-------
kriging realizations. Examination of these results indicate that monitoring wells in close
proximity to several other monitoring wells (red color coding on Figure 6.5) generally
provide relatively lesser amounts of information than do wells at greater distances from
other wells, or wells located in areas having limited numbers of monitoring points (blue
color coding on Figure 6.5). This is intuitively obvious, but the analysis allows the most
valuable and least valuable wells to be identified quantitatively. For example, Table 6.1
identifies the 21 wells (ranked 1-12) that provide the relative least amount of information
(potential candidates for removal from the monitoring program) and the 9 wells (ranked
34-42) that provide the greatest amount of information (candidates for retention in the
monitoring program) regarding the occurrence and distribution of TCE in groundwater in
the Upper Vashon Aquifer. Wells ranked from 23 to 32 fall in the "intermediate" range
and receive no recommendation for removal or retention.
6.3.2 Additional Well Analysis
The kriging predicted standard error map and plume delineation also can be used to
evaluate the addition of new wells to the monitoring program. Figure 6.5 shows the eight
Upper Vashon wells not included in the kriging ranking analysis, along with the predicted
standard error map for the kriging realization containing the "base-case" data from the
other 42 wells in the Upper Vashon aquifer. The map also shows an estimate of the
extent of the TCE plume (as defined by the 5-|ug/L isopleth) (USAGE, 2001). The lighter
yellow shading represents areas with less spatial uncertainty, and the darker shading
represents area with greater spatial uncertainty. Figure 6.4 shows that, with the potential
exception of well FL2, all of the new wells are located in areas with higher spatial
uncertainty. Once a final monitoring network program is established, a similar standard
error map could be created to determine the optimal locations for the new wells. For
example, well NEW-5 could provide better spatial information if it were shifted about
500 feet to the south, based on the predicted standard error map shown on Figure 6.4.
6-12
022/742479/Fort Lewis Draft Final.doc
-------
SECTION 7
SUMMARY OF THREE-TIERED MONITORING NETWORK
EVALUATION
The 83 wells included in the original and/or revised groundwater monitoring programs
at the Fort Lewis Logistics Center were evaluated using qualitative hydrogeologic and
RA knowledge, temporal statistical techniques, and spatial statistics. At each tier of the
evaluation, monitoring points that provide relatively greater amounts of information
regarding the occurrence and distribution of COCs in groundwater were identified, and
were distinguished from those monitoring points that provide relatively lesser amounts of
information. In this section, the results of the evaluations are combined to generate a
refined monitoring program that could potentially provide information sufficient to
address the primary objectives of monitoring, at reduced cost. Monitoring wells not
retained in the refined monitoring network could be removed from the monitoring
program with relatively little loss of information. The results of the evaluations were
combined and summarized in accordance with the following decision logic:
1. Each well retained in the monitoring network on the basis of the qualitative
hydrogeologic evaluation is recommended to be retained in the refined
monitoring program.
2. Those wells recommended for removal from the monitoring program on the
basis of all three evaluations, or on the basis of the qualitative and temporal
evaluations (with no recommendation resulting from the spatial evaluation)
should be removed from the monitoring program.
7-1
022/742479/Fort Lewis Draft Final.doc
-------
3. If a well is recommended for removal based on the qualitative evaluation and
recommended for retention based on the temporal or spatial evaluation, the final
recommendation is based on a case-by-case review of well information.
The results of the qualitative, temporal, and spatial evaluations are summarized in Table
7.1. These results indicate that 15 of the 83 monitoring wells could be removed from the
groundwater LTM program with little loss of information. The justification for the
recommendations for the 5 wells that fell into case 3 of the decision logic is as follows:
« LC-05 and LC-132 are both recommended for removal from the monitoring
program based on the qualitative and spatial evaluations (Tables 4.3 and 6.1), but
are recommended for retention based on the on the temporal analysis, which
showed their TCE concentrations to be increasing (Table 5.1). However, these
wells provide redundant data because of their relative physical proximity (about
500 feet; Figure 5.4). It is recommended that LC-132 be retained and that well
LC-05 be removed from the program because LC-132 has higher TCE
concentrations that are increasing at a faster rate than the concentrations at well
LC-05.
« Well LC-51 is recommended for removal from the monitoring program based on
the qualitative evaluation (Table 4.3) and for retention in the temporal analysis
due to increasing TCE trends (Table 5.1). The decision was made to recommend
removal of well LC-51 from the monitoring program because wells FL2 and LC-
53 provide adequate plume coverage in the area, and LC-53 also exhibits an
increasing TCE concentration trend, as well as higher TCE concentrations than
those measured at well LC-51.
« Well LC-73a is recommended for removal from the monitoring program based on
the qualitative and temporal analyses, and for retention on basis of the spatial
analysis. A decision was made to recommend removal of the well from the
monitoring program because the well is outside the area of pertinent spatial
information (i.e., outside of the 5-ug/L TCE isopleth and across Murray Creek).
7-2
022/742479/Fort Lewis Draft Final.doc
-------
TABLE 7.1
SUMMARY OF EVALUATION OF CURRENT GROUNDWATER MONITORING PROGRAM
MONITORING NETWORK OPTIMIZATION
FORT LEWIS, WASHINGTON
Well ID
Hydro logic
Unit"7
Revised
Sampling
Frequency b/
hialitative Evaluatio
Remove
Retain
Tern
Remove
poral
Retain
Spatial Evaluation
Remove
Retain
Summary
Remove
Retain
Recommended
Monitoring
Frequency
Original Monitoring Network Wells
LC-03
LC-05
LC-06
LC-14a
LC-19a
LC-19b
LC-19c
LC-26
LC-41a
LC-44a
LC-49
LC-51
LC-53
LC-64a
LC-66a
LC-66b
LC-73a
LC-108
LC-132
LC-136a
LC-136b
LC-137a
LC-137b
LC-149c
LC-149d
LC-165
PA-381
PA-383
T-04
T-08
T-12b
T-13b
LC-lllb
LC-116b
LC-122b
LC-128
LC-137c
LC-64b
UV
uv
UV
uv
uv
uv
uv
uv
uv
uv
uv
uv
uv
uv
uv
uv
uv
uv
uv
uv
uv
uv
uv
uv
uv
uv
uv
uv
uv
uv
uv
uv
LV
LV
LV
LV
LV
LV
Quarterly
Annually
Semi-Annually
Annually
Quarterly
None
None
Annually
Annually
None
Annually
None
Annually
Quarterly
None
Annually
None
None
None
Quarterly
Annually
None
Quarterly
Annually
None
None
Annually
Annually
Annually
Semi-Annually
Quarterly
Semi-Annually
Annually
Annually
Annually
Annually
Annually
Annually
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
NI"7
NI
NI
NI
NI
NI
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
Biennially
Annually
Annually
Annually
-
-
Annually
Annually
-
Annually
Quarterly
Annually
-
-
Annually
Quarterly
Annually
-
-
Biennially
Biennially
Biennially
Biennially
Annually
Semi-Annually
Biennially
Semi-Annually
Biennially
Annually
-
Annually
Annually
Annually
Extraction Wells
LX-1
LX-2
LX-3
LX-4
LX-5
LX-6
LX-7
LX-8
LX-9
LX-10
LX-11
LX-1 2
LX-1 3
LX-14
LX-1 5
LX-1 6
LX-1 7
LX-1 8
LX-1 9
LX-21
RW-1
EW
EW
EW
EW
EW
EW
EW
EW
EW
EW
EW
EW
EW
EW
EW
EW
EW
EW
EW
EW
EW
Annually
Annually
Annually
Annually
Annually
Annually
Annually
Annually
Annually
Annually
Annually
Annually
Annually
Annually
Annually
Quarterly
Quarterly
Quarterly
Quarterly
Quarterly
Quarterly
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
NI
NI
NI
NI
NI
NI
NI
NI
NI
NI
NI
NI
NI
NI
NI
NI
NI
NI
NI
NI
NI
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
Every 3 Years
Every 3 Years
Every 3 Years
Every 3 Years
Every 3 Years
Every 3 Years
Every 3 Years
Every 3 Years
Every 3 Years
Every 3 Years
Every 3 Years
Every 3 Years
Every 3 Years
Every 3 Years
Every 3 Years
Semi-Annually
Quarterly
Quarterly
Quarterly
Quarterly
Semi-Annually
Wells Added to Monitoring Network in December 2001
FL2
FL3
FL4B
FL6
LC-16
LC-20
LC-24
LC-34
LC-57
LC-61b
LC-167
NEW-1
NEW-2
NEW-3
UV
UV
UV
UV
UV
UV
UV
UV
UV
UV
UV
uv
uv
uv
Annually
Quarterly
Quarterly
Quarterly
Quarterly
Quarterly
Quarterly
Quarterly
Quarterly
Quarterly
Quarterly
Quarterly
Quarterly
Quarterly
V
V
V
V
V
V
V
V
V
V
V
V
V
V
NA"
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
Annually
-
Biennially
Biennially
Biennially
Biennially
Biennially
Biennially
Semi-Annually
Semi-Annually
Quarterly
Quarterly
Quarterly
022/742479/3-tieied Ft Lewis Tables.xls/Table 7.1
7-3
-------
TABLE 7.1 (Continued)
SUMMARY OF EVALUATION OF CURRENT GROUNDWATER MONITORING PROGRAM
MONITORING NETWORK OPTIMIZATION
FORT LEWIS, WASHINGTON
Well ID
NEW-4
NEW-5
NEW-6
T-06
T-llb
FL4A
LC-41b
MAMC1
MAMC6
T-10
Hydro logic
Unit17
UV
uv
UV
uv
uv
LV
LV
LV
LV
LV
Revised
Sampling
Frequency b/
Quarterly
Quarterly
Quarterly
Quarterly
Quarterly
Quarterly
Quarterly
Quarterly
Quarterly
Quarterly
hialitative Evaluatio
Remove
Retain
V
V
V
V
V
V
V
V
V
V
Tern
Remove
poral
Retain
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
Spatial Evaluation
Remove
Retain
NI
NI
NI
NI
NI
Summary
Remove
V
V
V
V
V
V
V
V
V
V
Retain
Recommended
Monitoring
Frequency
Quarterly
Quarterly
Quarterly
Quarterly
Quarterly
Biennially
Annually
Quarterly
Quarterly
Semi-Annually
Proposed Additional Well
LC-lSO
UV
None
ADD
NA
NI
ADD
Annually
UV=Upper Vashon Aquifer; LV= Lower Vashon Aquifer; EW=(Vashon Aquifer) Extraction Well.
Sampling frequency established by the remedial action monitoring network optimization report (USAGE, 2001).
c NA = Fewer than four samples; not applicable for temporal trend analysis.
^ NI = Extraction well or Lower Vashon well not included in spatial analysis.
022/742479/3-tiered Ft Lewis Tables.xls/Table 7.1
7-4
-------
A refined monitoring program, consisting of 69 wells (16 sampled quarterly, 7
sampled semi-annually, 17 sampled annually, 14 sampled biennially, and the 15 1-5
extraction wells sampled every 3 years) would be adequate to address the two primary
objectives of monitoring. This refined monitoring network would result in 107 sampling
events per year, compared to 180 events per year in the current LOGRAM monitoring
program and 236 yearly events in the original sampling program. Implementing these
recommendations for optimizing the RA monitoring program at the Fort Lewis
Logistics Center could reduce site monitoring costs by $36,500 a year (more than 40%)
from the LOGRAMLTM strategy, and $64,500 (approximately 55%) from the original
LTMprogram (based on a per sample cost of $500 (USAGE, 2001)). Additional cost
savings could be realized if groundwater samples collected from select wells (e.g., wells
along the lateral and upgradient plume margins) were analyzed for a short list of
halogenated VOCs using Method 802 IB instead of Method 8260B.
7-5
022/742479/Fort Lewis Draft Final.doc
-------
SECTION 8
REFERENCES
American Society of Civil Engineers (ASCE) Task Committee on Geostatistical
Techniques in Hydrology. 1990a. Review of geostatistics in geohydrology - I.
Basic concepts. Journal of Hydraulic Engineering 116(5):612-632.
ASCE Task Committee on Geostatistical Techniques in Hydrology. 1990b. Review of
geostatistics in geohydrology - II. Applications. Journal of Hydraulic
Engineering 116(6): 63 3-65 8.
Clark, I. 1987. Practical Geostatistics. Elsevier Applied Science, Inc., London.
Environmental Systems Research Institute, Inc. (ESRI). 2001. ArcGIS Geostatistical
Analyst Extension to ArcGIS 8 Software, Redlands, CA.
Gibbons, R.D. 1994. Statistical Methods for Groundwater Monitoring. John Wiley &
Sons, Inc., New York.
Rock, N.M.S. 1988. Numerical Geology. Springer-Verlag. New York, New York
US Army Corps of Engineers (USACE). 2001. Draft Logistics Center (FTLE-33)
Remedial Action Monitoring Network Optimization Report. May.
US ACE and URS Corporation. 2002. Draft Field Investigation Report Phase II Remedial
Investigation, East Gate Disposal Yard, Fort Lewis, Washington. DSERTS no.
FTLE-67. July.
US Environmental Protection Agency (USEPA). 1994. Methods for Monitoring Pump-
and-Treat Performance. Office of Research and Development. EPA/600/R-
94/123.
URS Corporation. 2000. Draft Engineering Evaluation/Cost Analysis, East Gate
Disposal Yard and Logistics Center, Fort Lewis, Washington. September.
Wiedemeier, T.H., and P.E. Haas. 2000. Designing Monitoring Programs to Effectively
Evaluate the Performance of Natural Attenuation. Air Force Center for
Environmental Excellence (AFCEE). August.
8-1
022/742479/Fort Lewis Draft Final.doc
-------
APPENDIX D-2
OPTIMIZATION OF MONITORING PROGRAM
AT
LONG PRAIRIE GROUND WATER CONTAMINATION
SUPERFUND SITE, MINNESOTA
-------
G-2236-15
2.0
Long Prairie, Minnesota
. -f>
*•
to
Air Force Center for Environmental
May 8, 2003
Groundwater Services, Inc.
2211 Norfolk, Suite 1000, Houston, Texas 77098-4044
-------
V
GROUNDWATER
SERVICES, INC.
MAROS 2.0 APPLICATION
UPPER OUTWASH AQUIFER MONITORING NETWORK OPTIMIZATION
LONG PRAIRIE SITE
Long Prairie, Minnesota
Prepared
by
Groundwater Services, Inc.
2211 Norfolk, Suite 1000
Houston, Texas 77098
(713)522-6300
GSI Job No. G-2236
Revision No. DRAFT
Date: 5/08/03
-------
GSI Job No. G-2236-15 GROUNDWATER
February 19, 2003 SERVICES, INC.
MAROS 2.0 APPLICATION
UPPER OUTWASH AQUIFER MONITORING NETWORK
OPTIMIZATION, LONG PRAIRIE SITE
Long Prairie, Minnesota
Table of Contents
Executive Summary 1
Project Objectives 1
Results 2
1.0 Introduction 4
1.1 Geology/Hydrogeology 5
1.2 Remedial Action 5
2.0 MAROS Methodology 7
2.1 MAROS Conceptual Model 7
2.2 Data Management 8
2.3 Site Details 8
2.4 Data Consolidation 9
2.5 Overview Statistics: Plume Trend Analysis 9
2.5.1 Mann-Kendall Analysis 10
2.5.2 Linear Regression Analysis 10
2.5.3 Overview Plume Analysis 11
2.5.4 Moment Analysis 12
2.6 Detailed Statistics: Optimization Analysis 13
2.6.1 Well Redundancy Analysis- Delaunay Method 14
2.6.2 Well Sufficiency Analysis - Delaunay Method 15
2.6.3 Sampling Frequency- Modified CES Method 15
2.6.4 Data Sufficiency - Power Analysis 17
3.0 Site Results 19
3.1 Data Consolidation 19
3.2 Overview Statistics: Plume Trend Analysis 20
3.2.1 Mann-Kendall/Linear Regression Analysis 20
3.2.2 Moment Analysis 21
3.2.3 Overview Plume Analysis 23
3.3 Detailed Statistics: Optimization Analysis 24
3.3.1 Well Redundancy Analysis 24
3.3.2 Well Sufficiency Analysis 25
3.3.3 Sampling Frequency Analysis 26
3.3.4 Data Sufficiency - Power Analysis 27
4.0 Summary and Recommendations 29
Long Prairie Site Q MAROS 2.0 Application
Long Prairie, Minnesota Monitoring Network Optimization
-------
GSIJobNo. G-2236-15
February 19, 2003
GROUNDWATER
SERVICES, INC.
Tables
Table 1 Sampling Locations Used in the MAROS Analysis
Table 2 Mann-Kendall Analysis Decision Matrix
Table 3 Linear Regression Analysis Decision Matrix
Table 4 Upper Outwash Aquifer Site-Specific Parameters
Table 5 Results of Upper Outwash Aquifer Trend Analysis
Table 6 Redundancy Analysis Results - Delaunay Method
Table 7 Sampling Frequency Analysis Results - Modified CES
Table 8 Selected Plume Centerline Wells for Risk-Based Site Cleanup Evaluation
- Power Analysis
Table 9 Plume Centerline Concentration Regression Results - Power Analysis
Table 10 Risk-Based Site Cleanup Evaluation Results - Power Analysis
Table 11 Summary of MAROS Sampling Optimization Results
Figures
Figure 1 Upper Outwash Aquifer Groundwater Monitoring Network
Figure 2 MAROS Decision Support Tool Flow Chart
Figure 3 MAROS Overview Statistics Trend Analysis Methodology
Figure 4 Decision Matrix for Determining Provisional Frequency
Figure 5 Upper Outwash Aquifer PCE Mann-Kendall Trend Results
Figure 6 Upper Outwash Aquifer PCE Linear Regression Trend Results
Figure 7 Upper Outwash Aquifer PCE Mann-Kendall Trend Results, Recovery
Wells
Figure 8 Upper Outwash Aquifer PCE Linear Regression Trend Results, Recovery
Wells
Figure 9 Upper Outwash Aquifer PCE First Moment (Center of Mass) Over Time
Figure 10 Upper Outwash Aquifer PCE plume contoured with 1999 and 2002 data:
With "B" Zone Wells Only
Figure 11 Upper Outwash Aquifer PCE plume contoured with 1999 data: before
optimization and after optimization
Figure 12 Upper Outwash Aquifer Well Sufficiency Results
Appendices
Appendix A: Upper Outwash Aquifer Long Prairie site Historical PCE Maps
Appendix B: Upper Outwash Aquifer Long Prairie site MAROS 2.0 Reports
Long Prairie Site
Long Prairie, Minnesota
MAROS 2.0 Application
Monitoring Network Optimization
-------
GROUNDWATER
February 19, 2003 SERVICES, INC.
MAROS 2.0 APPLICATION
UPPER OUTWASH AQUIFER MONITORING NETWORK OPTIMIZATION
LONG PRAIRIE SITE
EXECUTIVE SUMMARY
Long-term monitoring programs, whether applied for process control, performance
measurement, or compliance purposes, require large scale data collection effort and
time commitment, making their cumulative costs very high. With the increasing use of
risk-based goals and natural attenuation in recent years as well as the move toward
long-term closure upon completion of cleanup activities, the need for better-designed
long-term monitoring plans that are cost-effective, efficient, and protective of human and
ecological health has greatly increased. The Monitoring and Remediation Optimization
System (MAROS) methodology provides an optimal monitoring network solution, given
the parameters within a complicated groundwater system which will increase its
effectiveness. By applying statistical techniques to existing historical and current site
analytical data, as well as considering hydrogeologic factors and the location of potential
receptors, the software suggests an optimal plan along with an analysis of individual
monitoring wells for the current monitoring system. This report summarizes the findings
of an application of the MAROS 2.0 software to the Upper Outwash Aquifer long-term
monitoring well networks at the Long Prairie Site in Long Prairie, Minnesota.
The primary constituent of concern at the site is tetrachloroethylene (PCE) which is
analyzed at a total of 44 wells consisting of 31 monitoring wells, 3 city wells, and 10
extraction wells (Figure 1). Sampling frequency for these wells varies: extraction wells
were generally sampled quarterly while monitoring wells were generally sampled
semiannually or annually since the implementation of the long-term monitoring plan in
1996. For some wells, sampling was even terminated for 3 years before they were
sampled again in October 2002. This resulted in some monitoring wells having only 5 ~
7 data records during the 7-year period (from 1996 to 2002). The historical PCE data for
all or in some cases a subset of wells were analyzed using the MAROS 2.0 software in
order to: 1) gain an overall understanding of the plume stability, and 2) recommend
changes in sampling frequency and sampling locations without compromising the
effectiveness of the long-term monitoring network.
Project Objectives
The general objective of the project was to optimize the Long Prairie long-term
monitoring network and sampling plan applying the MAROS 2.0 statistical and decision
support methodology. The key objectives of the project included:
Determining the overall plume stability through trend analysis and moment
analysis;
Evaluating individual well PCE concentration trends over time;
Long Prairie Site 1 MAROS 2.0 Application
Long Prairie, Minnesota Monitoring Network Optimization
-------
GROUNDWATER
February 19, 2003 SERVICES, INC.
Addressing adequate and effective sampling through reduction of redundant
wells without information loss and addition of new wells for future sampling;
• Assessing future sampling frequency recommendations while maintaining
sufficient plume stability information;
Evaluating risk-based site cleanup status using data sufficiency analysis.
Results
The MAROS 2.0 sampling optimization software/methodology has been applied to the
Long Prairie's existing monitoring program as of October 2002. Results from the
temporal trend analysis, moment analysis, sampling location determination, sampling
frequency determination, and data sufficiency analysis indicate that:
• Site monitoring wells were divided into source wells and tail wells where source
wells are near the dry cleaner site or have historically elevated concentrations of
PCE.
• 2 out of 4 source wells and 24 out of 27 tail wells have a Probably Decreasing,
Decreasing, or Stable trend. Both of the statistical methods used to evaluate
trends (Mann-Kendall and Linear Regression) gave similar trend estimates for
each well.
• 7 out of 10 recovery wells have Probably Decreasing or Decreasing trends. Both
the Mann-Kendall and Linear Regression methods gave similar trend estimates
for each well.
• The dissolved mass shows stability over time, whereas the center of mass shows
increase in distance over time in relation to the source location, and the statistical
distribution of the plume in the x and y directions show a show a relatively stable
trend over time. The results from the moment analysis are dependent on a
changing dataset over time due to the change in the wells sampled over the
sampling period analyzed. Overall these results indicate that the plume is not
increasing in size.
• Overall plume stability results indicate that a monitoring system of "Moderate"
intensity is appropriate for this plume compared to "Limited" or "Extensive"
systems due to a stable Upper Outwash Aquifer plume.
• The well redundancy optimization tool, using the Delaunay method aided with a
qualitative analysis, indicates that 12 existing monitoring wells (27% of all) may
not be needed in the current monitoring system without compromising the
accuracy of the monitoring network.
