ft to 6 ft bis also occurred relatively rapidly at the injection
cluster. The lack of influence at certain ports closer to the point of injection and along the same general
line to more distant ports suggests subsurface heterogeneity and preferential flow paths. In all cases,
tracer concentrations were reduced to baseline concentrations within 2-2.5 days, indicating a rather quick
clearing out of trace gas from the subsurface under mitigated and non-mitigated conditions. In two
interior probes (SGP9 and SGP10), relatively high concentrations before and after injection at several
sample ports suggest either carry-over from previous injections or flow of helium from other nearby
injection ports.
12.3 Testing Utility of Consumer Grade Radon Device (Safety Siren Pro)
12.3.1 Introduction to the Use of Consumer Grade Radon Monitoring Equipment in
Vapor Intrusion
Schuver and Siegel (2011) have highlighted the role of radon as a potential "general tracer of soil-gas
entry" and pointed out that there are multiple benefits from minimizing soil gas entry (including
reductions in problems attributable to moisture/mold, radon, and methane as well as reduction in VOCs).
They also advocated the active involvement of homeowners in observing the building-specific aspects of
vapor intrusion at VOC sites, both as an educational tool (to help homeowners understand temporal and
spatial variability) and a way to an efficient solution to which all stakeholders agree.
To address this need we evaluated the ability of a widely available low-cost ($129) consumer-grade radon
detector based on an ionization chamber, the Safety Siren Pro Series 3 manufactured by Family Safety
Products Inc., to provide a continuous indication of soil gas entry into the structure.
12.3.2 Objective of Consumer-Grade Radon Device Testing
The objective of this research was to evaluate the ability of a widely available low cost ($129) consumer
grade radon detector based on an ionization chamber to provide a continuous indication of soil gas entry
into the structure.
12.3.3 Methods of Consumer-Grade Radon Device Testing
The Safety Siren Pro Series 3 is a consumer-grade radon detection instrument that provides continuous
real-time measurement based on an ionization chamber and requires little operator labor. In this test, we
sought to compare the performance of the Safety Siren to a we 11-accepted method (electrets). Secondarily,
we were able to compare it with the continuous AlphaGUARD data and the weekly electret data.
The following six stations were planned for testing: 422 south basement S, 422 north basement
(downstairs AlphaGUARD), 420 south basement, 422 first floor, 422 second floor (upstairs
12-15
-------
Section 12—Results and Discussion: Special Studies
AlphaGUARD), and 420 first floor (note that the 420 first floor device was stolen on October 17, 2012,
and not replaced). According to the instruction manual, the Safety Siren detector may be placed face up
on a tabletop, countertop, or any flat surface where the ventilation slots will not be blocked. The detector
also must be kept dust free and a proper airflow must be maintained through the detector to obtain an air
sample that is representative of the local environment. It was impractical to use the Safety Siren for
ambient measurement at this site because of temperature and power issues. The manual restricts the
operating environment to 0°C (32°F) to 40°C (104°F). An additional electret was added on December 28,
2011, to the 422 second floor office to include that location in the comparison between the radon
detectors and electrets.
The Safety Siren has two displays—the "short term" is an average over the previous 7 days, and the "long
term" is the average from time of last reset (up to 5 years). The numeric LED display shows the level of
radon gas in pCi/L. The display range is 0.0 to 999.9. Readings are available after a minimum of 48 hours
of operation. We manually recorded the short-term reading at each of six indoor stations weekly. Data
were assembled in spreadsheet form for comparison to electret and AlphaGUARD results. The audible
alarm was muted. Every 24 hours, the detector does a self-test. If there is a failure in this self-test, an error
message will appear in the display window.
12.3.4 Consumer Grade Radon Detector Results and Discussion
As shown in Figure 12-12 and Tables 12-2 and 12-3, the Safety Siren consumer-grade detector shows
reasonably good agreement with an accepted professional method (electrets) over a range (1 to 5 pCi/L)
useful for determining compliance with EPA's recommend radon action level (4 pCi/L). Above 5 pCi/L,
the Safety Siren detector tends to overestimate the radon concentration; however, the overestimation
appears less pronounced in the second phase of the project (Figure 12-12). Thus, this device would
provide an indication of soil gas entry at low concentrations useful for radon management, although in the
higher range it might overestimate radon concentrations. Thus, it would be useful in showing a
homeowner when radon was being effectively excluded, but it might create a somewhat exaggerated
impression of vapor intrusion variability if high concentration peaks occur.
Figures 12-13 and 12-14 show comparisons between the 422 office and 422 north basement Safety Sirens
and electrets, along with black and gray bars representing mitigation system on/off periods. From these
graphs it is apparent that both the Safety Sirens and the electrets react in a very similar manner to
mitigation system on/off cycles: during mitigation on periods, radon concentrations decrease, and during
off periods, or passive cycles, radon concentrations increase. At lower concentrations, the Safety Sirens
and electrets produce similar readings, but higher concentrations yield a greater absolute difference
between readings for a given point in time (Figures 12-13 and 12-14). Thus, the Safety Sirens can
successfully be used as a means of monitoring whether the mitigation system is functioning, if not as an
accurate indicator of the exact degree of mitigation.
12-16
-------
Section 12—Results and Discussion: Special Studies
Radon Comparision - Consumer Grade Detector to Electrets
20.00
16.00
12.00
1.00
4.00
0.00
0.00
4.00
8.00
12.00
16.00
20.00
Siren consumer grade Radon detector one week duration {pCi/I)
-2
0 2 4 6 8 10 12 14 16 18 20
Radon Siren Concentration (pCi/L)
linear Comparison between Electretand Radon Siren Data
+ #22 Qffit*
« «J6a»N
• 422 Bast*
# 420FirM
O 429 Rat* S
1:1 Una
Figure 12-12. Comparison of electret and Safety Siren results for first phase of the project (top
graph) and second project phase (bottom graph).
12-17
-------
Section 12—Results and Discussion: Special Studies
Table 12-2. Comparison of Safety Siren, AlphaGUARD, and Electret Radon Data
Electret
Duplicates
(pCi/L)
Location
Time
Date
Safety Siren
(pCi/L)
AlphaGUARD (pCi/L)
Electrets (pCi/L)
1st Week's Radon Comparison
422 Office
15:55
1/4/2012
6.6
~5
4.92
422 First
16:03
1/4/2012
5.4
4.86
422 Base N
16:10
1/4/2012
14.4
-10
10.22
10.35
422 Base S
16:08
1/4/2012
14.6
9.57
420 First
16:13
1/4/2012
1.4
1.09
420 Base S
16:13
1/4/2012
3.7
2.72
2nd Week's Radon Comparison
422 Office
13:59
1/11/2012
5.7
4.69
4.56
422 First
14:12
1/11/2012
5.8
4.37
422 Base N
14:18
1/11/2012
12.6
8.78
9.05
9.11
422 Base S
14:19
1/11/2012
18.6
8.70
420 First
14:21
1/11/2012
1.6
1.18
420 Base S
14:25
1/11/2012
3.7
3.50
3rd Week's Radon Comparison
422 Office
11:25
1/18/2012
6.9
5.09
4.88
422 First
11:26
1/18/2012
6.4
4.46
422 Base N
11:27
1/18/2012
13.7
9.73
9.34
9.73
422 Base S
11:28
1/18/2012
18.8
8.89
420 First
11:40
1/18/2012
1.9
0.98
420 Base S
11:42
1/18/2012
3.0
2.84
4th Week's Radon Comparison
422 Office
15:17
1/25/2012
5.7
4.79
4.74
422 First
15:18
1/25/2012
5.9
3.81
422 Base N
15:20
1/25/2012
12.2
8.52
7.83
7.98
422 Base S
15:21
1/25/2012
18.8
8.12
420 First
15:25
1/25/2012
1.9
1.74
420 Base S
15:26
1/25/2012
3.8
3.60
5th Week's Rn Comparison
422 Office
14:41
2/1/2012
5.7
4.46
4.15
422 First
14:40
2/1/2012
5.5
3.42
422 Base N
14:39
2/1/2012
12.6
7.71
8.24
8.03
422 Base S
14:39
2/1/2012
18.9
7.26
420 First
14:38
2/1/2012
1.0
0.25
420 Base S
14:36
2/1/2012
1.8
1.27
(continued)
12-18
-------
Section 12—Results and Discussion: Special Studies
Table 12-2. Comparison of Safety Siren, AlphaGUARD, and Electret Radon Data (cont.)
Electret
Duplicates
(pCi/L)
Location
Time
Date
Safety Siren
(pCi/L)
AlphaGUARD (pCi/L)
Electrets (pCi/L)
6th Week's Radon Comparison
422 Office
14:03
2/8/2012
5.2
4.78
4.58
422 First
14:04
2/8/2012
5.3
4.48
422 Base N
14:15
2/8/2012
13.3
8.68
8.60
8.62
422 Base S
14:15
2/8/2012
18.9
9.56
420 First
14:10
2/8/2012
2.3
1.09
420 Base S
14:11
2/8/2012
5.4
2.40
7th Week's Radon Comparison
422 Office
12:19
2/15/2012
5.6
4.80
4.41
422 First
12:20
2/15/2012
6.0
4.15
422 Base N
12:23
2/15/2012
13.3
8.44
8.28
7.47
422 Base S
12:25
2/15/2012
19.1
8.34
420 First
12:28
2/15/2012
1.4
0.36
420 Base S
12:30
2/15/2012
3.0
1.94
8th Week's Radon Comparison
422 Office
14:28
2/22/2012
4.8
4.30
3.68
422 First
14:29
2/22/2012
5.2
3.82
422 Base N
14:30
2/22/2012
12.0
7.74
6.08
5.82
422 Base S
14:31
2/22/2012
18.1
6.56
420 First
14:26
2/22/2012
1.4
0.42
420 Base S
14:25
2/22/2012
3.7
2.08
9th Week's Rn Comparison
422 Office
15:40
3/1/2012
6.1
4.74
3.97
422 First
15:40
3/1/2012
6.2
3.88
422 Base N
15:41
3/1/2012
12.7
8.48
9.00
9.00
422 Base S
15:42
3/1/2012
19.6
10.43
420 First
15:46
3/1/2012
1.4
0.45
420 Base S
15:47
3/1/2012
2.1
2.56
12-19
-------
Section 12—Results and Discussion: Special Studies
Table 12-3. Comparison of Safety Siren and Electret Radon Data
End Date
Electret 422
Office
Safety Siren
422 Office
Electret 422
First Floor
Safety Siren
422 First Floor
Electret 422
Basement N
Safety Siren 422
Basement N
10/10/2012
3.2
4.0
3.0
4.4
12
13
10/17/2012
2.4
2.8
3.1
9.2
11
10/25/2012
0.2
0.8
0.6
0.3
0.7
0.8
10/31/2012
0.4
0.4
0.3
0.3
0.7
0.6
11/07/2012
-0.9
0.9
0.5
0.7
0.6
0.9
11/14/2012
1.6
0.8
0.2
0.6
0.95
1.2
11/21/2012
3.7
4.2
3.5
3.7
6.5
7.6
11/28/2012
4.7
5.2
4.3
6
9.5
11
12/05/2012
3.2
4.2
3.3
3.8
7.5
8.0
12/12/2012
4.4
5.6
3.9
4.9
9.8
11
12/19/2012
0.5
0.6
0.4
0.7
0.9
1.1
12/26/2012
0.2
0.7
0.4
0.7
0.8
1.1
01/02/2013
1.7
2.6
1.6
2.4
3.9
4.5
01/09/2013
4.5
4.6
3.8
4.6
8.7
11
01/16/2013
4.7
5.1
4.3
5.9
9.4
11
01/23/2013
4.6
5.3
4.5
5
9.5
12
01/30/2013
4
5.2
3.8
5.3
8.2
10
02/06/2013
4.7
6.2
4.6
5.6
9.2
13
02/13/2013
0.2
0.8
0.3
0.6
0.7
1.3
02/20/2013
0.1
0.5
0.4
0.3
0.5
0.8
03/6/2013
0.05
0.5
0.4
0.6
0.5
0.7
03/14/2013
0.03
0.5
0.5
0.5
0.5
0.5
03/20/2013
0.2
0.5
0.4
0.3
0.4
1
03/27/2013
-0.03
0.4
0.3
0.4
0.5
0.6
04/03/2013
0.2
0.5
0.4
0.5
0.5
1.1
04/10/2013
0.06
0.4
0.7
0.5
0.5
0.6
05/01/2013
1.8
2.8
2.5
2.4
4.1
6.2
(continued)
12-20
-------
Section 12—Results and Discussion: Special Studies
Table 12-3. Comparison of Safety Siren and Electret Data (cont.)
End Date
Electret 422
Basement S
Safety Siren 422
Basement S
Electret 420
First Floor
Safety Siren
420 First Floor
Electret 420
Basement S
Safety Siren 420
Basement S
10/10/2012
10
12
2.7
2.6
5.1
5.3
10/17/2012
8.6
00
CO
2.4
2.8
3.5
3.8
10/25/2012
0.6
0.7
0.4
0.5
10/31/2012
0.9
0.5
2.4
0.4
11/07/2012
0.5
0.5
0.2
0.6
11/14/2012
1.0
1.2
0.7
0.9
11/21/2012
7.5
8.1
3.5
3.2
11/28/2012
10
13
5.2
3.8
12/05/2012
8.5
9.6
3.9
4.2
12/12/2012
9.5
12
5.7
5.1
12/19/2012
1
0.7
0.6
0.7
12/26/2012
0.9
0.6
0.6
0.4
01/02/2013
3.6
4.8
1.5
1.6
01/09/2013
8.4
10
3.0
2.7
01/16/2013
10
12
5.7
6.5
01/23/2013
9.9
13
3.6
4.0
01/30/2013
9.6
12
4.3
3.6
02/06/2013
9.7
13
3.2
4.2
02/13/2013
0.8
1.3
0.8
0.9
02/20/2013
0.7
0.7
0.6
0.4
03/06/2013
0.7
0.7
0.6
0.5
03/14/2013
0.6
0.8
0.5
0.3
03/20/2013
0.9
0.6
0.6
0.2
03/27/2013
0.6
0.8
0.5
0.4
04/03/2013
0.6
0.8
0.7
0.9
04/10/2013
0.6
0.5
0.6
0.6
05/01/2013
6.1
8.6
2.4
2.8
12-21
-------
Section 12—Results and Discussion: Special Studies
422 Office Radon Siren/Elect ret Comparison
v
3 4
C
o
+= 3
2
£ i
M
s
p ,
¦
¦
¦
,4
¦¦¦
¦I
m
m
* 1
~~~
1
~—
y
¦
~
~
—~-
f
*
~
~
m
*
¦
m
~
~
*
y '
~
¦ i
~TTSft
L*. J
¦ » -w
O
tlcctprl 42MMRa*
Rn 4£2
~
i j
_ m
¦ ~
~
¦
a
4
1
/ *
If
3p
ttoctnrt 4J2&h*M
«n Writ UZBauN
*Mi»'iUirnvi On
•"MltlfttKM Pillrrf
•H
3
N
f-t
*N
H
PM
iH
m
•H
PM
•HI
(0
fM
T
©*
PM
PO
Date
Figure 12-14. Comparison between the 422 basement N Safety Siren and electret.