• The well sufficiency optimization tool, based on the Delaunay method, indicates
that there are no areas within the existing monitoring network that have
Long Prairie Site 2 MAROS 2.0 Application
Long Prairie, Minnesota Monitoring Network Optimization
-------
GROUNDWATER
February 19, 2003 SERVICES, INC.
significantly high uncertainty in the PCE concentration estimation. Therefore, no
new monitoring wells are recommended.
• Application of the well sampling frequency determination tool, the Modified CES
method, leads to significant reduction in sampling frequency. Among the 44
wells in the current monitoring system, 19 are recommended for annual sampling
and 25 for biennial sampling. Considering only wells that have been sampled
consistently up to October 2002 (26 wells) and the sampling frequency reduction
alone, a reduction of approximately 57% in total samples each year can be
achieved.
• The MAROS Data Sufficiency (Power Analysis) application indicates that the
monitoring record has sufficient statistical power to conclude that the site has
attained cleanup goal at (or farther than) the "hypothetical statistical compliance
boundary" located near the most downgradient well at the site. As the plume
shrinks, this hypothetical statistical compliance boundary will move upgradient
gradually.
The recommended long-term monitoring strategy results in a significant reduction in
sampling costs and allows site personnel to develop a better understanding of plume
behavior over time. A reduction in the number of redundant wells will still maintain
adequate delineation of the plume as well as knowledge of the plume state over time.
The MAROS optimized plan results in a monitoring network of 32 wells: 16 sampled
annually, and 16 sampled biennially. The MAROS optimized plan would result in 24
samples per year, compared to 51 samples per year if all the monitoring wells in the
current network were sampled every year. Implementing these recommendations could
lead to a 52% reduction from the current monitoring plan in terms of the samples to be
collected per year. The reduction in the number of redundant wells and decreased
sampling frequency is expected to result in a moderate cost savings over the long-term
at the Long Prairie site. An approximate cost savings estimate range from $2,700 to
$7,560 per year (based on an average per sample cost range of $100 to $280) is
projected while still maintaining adequate delineation of the plume as well as knowledge
of the plume state over time.
Long Prairie Site 3 MAROS 2.0 Application
Long Prairie, Minnesota Monitoring Network Optimization
-------
GROUNDWATER
February 19, 2003 SERVICES, INC.
1.0 INTRODUCTION
Long-term monitoring programs, whether applied for process control, performance
measurement, or compliance purposes, require large scale data collection effort and
time commitment, making their cumulative costs very high. With the increasing use of
risk-based goals and natural attenuation in recent years as well as the move toward
long-term closure upon completion of cleanup activities, the need for better-designed
long-term monitoring plans that are cost-effective, efficient, and protective of human and
ecological health has greatly increased. AFCEE's Monitoring and Remediation
Optimization System (MAROS) methodology provides an optimal monitoring network
solution, given the parameters within a complicated groundwater system which will
increase its effectiveness. By applying statistical techniques to existing historical and
current site analytical data, as well as considering hydrogeologic factors and the location
of potential receptors, the software suggests an optimal plan along with an analysis of
individual monitoring wells for the current monitoring system. This report summarizes
the findings of an application of the MAROS 2.0 software to the Upper Outwash Aquifer
long-term monitoring well network at the Long Prairie site, Long Prairie, Minnesota.
1.1 Geology/Hydrogeology
The Long Prairie groundwater contamination site is a groundwater plume of chlorinated
organic compounds (mostly Tetrachlorothene (PCE)) located below portions of the city
of Long Prairie, Minnesota.
The subsurface in the vicinity of Long Prairie consists of a series of glacial till and
outwash deposits approximately 700 feet thick. The aquifer system near Long Prairie
consists of two water-bearing outwash units. However, in some areas the separating
aquitard is not present between the upper and lower outwash units within the outwash
valley. A sandy clay till aquitard separates the two outwash units on the eastern side of
the city. Leakage is inhibited primarily by the low vertical hydraulic conductivity of the till.
The lower outwash is in direct contact with the upper outwash unit below the western
side of Long Prairie (Barr, 2001). The uppermost geologic unit is silty sand with some
coarser sand and gravel (glacial outwash deposit), the most prolific aquifer in the area.
The aquifer is essentially a wedge of outwash sand and gravel approximately 60 feet
thick near municipal well #4 and thins to less than 5 feet to the southeast and maybe
locally absent beyond. Underlying the glacial outwash sediments is glacial till composed
of sandy clay with varying concentrations of gravel. The till extends to a depth of at least
200 ft bgs and appears to be continuous beneath the site. The backlot, where the PCE
release occurred, is located over an area where the till is present between the two
outwash units and it near the western limit of the till. The saturated thickness of the
upper outwash is only about 10 feet near the source, near MW-10.
The groundwater table in the vicinity of the source area is significantly higher possibly
due to a continuation of the water table as it comes into town from the south in a broad,
shallow valley that contains the Charlotte Lake. Just north of this area, the lower till
pinches out, so the aquifer is much thicker and water elevations are largely controlled by
Long Prairie Site 4 MAROS 2.0 Application
Long Prairie, Minnesota Monitoring Network Optimization
-------
GROUNDWATER
February 19, 2003 SERVICES, INC.
the Long Prairie River. The water table configuration suggests that the groundwater
discharges to the Long Prairie River from the area of well MW-7A and south
(approximately Northwesterly direction). North of well MW-5A, the groundwater flow
direction appears to be approximately parallel with the river (approximately Northerly
direction). Base wells, domestic production wells, extraction wells, and regional
pumping affect local groundwater flow directions. Groundwater elevations range from
1280 to 1290 ft msl at the northern and southern edges of the area, respectively. The
groundwater seepage velocity is approximately 472 ft/yr. For a detailed description of
site geology and hydrogeology refer to Barr (2001).
1.2 Remedial Action
The Long Prairie site is has an approximate 7,000 square foot source area with a one-
half mile long ground water plume located on Long Prairie, Todd County, Minnesota.
The chlorinated plume originates in the commercial area of Long Prairie and extends
through an older residential area of the city. The source of PCE in the groundwater was
a dry-cleaning facility which operated from 1978 until mid-1984, located near well MW-
10. The site was discovered in 1983 during the Minnesota Department of Health's VOC
state wide analysis of public water supplies. The municipal water supply of Long Prairie
was found to be contaminated with PCE and its degradation products trichloroethylene
(TCE) and cis-1,2-dichloroethylene (DCE). The United States Environmental Protection
Agency (U.S. EPA) placed the site on the National Priority List (NPL) in 1986. The
Remedial Investigation/Feasibility Study (RI/FS) and the Remedial Design/Remedial
Action (RD/RA) were conducted under a Multi-Site Cooperative Agreement between
U.S. EPA and the Minnesota Pollution Control Agency (MPCA).
A soil vapor extraction system was installed in 1997 to remove PCE in the soil above the
water table in the alley adjacent to the dry cleaning business, the source area. This
system operated until the end of 1999. A groundwater recovery system and GAC plant
was started up in 1996, with two additional recovery wells added in 1999. The system is
still operating today and consists of nine groundwater recovery wells in the plume area
for groundwater flow control. The objective of the remediation is to restore the Upper
Outwash aquifer to drinking water standards by reducing the PCE concentration to less
than the MCL (5 ppb) as well as preventing the spread of the plume to wells presently
unaffected, including the city of Long Prairie municipal supply well 6.
The groundwater long-term monitoring plan started in 1996 consists of 31 monitoring
wells, 3 city wells, and 10 extraction wells (Figure 1). The monitoring well naming
convention includes: "a" wells, shallow wells screened at the water table; "b" wells, mid-
depth wells screened at the base of the upper outwash; and "c" wells, deep wells
screened in the lower outwash. The monitoring system is used for performance
monitoring and compliance monitoring with the following goals: 1) plume containment
monitoring to confirm that the plume remains hydraulically controlled; and 2) plume
reduction monitoring to verify progress toward achieving cleanup goals.
The sampling frequency for the long-term monitoring wells varies: extraction wells have
generally been sampled quarterly while monitoring wells were generally sampled
Long Prairie Site 5 MAROS 2.0 Application
Long Prairie, Minnesota Monitoring Network Optimization
-------
GROUNDWATER
February 19, 2003 SERVICES, INC.
semiannually or annually since the implementation of the long-term monitoring plan in
1996. For some wells, sampling was even terminated for 3 years before they were
sampled again in October 2002. This resulted in some monitoring wells having only 5 ~
7 data records during the 7-year period (from 1996 to 2002). The historical PCE data for
all or in some cases a subset of wells were analyzed using the MAROS 2.0 software in
order to: 1) gain an overall understanding of the plume stability, and 2) recommend
changes in sampling frequency and sampling locations without compromising the
effectiveness of the long-term monitoring network.
Long Prairie Site 6 MAROS 2.0 Application
Long Prairie, Minnesota Monitoring Network Optimization
-------
GROUNDWATER
February 19, 2003 SERVICES, INC.
2.0 MAROS METHODOLOGY
The MAROS 2.0 software used to optimize the LTM network at the Long Prairie site is
explained in general terms in this section. MAROS is a collection of tools in one
software package that is used in an explanatory, non-linear fashion. The tool includes
models, statistics, heuristic rules, and empirical relationships to assist the user in
optimizing a groundwater monitoring network system while maintaining adequate
delineation of the plume as well as knowledge of the plume state over time. Different
users utilize the tool in different ways and interpret the results from a different viewpoint.
For a detailed description of the structure of the software and further utilities, refer to the
MAROS 2.0 Manual (Aziz et al. 2002).
2.1 MAROS Conceptual Model
In MAROS 2.0, two levels of analysis are used for optimizing long-term monitoring plans:
1) an overview statistical evaluation with interpretive trend analysis based on temporal
trend analysis and plume stability information; and 2) a more detailed statistical
optimization based on spatial and temporal redundancy reduction methods (see Figure 2
for further details). In general, the MAROS method applies to 2-D aquifers that have
relatively simple site hydrogeology. However, for a multi-aquifer (3-D) system, the user
could apply the statistical analysis layer-by-layer.
The overview statistics or interpretive trend analysis assesses the general monitoring
system category by considering individual well concentration trends, overall plume
stability, hydrogeologic factors (e.g., seepage velocity, and current plume length), and
the location of potential receptors (e.g., property boundaries or drinking water wells). The
analysis relies on temporal trend analysis to assess plume stability, which is then used
to determine the general monitoring system category. Since the temporal trend analysis
focuses on where the monitoring well is located, the site wells are divided into two
different zones: the source zone or the tail zone. The source zone includes areas with
non-aqueous phase liquids (NAPLs), contaminated vadose zone soils, and areas where
aqueous-phase releases have been introduced into ground water. The tail zone is
usually the area downgradient of the contaminant source zone. Although this
classification is a simplification of the well location, this broadness makes the user aware
on an individual well basis that the concentration trend results can have a different
interpretation depending on the well location in and around the plume. The location and
type of the individual wells allows further interpretation of the trend results, depending on
what type of well is being analyzed (e.g., remediation well, leading plume edge well, or
monitoring well). General recommendations for the monitoring network frequency and
density are suggested based on heuristic rules applied to the source and tail trend
results.
The detailed statistics level of analysis or sampling optimization, on the other hand,
consists of a well redundancy analysis and well sufficiency analysis using the Delaunay
method, a sampling frequency analysis using the Modified Cost Effective Sampling
(CES) method and a data sufficiency analysis using power analysis. The well
Long Prairie Site 7 MAROS 2.0 Application
Long Prairie, Minnesota Monitoring Network Optimization
-------
GROUNDWATER
February 19, 2003 SERVICES, INC.
redundancy analysis is designed to minimize monitoring locations and the Modified CES
method is designed to minimize the frequency of sampling. The data sufficiency
analysis uses power analysis to assess the sampling record to determine if the current
monitoring network and record is sufficient in terms of evaluating risk-based site target
level status.
2.2 Data Management
In MAROS, ground water monitoring data can be imported from simple database-format
Microsoft® Excel spreadsheets, Microsoft Access tables, previously created MAROS
database archive files, or entered manually. Compliance monitoring data interpretation in
MAROS is based on historical ground water monitoring data from a consistent set of
wells over a series of sampling events. Statistical validity of the concentration trend
analysis requires constraints on the minimum data input of at least four wells (ASTM
1998) in which COCs have been detected. Individual sampling locations need to include
data from at least the six most-recent sampling events. To ensure a meaningful
comparison of COC concentrations over time and space, both data quality and data
quantity need to be considered. Prior to statistical analysis, the user can consolidate
irregularly sampled data or smooth data that might result from seasonal fluctuations or a
change in site conditions.
Imported ground water monitoring data and the site-specific information entered in Site
Details can be archived and exported as MAROS archive files. These archive files can
be appended as new monitoring data becomes available, resulting in a dynamic long-
term monitoring database that reflects the changing conditions at the site (i.e.
biodegradation, compliance attainment, completion of remediation phase, etc.).
2.3 Site Details
Information needed for the MAROS analysis includes site-specific parameters such as
seepage velocity and current plume length. Part of the trend analysis methodology
applied in MAROS focuses on where the monitoring well is located, therefore the user
needs to divide site wells into two different zones: the source zone or the tail zone. The
source zone includes areas with non-aqueous phase liquids (NAPLs), contaminated
vadose zone soils, and areas where aqueous-phase releases have been introduced into
ground water. The source zone generally contains locations with historical high ground
water concentrations of the COCs. The tail zone is usually the area downgradient of the
contaminant source zone. It is up to the user to make further interpretation of the trend
results, depending on what type of well is being analyzed (e.g., remediation well, leading
plume edge well, or monitoring well).
MAROS allows the analysis of up to 5 COCs concurrently and users can pick COCs
from a list of compounds existing in the monitoring data, or select COCs based on
recommendations provided in MAROS based on toxicity, prevalence, and mobility of
compounds.
Long Prairie Site 8 MAROS 2.0 Application
Long Prairie, Minnesota Monitoring Network Optimization
-------
GROUNDWATER
February 19, 2003 SERVICES, INC.
2.4 Data Consolidation
Typically long-term monitoring raw data have been measured irregularly in time or
contain many non-detects, trace level results, and duplicates. Therefore, before the data
can be further analyzed, raw data are filtered, consolidated, transformed, and possibly
smoothed to allow for a consistent dataset meeting the minimum data requirements for
statistical analysis mentioned previously.
MAROS allows users to specify the period of interest in which data will be consolidated
(i.e., monthly, bi-monthly, quarterly, semi-annual, yearly, or a biennial basis). In
computing the representative value when consolidating, one of four statistics can be
used: median, geometric mean, mean, and maximum. Non-detects can be transformed
to one half the reporting or method detection limit (DL), the DL, or a fraction of the DL.
Trace level results can be represented by their actual values, one half of the DL, the DL,
or a fraction of their actual values. Duplicates are reduced in MAROS by one of three
ways: assigning the average, maximum, or first value. The reduced data for each COC
and each well can be viewed as a time series in a graphical form on a linear or semi-log
plot generated by the software.
2.5 Overview Statistics: Plume Trend Analysis
Within the MAROS software there are historical data analyses that support a conclusion
about plume stability (e.g., increasing plume, etc.) through statistical trend analysis of
historical monitoring data. Plume stability results are assessed from time-series
concentration data with the application of three statistical tools: Mann-Kendall Trend
analysis, linear regression trend analysis and moment analysis. The two trend methods
are used to estimate the concentration trend for each well and each COC based on a
statistical trend analysis of concentrations versus time at each well (Figure 2). These
trend analyses are then consolidated to give the user a general plume stability and
general monitoring frequency and density recommendations (see Figure 3 for further
step-by-step details). Both qualitative and quantitative plume information can be gained
by these evaluations of monitoring network historical data trends both spatially and
temporally. The MAROS Overview Statistics are the foundation the user needs to make
informed optimization decisions at the site. The Overview Statistics are designed to
allow site personnel to develop a better understanding of the plume behavior over time
and understand how the individual well concentration trends are spatially distributed
within the plume. This step allows the user to gain information that will support a more
informed decision to be made in the next level or detailed statistics optimization analysis
(Figure 2).
2.5.1 Mann-Kendall Analysis
The Mann-Kendall test is a non-parametric statistical procedure that is well suited for
analyzing trends in data over time. The Mann-Kendall test can be viewed as a
Long Prairie Site 9 MAROS 2.0 Application
Long Prairie, Minnesota Monitoring Network Optimization
-------
GROUNDWATER
February 19, 2003 SERVICES, INC.
nonparametric test for zero slope of the first-order regression of time-ordered
concentration data versus time. The Mann-Kendall test does not require any
assumptions as to the statistical distribution of the data (e.g. normal, lognormal, etc.)
and can be used with data sets which include irregular sampling intervals and missing
data. The Mann-Kendall test is designed for analyzing a single groundwater constituent,
multiple constituents are analyzed separately. The Mann-Kendall S statistic measures
the trend in the data: positive values indicate an increase in concentrations over time
and negative values indicate a decrease in concentrations over time. The strength of the
trend is proportional to the magnitude of the Mann-Kendall statistic (i.e., a large value
indicates a strong trend). The confidence in the trend is determined by consulting the S
statistic and the sample size n in a Kendall probability table such as the one reported in
Hollander and Wolfe (1973).
The concentration trend is determined for each well and each COC based on results of
the S statistic, the confidence in the trend, and the Coefficient of Variation (COV). The
decision matrix for this evaluation is shown in Table 2. A Mann-Kendall statistic that is
greater than 0 combined with a confidence of greater than 95% is categorized as an
Increasing trend while a Mann-Kendall statistic of less than 0 with a confidence between
90% and 95% is defined as a Probably Increasing trend, and so on.
Depending on statistical indicators, the concentration trend is classified into six
categories:
• Decreasing (D),
• Probably Decreasing (PD),
• Stable (S),
• No Trend (NT),
• Probably Increasing (PI)
• Increasing (I).
These trend estimates are then analyzed to identify the source and tail region overall
stability category (see Figure 2 for further details).
2.5.2 Linear Regression Analysis
Linear Regression is a parametric statistical procedure that is typically used for
analyzing trends in data over time. Using this type of analysis, a higher degree of
scatter simply corresponds to a wider confidence interval about the average log-slope.
Assuming the sign (i.e., positive or negative) of the estimated log-slope is correct, a level
of confidence that the slope is not zero can be easily determined. Thus, despite a poor
goodness of fit, the overall trend in the data may still be ascertained, where low levels of
confidence correspond to "Stable" or "No Trend" conditions (depending on the degree of
scatter) and higher levels of confidence indicate the stronger likelihood of a trend. The
linear regression analysis is based on the first-order linear regression of the log-
transformed concentration data versus time. The slope obtained from this log-
transformed regression, the confidence level for this log-slope, and the COV of the
untransformed data are used to determine the concentration trend. The decision matrix
for this evaluation is shown in Table 3. To estimate the confidence in the log-slope, the
Long Prairie Site 10 MAROS 2.0 Application
Long Prairie, Minnesota Monitoring Network Optimization
-------
GROUNDWATER
February 19, 2003 SERVICES, INC.
standard error of the log-slope is calculated. The coefficient of variation, defined as the
standard deviation divided by the average, is used as a secondary measure of scatter to
distinguish between "Stable" or "No Trend" conditions for negative slopes. The Linear
Regression Analysis is designed for analyzing a single groundwater constituent; multiple
constituents are analyzed separately, (up to five COCs simultaneously). For this
evaluation, a decision matrix developed by Groundwater Services, Inc. is also used to
determine the "Concentration Trend" category (plume stability) for each well.
Depending on statistical indicators, the concentration trend is classified into six
categories:
Decreasing (D),
Probably Decreasing (PD),
Stable (S),
No Trend (NT),
Probably Increasing (PI)
Increasing (I).
The resulting confidence in the trend, together with the log-slope and the COV of the
untransformed data, are used in the linear regression analysis decision matrix to
determine the concentration trend. For example, a positive log-slope with a confidence
of less than 90% is categorized as having No Trend whereas a negative log-slope is
considered Stable if the COV is less than 1 and categorized as No Trend if the COV is
greater than 1.
2.5.3 Overall Plume Analysis
General recommendations for the monitoring network frequency and density are
suggested based on heuristic rules applied to the source and tail trend results.
Individual well trend results are consolidated and weighted by the MAROS software
according to user input, and the direction and strength of contaminant concentration
trends in the source zone and tail zone for each COC are determined. Based on
i) the consolidated trend analysis,
ii) hydrogeologic factors (e.g., seepage velocity), and
iii) location of potential receptors (e.g., wells, discharge points, or property
boundaries),
the software suggests an general optimization plan for the current monitoring system in
order to efficiently effectively monitor in the future. A flow chart of the MAROS
methodology utilizing trend analysis results and other site-specific parameters to form a
general sampling frequency and well density recommendation is outlined in Figure 3. For
example, a generic plan for a shrinking petroleum hydrocarbon plume (BTEX) in a slow
hydrogeologic environment (silt) with no nearby receptors would entail minimal, low
frequency sampling of just a few indicators. On the other hand, the generic plan for a
chlorinated solvent plume in a fast hydrogeologic environment that is expanding but has
very erratic concentrations over time would entail more extensive, higher frequency
sampling. The generic plan is based on a heuristically derived algorithm for assessing
Long Prairie Site 11 MAROS 2.0 Application
Long Prairie, Minnesota Monitoring Network Optimization
-------
GROUNDWATER
February 19, 2003 SERVICES, INC.
future sampling duration, location and density that takes into consideration plume
stability. For a detailed description of the heuristic rules used in the MAROS software,
refer to the MAROS 2.0 Manual (Aziz et al. 2002).
2.5.3 Moment Analysis
An analysis of moments can help resolve plume trends, where the zeroth moment shows
change in dissolved mass vs. time, the first moment shows the center of mass location
vs. time, and the second moment shows the spread of the plume vs. time. Moment
calculations can predict how the plume will change in the future if further statistical
analysis is applied to the moments to identify a trend (in this case, Mann Kendall Trend
Analysis is applied). The trend analysis of moments can be summarized as:
• Zeroth Moment: Change in dissolved mass over time
• First Moment: Change in the center of mass location over time
• Second Moment: Spread of the plume over time
The role of moment analysis in MAROS is to provide a relative measure of plume
stability and condition. Plume stability may vary by constituent, therefore the MAROS
moment analysis can be used to evaluate multiple COCs simultaneously in order to
provide used to provide a quick way of comparing individual plume parameters to
determine the size and movement of constituents relative to one another. Moment
analysis in the MAROS software can also be used to assist the user in evaluating the
impact on plume delineation in future sampling events by removing identified
"redundant" wells from a long-term monitoring program (this analysis was not performed
as part of this study, for more details on this application of moment analysis refer to the
MAROS 2.0 Manual (Aziz et al. 2002).