12-22
-------
Section 13—Conclusions and Recommendations
Table of Contents
13.0 Conclusions and Recommendations 13-1
13.1 Conclusions 13-1
13.1.1 Mitigation System Performance 13-2
13.1.2 Meteorological Effects on Vapor Intrusion 13-4
13.1.3 Preferential Pathways and Revisions to Conceptual Site Model: Helium Tracer
and Geophysical Tests 13-5
13.1.4 Revisions to Conceptual Site Model: VOC Data 13-7
13.1.5 Temporal Trends 13-8
13.2 Practical Implications for Practitioners 13-9
13.2.1 Mitigation Design Implications 13-9
13.2.2 Sampling Planning Implications 13-9
13.2.3 Delineating Preferential Pathways 13-10
13.3 Recommendations 13-10
List of Tables
13-1. Summary of Lines of Evidence for Meteorological Factors Influencing Vapor Intrusion in
This Study 13-6
13-i
-------
Section 13—Conclusions and Recommendations
13-ii
-------
Section 13—Conclusions and Recommendations
13.0 Conclusions and Recommendations
Conclusions of this phase of the research (results through May 2013) should be considered as a work in
progress, because active work at this site is ongoing at the time of this report and additional reports will
update (and may change) the study's findings. Also any application of these and other conclusions and
results of this study, and especially the "practical implications for practitioners" discussed in Section 13.2,
should be carefully considered in light of the facts that (1) this report describes a study on one site of a
relatively old urban residence in the Midwest and (2) vapor intrusion is variable in terms of the site-
specific conditions that influence vapor intrusion processes and impacts.
13.1 Conclusions
The following summarizes the main conclusions that can be drawn from this study. Additional details can
be found in the following subsections.
¦ Mitigation system performance—radon. The mitigation system installed in the duplex met or
exceeded all conventional performance tests, as well as more comprehensive tests involving
pressure differentials and continuous indoor radon monitoring. Radon reductions greater than
90% were observed, and all measured radon levels were below 4 pCi/L with the mitigation
system on.
¦ Mitigation system performance—VOCs. The mitigation system did not perform as well with
VOCs as it did with radon. During 7 months of mitigation system operation, immediate VOC
reductions in indoor air were observed and were significant overall, but the system only achieved
a reduction of just over 60% of VOC indoor air concentrations before mitigation. However,
additional decreases in VOC levels were observed near the end of the monitoring period reported
in this document (May 2013). During these periods of mitigation system operation, the system
was also observed to increase soil gas levels below the slab and at depth below the duplex,
suggesting that VOCs are being redistributed by the mitigation system and that soil gas
concentrations close to the building may be enhanced by drawing higher concentrations of VOCs
from greater depths. In addition, several snow events corresponded to increases in indoor air
VOC levels during mitigation that were not observed for radon.
¦ Meteorological variables. Multiple meteorological variables likely interact in complex ways to
affect VOC vapor intrusion at this duplex. The evidence is overwhelming that cold temperatures
contribute to greater vapor intrusion in this duplex. This was expected from knowledge of the
stack effect mechanism. The evidence also indicates that both snowfall and snow/ice
accumulation can increase VOC vapor intrusion, although this effect is complex for VOCs and
may be absent for radon. There is relatively little evidence of rain effects on VOCs, but some
evidence suggests a rain effect on radon. Barometric pressure change appears to have effects on
radon and probably VOCs, although the interactions are complex and additional work on the time
series data is needed to determine how best to analyze the effects of barometric pumping on vapor
intrusion in the duplex. There is evidence of an association between winds from westerly
directions with vapor intrusion in the 422 portion of the duplex, but the evidence for an effect of
wind velocity is equivocal. Additional study is needed to assess how to best model the complex
interactions between meteorological variables and vapor intrusion at this site.
¦ Preferential pathways and revisions to conceptual site model: Helium tracer and geophysical
tests. Four helium tracer tests performed pairwise with common subsurface injection locations
yielded similar overall patterns of tracer distribution in soil gas outside the building with and
without mitigation system operating. The variability between paired tests (mitigation on and
mitigation off) was more pronounced beneath the building where the mitigation system would
13-1
-------
Section 13—Conclusions and Recommendations
have been expected to have the most significant influence on airflow. The similar patterns
between tests performed in different subsurface areas (i.e., different injection points) suggest
control by common features of soil stratigraphy or the building envelope. Helium tracer
concentrations suggest easy horizontal migration toward the building over distances of up to 20 ft
and rapid vertical migration from 13 ft to 6 ft bis at the injection cluster. However, lower helium
concentrations at certain ports suggest subsurface heterogeneity and preferential flow paths that
could not be fully mapped with only the four tests conducted. Geophysical tests confirmed the
location of many known features in and around the duplex, including the shallow, moist silty clay
layer overlying the deeper sand/gravel outwash layer and the shallower (7 to 7.5 ft bis) silt/clay
layer. Ground penetrating radar (GPR) results suggest that the concrete slab varies from 0.5 to 0.7
ft in thickness with an irregular undulating contact with the underlying fill material and resulting
gaps where soil gas may pool or move preferentially.
¦ Revisions to conceptual site model: VOC data. Although chloroform was detected in
groundwater, the measured concentrations were too low to account for the peak chloroform
concentrations observed in soil gas (see Section 11). This suggests that there may be other
sources of chloroform such as combined sewers28 or drinking water mains29 that leak below
grade, higher groundwater concentrations at some locations near the site, or chloroform mass
stored in the vadose zone from a historic release. For PCE, the results suggest a groundwater
source, but the narrow range of variability in PCE concentrations and stability of this variability
over time make it unlikely that variability in groundwater concentrations is the only source of the
observed changes in soil gas or indoor air concentrations observed in this study. The variability in
indoor air PCE concentrations is also influenced by subsurface, building-related, and
meteorological variables. The potential that other sources of PCE may exist in the vadose zone or
combined sewer lines cannot be ruled out at this point.
¦ Temporal variability. PCE levels in indoor air follow the general trend of starting higher at the
beginning of the project (January 2011), dropping to a low in early summer, and rising slightly
and leveling out through the end of the intensive pre-mitigation study period (February 2012).
This general trend was attributed primarily to temperature, because the winter of 2010-2011 was
much more severe than the winter of 2011-2012. During mitigation testing, which began in
October 2012, indoor air PCE concentrations rose to levels above those observed in the March
2011 to September 2012 time period. We postulate that VOCs can be moved close to the structure
either by a cumulative stack effect during a severe winter or by operation of an SSD mitigation
system. It is unknown whether this VOC migration effect toward the slab will be common at sites
with other geological formations or contaminant distributions. Because our mitigation testing
included several on-off cycles over one winter, we do not know whether more substantial
reductions in indoor air VOC concentrations would be achieved with continued operation of the
SSD mitigation system. Spatial patterns changed dramatically when mitigation was operating,
indicating that during initial operation the mitigation system was influencing both soil gas and
indoor air concentrations.
13.1.1 Mitigation System Performance
An active subslab depressurization system was installed in the duplex, which met conventional tests of
mitigation system performance used by practitioners:
28Chloroform can form in sewers that receive bleach-containing products.
29Groundwater chloroform concentrations at this duplex are lower than the mean and peak drinking water concentrations for Indianapolis (19 ppb
and 82 ppb).
13-2
-------
Section 13—Conclusions and Recommendations
¦ Depressurization was observed with handheld micro manometers at 10 locations, which is more
locations than are typically monitored for such a small structure. In many cases, the observed
depressurization substantially exceeded conventional design criteria (Section 5.1).
¦ VOC sampling in the week following installation showed a dramatic reduction in VOC
concentrations in indoor air as measured by weeklong passive samplers, using methods consistent
with those used to monitor concentrations prior to mitigation (Section 5.3).
¦ Radon concentrations in indoor air were immediately and dramatically reduced, both as measured
by directly reading instruments and by weeklong electret samples (Section 5.2).
¦ A U-tube manometer installed as a visual indication at the extraction points indicated that the
system was functional when it was turned on. No sign of difficulty was seen with the U-tube
manometer.
A program of SSD testing was then performed involving cycles of active operation, passive operation,
and shutoff, with each cycle extending for multiple weeks. Such testing cycles are not conventionally
done in installed systems and may not be economically or logistically feasible. An intensive program of
monitoring of mitigation system performance was also conducted, more extensive than would normally
be done by residential mitigation practitioners. Although this program showed that the mitigation system
continued to provide a significant benefit, some aspects of the mitigation performance differed from
conventional expectations.
¦ "Control" of differential pressure as indicated by micro manometers connected to a computerized
data acquisition system was constant for long periods of time. During much of the testing, subslab
to indoor air differential pressures were observed at a negative 15 Pa or more, indicating a strong
driving force out of the structure. However, pressure "control" was lost on periods of up to
several days. The pressure control loss during active SSD was observed on each side of the
duplex but not on the same dates. The losses of pressure control were generally contemporaneous
with either snow or heavy rain events. On at least two occasions, pressure swings of a total of at
least 30 Pa were observed. These relatively brief events are most likely not significant in the
context of exposures being compared with risk-based standards derived from a 30-year cancer
exposure period. However, these events suggest that the conventional design criteria of protection
against a 5 Pa differential pressure change may be insufficient if indoor air concentrations
attributable to vapor intrusion must be controlled with reference to risk-based standards
applicable to a 24-hour or less exposure period (Section 5.1).
¦ The weeklong integrated passive samples from the mitigation system for the third week of
operation, for example, suggested that the indoor air VOC concentrations had reached levels
substantially higher than those observed in the first week after operation began. This pattern was
not seen with the weeklong integrated radon samples. Initial monitoring of SSD systems in
practice is not normally continued for multiple weeks (Sections 5.3 and 6.1).
¦ Instantaneous VOC concentrations in indoor air, as measured by an on-site GC, exhibited several
prominent spikes with durations between 1 and 3 days during active mitigation system operations.
These spikes were associated with brief snow storm events and particular wind directions that
appear for this house to be connected to higher indoor air concentrations. These same spikes were
not clearly seen in any of the radon data sets (Sections 5.3 and 6.5.1).
¦ The brief excursions in pressure and VOC concentrations were not contemporaneous.
¦ The overall performance of the mitigation system in reducing indoor concentrations was better
for radon than for VOCs. This finding is of possible concern because VOC mitigation design
practices and standards are drawn from the radon literature. The estimated reduction for radon
was approximately 91%, and the mitigation system consistently controlled radon to well below
13-3
-------
Section 13—Conclusions and Recommendations
the 2 pCi/1 limit that EPA believes to be the practical limit of mitigation system effectiveness
(which is constrained by ambient levels). For the 7 months of on/off mitigation system operation,
average reductions of 68% in chloroform and 61% in PCE in indoor air were observed,
considerably lower than that achieved for radon (Section 5.3).
¦ Concentrations of VOCs in some subslab and soil gas ports rose after mitigation began to levels
not seen in more than a year of approximately weekly monitoring. This suggests that VOCs are
being redistributed by the mitigation system and that soil gas concentrations close to the building
may be enhanced by drawing higher concentrations of VOCs from greater depths (Sections 5.3
and 6.1.2).
¦ A significant drop off in VOC concentrations (by an order of magnitude or more from the peak
VOC levels) was observed for all indoor air concentrations during the last few weeks of
mitigation system operation (through mid-May 2013). Whether this trend would continue over a
longer mitigation period (e.g., 6 months) is unknown. It is possible that, although initial
mitigation system operation does influence and increase soil gas concentrations at this site, VOC
concentrations may reach lower values at steady-state with longer term system operation.
(Section 5.3).
¦ The flux of VOCs through the SSD system was dramatically higher than the flux through the
building when the mitigation system was not installed (Section 5.4).
Taken together, these observations suggest that a "typical radon style system" would not necessarily
achieve the same degree or consistency of protection for VOCs, at least over the initial months of
operation that were tested in this study. We caution, however, that the application of this result should
not be to simply increase the conservatism of the depressurization standard and thus the amount of
vacuum routinely applied. Applying more vacuum routinely could have the deleterious effect of drawing
more VOC mass near to the structure. These results may suggest additional evaluation to weigh the
benefits of the following alternatives:
¦ automated (pressure controlled) variable speed fans in the SSD system
¦ a higher powered active SSD system run continuously
¦ modifications of the system to provide additional extraction points (such as in the basement
walls) or to enhance the effectiveness of SSD operation by sealing additional entry paths.30
Future testing could also investigate the possibility of additional and more stable VOC reductions with
longer term continuous mitigation system operation than was possible under the time frame of this study.
13.1.2 Meteorological Effects on Vapor Intrusion
In this study, we used multiple analytical tools to assess the relationship between meteorological
parameters and vapor intrusion:
¦ exploratory data analysis through visual examination of the shape of temporal trends in stacked
plots of indoor air and certain soil gas ports over the full project period (Section 6)
¦ detailed examination of temporal trends on stacked plots during unusual differential pressure
events (Section 9.1)
¦ visual examination of XY graphs (Section 9.2)
¦ quantitative time series methods (Section 10).
30Some sealing activities were performed on this structure in spring 2011 (as described in Section 3.2.2 of U.S. EPA, 2012a), but additional
sealing was not performed during the SSD installation.
13-4
-------
Section 13—Conclusions and Recommendations
Not all methods of analysis have yet been completed for all possible meteorological variables. The lines
of evidence investigated for the meteorological variables are summarized in Table 13-1.
The evidence is overwhelming that cold temperatures contribute to greater vapor intrusion in this duplex.
This was expected from knowledge of the stack effect mechanism.
The evidence suggests that both snowfall and snow/ice accumulation have effects on VOC vapor
intrusion, although this effect is likely absent for radon. The snow effect is likely to be complex in that
snow varies dramatically in moisture content and thus air permeability from one snow event to another
and as a snow accumulation ages over time. There is relatively little evidence of rain effects on VOCs, but
there is more evidence of a rain effect on radon.
Barometric pressure appears to have effects on radon and probably VOCs, although it is likely a complex
function of barometric pressure. Additional work will be required to determine how best to assess the
power of barometric pumping from time series data.
There is relatively strong evidence both in the data and from theory to support an association between
winds from a range of westerly directions with vapor intrusion in the 422 portion of the duplex. This
effect is expected to be different in other structures, depending on their orientation to the wind and the
distribution of subsurface contaminants. The evidence for an effect of wind velocity is equivocal.
Thus, multiple meteorological variables likely control VOC vapor intrusion at this duplex. The
meteorological variables likely interact in complex ways that would make the system difficult to model
completely. Such multiple variable effects are also known in the radon vapor intrusion literature. In a
recent review, Lewis and Houle stated:
"This paper identified about thirteen factors that can affect radon variation in the soil
and house environment. The thirteen factors being soil moisture content, soil
permeability, wind, temperature, barometric pressure, rainfall, frozen ground, snow
cover, earth tides, atmospheric tides, occupancy factors, season and time of day. One can
see the complexity of understanding and studying radon variability in homes. "
Although radon is not expected to be a highly quantitative tracer for VOC vapor intrusion in most
situations, we should not expect VOC vapor intrusion to be a significantly simpler process.
The relationships among the variables may not be purely linear, they could be synergistic. For example,
the PCE concentration curve vs. cold temperature related variables appears to curve upward under the
most extreme winter conditions. This can be seen in the PCE plots vs. heating degree days (Figure 9-42)
and vs. stack effect driving force (Figure 10-10 in U.S. EPA, 2012a). That may reflect an additive or
synergistic effect between related variables such as cold temperature, frozen ground, and snow cover. In
other words, cold could influence vapor intrusion through separate physical mechanisms, such as
enhancement of the strength of the stack effect and formation of a lower permeability frozen ground layer
near the surface.
13.1.3 Preferential Pathways and Revisions to Conceptual Site Model: Helium Tracer
and Geophysical Tests
Two primary methods were used in this study in an attempt to better understand the effects of
stratigraphic features on vapor intrusion at this duplex helium tracer tests and geophysical studies.
13-5
-------
Table 13-1. Summary of Lines of Evidence for Meteorological Factors Influencing Vapor Intrusion in This Study
(Blank cells reflect types of analysis not completed for a given parameter)
Cold Exterior
Temperatures (or
substantial change
in temperatures)
Rain
Events/
rainfall
amount
Snow or Ice
Accumulation on
Ground
Barometric
Pressure
Changes
Snowfall
West to NW
Winds
High Wind
Velocity
Apparent Temporal Association with
VOC Concentrations in Indoor Air
(Section 6, also U.S. EPA, 2012a)
Yes
Yes
Yes
Possibly for
chloroform
Apparent Temporal Association with
VOC Concentrations in Wall Ports or
Subslab Ports (Section 6)
Yes
Yes
Weak
Some
Apparent Temporal Association with
Large Subslab to Indoor Differential
Pressure Events (Section 9.1)
Yes in
some
cases
Yes in some cases
Yes in some
cases
Yes in a few
cases
Yes in a few
cases
Apparent Trend in XY Graph of.
Meteorological Parameter vs.
Subslab/lndoor Differential Pressure
(Section 9.1 and U.S. EPA, 2012a )
Yes
No
Yes
No
Apparent Trend in XY Graph of
Meteorological Parameter vs. VOC
Concentration (Section 9.2)
Yes for PCE, not definitive
for chloroform
Yes
No clear
relationship
Not
definitive
Yes for PCE,
No for
Chloroform
No for PCE,
Yes for
Chloroform
Correlation with Radon in Quantitative
Time Series Analysis (Sections 10.1 to
10.4); 422 Basement and Office
No
No
Yes in most analyses
Yes in some
analyses
Yes in most
analyses
Yes in some
analyses
Correlation with Chloroform in
Quantitative Time Series Analysis
(Sections 10.5 and 10.7); 422
Basement
Yes in one of two cases
with opposite signs for the
coefficients ofthe current
and past weeks.
Yes
No
Yes in some
analyses
Correlation with PCE in Quantitative
Time Series Analysis (Sections 10.6
and10.8), 422 Basement
Yes in one of two cases
but with an unexpectedly
negative coefficient for the
current week.
Yes, although
coefficients are both
positive and negative
No
Yes
No
No
-------
Section 13—Conclusions and Recommendations
Helium tracer tests (Section 12.2) performed with common injection locations yielded similar overall
patterns of tracer distribution in the exterior soil gas clusters with and without mitigation system
operation. The variability between paired tests (mitigation on and mitigation off) was more pronounced
beneath the building where the mitigation system would have been expected to have the most significant
influence on airflow. The highest concentrations observed were usually those close to the 13-ft deep
injection depth, suggesting attenuation as the tracer migrated upward. The similar patterns between tests
performed in different subsurface areas (different injection wells) suggest control by common features of
soil stratigraphy or the building envelope because the tests would have these factors in common. The high
concentrations observed at the 9-ft and 13-ft depths for all the injections, despite the buoyancy of the
helium-enriched soil gas, suggest that transference of helium is easy horizontally toward the building over
distances of up to 20 ft typically within 2 days. Vertical migration from 13 ft to 6 ft bis also occurred
relatively rapidly at the injection cluster. The lack of influence at certain ports closer to the point of
injection and along the same general line to more distant ports suggests subsurface heterogeneity and
preferential flow paths. However, many more than the four tracer tests performed would likely be needed
to fully map the preferential flow paths.
Geophysical studies (Section 12.1) confirmed the location of many known features in and around the
duplex. The results suggested an electrically conductive unsaturated zone with irregular hummocky
layers overlaying a more regular horizontally stratified zone at approximately 7.5 feet deep, which likely
corresponds to the facies contrast between the underlying sand/gravel layer and the shallower silt/clay
layer. This contact may also form local perched water conditions which may also cause such a
geophysical response. These stratigraphic features may influence subsurface fluid fate and transport.
The interior ground penetrating radar (GPR) results suggest the concrete slab varies from 0.5 to 0.7 ft in
thickness with an irregular undulating contact with the underlying material. This underlying material is
electrically conductive and contains many discontinuities and non-horizontal interfaces that may govern
subsurface fluid fate and transport. This suggests the concrete floor was not poured on well-flattened fill
material and may provide pockets into which soil gas may pool or move preferentially. However, the
presence of fill material of some sort is evident at many locations beneath the slab. The GPR data suggest
there is a thicker concrete footing below the doorways, which separate the three rooms in each of the
duplex basements. The GPR response to the cistern and the covered sewer line is evident as are GPR
reflections likely due to fill material.
13.1.4 Revisions to Conceptual Site Model: VOC Data
The finding of chloroform in groundwater, but at current concentrations substantially too low to account
for the peak chloroform concentrations in soil gas (Section 11), would be consistent with several
alternative conceptual models:
20. Chloroform is primarily transported to the site through combined sewers that leak below grade.
Chloroform is stripped from these waters as they infiltrate down to the water table. It should be
noted that the concentrations observed in groundwater at this duplex are lower than even the
mean drinking water concentration for Indianapolis (19 ppb) and are far below the peak reported
drinking water concentration (82 ppb).31 Additional formation is likely in combined sewers that
receive bleach-containing products.
21. Chloroform is present in higher concentration in groundwater at some locations near the site that
has not yet been delineated by the three existing groundwater well clusters.
22. Chloroform mass is stored primarily in the vadose zone from a historic release.
31NY Times, May 16, 2012 http://projects.nytimes.com/toxic-waters/contaminants/in/marion/in5249004-indianapolis-water
13-7
-------
Section 13—Conclusions and Recommendations
23. Groundwater chloroform concentrations at the site were historically (or are periodically but
briefly) higher, and the observed soil gas concentrations are a legacy of that period.
Our results suggest a groundwater source for PCE, but the narrow range of variability in PCE
concentrations (below an order of magnitude) and stability of this variability overtime (see Section 11)
make it unlikely that variability in groundwater concentrations is strongly related to the changes in soil
gas or indoor air concentrations over the time scales observed in this study. As has been shown in this and
other studies, the variability in indoor air PCE concentrations is also influenced by subsurface, building-
related, and meteorological variables that affect the concentration of PCE as it migrates from the water
table, enters the building, and mixes with indoor air. However, other sources of PCE that may also exist
in the vadose zone or combined sewer lines cannot be ruled out at this point, and variability in those
sources could also influence the PCE concentrations in indoor air.
13.1.5 Temporal Trends
PCE levels at all six indoor air sampling stations follow the same general trend of starting higher at the
beginning of the project (January 2011), dropping to a low in early summer, and rising slightly and
leveling out through the end of the pre-mitigation study period (February 2012). This general trend was
attributed primarily to temperature, since the winter of 2010-2011 was much more severe than the winter
of 2011-2012 (Section 6).
Sampling was discontinued for the spring and summer of 2012 because of funding limitations. However,
the concentrations in October 2012 before the mitigation system was installed were very similar to those
observed in October 2011. which suggests that the intensive sampling conducted for this project did not
distort the observed indoor concentrations. Surprisingly, the concentration levels of PCE in the first
period of active mitigation rose to levels above those observed in the March 2011 to February 2012 time
period. The concentrations rose somewhat more after the mitigation system was switched into a passive
mode. The observation of higher PCE post-mitigation suggests that the geologic formation could yield
enough vapor intrusion-derived PCE to account for the January 2011 concentrations. We postulate that
VOCs can be moved close to the structure either by a cumulative stack effect during a severe winter or by
operation of an SSD mitigation system. It is unknown whether this VOC migration effect toward the slab
will be common at sites with other geological formations or contaminant distributions or whether it would
continue at this site with continued operation of the mitigation system. Because this project used an
experimental design with multiple on-off cycles for mitigation over the course of one winter, it is also
unknown whether more stable reductions in indoor air VOC concentrations would be achieved with
longer term continuous operation of the SSD mitigation system.
Soil gas concentrations at some sampling locations) vary only within a narrow range of two or three times
over 1 year, suggesting that multiple samples or time-integrated samples may have limited benefit.
However, at some locations, changes of about a factor of 10 in soil gas concentrations occur over 1 year,
suggesting that there would be significant additional information provided from additional soil gas
sampling rounds at these locations. However, these changes are gradual, suggesting that sampling rounds
should be widely spaced and that time-integrated soil gas sampling over periods of 8 hours to 1 week
would have limited benefit.
The soil gas samples showed much more spatial variability in winter than in summer when mitigation was
not being operated. Spatial patterns changed dramatically, as might be expected when mitigation was
operating.
13-8
-------
Section 13—Conclusions and Recommendations
13.2 Practical Implications for Practitioners
As described in the introduction to this section, the practical implications discussed below should be
couched with the caveats that (1) this is a work in progress subject to change as additional information is
gathered in the next phase of study and (2) it does represent an intensive study of a single house and
therefore any conclusions may not (or may) be representative of other chlorinated VOC vapor intrusion
sites in different environments.
13.2.1 Mitigation Design Implications
The results of our mitigation system testing suggest that the assumption that systems protective for radon
will always be equally protective for VOCs could be incorrect, at least during the first few months of
system operation. Although this house could be relatively unique (given its old age and very course
grained geology in the deeper layers) at least in this situation subslab to basement differential pressure
variations of at least 30 Pa were observed repeatedly but for a small percentage of the total period of
observation. Since these unusual pressure events typically lasted from 1 to 3 days, they may not
significantly impact design of systems designed to protect against chronic risks in indoor air, but they are
more likely to impact design for systems protecting against short-term risks.
Similarly, if short-term risks are a design driver and the mitigation systems will not have backup electrical
power (the majority of SSD systems in our experience), caution may be necessary about the potential for
the mitigation system to increase the subslab VOC concentration at least during early system operation.
For example, if the subslab VOC concentration increases, then indoor air concentrations could increase if
the mitigation system stops operating during a power failure.
It should be stressed that the observations of the effects of mitigation on soil gas and indoor air VOC
levels in this report are based on on/off operation cycles of the mitigation system for about 7 months after
installation. Also, indoor air VOC concentrations were steadily declining in all sampling locations just
before the sampling period in this report was completed (in May 2013). This suggests the possibility that
although initial mitigation system operation appeared to move higher VOC soil gas concentrations closer
to the building, resulting in lower VOC reductions than were observed for indoor air, continued
mitigation system operation could, over time, result in more stable reductions in indoor air levels once
more steady state conditions are reached in the subsurface. Alternatively, the decline in VOC
concentrations in May 2013 could be attributable to weather changes (indoor VOC concentrations have
been shown to decline in the spring time at this site). However, it is clear from these results that at least at
sites with similar conditions to this one, VOCs may not respond as rapidly or as completely to mitigation
as radon does.
13.2.2 Sampling Planning Implications
The results reported here provide little support of the common guidance that vapor intrusion sampling
must be timed around rain events greater than one half inch. While there may have been some effects on
vapor intrusion of major seasonal flooding events that changed the local water table by approximately 5
ft, there was not any apparent effect on indoor air concentrations from more moderate rain events. The
results reported here do suggest that snow events, snow cover, and/or frozen soils may temporarily and
significantly increase vapor intrusion.
Several variables have been shown here to most likely have an interactive effect on VOC vapor intrusion:
¦ cold temperatures
¦ snow/ice
¦ barometric pressure
13-9
-------
Section 13—Conclusions and Recommendations
¦ wind direction
Given this complexity and variability in vapor intrusion observed at this site, practitioners should not
expect to be able to explain in detail temporal patterns drawn from smaller data sets (for example, three or
four rounds of VOC sampling). However, as results from this study are confirmed in studies of other
buildings, it may be possible to develop recommendations to guide selection of "near worst case" indoor
air sampling conditions for specific sites based on the site's known characteristics such as climate and
stratigraphy.
13.2.3 Delineating Preferential Pathways
The results of our geophysics and tracer injection studies show that these tools can provide useful insights
in delineating structures beneath a building and ground surfaces. However, the level of effort required to
fully understand and map the features that influence subslab gas flow in even a small old building such as
this may be too high for routine application of those tools.
13.3 Recommendations
The results presented here suggest several fruitful lines of future inquiry:
24. The results suggest that current chemical vapor intrusion mitigation system designs, based on
radon system's experience, may produce designs that are highly effective for radon but not as
effective for VOCs, at least during the initial months of system operation. This finding suggests a
need for longer term confirmation of post-mitigation VOC concentrations and replication of this
study's findings in other environments. Specifically, buildings of other ages/designs in conditions
similar to this (15 to 20 ft to groundwater, moderate strength source, coarse deep geology) should
be tested. This finding should also prompt more intensive studies of long-term mitigation system
performance in commercial buildings and in other geographies. The current trend of TCE being
managed based on short-term exposure thresholds provides additional impetus for such studies,
because radon and other VOCs are not usually managed based on short-term health effects.
25. The finding that the current mitigation system was not at all times able to shut off VOC vapor
intrusion as well as it controlled radon leaves open several questions:
• Would a mitigation system continuously operated over a longer period and/or with a more
powerful fan provide additional protection and, ultimately, a reduction in soil gas VOC levels
or merely exacerbate the problem of VOCs being drawn toward the foundation?
• Would a mitigation system with a variable speed fan controlled based on differential pressure
provide better performance?
• Would a degree of sealing beyond what was needed to meet radon standards have produced
a more effective VOC mitigation system in this structure?
• Would a mitigation system that extracted air from behind the basement walls be more
effective?
• Was the on/off testing protocol used in this study unrealistic?
26. Additional statistical analyses could be performed on this data set to:
• Determine if any of the meteorological parameters cause consistent trends in soil gas
concentrations of a sufficient magnitude to effect site management decision making.
• Examine the optimum spacing of sampling events in indoor air (although the autocorrelation
and partial autocorrelation analysis provides some hints, this is not the proper type of
statistical analysis to address the question of optimum sample spacing).
13-10
-------
Section 13—Conclusions and Recommendations
• Assess if there are important diurnal effects in indoor air (the current analyses were
conducted on data aggregated over 24 hours, but 7 or more samples a day were taken with
the on-site GC).
• Examine whether combinations of predictor variables may predict more of the variance in
indoor VOCs than any one variable alone. The most likely first strategy to explore would be to
use a function of exterior temperature as a model term in all of the models and see which of
the other predictor variables best accounts for the residual variability after temperature (and
thus, the stack effect) was modeled.
• Predict the optimum sampling time based on meteorological conditions.
27. The finding that wind direction effects appear to fit the Abreu and Johnson (2005) model
predictions (e.g., U.S. EPA, 2012d) suggest that this data set could be a useful test for validating
that model and exploring its utility in site management decision making.