The zeroth moment is a mass estimate. The zeroth moment calculation can show high
variability over time, largely due to the fluctuating concentrations at the most
contaminated wells as well as varying monitoring well network. Plume analysis and
delineation based exclusively on concentration can exhibit a fluctuating degree of
temporal and spatial variability. The mass estimate is also sensitive to the extent of the
site monitoring well network over time. The zeroth moment trend over time is determined
by using the Mann-Kendall Trend Methodology. The zeroth Moment trend test allows
the user to understand how the plume mass has changed over time. Results for the
trend include: Increasing, Probably Increasing, No Trend, Stable, Probably Decreasing,
Decreasing or Not Applicable (Insufficient Data). When considering the results of the
Zeroth moment trend, the following factors should be considered which could effect the
calculation and interpretation of the plume mass over time: 1) Change in the spatial
distribution of the wells sampled historically 2) Different wells sampled within the well
network over time (addition and subtraction of well within the network). 3) Adequate
versus inadequate delineation of the plume over time.
The first moment estimates the center of mass, coordinates (Xc and Yc) for each
sample event and COC. The changing center of mass locations indicate the movement
Long Prairie Site 12 MAROS 2.0 Application
Long Prairie, Minnesota Monitoring Network Optimization
-------
GROUNDWATER
February 19, 2003 SERVICES, INC.
of the center of mass over time. Whereas, the distance from the original source location
to the center of mass locations indicate the movement of the center of mass over time
relative to the original source. Calculation of the first moment normalizes the spread by
the concentration indicating the center of mass. The first moment trend of the distance to
the center of mass over time shows movement of the plume in relation to the original
source location over time. Analysis of the movement of mass should be viewed as it
relates to 1) the original source location of contamination 2) the direction of groundwater
flow and/or 3) source removal or remediation. Spatial and temporal trends in the center
of mass can indicate spreading or shrinking or transient movement based on seasonal
variation in rainfall or other hydraulic considerations. No appreciable movement or a
neutral trend in the center of mass would indicate plume stability. However, changes in
the first moment over time do not necessarily completely characterize the changes in the
concentration distribution (and the mass) over time. Therefore, in order to fully
characterize the plume the First Moment trend should be compared to the Zeroth
moment trend (mass change over time).
The second moment indicates the spread of the contaminant about the center of mass
(Sxx and Syy), or the distance of contamination from the center of mass for a particular
COC and sample event. The Second Moment represents the spread of the plume over
time in both the x and y directions. The Second Moment trend indicates the spread of
the plume about the center of mass. Analysis of the spread of the plume should be
viewed as it relates to the direction of groundwater flow. An increasing trend in the
second moment indicates an expanding plume, whereas a declining trend in the plume
indicates a shrinking plume. No appreciable movement or a neutral trend in the center of
mass would indicate plume stability. The second moment provides a measure of the
spread of the concentration distribution about the plume's center of mass. However,
changes in the second moment over time do not necessarily completely characterize the
changes in the concentration distribution (and the mass) over time. Therefore, in order to
fully characterize the plume the Second Moment trend should be compared to the zeroth
moment trend (mass change over time).
2.6 Detailed Statistics: Optimization Analysis
Although the overall plume analysis shows a general recommendation regarding
sampling frequency reduction and general sampling density, a more detailed analysis is
also available with the MAROS 2.0 software in order to allow for further reductions on a
well-by-well basis for frequency, well redundancy, well sufficiency and sampling
sufficiency. The MAROS Detailed Statistics allows for a quantitative analysis for spatial
and temporal optimization of the well network on a well-by-well basis. The results from
the Overview Statistics should be considered along with the MAROS optimization
recommendations gained from the Detailed Statistical Analysis described previously.
The MAROS Detailed Statistics results should be reassessed in view of site knowledge
and regulatory requirements as well as in consideration of the Overview Statistics
(Figure 2).
Long Prairie Site 13 MAROS 2.0 Application
Long Prairie, Minnesota Monitoring Network Optimization
-------
GROUNDWATER
February 19, 2003 SERVICES, INC.
The Detailed Statistics or Sampling Optimization MAROS module can be used to
determine the minimal number of sampling locations and the lowest frequency of
sampling that can still meet the requirements of sampling spatially and temporally for an
existing monitoring program. It also provides an analysis of the sufficiency of data for
the monitoring program.
Sampling optimization in MAROS consists of four parts:
• Well redundancy analysis using the Delaunay method
• Well sufficiency analysis using the Delaunay method
• Sampling frequency determination using the Modified CES method
• Data sufficiency analysis using statistical power analysis.
The well redundancy analysis using the Delaunay method identifies and eliminates
redundant locations from the monitoring network. The well sufficiency analysis can
determine the areas where new sampling locations might be needed. The Modified CES
method determines the optimal sampling frequency for a sampling location based on the
direction, magnitude, and uncertainty in its concentration trend. The data sufficiency
analysis examines the risk-based site cleanup status and power and expected sample
size associated with the cleanup status evaluation.
2.6.1 Well Redundancy Analysis - Delaunav Method
The well redundancy analysis using the Delaunay method is designed to select the
minimum number of sampling locations based on the spatial analysis of the relative
importance of each sampling location in the monitoring network. The approach allows
elimination of sampling locations that have little impact on the historical characterization
of a contaminant plume. The delaunay methodology application assumes that the
current sampling network adequately delineates the plume (bounding wells have non-
detect values) and that if a hydraulic containment system is currently in operation, this
will continue. An extended method or wells sufficiency analysis, based on the Delaunay
method, can also be used for recommending new sampling locations. Details about the
Delaunay method can be found in Appendix A.2 of the MAROS Manual (AFCEE 2002).
Well redundancy analysis uses the Delaunay triangulation method to determine the
significance of the current sampling locations relative to the overall monitoring network.
The Delaunay method calculates the network Area and Average concentration of the
plume using data from multiple monitoring wells. A slope factor (SF) is calculated for
each well to indicate the significance of this well in the system (i.e. how removing a well
changes the average concentration.)
The well redundancy optimization process is performed in a stepwise fashion. Step one
involves assessing the significance of the well in the system, if a well has a small SF
(little significance to the network), the well may be removed from the monitoring network.
Step two involves evaluating the information loss of removing a well from the network. If
one well has a small SF, it may or may not be eliminated depending on whether the
Long Prairie Site 14 MAROS 2.0 Application
Long Prairie, Minnesota Monitoring Network Optimization
-------
GROUNDWATER
February 19, 2003 SERVICES, INC.
information loss is significant. If the information loss is not significant, the well can be
eliminated from the monitoring network and the process of optimization continues with
fewer wells. However if the well information loss is significant then the optimization
terminates. This sampling optimization process allows the user to assess "redundant"
wells that will not incur significant information loss on a constituent-by-constituent basis
for individual sampling events.
Before applying the Delaunay method for spatial redundancy analysis, it is important to
select the appropriate set of wells for analysis, i.e., only the wells that contribute to the
spatial delineation of the plume. For example, if wells are far from the plume and
contribute little or nothing to the delineation of the plume (e.g., some sentry wells or
background wells far from the plume), they should be excluded from the analysis. One
reason not to use these wells is that these wells usually are on the boundary of the
triangulation and are hard to be eliminated since the Delaunay method protects
boundary wells from being easily removed. The elimination status of these wells, in fact,
should be determined from the regulatory standpoint. Another well type that could be
excluded from analysis is one of a clustered well set because the Delaunay method is a
two-dimensional method. Generally, only one well is picked from the clustered well set to
represent the concentration at this point. This well can be the one that has the highest
concentration or is screened in the representative aquifer interval with the geologic unit.
Data from clustered wells can also be averaged to form a single sample and then used
in the Delaunay method.
2.6.2 Well Sufficiency Analysis - Delaunav Method
The well sufficiency analysis, using the Delaunay method, is designed to recommend
new sampling locations in areas within the existing monitoring network where there is a
high level uncertainty in plume concentration. Details about the well sufficiency analysis
can be found in Appendix A.2 of the MAROS Manual (AFCEE 2002).
In many cases, new sampling locations need to be added to the existing network to
enhance the spatial plume characterization. In MAROS, the method for determining new
sampling locations recommends the area for a possible new sampling location where
there is a high level of uncertainty in concentration estimation. The Slope Factor (SF)
values obtained from the redundancy reduction described above are used to calculate
the concentration estimation error at each triangle area formed in the Delaunay
triangulation. The estimated SF value at each triangle area is then classified into four
levels: Small, Moderate, Large, or Extremely large because the larger the estimated SF
value, the higher the estimation error at this area. Therefore, the triangle areas with the
estimated SF value at the Extremely large or Large level are candidate regions for new
sampling locations.
The results from the Delaunay method and the method for determining new sampling
locations are derived solely from the spatial configuration of the monitoring network and
the spatial pattern of the contaminant plume. No parameters such as the hydrogeologic
Long Prairie Site 15 MAROS 2.0 Application
Long Prairie, Minnesota Monitoring Network Optimization
-------
GROUNDWATER
February 19, 2003 SERVICES, INC.
conditions are considered in the analysis. Therefore, professional judgement and
regulatory considerations must be used to make final decisions.
2.6.3 Sampling Frequency Determination - Modified CES Method
The Modified Cost Effective Sampling (MCES) method optimizes sampling frequency for
each sampling location based on the magnitude, direction, and uncertainty of its
concentration trend derived from its recent and historical monitoring records. The MCES
estimates the lowest-frequency sampling schedule for a given groundwater monitoring
location yet still provide needed information for regulatory and remedial decision-making.
The Modified CES method was developed on the basis of the Cost Effective Sampling
(Ridley et al. 1995). Details about the Modified CES method can be found in Appendix
A.3 of the MAROS Manual (AFCEE 2002).
In order to estimate the least frequent sampling schedule for a monitoring location that
still provides enough information for regulatory and remedial decision-making, MCES
employs three steps to determine the sampling frequency. The first step involves
analyzing frequency based on recent trends (Figure 4). A preliminary location sampling
frequency (PLSF) is determined based on the trends determined by rates of change
from linear regression and Mann-Kendall analysis of the most recent monitoring data.
The variability of the sequential sampling data is accounted for by the Mann-Kendall
analysis. The PLSF is then adjusted based on overall trends. If the long-term history of
change is significantly greater than the recent trend, the frequency may be reduced by
one level. Otherwise, no change could be made. The final step in the analysis involves
reducing frequency based on risk. Since not all compounds in the target being
assessed are equally harmful, frequency is reduced by one level if recent maximum
concentration for compound of high risk is less than 1/2 of the Maximum Concentration
Limit (MCL). The result of applying this method is a suggested sampling frequency
based on recent sampling data trends and overall sampling data trends.
The finally determined sampling frequency from the Modified CES method can be
Quarterly, Semiannual, Annual, and Biennial. Users can further reduce the sampling
frequency to, for example, once every three years, if the trend estimated from Biennial
data (i.e., data drawn once every two years from the original data) is the same as that
estimated from the original data.
2.6.4 Data Sufficiency Analysis - Power Analysis
Statistical power analysis is a technique for interpreting the results of statistical tests. It
provides additional information about a statistical test: 1) the power of the statistical test,
i.e., the probability of finding a difference in the variable of interest when a difference
truly exists; and 2) the expected sample size of a future sampling plan given the
minimum detectable difference it is supposed to detect. For example, if the mean
concentration is lower than the cleanup goal but a statistical test cannot prove this, the
power and expected sample size can tell the reason and how many more samples are
Long Prairie Site 16 MAROS 2.0 Application
Long Prairie, Minnesota Monitoring Network Optimization
-------
February 19, 2003
If
GROUNDWATER
SERVICES, INC.
needed to result in a significant test. The additional samples can be obtained by a
longer period of sampling or an increased sampling frequency. Details about the data
sufficiency analysis can be found in Appendix A.6 of the MAROS Manual (AFCEE 2002).
When applying the MAROS power analysis method, a hypothetical statistical compliance
boundary (HSCB) is assigned to be a line perpendicular to the groundwater flow
direction (see figure below). Monitoring well concentrations are projected onto the
HSCB using the distance from each well to the compliance boundary along with a decay
coefficient. The projected concentrations from each well and each sampling event are
then used in the risk-based power analysis. Since there may be more than one sampling
event selected by the user, the risk-based power analysis results are given on an event-
by-event basis. This power analysis can then indicate if target are statistically achieved
at the HSCB. For instance, at a site where the historical monitoring record is short with
few wells, the HSCB would be distant; whereas, at a site with longer duration of
sampling with many wells, the HSCB would be close. Ultimately, at a site the goal would
be to have the HSCB coincide with or be within the actual compliance boundary
(typically the site property line).
^ ^.:::::::::::::::::::::::::::::::::::::::
L
o
Groundwater flow direction
Concentrations
projected to this
line
* -f
The nearest
downgradient
receptor
In order to perform a risk-based cleanup status evaluation for the whole site, a strategy
was developed as follows.
• Estimate concentration versus distance decay coefficient from plume centerline
wells.
• Extrapolate concentration versus distance for each well using this decay
coefficient.
• Comparing the extrapolated concentrations with the compliance concentration
using power analysis.
Long Prairie Site
Long Prairie, Minnesota
17
MAROS 2.0 Application
Monitoring Network Optimization
-------
GROUNDWATER
February 19, 2003 SERVICES, INC.
Results from this analysis can be Attained or Not Attained, providing a statistical
interpretation of whether the cleanup goal has been met on the site-scale from the risk-
based point of view. The results as a function of time can be used to evaluate if the
monitoring system has enough power at each step in the sampling record to indicate
certainty of compliance by the plume location and condition relative to the compliance
boundary. For example, if results are Not Attained at early sampling events but are
Attained in recent sampling events, it indicates that the recent sampling record provides
a powerful enough result to indicate compliance of the plume relative to the location of
the receptor or compliance boundary.
Long Prairie Site IS MAROS 2.0 Application
Long Prairie, Minnesota Monitoring Network Optimization
-------
GROUNDWATER
February 19, 2003 SERVICES, INC.
3.0 SITE RESULTS
The groundwater long-term monitoring plan for Long Prairie was started in 1996. The
monitoring plan consisted of performance monitoring and compliance monitoring with the
following goals:
1) plume containment monitoring to confirm that the PCE plume remains
hydraulically controlled; and
2) plume reduction monitoring to verify progress toward achieving cleanup goals.
31 monitoring wells, 3 city wells, and 10 extraction wells were included in the long-term
monitoring network as of 2002 (Figure 1). The monitoring well naming convention
includes: "a" wells, shallow wells screened at the water table; "b" wells, mid-depth wells
screened at the base of the upper outwash; and "c" wells, deep wells screened in the
lower outwash. The sampling frequency for the long-term monitoring wells varies:
extraction wells have generally been sampled quarterly while monitoring wells were
generally sampled semiannually or annually since the implementation of the long-term
monitoring plan in 1996. For some wells, sampling was even terminated for 3 years
before they were sampled again in October 2002. This resulted in some monitoring
wells having only 5 ~ 7 data records during the 7-year period (from 1996 to 2002).
Monitoring data from 1996 to 2002 were used for the detailed optimization analysis, with
a subset of this data used in some of the analyses.
In applying the MAROS methodology to develop a revised monitoring strategy for the
Long Prairie site, many site and dataset parameters were applied. General site
assumptions include:
• All wells that were part of the network in between 1996 and 2002 were
considered in the temporal concentration trend analysis.
• Four chemicals of concern (COCs) that have been historically present at the site:
tetrachloroethene (PCE), trichloroethene (TCE), cis-1,2-dichloroethene (cis-1,2-
DCE), and vinyl chloride (VC), however, PCE is the predominant chemical and
has been used as an indicator parameter in the MAROS analyses.
• All source/tail assignments were made based on the PCE plume. Source wells
were selected based on historically elevated concentrations of PCE, near the dry
cleaner site in the vicinity of well MW-10.
• Site-specific hydrogeologic parameters related to the upper outwash aquifer
including groundwater flow direction, seepage velocity, saturated thickness,
porosity, receptor locations, can be found in the Table 4.
• Monitoring data from 1996 to 2002 were used for the "overall" trend analysis in
the sampling frequency optimization and other analyses in the MAROS detailed
optimization analysis.
Long Prairie Site 19 MAROS 2.0 Application
Long Prairie, Minnesota Monitoring Network Optimization
-------
GROUNDWATER
February 19, 2003 SERVICES, INC.
3.1 Data Consolidation
In MAROS, ground water monitoring data can be imported from simple database-format
Microsoft® Excel spreadsheets, Microsoft® Access tables, previously created MAROS
database archive files, or entered manually. The historical monitoring data from Long
Prairie were received in Excel database format. The columns in the file where formatted
to the MAROS Access file import format and then imported into the MAROS software
using the import tool. The long-term monitoring raw data contained many non-detects,
trace level results, and duplicates. Therefore, in the MAROS software the raw data are
filtered, consolidated, and the period of interest was specified (i.e. monitoring data from
1996 to 2002) as well as the wells of interest for the zone of interest. For statistical
evaluation of the data, a representative value for each sample point in time is needed.
MAROS has many automated options to choose how these values are assigned. For
the Long Prairie data, non-detects values were chosen to be set to the minimum
detection limit, allowing for uniform detection limits over time. Trace level results were
chosen to be represented by their actual values and duplicates samples were chosen to
be assigned the average of the two samples. The reduced data for each well were
viewed as a time series in a graphical form on a linear or semi-log plot generated by the
software.
3.2 Overview Statistics: Plume Trend Analysis
3.2.1 Mann-Kendall/Linear Regression Analysis
The goal of the Mann-Kendall and Linear Regression temporal trend analysis is to
assess the historical trend in the concentrations over time. These trend estimates are
then analyzed to identify the source and tail region overall stability category as well as
gaining an understanding of the individual well concentrations overtime (see Figure 2 for
further details). The PCE historical data for monitoring wells the Upper Outwash Aquifer
as well as the recovery wells were assessed for trends. No data consolidation was
performed to condense the sampling into regular sample intervals.
Only 31 monitoring wells and 9 recovery wells had sufficient data within the time period
of 1996 to 2002 (at least 6 sample events) to assess the trends in the wells. Trend
results from the Mann-Kendall and Linear Regression temporal trend analysis for both
Upper Outwash Aquifer monitoring wells and extraction wells are given in Table 5. The
monitoring well trend results show that 2 out of 4 source wells and 24 out of 27 tail wells
have a Probably Decreasing, Decreasing, or Stable trend. Both methods gave similar
trend estimates for each well. The recovery well trend results show that 7 out of 10 wells
have a Probably Decreasing, Decreasing, or Stable trend. Both methods gave similar
trend estimates for each well. When considering the spatial distribution of the trend
results (Figures 5 and 6 - maps created in ArcGIS from MAROS results), the majority of
the decreasing trend results are located in the interior of the plume or near the source,
indicating a decreasing source. Another area of decreasing trends is in the vicinity of
the line of recovery or plume containment wells (Figures 7 and 8 - maps created in
ArcGIS from MAROS results).
Long Prairie Site 20 MAROS 2.0 Application
Long Prairie, Minnesota Monitoring Network Optimization
-------
February 19, 2003
If
GROUNDWATER
SERVICES, INC.
Well Type
Source
Tail
Recovery
Zone A MAROS Trend Analysis
PD, D, S
2 of 4 (50%)
24 of 27 (90%)
7 of 10 (70%)
I, PI
Oof 4(0%)
3 of 27 (10%)
Oof 10(0%)
Note: Decreasing (D), Probably Decreasing (PD), Stable (S), Probably Increasing (PI), and Increasing (I)
Although monitoring wells and recovery wells are present in the well network, these well
trend results need to be treated differently for the purpose of individual trend analysis
interpretation primarily due to the different course of action possible for the two types of
wells. For monitoring wells, strongly decreasing concentration trends may lead the site
manager to decrease their monitoring frequency, as well look at the well as possibly
attaining its remediation goal. Conversely, strongly decreasing concentration trends in
recovery wells may indicate ineffective or near-asymptotic contamination extraction,
which may in turn lead to either the shutting down of the well or a drastic change in the
extraction scheme. Other reasons favoring the separation of these two types of wells in
the trend analysis interpretation is the fact that they produce very different types of
samples. On average, the extraction wells possess screens that are twice as large and
extraction wells pull water from a much wider area than the average monitoring well.
Therefore, the potential for the dilution of extraction well samples is far greater than
monitoring well samples.
3.2.2 Moment Analysis
The moment analysis in the MAROS software was applied at the Long Prairie site in
order to gain a better understanding of the overall plume stability in the Upper Outwash
Aquifer. Monitoring well data from 1996 to 2002 were used for the moment analysis, the
wells utilized for the analysis are listed in Table 1. Sampling frequency for these wells
was very irregular, therefore, the spatial moment analyses were based on sampling
events redefined on a yearly basis, that is, data collected between January 1st and
December 31st of a year were treated as if from the same sampling event performed on
July 1st of that year, with the geometric mean result utilized for each location.
Moment trend results from the Zeroth, First, and Second Moment analyses for the Zone
A monitoring well network were varied. Moment Trend results from the moment trend
analysis for the selected Upper Outwash well dataset are given in the Moment Analysis
Report, Appendix B. Approximately 17 wells were used in the moment analysis. Wells
with redundant spatial concentration information were not utilized in the moment analysis
(i.e. MW-1A).
The zeroth moment analysis showed a stable trend (no change in dissolved mass) over
time (Appendix B). The zeroth moment or mass estimate can show high variability over
time, largely due to the fluctuating concentrations at the most contaminated wells as well
Long Prairie Site
Long Prairie, Minnesota
21
MAROS 2.0 Application
Monitoring Network Optimization
-------
February 19, 2003
If
GROUNDWATER
SERVICES, INC.
as a varying monitoring well network. In order to reduce the fluctuating factors that could
influence a mass trend, the data were consolidated to annual sampling and the zeroth
moment trend evaluated. Another factor to consider when interpreting the mass
increase over time is the change in the spatial distribution of the wells sampled
historically. At the Long Prairie site there were changes in the well distribution over time,
due to addition and subtraction of wells from the well network as well as changes in
sampling frequency.
Moment
Type
Zeroth
First
Second
Mann -Kendall Trend Analysis
Trend
Stable
Increasing
Stable to No Trend
Comment
The amount of dissolved mass has not fluctuated appreciably over time.
This matches results in Table 5, where 15% of wells had stable trends.
The center of mass moved away from the source area over along the
direction of groundwater flow.
Stable to no trend, indicating that wells representing very large areas
both on the tip and the sides of the plume show little change in
concentrations. The shape of the plume is relatively constant over time.
The first moment, or center of mass, for each sample event in the Upper Outwash
aquifer had an increasing distance relative to the approximate source location, see
Figure 9, as well as the MAROS First Moment Reports in Appendix B. The center of
mass showed some movement forward along the direction of groundwater flow. These
spatial and temporal trends in the center of mass distance from the source location can
indicate transient movement based on season variation in rainfall or other hydraulic
considerations. With appreciable movement in the center of mass as is the case at Long
Prairie as well as a stable to decreasing source (zeroth moment and individual well
trend analysis results), there is an indication that the near source area is remediating
faster (on a mass basis) than the other areas of the plume. So although the plume is
stable the relative concentrations in the source area are decreasing faster than the other
areas of the plume. This concentration decrease can be seen in comparing the1996
and 2002 maps in Appendix A.