28. Although the information collected under this and the previous project has shown that PCE
groundwater levels are high enough to serve as a source for the VOCs in soil gas and indoor air at
this site, other vadose zone PCE sources cannot be ruled out. Also, current chloroform
groundwater concentrations were not adequate to serve as a source for the soil gas and indoor air
concentrations under and in the duplex. Additional work is needed, including additional
subsurface borings and soil gas monitoring points, to more positively identify the sources
responsible for the vapor intrusion of these VOCs at this site.
13-11
-------
Section 13—Conclusions and Recommendations
13-12
-------
Section 14—References
14.0 References
Aaltonen, J., H. Pumpanen, T. Hakola, S. Vesala, R. Rasmus, and J. Back. 2012. Snowpack
concentrations and estimated fluxes of volatile organic compounds in a boreal forest. H.
Biogeosciences 9: 2033-2044.
Abreu, L.D., and P.C. Johnson. 2005. Effect of vapor source—building separation and building
construction on soil vapor intrusion as studied with a three-dimensional numerical model.
Environmental Science and Technology 39( 12):4550-4561.
Abreu, L.D., and P.C. Johnson. 2006. Simulating the effect of aerobic biodegradation on soil vapor
intrusion into buildings: influence of degradation rate, source concentrations. Environmental
Science and Technology 40(7): 23 04-2315.
AFCEE (Air Force Center for Environmental Excellence). 2004. Principles and Practices of Enhanced
Anaerobic Bioremediation of Chlorinated Solvents. Prepared for AFCEE, Brooks City-Base, TX.
Available at
http://costperformance.org/remediation/pdf/principles and practices bioremediation.pdf.
Allen, M.K., D. Grande, and L. Pansch. 2007. Evaluation of Passive Sampling Techniques for Monitoring
Roadway and Neighborhood Exposures to Benzene and Other Mobile Source VOCs. Wisconsin
Department of Natural Resources (WDNR) Publication AM-384 2007. WDNR, Madison, WI.
funded by U.S. EPA Community Scale Monitoring Grant. Available at
http://dnr.wi.gov/topic/AirOualitv/documents/RoadwavStudvFinal.pdf.
Armstrong, R.L., and E. Brun. 2008. Snow and Climate: Physical Processes, Surface Energy Exchange
and Modeling. Cambridge University Press.
Arvela, H. 2008. Radonsafe Foundation, Moisture Prevention and Air Exchange in a Healthy Building.
Finnish Research Programme on Environmental Health. Available at
http: //www. ktl. fl/sytty/ab stracts/arvela .htm.
ATSDR (Agency for Toxic Substances and Disease Registry). 1997. Toxic Substances Portal—
Toxicological Profile for Chloroform. Available at
http://www.atsdr.cdc.gov/toxproflles/tp.asp?id=53&tid=16.
Batterman, S., T. Metts, and P. Kalliokoski. 2002a. Diffusive uptake in passive and active absorbent
sampling using thermal desorption tubes. Journal of Environmental Monitoring 4:870.
Batterman, S., T. Metts, and P. Kalliokoski. 2002b. Low-flow active and passive sampling of VOCs using
thermal desorption tubes: Theory and application at an offset printing facility. Journal of
Environmental Monitoring 4:361-370.
Begerow, J., E. Jermann, T. Keles, and L. Dunemann. 1999. Performance of two different types of
passive samplers for the GC/ECD-FID determination of environmental VOC levels in air.
Fresenius Journal of Analytical Chemistry 363:399-403.
Bender., J.A. 1957. Air Permeability of Snow. Res. Rep. 37 Snow, Ice and Permafrost Research
Establishment (U.S. Army Corps of Engineers) Wilmette, IL. p. 46-62.
14-1
-------
Section 14—References
Bertoni, G., R. Tappa, and I. Allegrini. 2001. Internal consistence of the new "Analyst" diffusive sampler:
a long term field test. The Diffusive Monitor 12:2-5.
Bozkurt, O., K.G. Pennell, and E.M. Suuberg. 2007a. Evaluation of characterization techniques for vapor
intrusion scenarios using a three-dimensional computational fluid dynamics (CFD) model. Air
and Waste Management Association, Providence, RI, September.
Bozkurt, O., K.G. Pennell, and E.M. Suuberg. 2007b. Characterizing vapor intrusion scenarios using a
computational fluid dynamics model. ACS National Meeting. Boston, MA. August.
Brenner, D. 2007. Effect of meteorological and site conditions on, and relationship of diurnal variation to,
vapor intrusion at NASA Ames Research Center. Conference Proceeding from Vapor Intrusion:
Learning from the Challenges-Air and Waste Management Association (AWMA), Providence,
RI.
Brown, V., and D. Crump. 1993. Appropriate sampling strategies to characterize VOCs in indoor air
using diffusive samplers. Pp. 241-249 in Proceedings of International Conference on VOCs,
London, October 27-28.
Brown, V., and D. Crump. 1993. Appropriate sampling strategies to characterize VOCs in indoor air
using diffusive samplers. Pp. 241-249 in Proceedings of International Conference on VOCs,
London, October 27-28.
Bruno, P., M. Caputi, M. Caselli, G. de Gennaro, and M. de Rienzo. 2005. Reliability of a BTEX radial
diffusive sampler for thermal desorption: Field measurements. Atmospheric Environment
39:1347-1355.
California Environmental Protection Agency. 2012. Advisory Active Soil Gas Investigations. Department
of Toxic Substances Control, Los Angeles Regional Water Quality Control Board, San Francisco
Regional Water Quality Control Board, April.
Cao, X.L., and C.N. Hewitt. 1994. Study of the degradation by ozone of adsorbents and hydrocarbons
adsorbed during the passive sampling of air. Environmental Science and Technology 28:151-125.
Carr, D.B., P.G., Laurent, C. Levy, and A.H. Horneman. 2011. Stylistic modeling of vadose zone
transport insight into vapor intrusion processes. Presented at the AEHS Workshop: Addressing
Regulatory Challenges in Vapor Intrusion. A State of the Science Update. San Diego CA, March
15.
Case, J., and K. Gorder. 2006. The investigation of vapor intrusion at Hill AFB. Conference Proceeding
from Vapor Intrusion: The Next Great Environmental Challenge—An Update. Air and Waste
Management Association (AWMA), Los Angeles, CA.
Christ, Maj John A. 2010. Vapor intrusion from entrapped NAPL sources and groundwater plumes :
process understanding and improved modeling tools for pathway assessment. Presented at
SERDP Technical Exchange Meeting, August 16.
Cleveland, W.S, and S.J. Devlin. 1988. Locally-weighted regression: an approach to regression analysis
by local fitting. Journal of the American Statistical Association #J(403):596-610.
14-2
-------
Section 14—References
Cleveland, W.S. 1981. LOWESS: A program for smoothing scatterplots by robust locally weighted
regression. The American Statistician 35( 1):54.
Cocheo, C., C. Boaretto, D. Pagani, F. Quaglio, P. Sacco, L. Zaratin, and D. Cottica. 2009. Field
evaluation of thermal and chemical desorption BTEX radial diffusive sampler Radiello®
compared with active (pumped) samplers for ambient air measurements. Journal of
Environmental Monitoring 11:297-3 06.
Cochrane, D,. and G. H. Orcutt. 1949. Application of least squares regression to relationships containing
autocorrelated error terms. Journal of the American Statistical Association 44\32-61,
Cody, R.J., A. Lee, and R. Wiley. 2003. Use of radon measurements as a surrogate for relative entry rates
of volatile organic compounds in the soil gas. Presented at the A WMA Conference on Indoor Air
Quality Problems and Engineering Solutions. Research Triangle Park, NC, July 21-23.
Cohen, A. B. 1971. Concepts of Nuclear Physics. McGraw-Hill.
Colby College. 2009. Regression: Patterns of Variation. Colby College Biology Department. Available at
http://www.colbv.edu/biology/BI17x/regression.html.
Conway, H. and J. Abrahamson. 1984. Air permeability as a textural indicator of snow. Journal of
Glaciology 30(106):328-333.
Coyne, L. 2010. Personal communication. SKC, Inc. August 17.
Dawson, H. 2008. US EPA 's Vapor Intrusion Database—Preliminary Evaluation of Attenuation Factors
for Chlorinated VOCs in Residences. U.S. EPA Office of Superfund Remediation and
Technology Innovation (OSRTI).
Dawson, H., and H. Schuver. 2010. US EPA's Vapor Intrusion Database: Preliminary Evaluation of
Attenuation Factors for Chlorinated VOCs in Residences. Presented at the 20th Annual
International Conference on Soils, Sediments, Water and Energy, March 16.
DePersio, T., and J. Fitzgerald. 1995. Guidelines for the Design, Installation, and Operation of Sub-Slab
Depressurization Systems. Massachusetts Department of Environmental Protection, Northeast
Regional Office. Available at http://www.mass.gov/dep/cleanup/laws/ssdle.pdf.
DeVaull, G.E. 2012. Analysis and Interpretation of Time-Variable Vapor Intrusion Data: Pressure-
Driven Flow Across Building Foundations. Proceedings paper, AWMA Vapor Intrusion
Conference, 2012. Also DeVaull, G.E. 2013. Analysis and Interpretation of Time-Variable Vapor
Intrusion Data: Pressure-Driven Flow Across Building Foundations. Presented at 23rd Annual
International Conference on Soil, Water, Energy and Air, San Diego, CA, March 2013.
Dickson, J., A. Lonegran, and R. Stetson. 2010. Characterization of multiple chlorinated solvent plumes
due to the impact of TCE screening level reduction. International Journal of Soil, Sediment and
Water 3(2): Section 11.
Dietz, R.N., and E.A. Cote. 1982. Air infiltration measurements in a home using convenient
perfluorocarbon tracer technique. Environment International 8( 1-6):419—433.
14-3
-------
Section 14—References
Dietz, R.N., et al. 1986. Detailed description and performance of a passive perflurocarbon tracer system
for building ventilation and air exchange measurements. Pp. 203-254 in Measured Air Leakage
of Buildings: A Symposium. Heinz R. Trechsel, Peter L. Lagus, ASTM Committee E-6 on
Performance of Building Constructions Contributor Heinz R. Trechsel, Peter L. Lagus. Published
by ASTM International.
DiGiulio, D., C. Paul, R. Cody, R. Willey, S. Clifford, P. Kahn, R. Mosley, A. Lee, and K. Christensen.
2006. Assessment of Vapor Intrusion in Homes near the Raymark Superfund Site Using Basement
and Subslab Air Samples. EPA/600/R-05/147. U.S. EPA Office of Research and Development,
Ada, OK, March.
Distler, M., and P. Mazierski. 2010. Soil vapor migration through subsurface utilities. Proceedings 2010
AWMA Vapor Intrusion Specialty Conference.
Division of Spill Prevention and Response Contaminated Sites Program. 2012. Vapor Intrusion
Guidance for Contaminated Sites. October.
http://dec. alaska.gov/spar/csp/guidance/Vapor%20Intrusion%20Guidance.pdf.
Eklund, B., and D. Borrows. 2009. Prediction of indoor air quality form soil-gas data at industrial
buildings. Groundwater Monitoring and Remediation 29(1): 118-125.
Environmental Quality Management. 2004. Users Guide for Evaluating Subsurface Vapor Intrusion into
Buildings. Prepared for US EPA under EPA Contract Number: 68-W-02-33, WA No. 004, PN
030224.0002. Available at
http://www.epa.gov/oswer/riskassessment/airmodel/pdf/20Q4 0222 3phase users guide.pdf.
Falta, R. 2006. "Environmental Remediation of Volatile Organic Compounds" in C. Ho and S. Webb Gas
Transport in Porous Media p. 354-370. Netherlands: Springer.
Folkes, D., W. Wertz, J. Kurtz, and T. Kuehster. 2009. Observed spatial and temporal distributions of
CVOCs at Colorado and New York vapor intrusion sites. Ground Water Monitoring &
Remediation 29(l):70-80.
Forbes, J., J. Havlena, M. Burkhard, and K. Myers. 1993. Monitoring of VOCs in the deep vadose zone
using multi-port soil gas wells and multi-port soil gas/ground-water wells. Pp. 557-571 in
Proceedings of the Seventh National Symposium on Aquifer Restoration, Ground-water
Monitoring, and Geophysical Methods, National Ground Water Association, Dublin, OH.
Garbesi, K., and R.G. Sextro. 1989. Modeling and field evidence of pressure-driven entry of soil gas into
a house through permeable below-grade walls. Environmental Science and Technology
23(12):1481-1487.
Greiner, M.A., and R.M. Franza.2003. Barriers and bridges for successful environmental technology
transfer. The Journal of Technology Transfer 28(2): 167-177(11).
GSI Environmental. 2008. Final Report: Detailed Field Investigation of Vapor Intrusion Processes.
ESTCP Project ER-0423, September.
Hall, S.K., J. Chakraborty, and R.J. Ruch. 1997. Chemical Exposure and Toxic Responses. Boca Raton,
FL: CRC Press.
14-4
-------
Section 14—References
Hosoda, M., A. Sorimachia, S. Tokonamia, T. Ishikawaa, S.K. Sahooa, N.M.M. Hassanb, M. Fukushib,
and S. Uchidaa. 2008. Generation and control of radon from soil. Proceedings of 12th
International Congress of the International Radiation Protection Association. Available at
http://www.irpal2.org.ar/fullpapers/FP0971.pdf
Hass, B.S., and R. Herrmann. 1998. Tracing volatile organic compounds in sewers. Water Science and
Technology 37(1):295-301.
Hayes, H. 2008. Air Toxics Ltd., personal communication to Chris Lutes, ARCADIS, August.
Heggie, A.C., and B. Stavropoulos. 2019. Evaluating vapor intrusion risk using comparative dynamic and
passive flux chambers at a TCE impacted site in Sydney Australia.
http://events.awma.Org/education/Final%20Papers/7-Hcggie.pdf. Paper presented atAWMA 2010
Vapor Intrusion Specialty Conference.
Helsel, D.R. 2005. Nondetects and Data Analysis: Statistics for Censored Environmental Data. Hoboken,
NJ: Wiley.
Hers et al. 2011. Cold Climate Vapor Intrusion Research Study - Results of Seasonal Monitoring of
House at North Battleford, Saskatchewan.
http://indoorairproject.files.wordpress.com/2011/03/sgs-attachment-4.pdf cited as in preparation
in http://www.epa.gov/oust/cat/pvi/PVI Database Report.pdf.
Holton, C., et al. 2013. Multi-year monitoring of a house over a dilute CHC plume: implications for
pathway assessment using indoor air sampling and forced under-pressurization tests. Presentation
at AEHS conference, March. Available at
https://iavi.rti.org/attachments/WorkshopsAndConferences/05 Johnson 03-19-13.pdf.
Howard, P.H., R.S. Boethling, W.F. Jarvis, W.M. Meylan, and E.M. Michalenko. 1991. Handbook of
Environmental Degradation Rates. Chelsea, MI: Lewis Publishers, Inc.
Hui, S.C. 2003. Lecture: Air Movement and Natural Ventilation. Department of Mechanical Engineering,
The University of Hong Kong. Used by Permission from Dr. Hui. Available at
http://www.arch.hku.hk/teaching/lectures/airvent/sect03.htm.
IndyGov. 2012. My Neighborhood. Available at http://maps.indv.gov/mvneighborhood/.
ITRC (Interstate Technology and Regulatory Council). 2007. Vapor Intrusion Pathway: A Practical
Guideline. Prepared by The Interstate Technology & Regulatory Council Vapor Intrusion Team.
Available at http://www.itrcweb.org/documents/VI-l.pdf.
Johnson, P.C., et al. 2012. Vapor Intrusion above a Dilute CHC Plume: Lessons-Learned from Two Years
of Monitoring. Presentation at AEHS Conference, March. Available at
https://iavi.rti.org/attachments/WorkshopsAndConferences/09 P%20Johnson%20AEHS%20Talk
%20032112.pdf.
Johnson, P.C., and R.A. Ettinger. 1991. Heuristic model for predicting the intrusion rate of contaminant
vapors into buildings. Environmental Science and Technology 25:1445-1452.
Judge, G.G, W.E. Griffiths, R.C. Hill, H. Lutkepohl, and T.C. Lee. 1985. The Theory and Practice of
Econometrics, Second Edition. New York: Wiley.
14-5
-------
Section 14—References
Jury, W.A., D. Russo, G. Streile, and H.E. Abd. 1990. Evaluation of volatilization by organic chemicals
residing below the soil surface. Water Resources Research 26(1): 13-20.
Koontz, M.D., and H.E. Rector. 1995. Estimation of Distributions for Residential Air Exchange Rates.
EPA Contract No. 68-D9-0166, Work Assignment No. 3-19. U.S. EPA Office of Pollution
Prevention and Toxics. Washington, DC.
Krewski, D., R. Mallick, J.M. Zielinski, and E.G. Letourneau. 2005. Modeling seasonal variation in
indoor radon concentrations. Journal of Exposure Analysis & Environmental Epidemiology
7J(3):234-243.
Kuehster, T.E., D.J. Folkes, and E.J. Wannamaker. 2004. Seasonal variation of observed indoor air
concentrations due to vapor intrusion. Presented at the Midwestern States Risk Assessment
Symposium, Indianapolis, IN.
Kurtz, J.P., and D.J. Folkes. 2008. Empirical data on lag time for vapor intrusion from a chlorinated VOC
groundwater plume. Presented at AEHS San Diego Conference, March 12.
Lau, P. 2008. Calculation of flow rate from differential pressure devices - orifice plates. Presented to the
European Metrology Association for Thermal Energy Measurement, August 2-4.
Lee, A., K. Baylor, P. Reddy, and M. Plate. 2010. EPA Region 9's RARE opportunity to improve vapor
intrusion indoor air investigations. Presented at the 20th Annual International Conference on
Soils, Sediments, Water and Energy, San Diego, CA. March 16.
Lewis, R.K. and P.N. Houle. 1985. A Living Radon Reference Manual.
http://www.dep.state.pa.us/brp/radon di vision/Li vingRadon Reference Manual.htm reviewing
Nazaroff, W. and Doyle, S. Radon Entry into Houses Having a Crawl Space. Health Physics
45:265-281.
Linn, W., L. Appel, R. DeZeeuw, P. Doom, J. Doyon, T. Evanson, J. Farrell, D. Hanson, R. Jurgens, J.
So, C. Speigel, D. Switek, and S. Yankey. 2010. Conducting Contamination Assessment Work at
Drycleaning Sites. State Coalition for Remediation of Dry cleaners. Available at
http://www.drvcleancoalition.org/download/assessment.pdf.
Loureiro, C.O., and L.M. Abriola. 1990. Three-dimensional simulation of radon transport into houses
with basements under constant negative pressure. Environmental Science and Technology
24(9):1338-1348.
Luo, H., P. Dahlen, P. Johnson, T. Creamer, T. Peargin, P. Lundegard, B. Hartman, L. Abreau, and T.
McAlary. 2006. Spatial and temporal variability in hydrocarbon and oxygen concentrations
beneath a building above a shallow NAPL source. Presented at the Battelle Conference on
Remediation of Chlorinated and Recalcitrant Compounds, Monterey, CA, May.
Luo, H., P. Dahlen, P.C. Johnson, T. Peargin, and T. Creamer. 2009. Spatial variability of soil-gas
concentrations near and beneath a building overlying shallow petroleum hydrocarbon-impacted
soils. Ground Water Monitoring & Remediation 29( 1): 81—91.
Lutes, C., Cosky, B., B. Schumacher, J. Zimmerman, H. Hayes, R. Truesdale, S. Lin, and B. Hartman.
2012a. Long-term detailed study of vapor intrusion temporal variability: implications for indoor
air and soil gas sampling guidance. Oral Presentation and Short Paper: AWMA Symposium on Air
14-6
-------
Section 14—References
Quality Measurement Methods and Technology, Durham, NC. April.
Lutes, C., Cosky, B., L. Abreu, B. Schumacher, J. Zimmerman, R. Kapucinski, and R. Truesdale. 2012b.
Can short-term fan testing provide insight into vapor intrusion. Oral Presentation and Short Paper
AWMA Symposium on Air Quality Measurement Methods and Technology, Durham, NC. April.
Lutes, C., Cosky, B., Uppencamp, R., Abreu, L., Schumacher, B., Zimmerman, J. Truesdale, R., Lin, S,
Hayes, H. and Hartman, B. 2012c. Recent observations on spatial and temporal variability in the
field. Invited presentation at EPA Workshop: Recent Advances to Vapor Intrusion Application
and Implementation " at the 22nd Annual International Conference on Soil, Water, Energy & Air
(AEHS) San Diego, March.
Lutes, C., Cosky, B., Uppencamp, R., Abreu, L., Schumacher, B., Zimmerman, J. Truesdale, R., Lin, S,
Hayes, H. and Hartman, B. 2012d. Short-term variability, radon tracer, and long-term passive
sampler performance in the field. Invited presentation at EPA Workshop: Recent Advances to
Vapor Intrusion Application and Implementation at the 22nd Annual International Conference on
Soil, Water, Energy & Air (AEHS) San Diego, March.
Lutes, C.; R. Studebaker, andN. Weinberg. 2010b. Optimal design of sub slab depressurization systems
for volatile organic compound vapor intrusion (VI) mitigation. Poster presented at AWMA
Specialty Conference: Vapor Intrusion 2010, Chicago, IL, September 28-30. Available at
http://events.awma.org/education/Posters/Final/Lutes OptimalPoster.pdf.
Lutes, C.C., R. Uppencamp, C. Singer, L. Abreu, R. Mosley, sand D. Greenwell. 2009. Radon Tracer as
a Multipurpose Tool to Enhance Vapor Intrusion Assessment and Mitigation. Poster presented at
the Partners in Environmental Technology Technical Symposium and Workshop, Washington,
DC, December.
Lutes, C.C., R. Uppencamp, H. Hayes, D. Greenwell, and R. Mosley. 2010. Long-term integrated
samplers for indoor air and sub-slab soil gas at VI sites. Presented at the 2010 Battelle
Conference, Monterey, CA.
Lutes, C.C., R. Uppencamp, H. Hayes, R. Mosley and D. Greenwell. 2010b. Long-term integrating
samplers for indoor air and subslab soil gas at VI sites. Oral Presentation at AWMA Specialty
Conference: Vapor Intrusion 2010, September 28-30, 2010. Chicago, IL. paper submitted to
proceedings. Available at http://events.awma.org/education/Posters/Final/Lutes RadonPoster.pdf
or http://events.awma.org/education/Final%20Papers/7-Lutes.pdf
Lutes, C.C., R. Uppencamp, H. Hayes, R. Mosley and D. Greenwell. 2010. Long-term integrated
samplers for indoor air and subslab soil gas at VI sites. Oral presentation at Seventh International
Remediation of Chlorinated and Recalcitrant Compounds Conference, Monterey, CA; May 24-
27.
MacLeod, G., and J.M. Ames. 1986. Comparative assessment of the artifact background on thermal
desorption of Tenax GC and Tenax TA. Journal of Chromatography 355:393-398.
MADEP. 2002. Indoor Air Sampling and Evaluation Guide. Boston, MA: Commonwealth of
Massachusetts Executive Office of Environmental Affairs, Office of Research and Standards,
Department of Environmental Protection.
14-7
-------
Section 14—References
Marotta, L., M. Snow, and S. Varisco. 2012. Extending the hydrocarbon range above naphthalene for soil
vapor and air samples using automated thermal desorption/gas chromatography/mass
spectrometry (ATD/GC/MS). Presented at National Environmental Monitoring Conference,
Washington, DC, August 6-10.
McAlary, T.A., P. Dollar, P. de Haven, R. Moss, G. Wilkinson, J. Llewellyn, and D. Crump. 2002.
Assessment of subsurface vapour transport through Triassic sandstone and quarry fill into indoor
air in Weston Village, Runcorn. Presented at Indoor Air 2002, Monterey, CA.
McHugh, T. E., and Nickels, T. N. 2007. Evaluation of spatial and temporal variability in VOC
concentrations at vapor intrusion investigation sites. Conference Proceeding from Vapor
Intrusion: Learning from the Challenges-Air and Waste Management Association (AWMA),
Providence, RI.
McHugh, T., P. De Blanc, and R. Pokluda. 2006. Indoor air as a source of VOC contamination in shallow
soils below buildings. Soil and Sediment Contamination 75:103-22.
McHugh, T.E., T.N. Nichols, and S. Brock. 2007. Evaluation of spatial and temporal variability in VOC
concentrations at vapor intrusion investigation sites. Pp. 129-142 in Proceeding of Air & Waste
Management Association's Vapor Intrusion: Learning from the Challenges, Providence, RI,
September 26-28.
Middleditch, B.S. 1989. Analytical Artifacts: GC, MS, HPLC, TLC and PC. Journal of Chromatography
Library Volume 44. Elsevier, Amsterdam, p. 668-669.
Moses, H., H.F. Lucas and G.A. Zerbe. 1963. The effect of meteorological variables upon radon
concentration three feet above the ground. J. APCA. 73(1): 12-19.
Mosley, R.B. 2007. Use of Radon to Establish a Building-specific Sub-slab Attenuation Factor for
comparison with Similar Quantities measured for Other Vapor Intrusion Contaminants.
Presented at the National Environmental Monitoring Conference, Cambridge, MA, August 19-
25.
Mosley, R.B., D. Greenwell, A. Lee, K. Baylor, M. Plate, and C. Lutes. 2008. Use of Integrated Indoor
Radon and Volatile Organic Compounds (VOCs) to Distinguish Soil Sources from Above-Ground
Sources. Extended abstract and oral presentation, AWMA Symposium on Air Quality
Measurement and Technology, Chapel Hill NC, November 6.
Mosley, R.B., D. Greenwell, and C.C. Lutes. 2010. Use of integrated indoor concentrations of tracer
gases and volatile organic compounds (VOCs) to distinguish soil sources from above-ground
sources. Poster presented at the Seventh International Remediation of Chlorinated and
Recalcitrant Compounds Conference, Monterey, CA, May 24-27.
Murry, D.M., and D.E. Burmaster. 1995. Residential air exchange rates in the United States: Empirical
and estimated parametric distributions by season and climatic region. Risk Analysis 7J(4):459-65.
National Weather Service. 2012. Excessive Heat - An Underrated Problem. National Weather Service
Regional Office, Central Regional Headquarters. Available at
http://www.crh.noaa.gov/Image/lsx/wcm/Heat/ExcessiveHeat.pdf.
14-8
-------
Section 14—References
Nau, R. 2005a. Duke University, Class Notes for Decision 411 Forecasting. Available at
http://people.duke.edu/~rnau/41 ldiff.htm.
Nau, R. 2005b. Duke University, Class Notes for Decision 411 Forecasting. Available at
http: //people .duke. edu/~rnau/411 regou .htm.
Nazaroff, W.W. 1988. Predicting the rate of 222Rn entry from soil into the basement of a dwelling due to
pressure-driven air flow. Radiation Protection Dosimetry 22(1/4): 199-202.
Nazaroff, W.W., and A.V. Nero (Eds.). 1988. Radon and Its Decay Products in Indoor Air. New York:
John Wiley and Sons.
Nazaroff, W.W., S.R. Lewis, S.M. Doyle, B.A. Moed, and A.V. Nero, 1987. Experiments on pollutant
transport from soil into residential basements by pressure-driven airflow. Environmental Science
& Technology 27:459-466.
Neptune and Company. 2007. Report for 2007 Annual Vapor Intrusion Sampling for Buildings N210 and
19 and Additional Baseline Sampling for Building 16. Prepared for ISSi/SAIC NASA Ames
Research Center Moffett Field, California 94035-1100 Neptune Project No 09501.
New York Times. 2012. Toxic Waters: Indianapolis Water. May 26. Available at
http://proiects.nvtimes.com/toxic-waters/contaminants/in/marion/in52490Q4-indianapolis-water.
Newton, E., and R. Rudela. 2007. Estimating correlation with multiply censored data arising from the
adjustment of singly censored data. Environmental Science and Technology 47(l):221-228.
NJ DEP (New Jersey Department of Environmental Protection). 2012. Draft Vapor Intrusion Guidance.
January.
NJ DEP (New Jersey Department of Environmental Protection). 2013. Vapor Intrusion Technical
Guidance. January. Available at
http://www.ni .gov/dep/srp/guidance/vaporintrusion/vig main.pdf
Odabasi, M. 2008. Halogenated volatile organic compounds from the use of chlorine-bleach-containing
household products. Environmental Science and Technology 42:1445-1451.
Odencrantz, J.E., H. O'Neill, S.J. Steinmacher, J.D. Case, and P.C. Johnson. 2008. Residential vapor
intrusion evaluation: long duration passive sampling v. short duration active sampling.
Remediation Journal 18(4):49-54. Available at
htto://onlinelibrarv.wilev.com/doi/10.1002/rem.20181/abstract.
Oury, B., F. Lhuillier, J. Protois, and Y. Morele. 2006. Behavior of the GABIE, 3M3500, PerkinElmer
Tenax TA, and RADIELLO 145 diffusive samplers exposed over a long time to a low
concentration of VOCs. Journal of Occupational and Environmental Hygiene 3(10):547-557.
Parker, G.B. 1985. Measurement of air exchange rates in residential and commercial buildings in the
northwest: techniques and results. PNL-SA-13507. Presented at Conference on Conservation in
Buildings: Northwest Perspective, Butte MT, May 20. Abstract available at
http://www.osti.gov/energvcitations/product.biblio.isp7osti id=5567459.
14-9
-------
Section 14—References
Pennequin-Cardinala, A. et al. 2005. Performances of the Radiellos diffusive sampler for BTEX
measurements: Influence of environmental conditions and determination of modeled sampling
rates. Atmospheric Environment 39:2535-2544.
Perron, P. 1988. Trends and random walks in macroeconomic time series. Journal of Economic Dynamics
and Control 72:297-332.
R Development Core Team. 2012. R: A Language and Environment for Statistical Computing. R
Foundation for Statistical Computing, Vienna, Austria. Available at http://www.R-proiect.org.
Razali, N., and Y.B. Wah. 2011. Power comparisons of Shapiro-Wilk, Kolmogorov-Smirnov, Lilliefors
and Anderson-Darling tests. Journal of Statistical Modeling and Analytics 2( 1):21—33. Available
at http://instatmv.org.mv/downloads/e-iurnal%202/3.pdf. Retrieved 5 June 2012.