The second moment, or spread of the plume over time in both the x and y directions for
the sample events, showed a stable to no trend, Appendix B. The second moment
provides a measure of the spread of the concentration distribution about the plume's
center of mass. Analysis of the spread of the plume indicates stability in the plume in
the y direction and stable to no trend in the x direction, indicating that wells representing
very large areas both on the tip and the sides of the plume show little change in
concentrations over time and that the overall shape of the plume is relatively constant
with time. This stable trend in the spread of the plume strengthens the individual well
Mann-Kendall and Linear Regression trend analysis spatially, where most of the tail
wells showed a decreasing or probably decreasing PCE concentration trend.
Long Prairie Site
Long Prairie, Minnesota
22
MAROS 2.0 Application
Monitoring Network Optimization
-------
GROUNDWATER
February 19, 2003 SERVICES, INC.
3.2.3 Overview Statistics: Plume Analysis
In evaluating overall plume stability, the trend analysis results and all monitoring wells
were assigned "Medium" weights within the MAROS software (as described in Figure 3),
assuming equal importance for each well and each trend result in the overall analysis.
Overview Statistics Results:
• Overall trend for Source region: Stable,
• Overall trend for Tail region: near Stable,
• Overall results from moment analysis indicate a stable to decreasing plume,
• Overall monitoring intensity needed: Moderate.
These results matched with the judgment based on the visual comparison of PCE
plumes over time, as well as the Moment Analysis. The PCE concentrations observed
over the history of monitoring at the site are plotted in Appendix A. The PCE plume
observed in 1996 was very similar to that of 2002, indicating that the PCE plume is
relatively stable over time.
For a generic plume, the MAROS software indicates:
• No recommendation for sampling frequency
• Upper Outwash Aquifer may need 35 wells for the sampling network
These MAROS results are for a generic site, and are based on knowledge gained from
applying the MAROS Overview Statistics. There is no recommendation for frequency of
sampling for the whole monitoring network due to some uncertainty in the trends and the
presence of an active remediation system. Also, the recommended the number of wells
seems high when applied to the entire site. So, although the overall plume trend
analysis shows a stable plume, no general sampling frequency recommendation was
assessed by the MAROS software. Therefore, a more detailed analysis was performed
using the MAROS 2.0 software in order to allow for possible reductions on a well-by-well
basis, frequency and well redundancy analysis were conducted. These overview
statistics were also used when evaluating a final recommendation for each well after the
detailed statistical analysis was applied.
Long Prairie Site 23 MAROS 2.0 Application
Long Prairie, Minnesota Monitoring Network Optimization
-------
GROUNDWATER
February 19, 2003 SERVICES, INC.
3.3 Detailed Statistics: Optimization Analysis
From May 1996 to October 2002, a total of 44 wells were sampled, among which there
are 31 monitoring wells, 3 city wells, and 10 extraction wells (Table 1). Sampling
frequency for these wells varies: extraction wells were generally sampled quarterly while
monitoring wells were generally sampled semiannually or annually. A brief sampling
history for these wells is summarized in the last column of Table 1. All 44 wells were
used in the MAROS sampling optimization analysis. In the well redundancy and well
sufficiency analyses, the monitoring wells and some of the city wells were used (mostly
"B" wells, i.e., wells screened in the middle of the aquifer). In the sampling frequency
analysis, all 44 wells were analyzed. In data sufficiency analysis, only monitoring wells
were used. Results for well redundancy and sufficiency analyses, sampling frequency
analysis, and data sufficiency analysis are detailed in the following sub-sections.
3.3.1 Well Redundancy Analysis - Delaunav Method
The goal of the well redundancy analysis is to identify wells that are redundant within
monitoring network as candidates for removal from the sampling plan. The approach
allows elimination of sampling locations that have little impact on the historical
characterization of a groundwater plume. The analysis assumes that the current state of
hydraulic containment at the site will continue and than the monitoring network
adequately delineates the plume.
A monitoring network of 17 monitoring wells was used in the well redundancy analysis.
Clustered wells that are screened in different zones of the aquifer and had equivalent
duplicates or lower concentrations were excluded from the analysis (Table 1 lists the
wells excluded and the 17 wells used in the analysis). For example, wells MW-2A and
MW-2C are screened above and below the aquifer zone in which MW-2B is screened
(the middle zone or "B" zone), respectively, and both had concentrations lower than MW-
2B. Therefore, MW-2B instead of MW-2A and MW-2C was used in the well redundancy
analysis. In most cases, the "B" zone wells were used. But for well cluster MW-11A,
MW-1B, and MW-11C, MW-11C was used because it had more sampling records for
analysis and had the same concentration level as MW-2B. The well redundancy
analysis was conducted with the latest 3 years' sampling events (May 1999 to October
2002). The results show that no monitoring wells can be eliminated from this 17-well
network (Table 6).
However, after a qualitative consideration of the need for plume and site characterization
and the wells' concentration history, 9 monitoring wells (all "A" zone wells) can be
eliminated (Table 6). Using similar qualitative analysis, 3 extraction wells in the source
area can be eliminated since their concentrations were always below MCL or DL (Table
6). This resulted in a total of 12 wells recommended for removal, that is, a reduction of
27% in the well network (12 out of 44 wells).
Long Prairie Site 24 MAROS 2.0 Application
Long Prairie, Minnesota Monitoring Network Optimization
-------
February 19, 2003
If
GROUNDWATER
SERVICES, INC.
Well removal candidates include:
Monitoring wells
• MW-1A
• MW-2A
• MW-3A
• MW-4A
• Mw-5A
Monitoring wells
• MW-6A
• MW-11A
• MW-16A
• MW-18A
Extraction wells
• RW-1A
• RW-1B
• RW-1C
Eliminating the above 12 wells has negligible influence on spatial plume characterization
(Figure 10). For monitoring wells that have sampling ports at different levels ("A", "B", or
"C" zones), only "B" zone or "C" zone wells were used in the 2-D plume contouring. "B"
zone and "C" zone wells were generally higher in PCE concentrations than their
corresponding "A" zone wells, resulting in conservative estimates of the spatial plume
distribution. All monitoring wells that are candidates for elimination from the monitoring
network, generally "A" zone wells, plume contouring was not affected and therefore the
plumes before and after well elimination (Figure 10) are identical.
In considering the MAROS redundancy analysis results, other factors needed to be
taken into consideration before recommending well removals. Recent concentrations in
7 of the 9 wells were all below detection limits (MW-1A, MW-3A, MW-5A, MW-11A, MW-
15A, MW-16A, MW-18A), indicating that these wells do not and probably will not
contribute to the vertical plume delineation. In the case of MW-2A and MW-4A, these
were eliminated because their concentrations were all lower than their corresponding "B"
zone wells. Since the dissolved PCE plume was originated from DNAPL, without vertical
upward movement of groundwater, it typically tends to migrate downward vertically and
stay at the bottom of the aquifer. Therefore, although the "A" zone wells are candidates
for elimination, their corresponding "B" (or "C") zone wells were kept. Also, in plume
contouring for the purpose of delineating the plume extent horizontally, using "B" wells
can provide a more conservative estimate (i.e., a larger plume) than using "A" zone wells
due to their typically higher concentrations. For example, Figure 11 depicts the
approximate PCE plumes observed in May 1999 and October 2002 using data from "B"
zone wells only. In order to monitor the possible vertical migration of the plume in the
aquifer, all "C" zone wells were kept, ensuring enough information will be available for
plume characterization vertically. In considering the recovery wells, three wells were
identified for removal from the monitoring network because these 3 wells (RW-1A, RW-
1B, and RW-1C) have been extracting groundwater have been consistently below
detection limits or the MCL.
3.3.2 Well Sufficiency Analysis - Delaunav Method
The well sufficiency analysis for recommending new sampling locations was performed
using the same set of monitoring wells as in the well redundancy analysis. With this
analysis, areas within the monitoring well network where there is high uncertainty for
predicting concentrations are recommended for additional sample locations within the
Long Prairie Site
Long Prairie, Minnesota
25
MAROS 2.0 Application
Monitoring Network Optimization
-------
February 19, 2003
If
GROUNDWATER
SERVICES, INC.
existing well network. The SF values obtained from the well redundancy analysis were
used to generate Figure 12, which indicates the triangular regions for placing new
sampling locations. It is seen that almost all triangular regions (except one to the east of
the plume source area) have M (medium) estimation errors. The region with L (large)
result has a SF value of 0.602, which is not significantly larger than the M regions whose
estimated SF values range from 0.3 to 0.6. Considering that the plume is stable to
decreasing (Figure 10 and Section 3.2 results) and that the source of contamination has
been remediated (Barr, 2001), a new sampling location to the east of the plume source
area is not necessary. Therefore, no new locations were recommended.
3.3.3 Sampling Frequency Analysis- Modified CES Method
Results from the sampling frequency analysis for the 31 monitoring wells, 3 city wells,
and 10 extraction wells are given in Table 7. Some of the annual or quarterly sampling
frequency recommendations were due to insufficient overall or recent data (i.e., less
than 6 data records), which prevented the MAROS estimation of concentration trend
using overall or recent monitoring data for some wells. After considering the MAROS
results along with the historical and recent concentration levels at these wells, final
sampling frequency recommendations are provided in Table 7. For the monitoring well
system well, considering all 31 wells prior to the well redundancy analysis, 18 wells can
be sampled biennially, and 13 annually. All 3 city wells are recommended to be sampled
biennially. For the recovery well system, considering all 10 wells prior to the well
redundancy analysis, 4 can be sampled biennially, and 6 annually. If only considering
wells that have been sampled consistently up to October 2002 (27 wells) and the sample
frequency reduction alone, a reduction of approximately 44% in total samples per year
can be achieved (see the breakdown table below).
Current Sampling Frequency
Frequency
Quarterly
Annually
Total samples per year
Number of Wells
8
19
51
Recommended Sampling Frequency
Frequency
Biennial
Annual
Total samples per year
Number of Wells
9
18
22.5
In some cases, the frequency recommendations from the MAROS software were not
adopted due to data inadequacy. Sampling frequencies for some of the wells were
irregular, ranging from quarterly to annual during the period between 1996 and 2002. For
some wells, sampling was even terminated for 3 years before they were sampled again
in October 2002. This resulted in some monitoring wells having only 5 ~ 7 data records
available during the 7-year period (from 1996 to 2002). Because the minimum data
requirement for the sampling frequency trend analysis is 6 sampling events, the overall
or recent trends for many wells could not be estimated. This resulted in frequency
results that were solely estimated from the overall data trend. For instance, well MW-3B
has only 3 concentration records between May 1999 and October 2002, making the
estimation of recent trend impossible. In cases of data inadequacy, the MAROS
frequency analysis will assign conservative results, i.e., semiannual or annual instead of
Long Prairie Site
Long Prairie, Minnesota
26
MAROS 2.0 Application
Monitoring Network Optimization
-------
GROUNDWATER
February 19, 2003 SERVICES, INC.
annual or biennial. However, with the incorporation of a qualitative assessment of the
concentration levels and concentration history for these wells, more reasonable
sampling frequencies were recommended (Table 7). For example, well MW-3B was
suggested for a biennial sampling because all its historical concentrations were below
the detection limit. Considering the plume stability which remains stable according to the
overview statistical analysis, it is unlikely that the plume will show rapid changes over the
long-term. Therefore, keeping the frequency of wells at annual and biennial level will
continue to allow for adequate plume delineation.
3.3.4 Data Sufficiency - Power Analysis
In the MAROS data sufficiency analysis, statistical power analysis was used to assess
the sufficiency of monitoring plans for detecting difference between the mean
concentration and cleanup goal. Results from the analysis indicate remediation
progress from the risk-based standpoint at a hypothetical statistical compliance
boundary (HSCB). The power and expected sample size associated with the target level
evaluation may indicate the need for expansion or redundancy reduction of future
sampling plans.
In the risk-based site cleanup evaluation, two analyses were performed (see Appendix B
for all related MAROS reports). In the first analysis, the distance from the most
downgradient well (MW-15 A and B) to the nearest downgradient receptor (HSCB) was
assumed to be 10 ft (just upgradient of the Long Prairie river). The general groundwater
flow angle is to the North. Selected plume centerline wells are MW-15B, MW-16B, MW-
17B, MW-4B, and MW-2B (Table 8). Sampling events from May 1999 to October 2002
were selected for the analysis (Table 9). Among these only 5 sampling events have
sufficient data for plume centerline concentration regression. Regression coefficients for
the 5 sampling events range from 1.5 x 10"3 to 5.4 x 10"3 per ft, all with high confidence
(Table 9 and see Appendix B for individual well projected concentration values). The
second analysis used the same parameters except that the distance to receptor was
assumed to be -100 ft, i.e., assuming that the HSCB is 100 ft upgradient of the most
downgradient well.
Table 10 shows the risk-based site cleanup status at selected sampling events for both
analyses (i.e., HSCB at 10 ft and HSCB at -100 ft downgradient of the monitoring
system). The results show that the risk-based site cleanup status in most cases for both
analyses is "attained", i.e., the projected mean site concentration at the HSCB is
statistically significantly lower than the target level. Similarly, the associated power is
high and the expected sample size is relatively small. The "not attained" results, all have
low power (<0.5) but not S/E (significantly exceed) status, indicate that the projected
mean site concentration at the HSCB is lower than the target level but is not statistically
significant. The "not attained" results can be explained by the small dataset (18
samples) available for these sampling events (September 2000 and September 2001),
while all other sampling events have more than 20 samples for analysis. Therefore, the
results indicate that the site is clean at the HSCB that is -100 ft downgradient (equal to
100 ft upgradient of the monitoring system from the risk-based standpoint). Also, with
Long Prairie Site 27 MAROS 2.0 Application
Long Prairie, Minnesota Monitoring Network Optimization
-------
GROUNDWATER
February 19, 2003 SERVICES, INC.
the plume stable to shrinking over time, the HSCB will move upgradient gradually. The
HSCB is getting tighter and tighter as the monitoring record increases. In general,
monitoring networks become more powerful over time. This analysis indicates that the
monitoring system is working because it is powerful enough to accurately reflect the
location of the plume relative to the compliance point. Therefore, the current monitoring
network is sufficient in terms of evaluating risk-based site target level status, if the pump-
and-treat remedial system continues to contain the plume and keeps reducing the
contaminant concentration.
Long Prairie Site 28 MAROS 2.0 Application
Long Prairie, Minnesota Monitoring Network Optimization
-------
GROUNDWATER
February 19, 2003 SERVICES, INC.
4.0 SUMMARY AND RECOMMENDATIONS
In recent years, the high cost of long-term monitoring as part of active or passive
remediation of affected ground water has made the design of efficient and effective
ground water monitoring plans a pressing concern. Periodically updating and revising
long-term monitoring programs with changing conditions at the site can mean
considerable savings in site monitoring costs. The MAROS decision-support software
presented in this report assists in revising existing long-term monitoring plans based on
the historical and current monitoring data and plume behavior over time.
The MAROS 2.0 sampling optimization software/methodology has been applied to the
Long Prairie existing long-term monitoring program as of October 2002. The
optimization results and subsequent recommendations allow for optimization of the
spatial and temporal groundwater monitoring system in place at the Long Prairie site.
The current long-term monitoring network could be optimized through reduction in both
sampling locations and sampling frequency (results are summarized in Table 11 and
MAROS Reports in Appendix B).
Overview Statistics
Both the Mann-Kendall and Linear Regression temporal trend methods gave similar
trend estimates for each well. Results from the temporal trend analysis indicate that
90% of the plume tail and edge area monitoring wells in the Upper Outwash Aquifer
indicate a Probably Decreasing, Decreasing, or Stable PCE concentration trend,
whereas only about half of the wells in the source area have similar trends. The trend
results for the recovery wells along the centerline of the plume indicate most wells have
Probably Decreasing, or Decreasing concentrations over time. These temporal trend
results were applied to the
Results from the moment trend analysis give evidence of a stable plume as well, with the
dissolved mass showing stability over time, whereas the center of mass shows
movement away from the source area due to decreasing source area concentrations
and the plume spread shows stability over time. Overall plume stability temporal results
recommend a moderate monitoring strategy due to the stable PCE plume. The overview
results are relatively generic and not well-by-well specific, therefore, a detailed statistical
analysis with a well-by-well analysis was performed.
Detailed Statistics
Further analysis from the well redundancy analysis using the Delaunay method indicate
that 12 monitoring wells could be eliminated from the original monitoring network of 44
wells without any significant loss of plume information (Table 11). The well sufficiency
analysis indicated no need for adding new wells into the current monitoring system. The
resulting reduction in sampling locations would therefore be 27% (12 out of 44).
Long Prairie Site 29 MAROS 2.0 Application
Long Prairie, Minnesota Monitoring Network Optimization
-------
GROUNDWATER
February 19, 2003 SERVICES, INC.
The sampling frequency optimization analysis using the modified CES method, indicated
that the wells in the monitoring system could be sampled at annual or biennial
frequency, lower than the current sampling frequency overall. If only considering wells
that have been sampled consistently up to October 2002 (27 wells) and the sample
frequency reduction alone, this is a reduction of approximately 47% in total samples per
year.
Data sufficiency analysis using power analysis methods, shows that the site has reached
the target cleanup levels (0.005 mg/L for PCE) at the HSCB that is 100 ft upgradient
from the most downgradient well. With the plume stable to shrinking over time, the
HSCB will move upgradient gradually. This analysis indicates that the monitoring
system is working because it is powerful enough to accurately reflect the location of the
plume relative to the compliance boundary. This analysis shows the sufficiency of the
monitoring system in terms of evaluating risk-based site target level status if the pump-
and-treat remedial system continues to contain the plume and keeps reducing the
contaminant concentration in the aquifer.
The recommended long-term monitoring strategy results in a reduction in sampling costs
and allows site personnel to develop a better understanding of plume behavior over
time. A reduction in the number of redundant wells is expected to result in a moderate
cost savings over the long-term at the Long Prairie site. Overall, the MAROS optimized
plan consists of 32 wells: 16 sampled annually, and 16 sampled biennially ($6,720). The
MAROS optimized plan would result in 24 samples per year, compared to 51 samples
per year in the current monitoring program (Table 11). Implementing these
recommendations could lead to a 52% reduction from the current monitoring plan in
terms of the samples to be collected per year. An approximate cost savings estimate
range from $2,700 to $7,560 per year (based on an average per sample cost range of
$100 to $280) is projected while still maintaining adequate delineation of the plume as
well as knowledge of the plume state over time.
Long Prairie Site 30 MAROS 2.0 Application
Long Prairie, Minnesota Monitoring Network Optimization
-------
GROUNDWATER
February 19, 2003 SERVICES, INC.
CITED REFERENCES
Air Force Center for Environmental Excellence (AFCEE), 2002. Monitoring and
Remediation Optimization System (MAROS) 2.0 Software Users Guide.
Air Force Center for Environmental Excellence (AFCEE), 1997, AFCEE Long-Term
Monitoring Optimization Guide, http://www.afcee.brooks.af.mil.
Aziz, J, Ling, M., Newell, C., Rifai, H., and Gonzales, J., 2002, MAROS: a Decision
Support System for Optimizing Monitoring Plans, Ground Water, In Press.
Barr, 1999, 1997/1998 Annual Report, Long Prairie GW Remediation System July 1999
Barr, 2000, 1998/1999 Annual Report, Long Prairie GW Remediation System Feb 2000
Barr, 2001, 1999/2000 Annual Report, Long Prairie GW Remediation System March
2001
Barr, 2002, 2000/2001 Annual Report, Long Prairie GW Remediation System April 2002
Gilbert, R. O., 1987, Statistical Methods for Environmental Pollution Monitoring, Van
Nostrand Reinhold, New York, NY, ISBN 0-442-23050-8.
Ridley, M.N. et al., 1995. Cost-Effective Sampling of Groundwater Monitoring Wells, the
Regents of UC/LLNL, Lawrence Livermore National Laboratory.
Long Prairie Site 31 MAROS 2.0 Application
Long Prairie, Minnesota Monitoring Network Optimization
-------
February 19,2003
GSI Job No. G-2236-15
V
GROUNDWATER
SERVICES, INC.
MAROS 2.0 APPLICATION
MONITORING NETWORK OPTIMIZATION
Long Prairie Site
Long Prairie, Minnesota
TABLES
Table 1 Sampling Locations Used in the MAROS Analysis
Table 2 Mann-Kendall Analysis Decision Matrix
Table 3 Linear Regression Analysis Decision Matrix
Table 4 Upper Outwash Aquifer Aquifer Site-Specific Parameters
Table 5 Results of Upper Outwash Aquifer Trend Analysis
Table 6 Redundancy Analysis Results - Delaunay Method
Table 7 Sampling Frequency Analysis Results - Modified CES
Table 8 Selected Plume Centerline Wells for Risk-Based Site Cleanup Evaluation
- Power Analysis
Table 9 Plume Centerline Concentration Regression Results - Power Analysis
Table 10 Risk-Based Site Cleanup Evaluation Results - Power Analysis
Table 11 Summary of MAROS Sampling Optimization Results
-------
GSI Job No. G-2236-15
Issued: 2/19/03
Page 1 of 3
If
GROUNDWATER
SERVICES, INC.
Table 1
Sampling Locations Used in the MAROS Analysis
Long Prairie Site
Long Prairie, Minnesota
Well
Name
BAL2B
BAL2C
CW-3
CW-6
MW-1A
MW-1B
MW-2A
MW-2B
MW-2C
MW-3A
MW-3B
MW-4A
MW-4B
MW-4C
MW-5A
Used in Delaunay Analysis?
No, city well and far from the
plume
Yes
Yes
No, city well and far from the
plume
No, screened above and
duplicates MW-1B
Yes
No, screened above MW-2B
and lower in concentration
Yes
No, screened below MW-2B
and lower in concentration
No, screened above and
duplicates MW-3B
Yes
No, screened above MW-4B
and lower in concentration
Yes
No, screened below MW-4B
and lower in concentration
No, screened above and
duplicates MW-5B
Used in Modified
CES Analysis?
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Summary of Sampling History (sampling data
available since 1996)
Sampled only once in October 2002, below DL
Sampled semiannually until October 1999, then
sampled in October 2002; all records were below DL
Sampled quarterly on average; all records were
below DL
Sampled quarterly until March 2000, then sampled
quarterly since April 2002; all records were below DL
Sampled semiannually until October 1999, then
sampled in October 2002; all records were below DL
Sampled semiannually until October 1999, then
sampled in October 2002; all records were below DL
Sampled annually since 1997
Sampled annually since 1997
Sampled annually since 1 997 except when in 1999 it
was sampled several times
Sampled semiannually until October 1999, then
sampled in October 2002; all records were below
MCL or DL
Sampled semiannually until October 1999, then
sampled in October 2002; all records were below DL
Sampled only in 1999, then sampled in October
2002
Sampled annually since 1997
Sampled annually except when in 1999 it was
sampled several times
Sampled semiannually until October 1999, then
sampled in October 2002; all records were below DL
Notes:
MCL= the maximum contaminant level of PCE (0.005 mg/L)
DL = detection limit
-------
GSI Job No. G-2236-15
Issued: 2/19/03
Page 2 of 3
If
GROUNDWATER
SERVICES, INC.
Table 1
Sampling Locations Used in the MAROS Analysis
Long Prairie Site
Long Prairie, Minnesota
Welt
Name
MW-5B
MW-6A
MW-6B
MW-6C
MW-10
MW-11A
MW-11B
MW-11C
MW-13C
MW-14B
MW-14C
MW-15A
MW-15B
MW-16A
MW-16B
Used in Detaunay
Analysis?
Yes
No, screened above MW-6B
and generally lower in
concentration
Yes
No, screened below MW-6B
and lower in concentration
Yes
No, screened above and
duplicates MW-11C
No, screened above and
duplicates MW-11C
Yes, it has more records
than MW-11B
Yes
Yes
No, screened below MW-
14B and lower in
concentration
No, screened above and
duplicates MW-5B
Yes
No, screened above MW-5B
and lower in concentration
Yes
Used in Modified
CES Analysis?
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Summary of Sampling History (sampling data
available since 1996)
Sampled semiannually until October 1999, then
sampled in October 2002; all records were below DL
Sampled annually since 1997
Sampled annually since 1997
Sampled annually since 1 998 except when in 1999 it
was sampled several times
Sampled semiannually until October 1999, then
sampled in October 2002
Sampled only in 1999, then sampled in October
2002; all records were below DL
Sampled annually since 1999; all records were
below DL
Sampled annually since 1 997 except when in 1999 it
was sampled several times; all records were below
MCL or DL
Sampled semiannually until October 1999, then
sampled in October 2002; all records were below
MCL or DL
Sampled semiannually until October 1999, then
sampled annually
Sampled semiannually until October 1999, then
sampled annually; all records were below MCL or DL
Sampled semiannually since October 1998, then
sampled annually since October 1999; all records
were below DL
Sampled semiannually since October 1998, then
sampled annually since October 1999; all records
were below DL
Sampled 5 times between 1998 and 1999, then
sampled in October 2002; all records were below
MCL or DL
Sampled 5 times between 1998 and 1999, then
sampled annually
Notes:
MCL= the maximum contaminant level of PCE (0.005 mg/L)
DL = detection limit
-------
GSI Job No. G-2236-15
Issued: 2/19/03
Page 3 of 3
If
GROUNDWATER
SERVICES, INC.
Table 1
Sampling Locations Used in the MAROS Analysis
Long Prairie Site
Long Prairie, Minnesota
Welt
Name
MW-17B
MW-18A
MW-18B
MW-19B
RW-1A
RW-1B
RW-1C
RW-3
RW-4
RW-5
RW-6
RW-7
RW-8
RW-9
Used in Detaunay
Analysis?
Yes
No, screened above MW-
18B and lower in
concentration
Yes
Yes
No, recovery well
No, recovery well
No, recovery well
No, recovery well
No, recovery well
No, recovery well
No, recovery well
No, recovery well
No, recovery well
No, recovery well
Used in Modified
CES Analysis?
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Summary of Sampling History (sampling data
available since 1996)
Sampled 4 times between 1998 and 1999, then
sampled annually
Sampled 4 times between 1998 and 1999, then
sampled in October 2002; all records were below DL
Sampled 4 times between 1998 and 1999, then
sampled in October 2002; all records were below
MCL but higher than DL
Sampled 4 times between 1998 and 1999, then
sampled annually; all records were below DL
Sampled quarterly on average until 1999, then
sampled in October 2002
Sampled quarterly on average until 1999, then
sampled in October 2002
Sampled only once in October 2002
Sampled quarterly
Sampled quarterly until 99, then sampled annually
Sampled quarterly
Sampled quarterly
Sampled quarterly
Sampled quarterly since 1999
Sampled quarterly since 1999
Notes: MCL= the maximum contaminant level of PCE (0.005 mg/L)
DL = detection limit
-------
GSI Job No. G-2236-15
Issued 2/19/03
Page 1 of 1
If
GROUNDWATER
SERVICES, INC.
Mann-Kendall
Mann-Kendall
Statistic
S>0
S>0
S>0
S<0
S<0
S<0
S<0
Table 2
Analysis Decision Matrix
Confidence in the
Trend
> 95%
90 - 95%
< 90%
< 90% and COV > 1
< 90% and COV < 1
90 - 95%
> 95%
(Aziz, et. al., 2002)
Concentration Trend
Increasing
Probably Increasing
No Trend
No Trend
Stable
Probably Decreasing
Decreasing
Table 3
Linear Regression Analysis Decision Matrix (Aziz, et. al., 2002)
Confidence in the
Trend
Log-slope
Positive
Negative
< 90%
90 - 95%
> 95%
No Trend
Probably Increasing
Increasing
COV < 1 Stable
COV > 1 No Trend
Probably Decreasing
Decreasing
-------
*J
"S
E
5
re
CL
O
'o
0) 0
Q. ±±
0) 0
UJ ±i c
CD
Q- it
D)
O
0>
Q.
Q.
CO
CO
CM
0)
re
^
jlj
*s
E
re
re
Q.
,_
^v-
i= v°
ij± 0^
fS °
Source
_c
en
03
O
CD
>v CL
~ CL
CJ — i
O — '
"CD 03
> 03
CD . E
°* -"'X
S w o
CL /-, s_
CD C CL
CD O CL
CO CL <
CO
O
V
1
CD
c~
O
'^— '
8
o
^
o
CD
Saturated
_c
en
03
0
CD
CL
CL
£2
"™ en
E tn
x ®
2 5
a. o
CL jz
< 1-
,_
•z.
irection
Q
^
o
LL
^
"03
£Z
2
0
(General
en
CD
0
_g
CD
"CD
c
g
o>
c
en
c
g
•c
CD
"CD
o
_g
c
en
"5
en
CD
D
CL
CD
r-
'o
•2-
c
'o
CD
£
c
o>
c
'CL
0.
d
g
"o
.i
"2
Q
.2
CD
-o
c
g
o>
"CD
c
g
g
d
o
o
CD
outwash
CD
CD
^0
E
JD
'5
cr
CD
'o
CD
CL
c
0
CD
en
CD
JD
c
g
"CD
E
'x
CL
CL
C
CD
en
3
"en
CD
o>
.^
c
CD
CD
O
£•
CD
£
CD
"5
C
g
8
_g
CD
-^1
-o
CD
c
g
CD
i=
§
c
LU
O
CL
1
O
"en
^
g)
E
o
1
3
JD
0
CD
g
CD
§
O
in
•5
g
8
o
Oo
~5 CU CU
55 to c?
O «QL
-------
GSIJobNo. G-2236-15
Issued 2/19/03
Page 1 of 2
TABLE 5
Upper Outwash Aquifer Trend Analysis
Long Prairie Site
Long Prairie, Minnesota
Notes:
1. Consolidation of data included non-detect values set to the minium detection limit (0.001 mg/L)
and duplicate data for the quarter were averaged.
2. All wells that were part of the network in between 1996 and 2002 with more than 4 sample events were analyzed.
3. RW = Recovery Well; MW = Monitoring Well; CSW = City Supply Well
4. Decreasing (D), Probably Decreasing (PD), Stable (S), No Trend (NT), Probably Increasing (PI), and Increasing (I)
5. S = Source Zone Well; T = Tail Zone Well
6. Overall Trend is calculated from a weighted average of the Linear Regression and Mann-Kendall Trends.
For further details on this methodolgy refer to the MAROS Manual Appendix A.8.
mtOUNDWATEK
SE1¥!C1S, IMC
Well
CW-3
CW-6
BAL2C
MW-1A
MW-1B
MW-2A
MW-2B
MW-2C
MW-3A
MW-3B
MV\MB
MW-4C
MW-5A
MW-5B
MW-6A
MW-6B
MW-6C
MW-10
MW-11B
MW-11C
MW-13C
MW-14B
MW-14C
MW-15A
MW-15B
MW-16A
MW-16B
MW-17B
MW-18A
MW-18B
MW-19B
RW-1A
RW-1B
RW-3
RW-4
RW-5
RW-6
RW-7
RW-8
RW-9
Well
Type3
CSW
CSW
MW
MW
MW
MW
MW
MW
MW
MW
MW
MW
MW
MW
MW
MW
MW
MW
MW
MW
MW
MW
MW
MW
MW
MW
MW
MW
MW
MW
MW
RW
RW
RW
RW
RW
RW
RW
RW
RW
Well
Category ''
T
T
T
T
T
S
S
S
T
T
T
T
T
T
T
T
T
S
T
T
T
T
T
T
T
T
T
T
T
T
T
T
T
S
T
T
T
T
T
T
Mann-Kendall
Trend 4
S
S
S
S
S
NT
NT
S
NT
S
D
S
S
S
NT
D
PD
D
S
S
NT
PD
S
S
S
S
S
D
S
S
S
PD
NT
D
D
D
D
D
D
NT
Linear
Regression
Trend 4
S
I
S
S
S
NT
NT
S
NT
S
D
S
S
S
NT
D
D
D
S
PD
NT
D
S
D
D
S
NT
D
S
I
I
NT
NT
D
D
D
D
D
D
NT
Overall
Trend 6
S
PI
S
S
S
NT
NT
S
NT
S
D
S
S
S
NT
D
D
D
S
S
NT
D
S
PD
PD
S
S
D
S
PI
PI
S
NT
D
D
D
D
D
D
NT
Number
of
Samples
24
14
8
8
8
6
6
8
8
8
6
8
8
8
6
6
7
8
4
8
9
10
11
7
7
6
8
7
5
5
7
12
12
25
15
25
25
25
12
12
Number
of
Detects
0
0
0
0
0
6
6
5
1
0
6
8
0
0
5
6
7
8
0
1
2
10
1
0
0
1
8
7
0
5
0
11
1
25
6
25
25
25
12
12
-------
GSI Job No. G-2236-15
Issued: 2/19/03
Page 1 of 3
If
GROUNDWATER
SERVICES, INC.
Table 6
Well Redundancy Analysis Results - Delaunay Method
Long Prairie Site
Long Prairie, Minnesota
Well
Name
BAL2B
BAL2C
CW-3
CW-6
MW-1A
MW-1B
MW-2A
MW-2B
MW-2C
MW-3A
MW-3B
MW-4A
MW-4B
MW-4C
MW-5A
Well Used in
Analysis?
No, city well and far
from the plume
Yes
Yes
No, city well and far
from the plume
No, screened above
and duplicates MW-1B
Yes
No, screened above
MW-2B and lower in
concentration
Yes
No, screened below
MW-2B and lower in
concentration
No, screened above
and duplicates MW-3B
Yes
No, screened above
MW-4B and lower in
concentration
Yes
No, screened below
MW-4B and lower in
concentration
No, screened above
and duplicates MW-5B
MAROS Well
Redundancy
Analysis
Result
-
Keep
Keep
-
-
Keep
-
Keep
-
-
Keep
-
Keep
-
-
MAROS
Interpreted
Well
Redundancy
Keep
Keep
Keep
Keep
Eliminate
Keep
Eliminate
Keep
Keep
Eliminate
Keep
Eliminate
Keep
Keep
Eliminate
Comments
City well monitoring
Downgradient sentry well
City well monitoring
City well monitoring
Duplicates MW-1B and concentrations
below DL
Cross gradient sentry well
Monitors only the upper zone at MW-4 and
concentrations dropped to below MCL
Inside plume well and close the source
Monitors the lower zone at MW-2 for
possible downward or lower zone
contaminant migration
Duplicates MW-3B and historical
concentrations below MCL or DL
Downgradient of the source
Monitors only the upper zone at MW-4 and
lower in concentration than MW-4B
Inside plume well
Monitors the lower zone at MW-4 for
possible downward or lower zone
contaminant migration
Duplicates MW-5B and historical
concentrations below DL
Notes: Sampling events from May 1999 to October 2002 were used in the analysis
InsideSF = 0.1, HullSF = 0.01, AR = CR = 0.95
MCL = the maximum contaminant level of PCE (0.005 mg/L)
DL = detection limit
-------
GSI Job No. G-2236-15
Issued: 2/19/03
Page 2 of 3
If
GROUNDWATER
SERVICES, INC.
Table 6
Well Redundancy Analysis Results - Delaunay Method
Long Prairie Site
Long Prairie, Minnesota
Well
Name
MW-5B
MW-6A
MW-6B
MW-6C
MW-10
MW-11A
MW-11B
MW-11C
MW-13C
MW-14B
MW-14C
MW-15A
MW-15B
MW-16A
MW-16B
Well Used in Analysis?
Yes
No, screened above
MW-6B and generally
lower in concentration
Yes
No, screened below
MW-6B and lower in
concentration
Yes
No, screened above and
duplicates MW-1 1C
No, screened above and
duplicates MW-1 1C
Yes, it has more records
than MW-11B
Yes
Yes
No, screened below
MW-14B and lower in
concentration
No, screened above and
duplicates MW-5B
Yes
No, screened above
MW-5B and lower in
concentration
Yes
MAROS Well
Redundancy
Analysis
Result
Keep
-
Keep
-
Keep
-
-
Keep
Keep
Keep
-
-
Keep
-
Keep
MAROS
Interpreted
Well
Redundancy
Keep
Eliminate
Keep
Keep
Keep
Eliminate
Keep
Keep
Keep
Keep
Keep
Keep
Keep
Eliminate
Keep
Comments
Cross gradient sentry well
Monitors only the upper zone at MW-4
and lower in concentrations than MW-6B
Inside plume well
Monitors the lower zone at MW-6 for
possible downward or lower zone
contaminant migration
Continues to monitor the source for
abnormal conditions
Monitors only the upper zone at MW-1 1
and concentrations below DL
Monitors the middle zone at MW-1 1
Monitors the lower zone at MW-1 1 for
possible downward or lower zone
contaminant migration
Inside plume well
On plume edge
Monitors the lower zone at MW-1 4 for
possible downward or lower zone
contaminant migration
Monitors the upper zone at MW-1 5 since it
is just upgradient of the Long Prairie river
Downgradient sentry well just upgradient
of the Long Prairie river
Monitors only the upper zone at MW-1 6
and concentrations below MCL or DL
Downgradient well on plume edge
Notes: Sampling events from May 1999 to October 2002 were used in the analysis
InsideSF = 0.1, HullSF = 0.01, AR = CR = 0.95
MCL = the maximum contaminant level of PCE (0.005 mg/L)
DL = detection limit
-------
GSI Job No. G-2236-15
Issued: 2/19/03
Page 3 of 3
If
GROUNDWATER
SERVICES, INC.
Table 6
Well Redundancy Analysis Results - Delaunay Method
Long Prairie Site
Long Prairie, Minnesota
Well
Name
MW-17B
MW-18A
MW-18B
MW-19B
RW-1A
RW-1B
RW-1C
RW-3
RW-4
RW-5
RW-6
RW-7
RW-8
RW-9
Well Used in
Analysis?
Yes
No, screened above
MW-18B and lower in
concentration
Yes
Yes
No, recovery well
No, recovery well
No, recovery well
No, recovery well
No, recovery well
No, recovery well
No, recovery well
No, recovery well
No, recovery well
No, recovery well
MAROS Well
Redundancy
Analysis
Result
Keep
-
Keep
Keep
-
-
-
-
-
-
-
-
-
-
MAROS
Interpreted
Well
Redundancy
Keep
Eliminate
Keep
Keep
Eliminate
Eliminate
Eliminate
Keep
Keep
Keep
Keep
Keep
Keep
Keep
Comments
Inside plume well
Monitors only the upper zone at MW-18 and
concentrations below DL
Downgradient well on plume edge
Downgradient well close to plume tail
Source has been cleaned up and
concentrations below MCL or DL
Source has been cleaned up and
concentrations below MCL or DL
Source has been cleaned up and
concentrations below MCL or DL
Recovery well for performance monitoring
Recovery well for performance monitoring
Recovery well for performance monitoring
Recovery well for performance monitoring
Recovery well for performance monitoring
Recovery well for performance monitoring
Recovery well for performance monitoring
Notes: Sampling events from May 1999 to October 2002 were used in the analysis
InsideSF = 0.1, HullSF = 0.01, AR = CR = 0.95
MCL = the maximum contaminant level of PCE (0.005 mg/L)
DL = detection limit
-------
GSI Job No. G-2236-15
Issued: 2/19/03
Page 1 of 3
If
GROUNDWATER
SERVICES, INC.
Table 7
Sampling Frequency Analysis Results - Modified CES
Long Prairie Site
Long Prairie, Minnesota
Well
Name
BAL2B
BAL2C
CW-3
CW-6
MW-1A
MW-1B
MW-2A
MW-2B
MW-2C
MW-3A
MW-3B
MW-4A
MW-4B
MW-4C
MW-5A
MAROS
Frequency
Based on
Recent
Trend'11
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Quarterly
Annual
Annual
Annual
MAROS
Frequency
Based on
Overall
Trend121
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Quarterly
Annual
Annual
Annual
MAROS
Recommended
Frequency'31
Annual
Annual
Biennial
Biennial
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Quarterly
Annual
Annual
Annual
MAROS
Interpreted
Sampling
Frequency
Result
Biennial
Biennial
Biennial
Biennial
Biennial
Biennial
Annual
Annual
Annual
Biennial
Biennial
Annual
Annual
Annual
Biennial
Comments
Concentrations below DL (the MAROS
result was due to insufficient data)
Historical concentrations below DL (the
MAROS result was due to insufficient recent
data)
Historical concentrations below DL
Historical concentrations below DL
Historical concentrations below DL (the
MAROS result was due to insufficient recent
data)
Historical concentrations below DL (the
MAROS result was due to insufficient recent
data)
Recent concentrations above MCL
Recent concentrations above MCL
Recent concentrations below MCL but
above DL
Historical concentrations below MCL or DL
(the MAROS result was due to insufficient
recent data)
Historical concentrations below DL (the
MAROS result was due to insufficient recent
data)
Recent concentrations above MCL (the
MAROS result was due to insufficient data)
Recent concentrations above MCL
Recent concentrations above MCL
Historical concentrations below DL (the
MAROS result was due to insufficient recent
data)
Notes: 1) The frequency determined by MAROS based on the analysis of recent data (May 1999 ~ October 2002)
2) The frequency determined by MAROS based on the analysis of overall data (May 1996 ~ October 2002)
3) The frequency finally recommended by MAROS after considering recent and overall frequency results as well
as the rates of change in these trends
Rate parameters used are 0.5MCL/year, 1 .OMCL/year, and 2.0MCL/yearfor Low, Medium, and High rates,
respectively; MCL = the maximum contaminant level of PCE (0.005 mg/L); DL = detection limit
-------
GSI Job No. G-2236-15
Issued: 2/19/03
Page 2 of 3
If
GROUNDWATER
SERVICES, INC.
Table 7
Sampling Frequency Analysis Results - Modified CES
Long Prairie Site
Long Prairie, Minnesota
Well
Name
MW-5B
MW-6A
MW-6B
MW-6C
MW-10
MW-11A
MW-11B
MW-11C
MW-13C
MW-14B
MW-14C
MW-15A
MW-15B
MW-16A
MW-16B
MAROS
Frequency
Based on
Recent
Trend11'
Annual
Annual
Annual
Annual
Quarterly
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
MAROS
Frequency
Based on
Overall
Trend'2'
Annual
Annual
Annual
Annual
Quarterly
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
MAROS
Recommended
Frequency'3'
Annual
Annual
Annual
Annual
Quarterly
Annual
Biennial
Biennial
Biennial
Annual
Biennial
Biennial
Biennial
Biennial
Annual
MAROS
Interpreted
Sampling
Frequency
Result
Biennial
Annual
Annual
Annual
Annual
Biennial
Biennial
Biennial
Biennial
Annual
Biennial
Biennial
Biennial
Biennial
Annual
Comments
Historical concentrations below DL (the
MAROS result was due to insufficient recent
data)
Recent concentrations above MCL
Recent concentrations above MCL
Recent concentrations above MCL
Historical concentrations have dropped
significantly but recent concentrations still
above MCL (the MAROS result was due to
insufficient recent data)
Historical concentrations below DL (the
MAROS result was due to insufficient data)
Historical concentrations below DL
Historical concentrations below MCL or DL
Historical concentrations below MCL or DL
Recent concentrations above MCL
Historical concentrations below MCL or DL
Historical concentrations below DL
Historical concentrations below DL
Historical concentrations below MCL or DL
Recent concentrations above MCL
Notes: 1) The frequency determined by MAROS based on the analysis of recent data (May 1999 ~ October 2002)
2) The frequency determined by MAROS based on the analysis of overall data (May 1996 ~ October 2002)
3) The frequency finally recommended by MAROS after considering recent and overall frequency results as well
as the rates of change in these trends
Rate parameters used are 0.5MCL/year, 1 .OMCL/year, and 2.0MCL/yearfor Low, Medium, and High rates,
respectively; MCL = the maximum contaminant level of PCE (0.005 mg/L); DL = detection limit
-------
GSI Job No. G-2236-15
Issued: 2/19/03
Page 3 of 3
If
GROUNDWATER
SERVICES, INC.
Table 7
Sampling Frequency Analysis Results - Modified CES
Long Prairie Site
Long Prairie, Minnesota
Well
Name
MW-17B
MW-18A
MW-18B
MW-19B
RW-1A
RW-1B
RW-1C
RW-3
RW-4
RW-5
RW-6
RW-7
RW-8
RW-9
MAROS
Frequency
Based on
Recent
Trend11'
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
MAROS
Frequency
Based on
Overall
Trend'2'
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
MAROS
Recommended
Frequency'3'
Annual
Annual
Annual
Biennial
Annual
Annual
Annual
Annual
Biennial
Annual
Annual
Annual
Annual
Annual
MAROS
Interpreted
Sampling
Frequency
Result
Annual
Biennial
Biennial
Biennial
Biennial
Biennial
Biennial
Annual
Biennial
Annual
Annual
Annual
Annual
Annual
Comments
Recent concentrations above MCL
Historical concentrations below DL (the
MAROS result was due to insufficient data)
Historical concentrations below MCL (the
MAROS result was due to insufficient data)
Historical concentrations below DL
Historical concentrations below MCL or DL
(the MAROS result was due to insufficient
recent data)
Historical concentrations below DL (the
MAROS result was due to insufficient recent
data)
Recent concentrations below DL (the
MAROS result was due to insufficient data)
Recent concentrations above MCL
Recent concentrations below DL
Recent concentrations above MCL
Recent concentrations above MCL
Recent concentrations above MCL
Recent concentrations above MCL
Recent concentrations above MCL
Notes: 1) The frequency determined by MAROS based on the analysis of recent data (May 1999 ~ October 2002)
2) The frequency determined by MAROS based on the analysis of overall data (May 1996 ~ October 2002)
3) The frequency finally recommended by MAROS after considering recent and overall frequency results as well
as the rates of change in these trends
Rate parameters used are 0.5MCL/year, 1 .OMCL/year, and 2.0MCL/yearfor Low, Medium, and High rates,
respectively; MCL = the maximum contaminant level of PCE (0.005 mg/L); DL = detection limit
-------
GSI Job No. G-2236-15
Issued: 2/19/03
Page 1 of 1
If
GROUNDWATER
SERVICES, INC.