Revzan K.L., W.J. Fisk, and A.J. Gadgil. 1991. Modeling radon entry into houses with basements: model
description and verification. Indoor Air 2: 173-189.
Robinson A.L., and R.G. Sextro. 1997. Radon entry into buildings driven by atmospheric pressure
fluctuations. Environmental Science and Technology 31:1742-1748.
Robinson, A.L., R.G. Sextro, and W.J. Fisk. 1997. Soil-gas entry into an experimental basement driven
by atmospheric pressure fluctuations—Measurements, spectral analysis and model comparison.
Atmospheric Environment, 37(10): 1477-1485.
Robinson, A.L., R.G. Sextro, and W.J. Riley. 1997. Soil-gas entry into houses driven by atmospheric
pressure fluctuations - the influence of soil properties. Atmospheric Environment 31:1487-1495.
Rorech, G. J. 2001. Vacuum Extraction and Bioventing, Chapter 3. In E.K. Nyer, et al. (Ed.). In-situ
Treatment Technology, 2nd edition, Lewis.
Ryan, J.V, P.M. Lemieux, and W.T Preston. 1998. Near-real-time measurement of trace volatile organic
compounds from combustion processes using an on-line gas chromatograph. Waste Management
7S(6):403-410.
Said, S.E. and D.A. Dickey. 1984. Testing for unit roots in autoregressive-moving average models of
unknown order. Biometrika 71:599-601.
Scheeringa, K., and K. Hudson. 2011-2013.Monthly Climate Summaries. Indiana State Climate Office.
Available at http://iclimate.org/siimmarv.asp.
Schumacher, B., J. Zimmerman, C. Lutes, B. Cosky, R. Truesdale, R. Norberg and B. Hartman. 2013.
Indoor Air and Soil Gas Temporal Variability: Effects on Sampling Strategies Evidence from
Controlled and Uncontrolled Conditions. Oral Presentation at an EPA Organized Workshop,
Looking Beyond Natural Variation in Vapor Intrusion: Understanding, Controlling and
Addressing Site Variables for Improved Practices and Effective Protection Strategies, at the 23rd
Annual International Conference on Soil, Water, Energy and Air, San Diego, CA. March 19.
Schumacher, B., J. Zimmerman, G. Swanson, J. Elliot, and B. Hartman. 2010. Field observations on
ground covers/buildings. Presented at the 20th Annual International Conference on Soils,
Sediments, Water, and Energy, San Diego, CA. March 16.
14-10
-------
Section 14—References
Schuver, H. 2013. Radon as a tracer and risk driving co-contaminant in chemical vapor intrusion.
Presentation at RemTec 2013 Summit, Westminster, CO, March 6.
Schuver, H., and L. Siegel. 2011. Vapor intrusion: involved-stakeholder awareness of the uncertainty (and
multiple benefits of controls). Community Involvement Training Conference, Crystal City VA,
July. Available at
http://epa.gov/ciconference/download/presentationsAVed QpenTime%20Session Salon6 Schuve
rSiegel.pdf.
Schuver, H.J., and R.B. Mosley. 2009. Investigating vapor intrusion with confidence and efficiency (some
observations from indoor air-based radon intrusion studies). A WMA Vapor Intrusion 2009, San
Diego, CA.
Sigma-Aldrich. 2012. Radiello Manual, Volatile Organic Compounds (VOCs) Chemically Desorbed with
CS2, Supelco Edition. Available at http://www.sigmaaldrich.com/content/dam/sigma-
aldrich/docs/Supelco/Application Notes/radiello dl d6.pdf. accessed August 2012.
Smajstrla, A.G., and D.S. Harrison. 1998. Tensiometers for Soil Moisture Measurement and Irrigation
Scheduling. CIR487. Agricultural and Biological Engineering Department, Florida Cooperative
Extension Service, Institute of Food and Agricultural Sciences, University of Florida. Available
at http://edis.ifas.ufl.edu/ael46.
Steck, D.J. 2012. An update on the draft of "Lessons from radon for vapor intrusion research and
programs". Presented at EPA-AEHS Workshop, Recent Advances to VI Application &
Implementation, San Diego, CA. March.
The Polis Center. 2012. Study Neighborhoods from the Project on Religion and Urban Culture:
Mapleton-Fall Creek. Footnote 2. The Polis Center, Indianapolis, IN. Available at
http://ww\\ .polis. iupui.cdu/RUC7Ncighborhoods/MaplctonFallCrcck/MFCNarrativc.htm# ftn2.
Thomas, R.G. 1990. Volatilization from Soil, Chapter 16. In W.J. Lyman, et al. (Eds.). Handbook of
Chemical Property Estimation Methods. American Chemical Society.
Thomson, N.R. and D.J. Flynn. 2000. Soil vacuum extraction of perchloroethylene from the Borden
aquifer. Groundwater 38(5): 673-688.
Truesdale, R., H. Dawson, and I. Hers. 2005. Vapor intrusion database status and updates. Presented at
Specialty Workshop on Integrating Observed and Modeled Vapor Attenuation, AEHS 15th
Annual West Coast Conference on Soils, Sediments and Water, San Diego, CA. March 14.
Available at http://iavi.rti.org/attachments/Resources/1045 -
Truesdale Vapor Intrusion Database Status and Updates.pdf.
U.S. DOE (Department of Energy). Innovative Technology Summary Report: Frozen Soil Barrier.
DOE/EM-0483 October 1999.
U.S. EPA (Environmental Protection Agency). 1990. NAREL Standard Operating Procedures for Radon-
222 Measurement Using Diffusion Barrier Charcoal Canisters. EPA 520/5-90/032 Stock Number
PB91-179002. U.S. EPA National Air and Radiation Environmental Lab., Montgomery AL.
14-11
-------
Section 14—References
U.S. EPA (Environmental Protection Agency). 1992. Indoor Radon and Radon Decay Product
Measurement, Device Protocols. EPA 402-R-92-004, Office of Radiation Programs, Washington,
DC, July (revised).
U.S. EPA (Environmental Protection Agency). 1993a. Protocols for Radon and Radon Decay Product
Measurements in Homes. EPA-402-R-92-003, May 1993a.
U.S. EPA (Environmental Protection Agency). 1993b. Radon Reduction Techniques for Existing
Detached Houses, Technical Guidance (third edition) for Active Soil Depressurization Systems.
EPA/625/R-93/011.
U.S. EPA (Environmental Protection Agency). 1999a. Compendium of Methods for the Determination of
Toxic Organic Compounds in Ambient Air, Second Edition, Compendium Method TO-14Ab :
Determination of Volatile Organic Compounds (VOCs) in Ambient Air Using Specially Prepared
Canisters With Subsequent Analysis By Gas Chromatography. EPA/625/R-96/010b. Available at
http://www.epa.gov/ttnamtil/files/ambient/airtox/to-14ar.pdf.
U.S. EPA (Environmental Protection Agency). 1999b. Compendium of Methods for the Determination of
Toxic Organic Compounds in Ambient Air Second Edition Compendium Method TO-15
Determination of Volatile Organic Compounds (VOCs) in Air Collected in Specially-Prepared
Canisters and Analyzed by Gas Chromatography Mass Spectrometry (GC/MS). EPA/625/R-
96/010b. Available at http://www.epa.gov/ttnamti 1 /files/ambient/airtox/to-15 r .pdf.
U.S. EPA (Environmental Protection Agency). 1999c. Compendium of Methods for the Determination of
Toxic Organic Compounds in Ambient Air, Second Edition, Compendium Method TO-17,
Determination of Volatile Organic Compounds in Ambient Air Using Active Sampling Onto
Sorbent Tubes. EPA/625/R-96/010b. Available at
http: //www .epa. gov/ttnamti 1 /file s/ambient/airtox/to-17r.pdf.
U.S. EPA (Environmental Protection Agency). 2002a. OSWER Draft Guidance for Evaluating the Vapor
Intrusion to Indoor Air Pathway from Groundwater and Soils (Subsurface Vapor Intrusion
Guidance). EPA530-D-02-004. Available at http://www.epa.gov/correctiveaction/eis/vapor.htm
U.S. EPA (Environmental Protection Agency). 2002b. Guidance on Environmental Data Verification and
Validation. EPA QA/G-8 EPA/240/R-02/004. November. Available at
http: //www .epa. gov/qualitv/q s-docs/g 8 -final .pdf.
U.S. EPA (Environmental Protection Agency). 2003. A Standardized EPA Protocol for Characterization
of Indoor Air Quality in Large Office Buildings. U.S. EPA Indoor Air Division and Atmospheric
Research and Exposure Assessment Laboratory. Available at http://www.epa.gov/iaq/
base/pdfs/2003_base_protocol.pdf.
U.S. EPA (Environmental Protection Agency). 2005a. DRAFT Standard Operating Procedure (SOP) for
Installation of Sub-Slab Vapor Probes and Sampling Using EPA Method TO-15 to Support Vapor
Intrusion Investigations. Available at http://www.epa.gov/region8/r8risk/pdf/epa sub-
slabvapor.pdf.
U.S. EPA (Environmental Protection Agency). 2005b. DRAFT Assessment of Vapor Intrusion in Homes
Near the Former Raymark Superfund Site - Recommendations for Testing at Other Sites.
14-12
-------
Section 14—References
U.S. EPA (Environmental Protection Agency). 2008. Engineering Issue: Indoor Air Vapor Intrusion
Mitigation Approaches. EPA/600/R-08-115. National Risk Management Research Laboratory,
Cincinnati, OH. http://www.clu-in.org/download/char/600r08115 .pdf
U.S. EPA (Environmental Protection Agency). 2008b. U.S. EPA Contract Laboratory Program National
Functional Guidelines for Superfund Organic Methods Data Review. EPA-540-R-08-01. Office
of Superfund Remediation and Technology Innovation, Washington, DC. June. Available at
http ://epa. gov/supcrfund/prograins/clp/download/soinnfg.pdf.
U.S. EPA (Environmental Protection Agency). 2011. Exposure Factors Handbook 2011 Edition (Final).
EPA/600/R-09/052F. 2011. Office of Research and Development, Washington, DC.
U.S. EPA (Environmental Protection Agency). 2012a. Fluctuation of Indoor Radon and VOC
Concentrations Due to Seasonal Variability. EPA/600/R-12/673. Office of Research and
Development, Las Vegas, NV. September. Available at http://www.clu-
in.org/download/contaminantfocus/vi/VI-EPA600-R-09-073.pdf
U.S. EPA (Environmental Protection Agency). 2012b. EPA On-line Tools for Site Assessment
Calculation: Johnson and Ettinger Attenuation Factor. U.S. EPA Ecosystems Research, Athens,
GA. Available at http ://www. epa. gov/athens/learn2model/part-two/onsite/i ne alpha.htm.
U.S. EPA (Environmental Protection Agency). 2012c. EPA's Vapor Intrusion Database: Evaluation and
Characterization of Attenuation Factors for Chlorinated Volatile Organic Compounds and
Residential Buildings. EPA 530-R-10-002. Office of Solid Waste and Emergency Response,
Washington, DC. March. Available at
http://www.epa.gov/oswer/vaporintrusion/documents/OSWER 2010 Database Report 03-16-
2012 Final witherratum 508.pdf
U.S. EPA (Environmental Protection Agency). 2012d. Conceptual Model Scenarios for the Vapor
Intrusion Pathway. Office of Solid Waste and Emergency Response, Washington, DC. Available
at http://www.epa.gov/oswer/vaporintrusion/documents/vi-cms-vllfinal-2-24-2012.pdf.
U.S. EPA (Environmental Protection Agency). 2012e. Drinking Water Contaminants. Available at
http://water.epa.gov/drink/contaminants/index.cfm.
U.S. EPA (Environmental Protection Agency). 2012f. Petroleum Hydrocarbons and Chlorinated
Hydrocarbons Differ In Their Potential for Vapor Intrusion. Information Paper. Office of
Underground Storage Tanks, Washington, DC. March. Available at
http: //www .epa. gov/oust/cat/pvi/pvicvi .pdf.
U.S. EPA (Environmental Protection Agency). 2012g. A Citizen's Guide to Radon. EPA 402-K-12-002.
Office of Air and Radiation, Indoor Environments Division. Washington, DC. May.
http: //www .epa. gov/radon/pubs/citguide .html.
U.S. EPA (Environmental Protection Agency). In press. Geophysical Investigations at the Indianapolis
Vapor Intrusion Study House, 420 and 422 East 28th St., Indianapolis, IN. EPA/600/R-13/184.
Office of Research and Development, National Exposure Research Laboratory. Las Vegas, NV.
University of Minnesota. 2008. Re-Arch: The Initiative for Renewable Energy in Architecture. Fact Sheet.
University of Minnesota Initiative for Renewable Energy in Architecture (re-ARCH). Available
at http://www.rearch.umn.edu/factsheets/VentilationFactSheet.pdf.
14-13
-------
Section 14—References
Wang, F., and I.C. Ward. 2000. The development of a radon entry model for a house with a cellar.
Building and Environment 35:615-631.
Wania, F., J.T. Hoff, C.Q. Jia, and D. Mackay. 1998. The effects of snow and ice on the environmental
behaviour ofhydrophobic organic chemicals. Environmental Pollution 102(l):25-4\.
Wertz, W., and T. Festa. 2007. The patchy fog model of vapor intrusion. Pp. 28-36 in Proceedings of
AWMA Conference on Vapor Intrusion: Learning from the Challenges, Providence, RI,
September 26-28. Pittsburgh, PA: Air and Waste Management Association.
White, R.G. 1964. Determination of carbon disulfide in benzene by ultraviolet spectrophotometry.
Applied Spectroscopy 18(4): 112-113.
Whitmore, A., and R.L. Corsi. 1994. Measurement of gas-liquid mass transfer coefficients for volatile
organic compounds in sewers. Environmental Progress May: 114-123.
Wickham, H. 2009. ggplot2: Elegant Graphics for Data Analysis. New York: Springer.
Wilson, L.G., L.G. Everett, and S.J. Cullen (Eds.). 1995. Handbook of Vadose Zone Characterization and
Monitoring. Boca Raton, FL: Lewis Publishers.
Winberry, W. T., L. Forehand, N.T. Murphy, A. Ceroli, B. Phinney, and A. Evans. 1990. Compendium of
Methods for the Determination of Air Pollutants in Indoor Air. EPA/600/4-90/010. U.S.
Environmental Protection Agency, Office of Research and Development, Research Triangle Park,
NC. April.
Wisbeck D., C. Sharpe, A. Frizzell, C. Lutes, and N. Weinberg. 2006. Using naturally occurring radon as
a tracer for vapor intrusion: a case study. ARCADIS, presented at the 2006 Society of Risk
Analysis (SRA) Annual Meeting, Baltimore, MD.
Yamazawaa, H., T. Miyazakia, J. Moriizumia, T. Iidaa, S. Takedab, S. Nagarab, K. Satob, and T.
Tokizawab. 2005. High levels of natural radiation and radon areas: radiation dose and health
effects. Proceedings of the 6th International Conference on High Levels of Natural Radiation and
Radon Areas. International Congress Series. 1276:221-222. February. Available at
http://www.sciencedirect.com/science/article/pii/S0531513104017947
Yao, Y., K.G. Pennell, and E.M. Suuberg. 2010.Proceedings of the Air & Waste Management
Association's Vapor Intrusion 2010 Conference. September 29-30. Available at http://www.clu-
in.org/download/contaminantfocus/vi/The%20Influence%20of%20Transient%20Processes.pdf
Yeates, G.L., and D.M. Nielsen. 1987. Design and implementation of an effective soil gas monitoring
program for four-dimensional monitoring of volatile organics in the subsurface. In Proceedings of
the NWWA Focus Conference on Ground Water Issues. Indianapolis, IN. April 21-23.