Table 8
Selected Plume Centerline Wells
Risk-Based Site Cleanup Evaluation - Power Analysis
Long Prairie Site
Long Prairie, Minnesota
Well Name
MW-15B
MW-16B
MW-17B
MW-4B
MW-2B
Distance from Well to Receptor (ft)
HSCB = 10ft
10.0
520.6
1063.2
1998.1
2897.1
HSCB = -100ft
-100
410.6
953.2
1888.1
2787.1
Notes: Groundwater flow angle is to the north/northwest (assumed 90 degrees
counterclockwise from East in this analysis); Distance from Well to Receptor
refers to the most downgradient well's Distance to the Hypothetical Statistical
Compliance Boundary (HSCB).
-------
GSI Job No. G-2236-15
Issued: 2/19/03
Page 1 of 1
If
GROUNDWATER
SERVICES, INC.
Table 9
Plume Centerline Concentration
Regression Results - Power Analysis
Long Prairie Site
Long Prairie, Minnesota
Sampling Event
May 1999
July 1999
September 1999
October 1999
March 2000
June 2000
September 2000
October 2000
December 2000
March 2001
May 2001
September 2001
November 2001
January 2002
April 2002
July 2002
October 2002
Number of
Center! ine Welts
5
0
2
3
0
0
5
0
0
0
0
5
0
0
0
0
5
Regression
Coefficient (1/ft)
-2.31 E-03
-
-
-5.38E-03
-
-
-1.50E-03
-
-
-
-
-1.51 E-03
-
-
-
-
-1.24E-03
Confidence in
Coefficient
97.9%
-
-
91.5%
-
-
92.7%
-
-
-
-
92.1%
-
-
-
-
91.3%
Notes: Regression is on natural log concentration of PCE versus distance from source
centerline wells shown in Table M4; no regression was performed for sampling event
with less than 3 centerline wells.
-------
GSI Job No. G-2236-15
Issued: 2/19/03
Page 1 of 1
If
GROUNDWATER
SERVICES, INC.
Table 10
Risk-Based Site Cleanup Evaluation Results - Power Analysis
Long Prairie Site
Long Prairie, Minnesota
Sampling Event
May 1999
October 1999
September 2000
September 2001
October 2002
Sample
Size
30
25
18
18
34
Distance to HSCB = 10 ft
Cleanup
Status
Attained
Attained
Not Attained
Attained
Attained
Power
1.000
1.000
0.414
0.685
1.000
Expected
Sample Size
7
<=3
54
25
8
Distance to HSCB = -100 ft
Cleanup
Status
Attained
Attained
Not Attained
Not Attained
Attained
Power
0.992
1.000
0.211
0.432
0.998
Expected
Sample Size
12
<=3
>100
50
11
Notes: The power analysis used for this application assumes normality of data. Distance to the Hypothetical
Statistical Compliance Boundary (HSCB) is the distance from the most downgradient well to the
HSCB; S/E = extrapolated result significantly exceeds the target level (0.005 mg/L).
-------
GSI Job No. G-2236-15
Issued: 2/19/03
Page 1 of 3
If
GROUNDWATER
SERVICES, INC.
Table 11
Summary of MAROS Sampling Optimization Results
Long Prairie Site
Long Prairie, Minnesota
Well
Name
BAL2B
BAL2C
CW-3
CW-6
MW-1A
MW-1B
MW-2A
MW-2B
MW-2C
MW-3A
MW-3B
MW-4A
MW-4B
MW-4C
MW-5A
MAROS
Welt
Category"1
T
T
T
T
T
T
S
S
S
T
T
T
T
T
T
Current
Sampling
Frequency'2'
Biennial*
Annual
Quarterly
Quarterly
Biennial*
Biennial*
Annual
Annual
Annual
Biennial*
Biennial*
Biennial*
Annual
Annual
Biennial*
MAROS
Trend
Result'31
NA
S
S
PI
S
S
NT
NT
S
NT
S
NA
D
S
S
MAROS
Interpreted Well
Redundancy
and Well
Sufficiency
Results
Keep
Keep
Keep
Keep
Eliminate
Keep
Eliminate
Keep
Keep
Eliminate
Keep
Eliminate
Keep
Keep
Eliminate
MAROS
Interpreted
Sampling
Frequency
Results
Biennial
Biennial
Biennial
Biennial
-
Biennial
-
Annual
Annual
-
Biennial
-
Annual
Annual
-
Comments
Concentration below DL & city well
monitoring
Historical concentrations below DL &
downgradient sentry well
Historical concentrations below DL & city
well monitoring
Historical concentrations below DL & city
well monitoring
Duplicates MW-1B & historical
concentrations below DL
Historical concentrations below DL &
cross gradient sentry well
Monitors only the upper zone at MW-4 and
concentrations dropped to below MCL
Recent concentrations above MCL &
inside plume well
Recent concentrations below MCL but
above DL & monitors the lower zone at
MW-2
Duplicates MW-3B & historical
concentrations below MCL or DL
Historical concentrations below DL &
downgradient of the source
Monitors only the upper zone at MW-4 and
lower in concentration than MW-4B
Recent concentrations above MCL &
inside plume well
Recent concentrations above MCL &
monitors the lower zone at MW-4
Duplicates MW-5B & historical
concentrations below DL
Notes: (1) S = Source well, T = Tail well
(2) Sampling frequency based on recent sampling results, 'assumed biennial due to lack of data in 2000 and 2001
(3) D = Decreasing, PD = Probably Decreasing, S = Stable, NT = No Trend, PI = Probably Increasing, I = Increasing
MCL = the maximum contaminant level of PCE (0.005 mg/L), DL = detection limit,"-" = Not Applicable
-------
GSI Job No. G-2236-15
Issued: 2/19/03
Page 2 of 3
If
GROUNDWATER
SERVICES, INC.
Table 11
Summary of MAROS Sampling Optimization Results
Long Prairie Site
Long Prairie, Minnesota
Welt
Name
MW-5B
MW-6A
MW-6B
MW-6C
MW-10
MW-11A
MW-11B
MW-11C
MW-13C
MW-14B
MW-14C
MW-15A
MW-15B
MW-16A
MW-16B
MAROS
Welt
Category'11
T
T
T
T
S
T
T
T
T
T
T
T
T
T
T
Current
Sampling
Frequency'2'
Biennial*
Annual
Annual
Annual
Annual
Biennial*
Annual
Annual
Biennial*
Annual
Annual
Annual
Annual
Biennial*
Annual
MAROS
Trend
Result'3'
S
NT
D
D
D
NA
S
PD
NT
D
S
PD
PD
S
S
MAROS
Interpreted Well
Redundancy and
Well Sufficiency
Results
Keep
Eliminate
Keep
Keep
Keep
Eliminate
Keep
Keep
Keep
Keep
Keep
Keep
Keep
Eliminate
Keep
MAROS
Interpreted
Sampling
Frequency
Results
Biennial
-
Annual
Annual
Annual
-
Biennial
Biennial
Biennial
Annual
Biennial
Biennial
Biennial
-
Annual
Comments
Historical concentrations below DL & cross
gradient sentry well
Monitors only the upper zone at MW-4 and
lower in concentrations than MW-6B
Recent concentrations above MCL &
inside plume well
Recent concentrations above MCL &
monitors the lower zone at MW-6
Recent concentrations above MCL &
monitors the source
Monitors only the upper zone at MW-1 1
and concentrations below DL
Historical concentrations below DL &
monitors the middle zone at MW-1 1
Historical concentrations below MCL or DL
& Monitors the lower zone at MW-1 1
Historical concentrations below MCL or DL
& Inside plume well
Recent concentrations above MCL & on
plume edge
Historical concentrations below MCL or DL
& monitors the lower zone at MW-1 4
Historical concentrations below DL &
monitors the upper zone at MW-1 5
Historical concentrations below DL &
downgradient sentry well just upgradient of
the Long Prairie river
Monitors only the upper zone at MW-1 6 &
concentrations below MCL or DL
Recent concentrations above MCL &
downgradient well on plume edge
Notes: (1) S = Source well, T = Tail well
(2) Sampling frequency based on recent sampling results, 'assumed biennial due to lack of data in 2000 and 2001
(3) D = Decreasing, PD = Probably Decreasing, S = Stable, NT = No Trend, PI = Probably Increasing, I = Increasing
MCL = the maximum contaminant level of PCE (0.005 mg/L), DL = detection limit
-------
GSI Job No. G-2236-15
Issued: 2/19/03
Page 3 of 3
If
GROUNDWATER
SERVICES, INC.
Table 11
Summary of MAROS Sampling Optimization Results
Long Prairie Site
Long Prairie, Minnesota
Welt
Name
MW-17B
MW-18A
MW-18B
MW-19B
RW-1A
RW-1B
RW-1C
RW-3
RW-4
RW-5
RW-6
RW-7
RW-8
RW-9
MAROS
Welt
Category'11
T
T
T
T
T
T
T
S
T
T
T
T
T
T
Current
Sampling
Frequency'2'
Annual
Biennial*
Biennial*
Annual
Biennial*
Biennial*
Biennial*
Quarterly
Annual
Quarterly
Quarterly
Quarterly
Quarterly
Quarterly
MAROS
Trend
Result'3'
D
S
PI
PI
S
NT
NA
D
D
D
D
D
D
NT
MAROS
Interpreted Well
Redundancy and
Well Sufficiency
Results
Keep
Eliminate
Keep
Keep
Eliminate
Eliminate
Eliminate
Keep
Keep
Keep
Keep
Keep
Keep
Keep
MAROS
Interpreted
Sampling
Frequency
Results
Annual
-
Biennial
Biennial
-
-
-
Annual
Biennial
Annual
Annual
Annual
Annual
Annual
Comments
Recent concentrations above MCL &
inside plume well
Monitors only the upper zone at MW-18 &
concentrations below DL
Historical concentrations below MCL &
downgradient well on plume edge
Historical concentrations below DL &
downgradient well close to plume tail
Source has been cleaned up and
concentrations below MCL or DL
Source has been cleaned up and
concentrations below MCL or DL
Source has been cleaned up and
concentrations below MCL or DL
Recent concentrations above MCL &
recovery well for performance monitoring
Recent concentrations below DL &
recovery well for performance monitoring
Recent concentrations above MCL &
recovery well for performance monitoring
Recent concentrations above MCL &
recovery well for performance monitoring
Recent concentrations above MCL &
recovery well for performance monitoring
Recent concentrations above MCL &
recovery well for performance monitoring
Recent concentrations above MCL &
recovery well for performance monitoring
Notes: (1) S = Source well, T = Tail well
(2) Sampling frequency based on recent sampling results, 'assumed biennial due to lack of data in 2000 and 2001
(3) D = Decreasing, PD = Probably Decreasing, S = Stable, NT = No Trend, PI = Probably Increasing, I = Increasing
MCL = the maximum contaminant level of PCE (0.005 mg/L), DL = detection limit
-------
February 19 2003 GROUNDWATER
GSI Job No. G-2236-15 SERVICES, INC.
FIGURES
MAROS 2.0 APPLICATION
MONITORING NETWORK OPTIMIZATION
Long Prairie Site
Long Prairie, Minnesota
Figure 1 Upper Outwash Aquifer Groundwater Monitoring Network
Figure 2 MAROS Decision Support Tool Flow Chart
Figure 3 MAROS Overview Statistics Trend Analysis Methodology
Figure 4 Decision Matrix for Determining Provisional Frequency
Figure 5 Upper Outwash Aquifer PCE Mann-Kendall Trend Results
Figure 6 Upper Outwash Aquifer PCE Linear Regression Trend Results
Figure 7 Upper Outwash Aquifer PCE Mann-Kendall Trend Results, Recovery
Wells
Figure 8 Upper Outwash Aquifer PCE Linear Regression Trend Results, Recovery
Wells
Figure 9 Upper Outwash Aquifer PCE First Moment (Center of Mass) Over Time
Figure 10 Upper Outwash Aquifer PCE plume contoured with 1999 and 2002 data:
With "B" Zone Wells Only
Figure 11 Upper Outwash Aquifer PCE plume contoured with 1999 data: before
optimization and after optimization
Figure 12 Upper Outwash Aquifer Well Sufficiency Results
-------
.
-------
GSI Job No. G-2236-15
Issued: 2/19/03 GROUNDWATER
Page 1 of 1 SERVICES, INC.
MAROS: Decision Support Tool
MAROS is a collection of tools in one software package that is used in an explanatory, non-linear fashion. The tool
includes models, geostatistics, heuristic rules, and empirical relationships to assist the user in optimizing a
groundwater monitoring network system while maintaining adequate delineation of the plume as well as knowledge
of the plume state over time. Different users utilize the tool in different ways and interpret the results from a different
viewpoint.
Overview Statistics
What it is: Simple, qualitative and quantitative plume information can be gained through evaluation of monitoring
network historical data trends both spatially and temporally. The MAROS Overview Statistics are the foundation the
user needs to make informed optimization decisions at the site.
What it does: The Overview Statistics are designed to allow site personnel to develop a better understanding of the
plume behavior over time and understand how the individual well concentration trends are spatially distributed within
the plume. This step allows the user to gain information that will support a more informed decision to be made in the
next level of optimization analysis.
What are the tools: Overview Statistics includes two analytical tools:
1) Trend Analysis: includes Mann-Kendall and Linear Regression statistics for individual wells and results in
general heuristically-derived monitoring categories with a suggested sampling density and monitoring
frequency.
2) Moment Analysis: includes dissolved mass estimation (0th Moment), center of mass (1st Moment), and
plume spread (2nd Moment) over time. Trends of these moments show the user another piece of
information about the plume stability over time.
What is the product: A first-cut blueprint for a future long-term monitoring program that is intended to be a
foundation for more detailed statistical analysis.
T
Detailed Statistics
What it is: The MAROS Detailed Statistics allows for a quantitative analysis for spatial and temporal optimization of
the well network on a well-by-well basis.
What it does: The results from the Overview Statistics should be considered along side the MAROS optimization
recommendations gained from the Detailed Statistical Analysis. The MAROS Detailed Statistics results should be
reassessed in view of site knowledge and regulatory requirements as well as the Overview Statistics.
What are the tools: Detailed Statistics includes four analytical tools:
1) Sampling Frequency Optimization: uses the Modified CES method to establish a recommended future
sampling frequency.
2) Well Redundancy Analysis: uses the Delaunay Method to evaluate if any wells within the monitoring
network are redundant and can be eliminated without any significant loss of plume information.
3) Well Sufficiency Analysis: uses the Delaunay Method to evaluate areas where new wells are
recommended within the monitoring network due to high levels of concentration uncertainty.
4) Data Sufficiency Analysis: uses Power Analysis to assess if the historical monitoring data record has
sufficient power to accurately reflect the location of the plume relative to the nearest receptor or
compliance point.
What is the product: List of wells to remove from the monitoring program, locations where monitoring wells may
need to be added, recommended frequency of sampling for each well, analysis if the overall system is statistically
powerful to monitor the plume.
Figure 2. MAROS Decision Support Tool Flow Chart
-------
GSI Job No. G-2236-15
Issued: 2/19/03
Page 1 of 1
If
GROUNDWATER
SERVICES, INC.
In Ml
i
(21 S>* of
V fiy
of
tor lip wtll Had by
C IK
• fi '
« (HJ ——
» _____________
|S|' -
(DJ
for Will On All
tor
Tall
(I
IPI
(MI)
PI
Pi
Md
^1
faj
"J
Im Wy
I
mm
*
»
•
Figure 3:
MAROS Overview Statistics Trend Analysis Methodology
CtmsHJtttlon
Fvei SsSwent
-------
GSI Job No. G-2236-15
Issued: 2/19/03
Page 1 of 1
INC.
1 i
I i
| \ |
i I £
I A: i ^
: : C
5
C
C
-------
o
o
CD
-a ?
c i
-------
o
o
CD
1 '
§ £
o i-
-
(D Q_
O) £?
CU ^
o: .3
CO (U
CD -ti
c (/)
^ CD
to
O
w
O
O ™
CL CL
(D g>
CU O
^ """
O)
COD
O)
si
0) 2
§ —
CO -a
w
to
t©
O
D
0. ^>
"-'CL
O)
C O)
'i/5 .g
CO CO
CD CO
O 2
CD O
CD CO
CO C
-------
-a ?
_ C o Q
* g, *=
o * e
Q
Q.
O O O
o
.9>
CO
o
CO
o
o
CD
O
O
CO
(U
C CO
£ O
= I
JD C
£> (U
(U '^
S 2
O Q_
(U
o:
= O
CO -I
?
fU r n
^ VJ
±2
O)
c
LJJ x'
O q
Q_ -1
(U
O)
O)
si
0) 2
fj =
W -a
t-^-"
O)
c a)
'i/5 .g
CO CO
CD CO
b FH
CD O
H Q
CD CO
CO C
CN ~t
™ ^
O CN
6 -a
.a
o
O
1
B
Q
-Q B
5 co
££
.t T3
CO C
CO CD
§^
CD O
Q Z
-------
1 '
•Z- § £
o i-
Q
Q.
CO
0. _
02
CO
o
CO
o
o
CD
O
O
CO
(U
C£
I- O
= CO
m a
> c
S «
O .b=
O (0
(U >r-
o: °-
«= ?
.Q o
0) ^
^ ^
tz .|
ro 2
(U Q_
5 g1
LU 9
o
Q_
CO
0)
O)
W co
0) 2
"co =
CO -a
0. ^>
O)
0)
CO CO
CD CO
CD O
CD CO
CO C
CN ~t
™ ^
O CN
6 -a
O
1
B
Q
-Q B
ro co
-Q o
2 2
Q-oI
.t T3
CO C
CO CD
§^
CD O
Q Z
-------
C
0)
O)
a)
0)
'E
0)
O
o
co
co
o
W
o
o
CD
O
O
co
(U
(U
O
CD CO
a p
™ L
O CN
6 -a
Z CD
S§
.2
CD "CO
CD
2 e
Q. 3
Q. O
< w
(C
O
-2
(U
O
c
o 02
S:l
"W Q_
LJJ q
O -1
Q_
-------
e «
z u
11
CD
CO
CM
CM
e>8
o -d •<-
~5 CU CU
55 « §>
O J2 D.
w
"co
CO
o1 <»
|1
ij
II
O)
0 T3
ti T3
CO 0
8
§1
It
S 0
CM
O 0
O >
CN >
i_ O)
0 c
^D "C
O O
"1
O o
"O 0
5 CD
_O c
co
CD -
E>< c
*= o
- o '•=
a. i 2
LU „ c
0
O) O
c o
-o o 2
i §^
2 °l
.
O CD
0)0)
-
2
5.
CD
-------
O W
I °
i
CD
CO
CM
CM
e>8
™ °
~> 03 CU
OT 3> §>
O J2 D.
0
0 _g
CO Q.
_W 0
"0 £
O) "5
E -°
| o
o?
O)
CN _W
s ^
CN _0
0 C
^3 O
O :J3
O •£
0
05 t
T- T3
>< O)
CO C
C 0
— w
W 0
Q. W
LU ^
O O
D- ^
o 0
E^
3 _C
o ±±
0)5
0 w
II
< " T3
>> C
• ^D CO
^ T3 CO
£«g
«&§
LL T3 O
-------
GSI JobNo. G-2236-15
Issued: 2/19/03
Page 1 of 1
IMC
Potential areas for
new locations are
indicated by triangles
with a high SF level.
Estimated SF Lew el:
S -
M -
L -
MSB
t\\ \ -C^
j\\ \ M-
/ I \
-i ee
M
BAL2C
' M llMW-SB
,' / \ "--^
\ I
^ i
\ I
M I I
M/»f
/ \
M ,;
\ ¥
£4
CW-3
Figure 12. Well Sufficiency Analysis for possible new sampling locations. Areas with L
or E symbols are candidate regions for placing new wells.
-------
February 19 2003 GROUNDWATER
GSI Job No. G-2236-15 SERVICES, INC.
MAROS 2.0 APPLICATION
MONITORING NETWORK OPTIMIZATION
Long Prairie Site
Long Prairie, Minnesota
APPENDICES
Appendix A: Upper Outwash Aquifer Long Prairie site Historical PCE Maps
Appendix B: Upper Outwash Aquifer Long Prairie site MAROS 2.0 Reports
-------
February 19 2003 GROUNDWATER
GSI Job No. G-2236-15 SERVICES, INC.
MAROS 2.0 APPLICATION
MONITORING NETWORK OPTIMIZATION
Long Prairie Site
Long Prairie, Minnesota
APPENDIX A: Upper Outwash Aquifer Long Prairie Historical PCE Maps
Appendix A: Upper Outwash Aquifer Long Prairie site Historical PCE Maps
-------
O)
q CD
10 o
C O
Q) O
O O T-; IO O
V A A A A
s
©
©
o
0©
CD CO
CO <
CN ~t
™ L
O CN
6 -a
O
-2
[U "5
Oco
_ CD
Q- c
. . c
onitor
Prairie,
cu
-I—•
CO
o .^
CD (t
^~ O
CD i-
™ ro
CO CD
.2 .9
1 1
§ §
<§§
t \ CC
O CD
| o
ll
CD o
^? CD
CD
QL
CD
•5
-------
c o
;J ;J
S- ° ^ ° o)
V A A A A
O
O
CD
©
©
©
©
©
CD CO
CO C
-------
o
10
C O
©
-2
[U "5
Oco
_ CD
Q- c
. . c
onitor
Prairie,
cu
-I—•
CD
O .^
CO Q!
^~ O
e
CO
CO "-
II
11
§ §
<§§
t \ CC
O CD
| o
ll
CD o
£ cu
CD
a:
CD
•5
-------
o
10
C O
-------
o
"ra
o>
-o £
c o
CN £
m s_
Co
CD CD
CO *-
U
c38
* \ CO
O CD
CD o
<£ CD
CD
a:
CD
•5
-------
o
I
"c
a)
_ o
7 c
C o
CN S
m s_
Co
CD CD
CO *-
U
c38
* \ CO
O CD
CD o
<£ CD
CD
a:
CD
•5
-------
_ o
= I§
m .,
0) O X
—I 0.
O
CD
o
A
O
A
O
A
0O0
s
CO
H
©
©
©
©
CD CO
CO C
CN ~t
™ L
O CN
6 -a
.Q
o
O
-2
LJJ "5
Oco
_
-------
February 19 2003 GROUNDWATER
GSI Job No. G-2236-15 SERVICES, INC.