Zhao, Y., and C. Frey. 2006. Uncertainty for data with non-detects: air toxic emissions from combustion.
Human and Ecological Risk Assessment 12:1171-1191.
14-14
-------
Appendix A—Radon Mitigation System Photos and Field Diagnostic Report
Appendix A.
Radon Mitigation System Photos and Field Diagnostic Report
A-l
-------
Appendix A—Radon Mitigation System Photos and Field Diagnostic Report
A-2
-------
Appendix A—Radon Mitigation System Photos and Field Diagnostic Report
MITIGATION PHOTOS
#5 #6
422 East 28th Street, Indianapolis, IN 46205
PDF created with pdfFactory trial version www.pdffactorv.com
PDF created with pdfFactory trial version www.odffactorv.com
A-3
-------
Appendix A—Radon Mitigation System Photos and Field Diagnostic Report
MITIGATION PHOTOS
2 422 East 28th Street, Indianapolis, IN 46205
PDF created with pdfFactory trial version www.pdffactorv.com
PDF created with pdfFactory trial version www.pdffactorv.coni
A-4
-------
Appendix A—Radon Mitigation System Photos and Field Diagnostic Report
MITIGATION PHOTOS
422 East 28th Street, Indianapolis, IN 46205
PDF created with pdfFactory trial version www.pdffactorv.com
PDF created with pdfFactory trial version www.pdffactorv.coni
A-5
-------
Appendix A—Radon Mitigation System Photos and Field Diagnostic Report
MITIGATION PHOTOS
422 East 28th Street, Indianapolis, IN 46205
PDF created with pdfFactory trial version www.pdffactorv.com
PDF created with pdfFactory trial version www.pdffactorv.coni
A-6
-------
Appendix A—Radon Mitigation System Photos and Field Diagnostic Report
Pressure Field Diagnostic Report
ARCADIS422 East 28th Street, Indianapolis, IN 46205
Notto scale
SUfJ SLAB SUCTION PIT
SUB SLAB PRESSURE TEST PORT
RP 265 VAPOR MITIGATION FAN
VAPOR EXHAUST PIPING
SYSTEM FAILURE INDICATOR
Diagnostics with two suction lines activated on the North end of the building.
1) -.155", 2) -.058", 3) -.020", 4) -.018", 5) -J006", 6) J035", 7) -.038", 8) JD17",9) -D11", 10)-.003"
Diagnostics with all four suction lines activated.
1) -.092", 2) -.089", 3) -.046", 4) -.046", 5) .009", 6) .065", 7) D66", 8) -D40",9) -D35", 10)-.006"
Guaranteed Radon Reduction
Protecting Indiana Homeowners Since 1988
PO Box 4146 - CarmeJ, IN 46C62-4146 • |317> 843-OB04
www.radonen * iron mentat.com
This form created on WD I FX™
Copyrlgl tzi 12 O- New Image Soma re- All rlglfc reamed.
PDF created with pdfFactory trial version wwiw.pdffactotv.com
PDF created with pdfFactory trial version www.Ddffactorv.corn
A-7
-------
Appendix A—Radon Mitigation System Photos and Field Diagnostic Report
Air Pressure Field Diagnostic Revised Report
1) The sequence 'you went through of changing the blower a nd the reasoning for It?
to determine what style of fan unit to install for our mitigation system we lake into account several factors.
We look at the age of the building, soil type under the slab, jir pressure field diagnostics arid the air
pressure reading on the system failure indicator (Mini U-TubeS- Before arty *ork is performed, we make
educated guesses on what type of fan unit by the age of the building, Usually older buildings prove to be
more difficult in moving air under the slab. This lack of air flow is typically clue to the lack of gravel, and
perforated drain tile, Also an older building will usually have tigfrtier soil under the slab, The combmation
of these factors usually means we will install a high pressure Ian unit, which is d«tgn«i to handle a greater
air pressure/work load to pull air through the tightier soil,
in this particular installation process, we had a home built in the early 1900's with Sand/Dirt/Coal Slag as
the soil mix. These initial factors normally indicate a high pressure fan will be needed. Chris Jordan, the
lead technician installed a high pressure fan in the beginning of the installation process, which seemed to
be the right choice with only two suction points according to our air pressure field diagnostics and more
importantly our U-Tube pressure readings (Inches Per Water Column). However, Chris discovered that after
installing the third and fourth suction points, a high pressure fan was no longer needed as the additional
suction points increased the negative air pressures and allowed the fan unit to not have to work as hard to
move air through the PVC piping (meaning the pressure readings on the Mini U-Tube monitor dropped
below 2"),
When the pressure reading on the Mini U-Tube monitor drops below 2 inches per water column (2" is a
measurement we use as a company guideline for our installation process. A fan manufacturer lists the max
pressure a fan unit can handle and most of the high flow fan units can handle greater pressures than 2", but
our experience has found they tend to not have a very long life span if they are running above 2" on the U-
Tube monitor), then a high pressure fan unit is no longer required. Chris at that point installed a larger
high flow Radon Away fan unit (RP 26S max pressure of 2.5"/water column). The initial high pressure fan
unit installed was a Radon Away GP 501 with a max pressure limit of 4.2". The high flow fans move more
air than the high pressure fan units. Ultimately the more air we can move under the slab the better chance
of success we will have in lower the radon and vapor levels. The pressure readings for each U-Tube after
the full installation was complete are written on our company's labels next to the U-Tubes,
2) What instrument was being used?
To measure air pressure under the slab we use an Infiltec Digital Micro-Manometer and to measure the
pressures of air flow through the PVC piping we use a Mini U-Tube Failure Indicator Monitor.
3) What date the measurements were matte on?
4) Whether these are single data points or averages over some time?
The installation process began on Tuesday, Ocotber 16th, 2012. That first day we performed a
preliminary air pressure field diagnostic test and air pressure readings after each suction point was
installed. The report has two sets of measurements listed, The first set of readings arc* air pressure
measurements after the second suction line was installed, The diagnostic measurements help us guage
PDF created with pdfFactory trial version www.pdffectorv.com
A-8
-------
Appendix A—Radon Mitigation System Photos and Field Diagnostic Report
how much dirt/gravel to dig out from each suction point and how many suction points are needed.
Wednesday, October 17th was the final day of the installation. The air pressure readings on
October 17th were taken after all four suction lines were installed. The measurements provided are not
an average, rather one last single reading after the technician was satisfied with the design and
performance of the system.
PDF created with pdfFaclory trial version www.pdffoctorv.com
A-9
-------
Appendix A—Radon Mitigation System Photos and Field Diagnostic Report
A-10
-------
Appendix B—Plots Comparing Onsite GC with Weekly Radiello Indoor
Air Results by Chemical, Sampling Period, and Sample Location
Appendix B.
Plots Comparing Onsite GC with Weekly Radiello Indoor Air
Results by Chemical, Sampling Period, and Sample Location
B-l
-------
Appendix B—Plots Comparing Onsite GC with Weekly Radiello Indoor
Air Results by Chemical, Sampling Period, and Sample Location
B-2
-------
Appendix B—Plots Comparing Onsite GC with Weekly Radiello Indoor
Air Results by Chemical Sampling Period, and Sample Location
Hartmanl - Chloroform
CD
E 1.0
?
c
o
OS
g 0.5
o
O
420BaseS
A
A
Cf
#"
*
Oc
Missing values
in GC = 1/2 MDL
Missing values _ Missing Values _ Darii_M_
in GC = MDL in GC = NA
~GC
B-3
-------
Appendix B—Plots Comparing Onsite GC with Weekly Radiello Indoor
Air Results by Chemical Sampling Period, and Sample Location
Hartmanl - Chloroform
420First
Missing values
in GC = 1/2 MDL
Missing values
in GC = MDL
Missing Values
in GC = IMA
Radiello
GC
B-4
-------
Appendix B—Plots Comparing Onsite GC with Weekly Radiello Indoor
Air Results by Chemical Sampling Period, and Sample Location
1.5
cn
E
5
c
o
*00
o
1.0
O 0-5
0.0
Hartmanl - Chloroform
\
422BaseS
&
s
a
&
<9
s
\
oc:
SMissing values
in GC = 1/2 MDL
Missing values
in GC = MDL
Missing Values-Radje||o
in GC = NA
B-5
-------
Appendix B—Plots Comparing Onsite GC with Weekly Radiello Indoor
Air Results by Chemical Sampling Period, and Sample Location
1.00-
cn
E 0.75'
CD
O
"-5 °-50'
to
-------
Appendix B—Plots Comparing Onsite GC with Weekly Radiello Indoor
Air Results by Chemical Sampling Period, and Sample Location
CD
1.0
5
c
.2
C3
<1)
o 0.5
c
o
O
0.0
Hartman2 - Chloroform
420BaseS
111
J IL
&
&
&
$
(fr
S Missing values Missing values Missing Values RaHjp|in
in GC = 1/2 MDL in GC = MDL in GC = NA
~gc
B-7
-------
Appendix B—Plots Comparing Onsite GC with Weekly Radiello Indoor
Air Results by Chemical, Sampling Period, and Sample Location
Hartrnan2 - Chloroform
420First
B Missing values Missing values Missing Values Dariipiin
in GC = 1/2 MDL " in GC = MDL in GC = NA
0GC
B-8
-------
Appendix B—Plots Comparing Onsite GC with Weekly Radiello Indoor
Air Results by Chemical Sampling Period, and Sample Location
1.5
CD
E
?
c
_o
CO
1.2 -
-------
Appendix B—Plots Comparing Onsite GC with Weekly Radiello Indoor
Air Results by Chemical Sampling Period, and Sample Location
Hartman2 - Chloroform
422First
*
i 'toi,
0Missing values
in GC = 1/2 MDL
Missing values
in GC = MDL
Missing Values
in GC = NA
Radiello
GC
B-10
-------
Appendix B—Plots Comparing Onsite GC with Weekly Radiello Indoor
Air Results by Chemical, Sampling Period, and Sample Location
Hartrnan2 - Chloroform
Outside
3"
co
E
3
c
o
CO
0
o
c
o
O
2-
0-
I III I
$
&
&
$>
#
&
*
-------
Appendix B—Plots Comparing Onsite GC with Weekly Radiello Indoor
Air Results by Chemical Sampling Period, and Sample Location
Hartman3 - Chloroform
420BaseS
>*>
0e° ^ <<$> ^ ^
Missing values
in GC = 1/2 MDL
Missing values
in GC = MDL
GC
B-12
-------
Appendix B—Plots Comparing Onsite GC with Weekly Radiello Indoor
Air Results by Chemical Sampling Period, and Sample Location
Hartman3 - Chloroform
4 2 0 F i rst
0.6
cn
o
o
0.4
o
O 0.2
^ ^ ^
^ ^
BMissing values
in GC = 1/2 MDL
Missing values Missing Values
in GC = MDL in GC = NA
Radiello
GC
B-13
-------
Appendix B—Plots Comparing Onsite GC with Weekly Radiello Indoor
Air Results by Chemical Sampling Period, and Sample Location
Hartman3 - Chloroform
422BaseS
en
j= 2
5
c
_o
0) .
O T
c
o
O
O
$
«#Vt *
HMissing values Missing values Missing Values RaHjP||n
in GC = 1/2 MDL in GC = MDL in GC = NA
E GC
B-14
-------
Appendix B—Plots Comparing Onsite GC with Weekly Radiello Indoor
Air Results by Chemical Sampling Period, and Sample Location
HartmanS - Chloroform
422First
I I I II
^ <$>
^ <;& ^
0Missing values
in GC = 1/2 MDL
Missing values
in GC = MDL
Missing Values
in GC = NA
Radiello
GC
B-15
-------
Appendix B—Plots Comparing Onsite GC with Weekly Radiello Indoor
Air Results by Chemical Sampling Period, and Sample Location
Hartman3 - Chloroform
Outside
1 00
-------
Appendix B—Plots Comparing Onsite GC with Weekly Radiello Indoor
Air Results by Chemical Sampling Period, and Sample Location
Hartmanl - Tetrachloroethene
——I * *
• i ** I
~ * ••••>;• • *****
' •• ~ * -M. V. -
H Missing values Missing values Missing Values ~ar|jfl||ri
in GC = 1/2 MDL in GC = MDL in GC = NA
~gc
B-17
-------
Appendix B—Plots Comparing Onsite GC with Weekly Radiello Indoor
Air Results by Chemical Sampling Period, and Sample Location
Hartmanl - Tetrachloroethene
420First
_A_
v
4*
*
• *
TR—-
*
L ~ ~
:/
4
V
~
IIII
II I
$
\
0C
oc
H Missing values Missing values Missing Values ~ar|jfl||ri
in GC = 1/2 MDL in GC = MDL in GC = NA
~gc
B-18
-------
Appendix B—Plots Comparing Onsite GC with Weekly Radiello Indoor
Air Results by Chemical, Sampling Period, and Sample Location
Hartmanl - Tetrachloroethene
™ 1.0
S Missing values Missing values , Missing Values RaHipnn
in GC = 1/2 MDL in GC = MDL in GC - NA
GC
B-19
-------
Appendix B—Plots Comparing Onsite GC with Weekly Radiello Indoor
Air Results by Chemical Sampling Period, and Sample Location
H artman 1- T etrachloroethene
422First
%
<9
\
\
• * w.
~ *
~ • * ,
~ ~
! t
%
* *J
i ¦ •«
Mb,
v.
* JW1
liii
¥
A
0l
£
€
A
0C
Missing values
in GC = 1/2 MDL
Missing values
in GC = MDL
Missing Values
in GC = NA
Radiello
GC
B-20
-------
Appendix B—Plots Comparing Onsite GC with Weekly Radiello Indoor
Air Results by Chemical Sampling Period, and Sample Location
Hartman2 - Tetrachloroethene
420B
aseS
•
~
•
•>
•
%
*
•
#
fr ~~
t
fr * ~
a th * ^ i
t *
Jfk
1
~i * ~
—•—*
1
*
> *
%
~
*%
* £ m *
TTT!
* it •
i fr * « *
r I
%
• • <
» i «?• *
~ <
hV.U fr
* * W £
*
II
—u 1,
—
V
&
&
&
&
&
{&
0Missing values
in GC = 1/2 MDL
Missing values
in GC = MDL
Missing Values
in GC = NA
Radiello
GC
B-21
-------
Appendix B—Plots Comparing Onsite GC with Weekly Radiello Indoor
Air Results by Chemical Sampling Period, and Sample Location
Hartman2^ T etrachloroethene
420First
*
*
%
Jw
&
£
J_L_
A
&
$
s&
$
&
Missing values Missing values Missing Values Dar|ifa||n
in GC = 1/2 MDL in GC = MDL in GC = NA
Egc
B-22
-------
Appendix B—Plots Comparing Onsite GC with Weekly Radiello Indoor
Air Results by Chemical Sampling Period, and Sample Location
Hartman2 - Tetrachloroethene
Missing values Missing values Missing Values paii-Hn
in GC = 1/2 MDL I in GC = MDL in GC = NA Kaaieno
GC
B-23
-------
Appendix B—Plots Comparing Onsite GC with Weekly Radiello Indoor
Air Results by Chemical Sampling Period, and Sample Location
Hartman2 - Tetrachloroethene
422First
~
t
J
~
•
# a
*
t
*
*
I
I
~ .