MAROS 2.0 APPLICATION
MONITORING NETWORK OPTIMIZATION
Long Prairie Site
Long Prairie, Minnesota
APPENDIX B: Upper Outwash Aquifer Long Prairie MAROS 2.0 Reports
Linear Regression Statistics Summary
Mann-Kendall Statistics Summary
Spatial Moment Analysis Summary
Zeroth, First, and Second Moment Reports
Plume Analysis Summary
Site Results Summary
Sampling Location Optimization Results
Sampling Frequency Optimization Results
Risk-Based Power Analysis - Plume Centerline Concentrations
Risk-Based Power Analysis - Site Cleanup Status
-------
MAROS Linear Regression Statistics Summary
Project: Long Prairie Site
Location: Todd County
Julia Aziz
Minnesota
Time Period: 5/20/1996 to 10/14/2002
Consolidation Period: No Time Consolidation
Consolidation Type: Median
Duplicate Consolidation: Average
ND Values: Specified Detection Limit
J Flag Values : Actual Value
Well
Average
Source/ Cone
Tail (mg/L)
Median
Cone
(mg/L)
Standard
Deviation
All
Samples
"ND" ?
Coefficient
Ln Slope of Variation
Confidence Concentration
in Trend Trend
TETRACHLOROETHYLENE(PCE)
MW-2A
MW-2B
MW-2C
RW-3
MW-10
MW-11B
MW-17B
MW-16B
MW-16A
MW-15B
MW-15A
MW-14C
MW-14B
MW-18B
MW-11C
MW-19B
MW-11A
CW-6
CW-3
BAL2C
BAL2B
MW-13C
MW-6A
RW-8
RW-7
RW-6
RW-5
RW-4
RW-1C
RW-1B
RW-1A
MW-18A
MW-6B
RW-9
MW-5B
MW-5A
MW-4C
MW-4B
MW-4A
s
s
s
s
s
T
T
T
T
T
T
T
T
T
T
T
T
T
T
T
T
T
T
T
T
T
T
T
T
T
T
T
T
T
T
T
T
T
T
1 .8E-02
4.0E-01
4.2E-03
8.6E-02
2.0E+01
4.0E-04
1.3E-01
1.1E-02
5.2E-04
4.0E-04
4.0E-04
4.6E-04
2.2E-01
2.2E-03
5.3E-04
4.0E-04
4.0E-04
4.0E-04
4.0E-04
4.0E-04
4.0E-04
8.1E-04
8.6E-02
2.5E-02
5.2E-02
4.3E-02
1.6E-01
9.8E-04
4.0E-04
8.7E-04
2.5E-03
4.0E-04
2.8E-01
1 .4E-02
4.0E-04
4.0E-04
1.1E-01
2.2E-01
2.0E-02
7.3E-03
6.7E-02
4.3E-03
8.7E-02
9.3E-01
4.0E-04
1 .4E-01
1 .1 E-02
4.0E-04
4.0E-04
4.0E-04
4.0E-04
2.1E-01
2.4E-03
4.0E-04
4.0E-04
4.0E-04
4.0E-04
4.0E-04
4.0E-04
4.0E-04
4.0E-04
4.1 E-02
2.3E-02
4.0E-02
4.1 E-02
1 .6E-01
4.0E-04
4.0E-04
4.0E-04
1 .7E-03
4.0E-04
1 .5E-01
1 .3E-02
4.0E-04
4.0E-04
1 .2E-01
1 .9E-01
2.0E-02
2.9E-02
5.4E-01
3.6E-03
4.7E-02
5.3E+01
O.OE+00
5.5E-02
7.4E-03
2.9E-04
O.OE+00
O.OE+00
2.1E-04
9.2E-02
5.5E-04
3.5E-04
O.OE+00
O.OE+00
O.OE+00
8.6E-12
O.OE+00
O.OE+00
8.3E-04
1 .4E-01
1 .2E-02
3.0E-02
3.2E-02
7.0E-02
8.3E-04
O.OE+00
1 .6E-03
3.4E-03
O.OE+00
3.8E-01
2.3E-03
O.OE+00
O.OE+00
4.1 E-02
1 .3E-01
1 .8E-02
No
No
No
No
No
Yes
No
No
No
Yes
Yes
No
No
No
No
Yes
Yes
Yes
Yes
Yes
Yes
No
No
No
No
No
No
No
Yes
No
No
Yes
No
No
Yes
Yes
No
No
No
-9.4E-04
-1 .2E-03
-8.6E-04
-8.3E-04
-3.4E-03
O.OE+00
-8.8E-04
5.5E-05
-2.6E-04
-2.6E-34
-2.6E-34
-5.0E-05
-3.7E-04
2.6E-06
-4.2E-04
1.1E-34
O.OE+00
1 .6E-34
O.OE+00
O.OE+00
O.OE+00
-2.4E-04
-1.1E-03
-1 .OE-03
-7.2E-04
-1.1E-03
-3.5E-04
-8.4E-04
O.OE+00
-4.9E-04
-4.6E-04
O.OE+00
-2.5E-03
1 .7E-04
O.OE+00
O.OE+00
-3.1E-04
-7.9E-04
O.OE+00
1.59
1.35
0.86
0.55
2.63
0.00
0.43
0.66
0.55
0.00
0.00
0.46
0.41
0.26
0.67
0.00
0.00
0.00
0.00
0.00
0.00
1.03
1.66
0.48
0.57
0.75
0.44
0.84
0.00
1.87
1.39
0.00
1.35
0.17
0.00
0.00
0.39
0.57
0.00
84.0%
86.6%
82.0%
100.0%
99.6%
100.0%
99.8%
53.2%
74.6%
100.0%
100.0%
62.4%
97.4%
100.0%
94.1%
100.0%
0.0%
100.0%
100.0%
100.0%
0.0%
69.5%
74.8%
100.0%
100.0%
100.0%
99.6%
99.9%
0.0%
89.4%
87.1%
100.0%
100.0%
86.1%
100.0%
100.0%
87.5%
99.8%
0.0%
NT
NT
S
D
D
S
D
NT
S
D
D
S
D
I
PD
I
N/A
I
S
S
N/A
NT
NT
D
D
D
D
D
N/A
NT
NT
S
D
NT
S
S
S
D
N/A
MAROS Version 2, 2002, AFCEE
Thursday, February 13, 2003
Page 1 of 2
-------
project; Long Prairie Site
Julia Aziz
Todd County
Minnesota
Well
Average
Source/ Cone
Tail (mg/L)
Median
Cone
(mg/L)
Standard
Deviation
All
Samples
"ND" ?
Coefficient
Ln Slope of Variation
Confidence Concentration
in Trend Trend
TETRACHLOROETHYLENE(PCE)
MW-3B
MW-3A
MW-1B
MW-1A
MW-6C
T
T
T
T
T
4.0E-04
9.3E-04
4.0E-04
4.0E-04
8.4E-02
4.0E-04
4.0E-04
4.0E-04
4.0E-04
1 .OE-01
O.OE+00
1 .5E-03
O.OE+00
O.OE+00
7.1E-02
Yes
No
Yes
Yes
No
O.OE+00
-5.8E-04
O.OE+00
O.OE+00
-1 .6E-03
0.00
1.61
0.00
0.00
0.84
100.0%
87.2%
100.0%
100.0%
98.2%
S
NT
S
S
D
Note: Increasing (I); Probably Increasing (PI); Stable (S); Probably Decreasing (PD); Decreasing (D); No Trend (NT); Not Applicable (N/A) -
Due to insufficient Data (< 4 sampling events); COV = Coefficient of Variation
MAROS Version 2, 2002, AFCEE
Thursday, February 13, 2003
Page 2 of 2
-------
MAROS Mann-Kendall Statistics Summary
Long Prairie Site
Location: Todd County
Julia Aziz
Minnesota
Time Period: 5/20/1996 to 10/14/2002
Consolidation Period: No Time Consolidation
Consolidation Type: Median
Duplicate Consolidation: Average
ND Values: Specified Detection Limit
J Flag Values : Actual Value
Well
Source/
Tail
Number of
Samples
Number of
Detects
Coefficient
of Variation
Mann-Kendall
Statistic
Confidence
in Trend
All
Samples Concentration
"ND" ? Trend
TETRACHLOROETHYLENE(PCE)
MW-2B
MW-10
MW-2A
MW-2C
RW-3
MW-13C
MW-18B
MW-18A
MW-17B
MW-16B
MW-16A
MW-15B
MW-15A
MW-1B
MW-14B
BAL2B
MW-11C
MW-1 1 B
MW-11A
RW-8
CW-6
CW-3
BAL2C
MW-14C
MW-4C
RW-5
RW-1C
RW-1B
RW-1A
MW-6C
MW-6B
MW-6A
MW-19B
MW-5A
MW-1 A
MW-4B
MW-4A
MW-3B
s
s
s
s
s
T
T
T
T
T
T
T
T
T
T
T
T
T
T
T
T
T
T
T
T
T
T
T
T
T
T
T
T
T
T
T
T
T
6
8
6
8
25
9
5
5
7
8
6
7
7
8
10
1
8
4
2
12
14
24
8
11
8
25
1
12
12
7
6
6
7
8
8
6
2
8
6
8
6
5
25
2
5
0
7
8
1
0
0
0
10
0
1
0
0
12
0
0
0
1
8
25
0
1
11
7
6
5
0
0
0
6
2
0
1.35
2.63
1.59
0.86
0.55
1.03
0.26
0.00
0.43
0.66
0.55
0.00
0.00
0.00
0.41
0.00
0.67
0.00
0.00
0.48
0.00
0.00
0.00
0.46
0.39
0.44
0.00
1.87
1.39
0.84
1.35
1.66
0.00
0.00
0.00
0.57
0.00
0.00
-7
-28
-5
-11
-173
-3
-3
0
-14
-4
-3
0
0
0
-18
0
-7
0
0
^6
0
0
0
-2
-11
-132
0
-11
-24
-10
-15
0
0
0
0
-12
0
0
86.4%
100.0%
76.5%
88.7%
100.0%
58.0%
67.5%
40.8%
97.5%
64.0%
64.0%
43.7%
43.7%
45.2%
93.4%
0.0%
76.4%
37.5%
0.0%
100.0%
47.8%
49.0%
45.2%
53.0%
88.7%
99.9%
0.0%
74.9%
94.2%
90.7%
99.9%
42.3%
43.7%
45.2%
45.2%
98.2%
0.0%
45.2%
No
No
No
No
No
No
No
Yes
No
No
No
Yes
Yes
Yes
No
Yes
No
Yes
Yes
No
Yes
Yes
Yes
No
No
No
Yes
No
No
No
No
No
Yes
Yes
Yes
No
No
Yes
NT
D
NT
S
D
NT
S
S
D
S
S
S
S
S
PD
N/A
S
S
N/A
D
S
S
S
S
S
D
N/A
NT
PD
PD
D
NT
S
S
S
D
N/A
S
MAROS Version 2, 2002, AFCEE
Thursday, February 13, 2003
Page 1 of 2
-------
Project: Long Prairie Site
Location: Todd County
Julia Aziz
Minnesota
Source/
Well Tail
Number of
Samples
Number of
Detects
Coefficient
of Variation
Mann-Kendall
Statistic
Confidence
in Trend
All
Samples
"ND" ?
Concentration
Trend
TETRACHLOROETHYLENE(PCE)
MW-3A T
RW-6 T
RW-9 T
RW-7 T
RW-4 T
MW-5B T
8
25
12
25
15
8
1
25
12
25
6
0
1.61
0.75
0.17
0.57
0.84
0.00
-7
-234
2
-255
-61
0
76.4%
100.0%
52.7%
100.0%
99.9%
45.2%
No
No
No
No
No
Yes
NT
D
NT
D
D
S
Note: Increasing (I); Probably Increasing (PI); Stable (S); Probably Decreasing (PD); Decreasing (D); No Trend (NT); Not Applicable (N/A)-
Due to insufficient Data (< 4 sampling events); Source/Tail (S/T)
The Number of Samples and Number of Detects shown above are post-consolidation values.
MAROS Version 2, 2002, AFCEE
Thursday, February 13, 2003
Page 2 of 2
-------
MAROS Statistical Trend Analysis Summary
Project: Long Prairie Site
Location: Todd County
Julia Aziz
Minnesota
Time Period: 5/20/1996 to 10/14/2002
Consolidation Period: No Time Consolidation
Consolidation Type: Median
Duplicate Consolidation: Average
ND Values: Specified Detection Limit
J Flag Values : Actual Value
Well
Source/
Tail
Number Number
of of
Samples Detects
Average Median
Cone. Cone.
(mg/L) (mg/L)
All
Samples
"ND" ?
Mann-
Kendall
Trend
Linear
Regression
Trend
TETRACHLOROETHYLENE(PCE)
BAL2B
BAL2C
CW-3
CW-6
MW-10
MW-11A
MW-11B
MW-11C
MW-13C
MW-14B
MW-14C
MW-15A
MW-15B
MW-16A
MW-16B
MW-17B
MW-18A
MW-18B
MW-19B
MW-1A
MW-1B
MW-2A
MW-2B
MW-2C
MW-3A
MW-3B
MW-4A
MW-4B
MW-4C
MW-5A
MW-5B
MW-6A
MW-6B
MW-6C
RW-1A
T
T
T
T
S
T
T
T
T
T
T
T
T
T
T
T
T
T
T
T
T
S
S
S
T
T
T
T
T
T
T
T
T
T
T
1
8
24
14
8
2
4
8
9
10
11
7
7
6
8
7
5
5
7
8
8
6
6
8
8
8
2
6
8
8
8
6
6
7
12
0
0
0
0
8
0
0
1
2
10
1
0
0
1
8
7
0
5
0
0
0
6
6
5
1
0
2
6
8
0
0
5
6
7
11
4.0E-04
4.0E-04
4.0E-04
4.0E-04
2.0E+01
4.0E-04
4.0E-04
5.3E-04
8.1E-04
2.2E-01
4.6E-04
4.0E-04
4.0E-04
5.2E-04
1.1E-02
1 .3E-01
4.0E-04
2.2E-03
4.0E-04
4.0E-04
4.0E-04
1 .8E-02
4.0E-01
4.2E-03
9.3E-04
4.0E-04
2.0E-02
2.2E-01
1.1E-01
4.0E-04
4.0E-04
8.6E-02
2.8E-01
8.4E-02
2.5E-03
4.0E-04
4.0E-04
4.0E-04
4.0E-04
9.3E-01
4.0E-04
4.0E-04
4.0E-04
4.0E-04
2.1E-01
4.0E-04
4.0E-04
4.0E-04
4.0E-04
1.1E-02
1.4E-01
4.0E-04
2.4E-03
4.0E-04
4.0E-04
4.0E-04
7.3E-03
6.7E-02
4.3E-03
4.0E-04
4.0E-04
2.0E-02
1.9E-01
1.2E-01
4.0E-04
4.0E-04
4.1E-02
1.5E-01
1 .OE-01
1 .7E-03
Yes
Yes
Yes
Yes
No
Yes
Yes
No
No
No
No
Yes
Yes
No
No
No
Yes
No
Yes
Yes
Yes
No
No
No
No
Yes
No
No
No
Yes
Yes
No
No
No
No
N/A
S
S
S
D
N/A
S
S
NT
PD
S
S
S
S
S
D
S
S
S
S
S
NT
NT
S
NT
S
N/A
D
S
S
S
NT
D
PD
PD
N/A
S
S
I
D
N/A
S
PD
NT
D
S
D
D
S
NT
D
S
I
|
S
S
NT
NT
S
NT
S
N/A
D
S
S
S
NT
D
D
NT
MAROS Version 2, 2002, AFCEE
Thursday, February 13, 2003
Page 1 of 2
-------
MAROS Statistical Trend Analysis Summary
Well
Number Number Average Median AM Mann-
Source/ of of Cone. Cone. Samples Kendall
Tail Samples Detects (mg/L) (mg/L) "ND" ? Trend
Linear
Regression
Trend
TETRACHLOROETHYLENE(PCE)
RW-1B
RW-1C
RW-3
RW-4
RW-5
RW-6
RW-7
RW-8
RW-9
T
T
S
T
T
T
T
T
T
12
1
25
15
25
25
25
12
12
1
0
25
6
25
25
25
12
12
8.7E-04
4.0E-04
8.6E-02
9.8E-04
1 .6E-01
4.3E-02
5.2E-02
2.5E-02
1 .4E-02
4.0E-04
4.0E-04
8.7E-02
4.0E-04
1 .6E-01
4.1E-02
4.0E-02
2.3E-02
1 .3E-02
No
Yes
No
No
No
No
No
No
No
NT
N/A
D
D
D
D
D
D
NT
NT
N/A
D
D
D
D
D
D
NT
Note: Increasing (I); Probably Increasing (PI); Stable (S); Probably Decreasing (PD); Decreasing (D); No Trend (NT); Not Applicable
(N/A); Not Applicable (N/A) - Due to insufficient Data (< 4 sampling events); No Detectable Concentration (NDC)
The Number of Samples and Number of Detects shown above are post-consolidation values.
MAROS Version 2, 2002, AFCEE
Thursday, February 13, 2003
Page 2 of 2
-------
MAROS Spatial Moment Analysis Summary
Project; Long Prairie Site
Location: Todd County
Julia Aziz
Minnesota
Effective Date
Oth Moment
1st (Center of
Estimated
Mass (Kg) Xc (ft)
Source
Yc (ft) Distance (ft)
2nd Moment
Sigma XX
(sq ft)
(Spread)
Sigma YY
(sq ft)
Number of
Wells
TETRACHLOROETHYLENE(PCE)
7/1/1996
7/1/1997
7/1/1998
7/1/1999
7/1/2000
7/1/2001
7/1/2002
4.7E+00
2.4E+01
2.9E+01
2.1E+01
1.5E+01
1.3E+01
7.3E+00
510,952
510,886
510,757
510,842
510,698
510,705
510,983
173,519
173,907
174,134
174,212
174,623
174,626
174,558
1,199
1,342
1,408
1,523
1,779
1,786
1,880
865,855
458,802
484,020
483,456
516,718
477,288
537,329
2,257,891
271,873
567,076
544,788
488,253
427,496
849,155
8
12
17
17
10
10
17
MAROS Version 2, 2002, AFCEE
Thursday, February 13, 2003
Page 1 of 2
-------
Long Prairie Site
Location: Todd County
Julia Aziz
Minnesota
Moment Type Consituent
Zeroth Moment: Mass
TETRACHLOROETHYLENE(PCE)
1st Moment: Distance to Source
TETRACHLOROETHYLENE(PCE)
2nd Moment: Sigma XX
TETRACHLOROETHYLENE(PCE)
2nd Moment: Sigma YY
TETRACHLOROETHYLENE(PCE)
Coefficient
of Variation
0.54
0.17
0.26
0.88
Mann-Kendall
S Statistic
-7
21
1
-3
Confidence
in Trend
80.9%
100.0%
50.0%
61.4%
Moment
Trend
S
I
NT
S
Note: The following assumptions were applied for the calculation of the Zeroth Moment:
Porosity: 0.30 Saturated Thickness: Uniform: 60 ft
Mann-Kendall Trend test performed on all sample events for each constituent. Increasing (I); Probably Increasing (PI); Stable (S);
Probably Decreasing (PD); Decreasing (D); No Trend (NT); Not Applicable (N/A)-Due to insufficient Data (< 4 sampling events).
Note: The Sigma XX and Sigma YY components are estimated using the given field coordinate system and then rotated to align with the
estimated groundwater flow direction. Moments are not calculated for sample events with less than 6 wells.
MAROS Version 2, 2002, AFCEE
Thursday, February 13, 2003
Page 2 of 2
-------
MAROS Zeroth Moment Analysis
Project: Long Prairie Site
Location: Todd County
Julia Aziz
Minnesota
COC: TETRACHLOROETHYLENE(PCE)
Change in Dissolved Mass Over Time
Date
/ / /
3.5E+01
2.5E+01 -
*-»,
S 2.0E+01
Sg 1.5E+01
S
1.0E+01
5.0E+00 ^
O.OE+00
Porosity: 0.30
Saturated Thickness:
Uniform: 60 ft
Mann Kendall S Statistic:
-7
Confidence in
Trend:
I 80.9%
Coefficient of Variation:
I 0.54
Zeroth Moment
Trend:
Data Table:
Effective Date
7/1/1996
7/1/1997
7/1/1998
7/1/1999
7/1/2000
7/1/2001
7/1/2002
Constituent
TETRACHLOROETHYLENE(PCE)
TETRACHLOROETHYLENE(PCE)
TETRACHLOROETHYLENE(PCE)
TETRACHLOROETHYLENE(PCE)
TETRACHLOROETHYLENE(PCE)
TETRACHLOROETHYLENE(PCE)
TETRACHLOROETHYLENE(PCE)
Estimated
Mass (Kg)
4.7E+00
2.4E+01
2.9E+01
2.1E+01
1.5E+01
1.3E+01
7.3E+00
Number of Wells
8
12
17
17
10
10
17
Note: Increasing (I); Probably Increasing (PI); Stable (S); Probably Decreasing (PD); Decreasing (D); No Trend (NT); Not Applicable (N/A) -
Due to insufficient Data (< 4 sampling events); ND = Non-detect. Moments are not calculated for sample events with less than 6 wells.
MAROS Version 2, 2002, AFCEE
2/13/2003
Page 1 of 1
-------
-------
MAROS First Moment Analysis
Project: Long Prairie Site
Location: Todd County
Julia Aziz
State: Minnesota
COC: TETRACHLOROETHYLENE(PCE)
Change in Location of Center of Mass Over Time
I/ H / UU '
174600 •
174500 •
174400 •
32 174300 •
>_ 174200 •
174100 •
174000 •
173900 •
173 800 •
•»*OWSQO
»07
* 07/98
* 07/97
• 07/96
Flow Direction:
,01 w
\
\
Source
Coordinate:
X: I 509,844
1
Y: j 173,062
510650 510700 510750 510800 510850 510900 510950 511000
Xc (ft)
Effective Date Constituent Xc (ft) Yc (ft) Distance from Source (ft) Number of Wells
7/1/1996 TETRACHLOROETHYLENE(P 510,952 173,519 1,199 8
7/1/1997 TETRACHLOROETHYLENE(P 510,886 173,907 1,342 12
7/1/1998 TETRACHLOROETHYLENE(P 510,757 174,134 1,408 17
7/1/1999 TETRACHLOROETHYLENE(P 510,842 174,212 1,523 17
7/1/2000 TETRACHLOROETHYLENE(P 510,698 174,623 1,779 10
7/1/2001 TETRACHLOROETHYLENE(P 510,705 174,626 1,786 10
7/1/2002 TETRACHLOROETHYLENE(P 510,983 174,558 1,880 17
Note: Increasing (I); Probably Increasing (PI); Stable (S); Probably Decreasing (PD); Decreasing (D); No Trend (NT); Not Applicable (N/A) -
Due to insufficient Data (< 4 sampling events). Moments are not calculated for sample events with less than 6 wells.
MAROS Version 2, 2002, AFCEE
2/13/2003
Page 1 of 1
-------
MAROS First Moment Analysis
Project: Long Prairie Site
Location: Todd County
COC: TETRACHLOROETHYLENE(PCE)
Julia Aziz
Minnesota
Distance from Source to Center of Mass
Mann Kendall S Statistic:
Date
g
t
3
O
CO
E
O
(0
to
5
/ / / / / / /
1.8E+03 •
1.6E+03 -
1.4E+03 -
1.2E+03 -
1 np-i-n^
'
8.0E+02 -
6.0E+02 •
4.0E+02 -
2.0E+02 -
n np+nn .