* *
H
•
•
~ ~
%
W
4 *
V
%
— - *
t*
-
~
~
~»
¦W !
* * « •
"X
J* s
is?
•
¦*1B
1#
ff
ii~ '
'
TfTM
~
r-
.
-------
Appendix B—Plots Comparing Onsite GC with Weekly Radiello Indoor
Air Results by Chemical Sampling Period, and Sample Location
Hartman2 - Tetrachloroethene
Outsid
e
1
~
~
iUl
t 1
r —
f
-------
Appendix B—Plots Comparing Onsite GC with Weekly Radiello Indoor
Air Results by Chemical Sampling Period, and Sample Location
Hartman3 - Tetrachloroethene
420BaseS
*
«»
•
•
#
*
•
* t
~
•
#
#
*
~
* J. »' .' i MLri
~ *
'V
—I 1 ,
,
&
,¦
-------
Appendix B—Plots Comparing Onsite GC with Weekly Radiello Indoor
Air Results by Chemical Sampling Period, and Sample Location
Hartman3 - Tetrachloroethene
420 First
*
$> <<& ^ ^
H Missing values Missing values Missing Values
in GC = 1/2 MDL in GC = MDL in GC = NA
0
Radiello
GC
B-27
-------
Appendix B—Plots Comparing Onsite GC with Weekly Radiello Indoor
Air Results by Chemical Sampling Period, and Sample Location
Hairtman3 - Tetrachloroethene
E 10
a
-------
Appendix B—Plots Comparing Onsite GC with Weekly Radiello Indoor
Air Results by Chemical Sampling Period, and Sample Location
Hartman3 - Tetrachloroethene
422First
~
•
~
#
*
~ •
~
•
*
~
*
*
* 1
* t H
~ t
»'!
:
#
•
1
*
•
•
4
~
ijj
• •*-
$
>**
~
«•
~ .
* H
*
*
•
%
.H
a
1 H —
i
51
iTv *
r ^
~
[>T *
~ V
CO
E
o>
o
c
o
O
0
&
s^
&
& , ^
H Missing values Missing values Missing Values Rar|iP||n
in GC = 1/2 MDL in GC = MDL in GC - NA
~¦GC
B-29
-------
Appendix B—Plots Comparing Onsite GC with Weekly Radiello Indoor
Air Results by Chemical, Sampling Period, and Sample Location
B-30
-------
Appendix C—Additional Statistical Analysis of Effect of
Mitigation on Indoor VOC Concentrations
Appendix C
Additional Statistical Analysis of Effect of Mitigation on
Indoor VOC Concentrations
c-i
-------
Appendix C—Additional Statistical Analysis of Effect of
Mitigation on Indoor VOC Concentrations
C-2
-------
Appendix C—Additional Statistical Analysis of Effect of
Mitigation on Indoor VOC Concentrations
Last ran on 09/14/2013
How much do indoor air concentrations differ between mitigation "On" and mitigation
"Off"?
To address this question we used the standard VOC measurement in our study: weeklong Radiello passive
samplers.
The mitigation system was installed on October 16, 2012 and run on 1 of 3 possible settings: "On." "Off,"
or "Passive." Based on analyses elsewhere in this report we consider "Not Yet Installed," "Passive," and
"Off to be equivalent. These all appear as "Off' in this section. It is also important to consider that the
heating system was used during some, but not all, of the mitigation testing. Only the 422 side was heated
however, allowing its effect to be isolated. The on/off dates for both mitigation and heating are shown in
calendar format in Figure C-l below.
Heating a Off On
Mitigation Status: HOff 171 On
September 2012
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
December 2012
1
2
3
4
5
6
7
8
9
10
11
12
m
¦
m
¦
31
March 2013
October 2012
1
2
3
4
5
_6_
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
January 2013
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
o
CM
24
25
26
April 2013
mum
November 2012
E
0
~
¦
15
16
17
18
22
23
24
25
26 27
29
30
i i
February 2013
I
1 2
n
n
n
May 2013
I
1
2
3
4
5
6
7
8
9
10
11
1213
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31i
Figure C-1. Calendar of mitigation testing and heating status.
The Radiello samples were taken from several locations on each side of the duplex. The observed
chloroform concentrations at each location are graphed in Figure C-2 and the PCE concentrations are
C-3
-------
Appendix C—Additional Statistical Analysis of Effect of
Mitigation on Indoor VOC Concentrations
plotted in Figure C-3. Note that both chloroform and PCE show a marked and continuing drop off in all
indoor air concentrations corresponding to the last period of mitigation on (February through April).
Chloroform
0.1
CO
Q5
£
O
"S3
<5 0.1
o
c
o
O
1:
0.1 -
420
ii
W,
Hf
~i 1 1 1 1 1 r
422
-
I 4
Tin rr
in iu
17
u
n
TU
~i 1 1 1 1 1 r
~D
Cj
GO
ns
ro
£1)
on
a>
GO
Mitigation
pOff
On
j!
—l"
GO
2012 -2013
Figure C-2. Radiello chloroform results during mitigation testing.
C-4
-------
Appendix C—Additional Statistical Analysis of Effect of
Mitigation on Indoor VOC Concentrations
CD
0.1
(0
s—
C
o>
c 0
o
O
.1
0.1
Tetrachloroethene
420
¦ i i r
"i 1 r
422
TP
rap
IE
-i 1 r 1 1 1 1 r
2012 -2013
Mitigation
— Off
On
Figure C-3. Radiello PCE results during mitigation testing.
Boxplots juxtaposing the chloroform and PCE concentration values with mitigation "On" and "Off' at
each location are given below (Figure C-4). In all cases the distribution of mitigated values is lower than
for non-mitigated conditions.
C-5
-------
Appendix C—Additional Statistical Analysis of Effect of
Mitigation on Indoor VOC Concentrations
Effect of Mitigation
on
Indoor Air VOC Concentration
1 =
0.11
E
1 =
c
o
03
0.1 =
1 =
-------
Appendix C—Additional Statistical Analysis of Effect of
Mitigation on Indoor VOC Concentrations
The two populations are normally distributed
Chemical concentration data are rarely normal, but often log-normally distributed. This is a simple
transformation of the data that is has a relevant interpretation, so we tested for both normality and log-
normality using the Shapiro-Wilk test (Table C-l).
Table C-1. Results of Shapiro-Wilk Normality Testing
Side
Location
Mitigation
Compound
N
Percent J Flags
P.Lognormal
422
BaseS
Off
CHCb
13
0
0.685
422
BaseS
Off
PCE
13
0
0.153
420
First
Off
CHCb
13
69.2
0.003
420
First
Off
PCE
13
23.1
0.315
420
BaseS
Off
CHCb
17
17.6
0.004
420
BaseS
Off
PCE
17
5.9
0.101
422
BaseN
Off
CHCb
13
0
1
422
BaseN
Off
PCE
13
0
0.089
422
First
Off
CHCb
13
0
0.18
422
First
Off
PCE
13
0
0.31
420
BaseN
Off
CHCb
13
30.8
0.134
420
BaseN
Off
PCE
13
7.7
0.719
420
First
On
CHCb
15
26.7
0.029
420
First
On
PCE
15
46.7
0.738
420
BaseS
On
CHCb
29
31
0.003
420
BaseS
On
PCE
29
13.8
0.071
420
BaseN
On
CHCb
15
26.7
0.008
420
BaseN
On
PCE
15
13.3
0.69
422
BaseN
On
CHCb
15
33.3
0.53
422
BaseN
On
PCE
15
6.7
0.043
422
First
On
CHCb
15
46.7
0.646
422
First
On
PCE
15
20
0.174
422
BaseS
On
CHCb
15
13.3
0.143
422
BaseS
On
PCE
15
6.7
0.075
There are six subsets that "failed" the test (those with p < 0.05 are unlikely to have come from a log-
normal distribution), but most locations and conditions (75%) do not. These results are encouraging and
we will proceed with the assumption that log-normality is a defensible assumption for these data.
The two populations have the same variance
An F-test can be used to test this for each mitigation "OnT'Off' comparison (Table C-2). The
assumption of equal variances holds in every case.
C-7
-------
Appendix C—Additional Statistical Analysis of Effect of
Mitigation on Indoor VOC Concentrations
Table C-2. Results of F-test for Equal Variance
Location
Side
Compound
P
V2
BaseN
420
CHCb
1
3
BaseS
420
CHCb
0
3
First
420
CHCb
1
3
BaseN
422
CHCb
1
3
BaseS
422
CHCb
1
3
First
422
CHCb
1
3
BaseN
420
PCE
0
3
BaseS
420
PCE
0
3
First
420
PCE
0
3
BaseN
422
PCE
1
3
BaseS
422
PCE
0
3
First
422
PCE
1
3
The sampled values are independent from one another
This is the hardest of the three assumptions to validate. This may not hold if the autocorrelation of the
VOC concentrations is longer than a week. We believe that VOC concentrations from consecutive weeks
are not autocorrelated due to the known air exchange rate at this house and autocorrelation results from
other vapor intrusion studies, but significant autocorrelation beyond a week was found in our data in some
cases (see Chapter 10). However, because the data being examined in this section do not span an entire
year, the data cannot be detrended and this may contribute to the observed autocorrelation. If the data
were indeed autocorrelated across weeks, the data in each of the two populations considered for each
comparison (the two populations are mitigation "On" observations and mitigation "Off' observations)
would be more similar amongst themselves than truly randomly chosen observations from each
population would be and it would be possible to come to an incorrect conclusion as to the significance of
the difference between the unmitigated and mitigated datasets.
For example, there are more mitigation "Off' samples preceded by other mitigation "Off' samples than
there are preceded by mitigation "On" samples. If the concentration the previous week influenced the
current week, the population of mitigation "Off' samples would be artificially heterogeneous causing its
variance to be underestimated. This could potentially result in erroneously declaring the difference
between two populations significant when it was not significant.
That being said, the results are quite convincing. We used a two-sided two-sample t-test to test the
difference between the log-concentrations with mitigation "On" and "Off' with the null hypothesis that
the difference between the two populations is 0 (that is to say that the null hypothesis is that mitigation
has no effect). This provides a p-value for that hypothesis and an estimate of the difference and a standard
error for that estimate which we used to calculate a 95% confidence interval for the estimated differences
between populations. The raw estimates and upper/lower confidence limits of the log transformed data are
not particularly valuable, but exponentiating them recovers the factor by which mitigation increases or
decreases VOC concentration in indoor air. The exponentiated results are presented in Table C-3 and
Figure C-5.
C-8
-------
Appendix C—Additional Statistical Analysis of Effect of
Mitigation on Indoor VOC Concentrations
Table C-3. Estimates of Concentration Reduction Factor with Mitigation On with 95% Confidence
Intervals on the Factor
Location
Side
Compound
P
low95
estimate
up95
Warning
BaseN
420
CHCb
0
0.327
0.443
0.601
*
BaseS
420
CHCb
0
0.416
0.527
0.669
*
First
420
CHCb
0.001
0.459
0.609
0.81
*
BaseN
422
CHCb
0
0.179
0.247
0.341
BaseS
422
CHCb
0
0.196
0.288
0.422
First
422
CHCb
0
0.254
0.343
0.465
BaseN
420
PCE
0.009
0.3
0.498
0.827
BaseS
420
PCE
0
0.328
0.479
0.7
First
420
PCE
0.019
0.378
0.585
0.908
BaseN
422
PCE
0.004
0.21
0.389
0.721
*
BaseS
422
PCE
0.014
0.19
0.393
0.812
First
422
PCE
0.007
0.228
0.418
0.767
Log-normality assumption potentially violated
Effect of Mitigation on Indoor Air VOC Concentrati
and 95% Conf dence Intervals
1.00
0.75
0.50-
0.25-
PCE
CHCI3
~l 1 1 1 1 1 1 1 1 1 1 1
-------
Appendix C—Additional Statistical Analysis of Effect of
Mitigation on Indoor VOC Concentrations
In Figure C-5 we include a red horizontal line at 1 because this represents the null hypothesis that turning
the mitigation on will have no effect. An example of how one might interpret this figure is:
"We are 95% confident that mitigation decreases indoor air chloroform concentration at 420BaseN by a
factor of 0.33 to 0.60."
Notice the confidence bands are not symmetrical on either side of the point estimate. This is because the
confidence intervals are computed in log space and then transformed back into real space.
Although the log-normality assumption is potentially violated in some of the tests and an argument could
be made that the independence assumption is violated as well, the results are conclusive. All of the point
estimates are below 1 and none of the confidence intervals include 1. A violation of the assumptions
might result in a modest increase in the width of the confidence intervals, but certainly not enough to
change the overall conclusion that mitigation does reduce indoor air VOC concentrations. Keep in mind
that each test is done individually and does not take into account the conclusion of the other tests. The fact
that 12 separate tests each corroborate one another inspires more overall confidence than the potential
violation of the assumptions inspires doubt.
A summary of the PCE and chloroform geometric means, the estimated effect of mitigation, and the
accompanying p-value are presented in Table C-4. Note that the difference between the mitigation off
and mitigation on geometric mean VOC concentrations are significant in every case, with p-values well
below 0.05.
Table C-4. Summary of PCE and Chloroform Geometric Means and Estimated Effect of
Mitigation
Geo Mean
Mitigation Off
GeoMean
Mitigation On
Effect
(On/Off)
Compound
Side
Location
p-value
PCE
420
BaseN
0.484
0.2409
0.498
0.008937
PCE
420
BaseS
0.5294
0.2536
0.479
0.0003092
PCE
420
First
0.3171
0.1856
0.585
0.01871
PCE
422
BaseN
1.424
0.5533
0.389
0.004114
PCE
422
BaseS
1.94
0.7617
0.393
0.01366
PCE
422
First
0.8455
0.3537
0.418
0.006519
CHCI3
420
BaseN
0.2055
0.09112
0.443
9.499e-06
CHCI3
420
BaseS
0.1951
0.1029
0.527
2.328e-06
CHCI3
420
First
0.1491
0.09086
0.609
0.001377
CHCI3
422
BaseN
0.5283
0.1305
0.247
2.31e-09
CHCI3
422
BaseS
0.635
0.1829
0.288
4.3e-07
CHCI3
422
First
0.3505
0.1203
0.343
1.026e-07
C-10
-------
-------
v>EPA
United States
Environmental Protection
Agency
Please make all necessary changes on the below label, detach
or copy and return to the address in the upper left hand corner.
If you do not wish to receive these reports CHECK HERE
detach, or copy this cover, and return to the address in the
upper left hand corner.
PRESORTED STANDARD
POSTAGE & FEES PAID
EPA PERMIT No, G-35
Office of Research
and Development (8101R)
Washington, DC 20460
Official Business
Penalty for Private Use
$300
EPA/600/R-13/241
June 2015
www.epa.gov
Recycled/Recyclable
Printed with vegetable-based ink on
paper that contains a minimum of
50% post-consumer fiber content
processed chlorine free
------- |