*
_
»
*
A
I 21
Confidence in
Trend:
100.0%
Coefficient of Variation:
First Moment Trend:
Data Table:
Effective Date
7/1/1996
7/1/1997
7/1/1998
7/1/1999
7/1/2000
7/1/2001
7/1/2002
Constituent
TETRACHLOROETHYLENE(P
TETRACHLOROETHYLENE(P
TETRACHLOROETHYLENE(P
TETRACHLOROETHYLENE(P
TETRACHLOROETHYLENE(P
TETRACHLOROETHYLENE(P
TETRACHLOROETHYLENE(P
Xc (ft)
510,952
510,886
510,757
510,842
510,698
510,705
510,983
Yc (ft)
173,519
173,907
174,134
174,212
174,623
174,626
174,558
Distance from Source (ft)
1,199
1,342
1,408
1,523
1,779
1,786
1,880
Number of Wells
8
12
17
17
10
10
17
Note: Increasing (I); Probably Increasing (PI); Stable (S); Probably Decreasing (PD); Decreasing (D); No Trend (NT); Not Applicable (N/A) -
Due to insufficient Data (< 4 sampling events). Moments are not calculated for sample events with less than 6 wells.
MAROS Version 2, 2002, AFCEE
2/13/2003
Page 1 of 1
-------
MAROS Second Moment Analysis
Project: Long Prairie Site
Todd County
COC: TETRACHLOROETHYLENE(PCE)
Change in Pli
10000000
1000000
£- 100000
JE 10000
CM
V 1000
w
100
10
1
mnnnnn _i_
tr
-2.
?• 100000 -
1
innnn - -
Data Table:
Effective Date
jme Spread Over Time
Date
*
; . • * * * *
Date
c?> <^ <$> <§> <§* «?> c?
%r %r 4r %r %r 4r %r
\y \y \y \y \y \y \y
•
******
Julia Aziz
Minnesota
Mann Kendall S Statistic:
? i -3
Confidence in
Trend:
| 61.4%
Coefficient of Variation:
|| 0.88
Second Moment
Trend:
I S
Mann Kendall S Statistic:
Confidence in
Trend:
| 50.0%
Coefficient of Variation:
j 0.26
Second Moment
Trend:
1 NT
i
Constituent Sigma XX (sq ft) Sigma YY (sq ft) Number of Wells
7/1/1996 TETRACHLOROETHYLENE(P 865,855 2,257,891 8
7/1/1997 TETRACHLOROETHYLENE(P 458,802 271,873 12
7/1/1998 TETRACHLOROETHYLENE(P 484,020 567,076 17
7/1/1999 TETRACHLOROETHYLENE(P 483,456 544,788 17
7/1/2000 TETRACHLOROETHYLENE(P 516,718 488,253 10
7/1/2001 TETRACHLOROETHYLENE(P 477,288 427,496 10
7/1/2002 TETRACHLOROETHYLENE(P 537,329 849,155 17
Note: Increasing (I); Probably Increasing (PI); Stable (S); Probably Decreasing (PD); Decreasing (D); No Trend (NT); Not Applicable (N/A) -
Due to insufficient Data (< 4 sampling events)
The Sigma XX and Sigma YY components are estimated using the given field coordinate system and then rotated to align with the
estimated groundwater flow direction. Moments are not calculated for sample events with less than 6 wells.
MAROS Version 2, 2002, AFCEE
2/13/2003
Page 1 of 1
-------
MAROS Site Results
Project: Long Prairie Site
Location: Todd County
User Defined Site and Data Assumptions:
Julia Aziz
Minnesota
Hydrogeology and Plume Information:
Groundwater
Seepage Velocity: 475 ft/yr
Current Plume Length: 2100 ft
Current Plume Width 1000 ft
Number of Tail Wells: 15
Number of Source Wells: 2
Source Information:
Source Treatment: SVG
NAPL is not at this site.
Down-gradient Information:
Distance from Edge of Tail to Nearest:
Down-gradient receptor: 1000 ft
Down-gradient property: 1 ft
Distance from Source to Nearest:
Down-gradient receptor: 3000ft
Down-gradient property: 1 ft
Consolidation Assumptions:
Time Period: 10/31/1996 to 10/14/2002
Consolidation Period: Yearly
Consolidation Type: Geometric Mean
Duplicate Consolidation: Average
ND Values: Specified Detection Limit
J Flag Values : Actual Value
Plume Information Weighting Assumptions:
Consolidation Step 1. Weight Plume Information by Chemical
Summary Weighting: Weighting Applied to All Chemicals Equally
Consolidation Step 2. Weight Well Information by Chemical
Well Weighting: No Weighting of Wells was Applied.
Chemical Weighting: No Weighting of Chemicals was Applied.
Note: These assumptions made when consolidating the historical mentoring lumping the Wells and COCs.
1.
Preliminary Monitoring System Optimization Results: Based on site classification, source treatment and Monitoring System
Category the following suggestions are made for site Sampling Frequency, Duration of Sampling, and Well Density. These
criteria take into consideration: Plume Stability, Type of Plume, and Groundwater Velocity.
coc
Tail Source Level of
Stability Stability Effort
Sampling
Duration
Sampling
Frequency
Sampling
Density
TETRACHLOROETHYLENE(PCE)
M
35
Remove treatment No Recommendation
system if previously
reducing concentation
Note:
Plume Status: (I) Increasing; (Pl)Probably Increasing; (S) Stable; (NT) No Trend; (PD) Probably Decreasing; (D) Decreasing
Design Categories: (E) Extensive; (M) Moderate; (L) Limited (N/A) Not Applicable, Insufficient Data Available
Level of Monitoring Effort Indicated by Analysi I Moderate
2,
MAROS Version 2, 2002, AFCEE
Thursday, February 13, 2003
Page 1 of 2
-------
Coefficient Mann-Kendall Confidence
Moment Type Consituent of Variation S Statistic in Trend
Moment
Trend
Zeroth Moment: Mass
TETRACHLOROETHYLENE(PCE) 0.54 -7 80.9%
S
1st Moment: Distance to Source
2nd
2nd
TETRACHLOROETHYLENE(PCE) 0.17 21 100.0%
Moment: Sigma XX
TETRACHLOROETHYLENE(PCE) 0.26 1 50.0%
Moment: Sigma YY
TETRACHLOROETHYLENE(PCE) 0.88 -3 61.4%
I
NT
S
Note: The following assumptions were applied for the calculation of the Zeroth Moment:
Porosity: 0.30 Saturated Thickness: Uniform: 60 ft
Mann-Kendall Trend test performed on all sample events for each constituent. Increasing (I); Probably Increasing (PI); Stable (S);
Probably Decreasing (PD); Decreasing (D); No Trend (NT); Not Applicable (N/A)-Due to insufficient Data (< 4 sampling events).
MAROS Version 2, 2002, AFCEE
Thursday, February 13, 2003
Page 2 of 2
-------
1 MAROS Sampling Location Optimization Re suits 1
Long Prairie site
Location: Todd County
Sampling Events Analyzed: From May 1999
5/20/1999
Parameters used: Constituent
TETRACHLOROETHYLENE(PCE
Meng
Minnesota
to October 2002
10/14/2002
Inside SF Hull SF Area Ratio Cone. Ratio
0.1 0.01 0.95 0.95
Average Minimum Maximum
Well X (feet) Y (feet) Removable? Slope Factor* Slope Factor* Slope Factor*
Eliminated?
TETRACHLOROETHYLENE(PCE)
BAL2C 511820.13 175912.75 0
CW-3 511983.13 173568.70 0
MW-10 509843.81 173061.80 0
MW-11C 511661.88 175323.95 0
MW-13C 510827.19 174770.72 0
MW-14B 511182.03 174835.77 0
MW-15B 510155.53 176497.39 0
MW-16B 510567.59 175986.83 0
MW-17B 510889.94 175444.19 0
MW-18B 510844.34 176150.06 0
MW-19B 510391.47 175606.44 0
MW-1B 509586.91 173544.45 0
MW-2B 510177.97 173610.33 0
MW-3B 510256.16 173355.97 0
MW-4B 510892.31 174509.33 0
MW-5B 510198.19 175015.45 0
MW-6B 510524.09 174085.47 0
0.488 0.463 0.519
0.628 0.470 0.720
0.498 0.338 0.752
0.678 0.629 0.730
0.726 0.690 0.770
0.450 0.288 0.619
0.474 0.291 0.588
0.269 0.011 0.437
0.506 0.378 0.627
0.270 0.244 0.296
0.624 0.556 0.704
0.685 0.587 0.776
0.463 0.281 0.588
0.680 0.532 0.793
0.357 0.225 0.468
0.490 0.409 0.637
0.206 0.090 0.447
D
D
D
D
D
D
D
D
D
D
D
D
D
D
D
D
D
Note: The Slope Factor indicates the relative importance of a well in the monitoring network at a given sampling event; the larger the SF
value of a well, the more important the well is and vice versa; the Average Slope Factor measures the overall well importance in the
selected time period; the state coordinates system (i.e., X and Y refer to Easting and Northing respectively) or local coordinates systems
may be used; wells that are NOT selected for analysis are not shown above.
* When the report is generated after running the Excel module, SF values will NOT be shown above.
MAROS Version 2, 2002, AFCEE
Wednesday, February 12, 2003
Page 1 of 1
-------
MAROS Sampling Frequency Optimization Results
Project: Long Prairie site
Location: ToddCounty
The Overall Number of Sampling Events: 28
"Recent Period" defined by events: From May 1999
5/20/1999
Meng
Minnesota
To October 2002
10/14/2002
"Rate of Change" parameters used:
Well
Constituent Cleanup Goal Low Rate Medium Rate High Rate
TETRACHLOROETHYLENE(PCE 0.005
Units: Cleanup Goal is in mg/L; all rate parameters are
Recommended
Sampling Frequency
0.0025 0.005 0.01
in mg/L/year.
Frequency Based Frequency Based
on Recent Data on Overall Data
TETRACHLOROETHYLENE(PCE)
BAL2B
BAL2C
CW-3
CW-6
MW-10
MW-11A
MW-11B
MW-11C
MW-13C
MW-14B
MW-14C
MW-15A
MW-15B
MW-16A
MW-16B
MW-1 7B
MW-18A
MW-18B
MW-19B
MW-1 A
MW-1B
MW-2A
MW-2B
MW-2C
Annual
Annual
Biennial
Biennial
Quarterly
Annual
Biennial
Biennial
Biennial
Annual
Biennial
Biennial
Biennial
Biennial
Annual
Annual
Annual
Annual
Biennial
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Quarterly Quarterly
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
MAROS Version 2, 2002, AFCEE
Wednesday, February 12, 2003
Page 1 of2
-------
Project: Long Prairie site
Location: ToddCounty
Meng
State: Minnesota
Well
MW-3A
MW-3B
MW-4A
MW-4B
MW-4C
MW-5A
MW-5B
MW-6A
MW-6B
MW-6C
RW-1A
RW-1B
RW-1C
RW-3
RW-4
RW-5
RW-6
RW-7
RW-8
RW-9
Recommended
Sampling Frequency
Annual
Annual
Quarterly
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Biennial
Annual
Annual
Annual
Annual
Annual
Frequency Based
on Recent Data
Annual
Annual
Quarterly
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Frequency Based
on Overall Data
Annual
Annual
Quarterly
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Note: Sampling frequency is determined considering both recent and overall concentration trends. Sampling Frequency is the
final recommendation; Frequency Based on Recent Data is the frequency determined using recent (short) period of monitoring
data; Frequency Based on Overall Data is the frequency determined using overall (long) period of monitoring data. If the "recent
period" is defined using a different series of sampling events, the results could be different.
MAROS Version 2, 2002, AFCEE
Wednesday, February 12, 2003
Page 2 of2
-------
[Regression of Plume Centerline Concentrations |
Long Prairie site
Location: Todd County
Groundwater Flow Direction: 90 degrees
From Period: 5/20/1999 to 10/14/2002
Selected Plume We||
Centerline Wells:
MW-15B
MW-16B
MW-1 7B
MW-4B
MW-2B
Meng
Minnesota
Distance to Receptor: 1 0 feet
Distance to Receptor (feet)
10.0
520.6
1063.2
1998.1
2897.1
The distance is measured in the Groundwater Flow Angle
from the well to the compliance boundary.
Sample Event Effective Date
Number of Regression Confidence in
Centerline Wells Coefficient (1 /ft) Coefficient
TETRACHLOROETHYLENE(PCE)
May 1999 5/20/1999
July 1999 7/7/1999
September 1 999 9/1 5/1 999
October 1999 10/26/1999
March 2000 3/15/2000
June 2000 7/1/2000
September 2000 9/10/2000
October 2000 10/26/2000
December 2000 12/31/2000
March 2001 3/23/2001
May 2001 5/22/2001
September 2001 9/15/2001
November 2001 11/29/2001
January 2002 1/31/2002
April 2002 4/3/2002
July 2002 7/25/2002
October 2002 10/14/2002
5 -2.31 E-03
0 O.OOE+00
2 O.OOE+00
3 -5.38E-03
0 O.OOE+00
0 O.OOE+00
5 -1 .50E-03
0 O.OOE+00
0 O.OOE+00
0 O.OOE+00
0 O.OOE+00
5 -1.51 E-03
0 O.OOE+00
0 O.OOE+00
0 O.OOE+00
0 O.OOE+00
5 -1 .24E-03
97.9%
0.0%
0.0%
91.5%
0.0%
0.0%
92.7%
0.0%
0.0%
0.0%
0.0%
92.1%
0.0%
0.0%
0.0%
0.0%
91 .3%
Note: when the number of plume centerline wells is less than 3, no analysis is performed and all related values
are set to ZERO; Confidence in Coefficient is the statistical confidence that the estimated coefficient is
different from ZERO (for details, please refer to "Conference in Trend" in Linear Regression Analysis).
MAROS Version 2, 2002, AFCEE
Wednesday, February 12, 2003
Page 1 of 1
-------
Risk-Based Power Analysis — Projected Concentrations
Long Prairie site
Location: Todd County
From Period:
Sampling
Event
5/20/1999 to
Effective
Date
10/14/2002
Well
Meng
Minnesota
Distance from the most downgradient well to receptor: 10 feet
Observed
Concentration
(mg/L)
Distance Down
Centerline (ft)
Regression
Coefficient
(1/ft)
Projected
Concentration
(mg/L)
Below
Detection
Limit?
Used in
Analysis?
TETRACHLOROETHYLENE(PCE)
May 1999
May 1999
May 1999
May 1999
May 1999
May 1999
May 1999
May 1999
May 1999
May 1999
May 1999
May 1999
May 1999
May 1999
May 1999
May 1999
May 1999
May 1999
May 1999
May 1999
May 1999
May 1999
May 1999
May 1999
May 1999
May 1999
May 1999
May 1999
May 1999
May 1999
May 1999
May 1999
5/20/1999
5/20/1999
5/20/1999
5/20/1999
5/20/1999
5/20/1999
5/20/1999
5/20/1999
5/20/1999
5/20/1999
5/20/1999
5/20/1999
5/20/1999
5/20/1999
5/20/1999
5/20/1999
5/20/1999
5/20/1999
5/20/1999
5/20/1999
5/20/1999
5/20/1999
5/20/1999
5/20/1999
5/20/1999
5/20/1999
5/20/1999
5/20/1999
5/20/1999
5/20/1999
5/20/1999
5/20/1999
BAL2C
CW-3
CW-6
MW-10
MW-11C
MW-13C
MW-14B
MW-14C
MW-15A
MW-15B
MW-16A
MW-16B
MW-17B
MW-18A
MW-18B
MW-1A
MW-1B
MW-2A
MW-2B
MW-2C
MW-3A
MW-3B
MW-4A
MW-4B
MW-4C
MW-5A
MW-5B
MW-6A
MW-6B
MW-6C
RW-1A
RW-1B
5.000E-04
5.000E-04
5.000E-04
9.900E-02
5.000E-04
5.000E-04
3.500E-01
5.000E-04
5.000E-04
5.000E-04
5.000E-04
1 .450E-02
2.000E-01
5.000E-04
2.200E-03
5.000E-04
5.000E-04
7.600E-02
9.700E-01
5.100E-03
5.000E-04
5.000E-04
7.400E-03
2.300E-01
1.700E-01
5.000E-04
5.000E-04
3.750E-01
2.600E-01
2.100E-01
1 .500E-03
5.000E-04
594.6
2938.7
430.4
3445.6
1183.4
1736.7
1671.6
1671.6
10.0
10.0
520.6
520.6
1063.2
357.3
357.3
2962.9
2962.9
2897.1
2897.1
2897.1
3151.4
3151.4
1998.1
1998.1
1998.1
1491.9
1491.9
2421.9
2421.9
2421.9
3468.7
3409.4
-2.31 E-03
-2.31 E-03
-2.31 E-03
-2.31 E-03
-2.31 E-03
-2.31 E-03
-2.31 E-03
-2.31 E-03
-2.31 E-03
-2.31 E-03
-2.31 E-03
-2.31 E-03
-2.31 E-03
-2.31 E-03
-2.31 E-03
-2.31 E-03
-2.31 E-03
-2.31 E-03
-2.31 E-03
-2.31 E-03
-2.31 E-03
-2.31 E-03
-2.31 E-03
-2.31 E-03
-2.31 E-03
-2.31 E-03
-2.31 E-03
-2.31 E-03
-2.31 E-03
-2.31 E-03
-2.31 E-03
-2.31 E-03
1 .268E-04
5.684E-07
1 .852E-04
3.495E-05
3.260E-05
9.099E-06
7.400E-03
1 .057E-05
4.886E-04
4.886E-04
1 .505E-04
4.363E-03
1.721E-02
2.193E-04
9.647E-04
5.375E-07
5.375E-07
9.511E-05
1.214E-03
6.382E-06
3.479E-07
3.479E-07
7.368E-05
2.290E-03
1 .693E-03
1 .600E-05
1 .600E-05
1 .404E-03
9.737E-04
7.864E-04
5.020E-07
1.919E-07
Yes
Yes
Yes
Yes
Yes
Yes
No
Yes
Yes
Yes
Yes
No
No
Yes
No
Yes
Yes
Yes
No
Yes
Yes
Yes
Yes
No
No
Yes
Yes
No
No
No
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
No
No
MAROS Version 2, 2002, AFCEE
Wednesday, February 12, 2003
Page 1 of 5
-------
Project: Long Prairie site
Location: Todd County
Meng
State: Minnesota
Sampling
Event
Effective
Date
Well
Concentration
(mg/L)
Distance Down
Centerline (ft)
Coefficient
(1/ft)
Concentration
(mg/L)
Detection
Limit?
Used in
Analysis?
TETRACHLOROETHYLENE(PCE)
May 1999
May 1999
May 1999
May 1999
May 1999
October 1 999
October 1 999
October 1 999
October 1 999
October 1 999
October 1 999
October 1 999
October 1 999
October 1 999
October 1 999
October 1 999
October 1 999
October 1 999
October 1 999
October 1 999
October 1 999
October 1 999
October 1 999
October 1 999
October 1 999
October 1 999
October 1 999
October 1 999
October 1 999
October 1 999
October 1 999
October 1 999
October 1 999
October 1 999
October 1 999
October 1 999
5/20/1999
5/20/1999
5/20/1999
5/20/1999
5/20/1999
10/26/1999
10/26/1999
10/26/1999
10/26/1999
10/26/1999
10/26/1999
10/26/1999
10/26/1999
10/26/1999
10/26/1999
10/26/1999
10/26/1999
10/26/1999
10/26/1999
10/26/1999
10/26/1999
10/26/1999
10/26/1999
10/26/1999
10/26/1999
10/26/1999
10/26/1999
10/26/1999
10/26/1999
10/26/1999
10/26/1999
10/26/1999
10/26/1999
10/26/1999
10/26/1999
10/26/1999
RW-3
RW-4
RW-5
RW-6
RW-7
BAL2C
CW-3
CW-6
MW-10
MW-11C
MW-13C
MW-14B
MW-14C
MW-15A
MW-15B
MW-16A
MW-16B
MW-17B
MW-18A
MW-18B
MW-19B
MW-1A
MW-1B
MW-2C
MW-3A
MW-3B
MV\MC
MW-5A
MW-5B
MW-6C
RW-1A
RW-1B
RW-3
RW-4
RW-5
RW-6
1.350E-01
5.000E-04
1.900E-01
4.100E-02
4.000E-02
5.000E-04
5.000E-04
5.000E-04
6.500E-02
5.000E-04
5.000E-04
2.100E-01
5.000E-04
5.000E-04
5.000E-04
5.000E-04
2.000E-03
1.400E-01
5.000E-04
1 .200E-03
5.000E-04
5.000E-04
5.000E-04
3.400E-03
5.000E-04
5.000E-04
1.400E-01
5.000E-04
5.000E-04
1.100E-01
1 .OOOE-03
5.000E-04
8.750E-02
5.000E-04
2.200E-01
4.400E-02
2922.2
2414.7
1730.8
2364.6
1686.5
594.6
2938.7
430.4
3445.6
1183.4
1736.7
1671.6
1671.6
10.0
10.0
520.6
520.6
1063.2
357.3
357.3
901.0
2962.9
2962.9
2897.1
3151.4
3151.4
1998.1
1491.9
1491.9
2421.9
3468.7
3409.4
2922.2
2414.7
1730.8
2364.6
-2.31 E-03
-2.31 E-03
-2.31 E-03
-2.31 E-03
-2.31 E-03
-5.38E-03
-5.38E-03
-5.38E-03
-5.38E-03
-5.38E-03
-5.38E-03
-5.38E-03
-5.38E-03
-5.38E-03
-5.38E-03
-5.38E-03
-5.38E-03
-5.38E-03
-5.38E-03
-5.38E-03
-5.38E-03
-5.38E-03
-5.38E-03
-5.38E-03
-5.38E-03
-5.38E-03
-5.38E-03
-5.38E-03
-5.38E-03
-5.38E-03
-5.38E-03
-5.38E-03
-5.38E-03
-5.38E-03
-5.38E-03
-5.38E-03
1 .594E-04
1 .904E-06
3.505E-03
1 .753E-04
8.172E-04
2.045E-05
6.882E-11
4.943E-05
5.863E-10
8.627E-07
4.407E-08
2.626E-05
6.253E-08
4.738E-04
4.738E-04
3.045E-05
1.218E-04
4.611E-04
7.323E-05
1 .757E-04
3.939E-06
6.041 E-11
6.041 E-11
5.854E-10
2.193E-11
2.193E-11
3.027E-06
1 .643E-07
1 .643E-07
2.436E-07
7.966E-12
5.478E-12
1.316E-08
1.151E-09
2.002E-05
1 .326E-07
Yes
Yes
No
Yes
No
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
No
No
No
No
No
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
No
No
No
No
No
No
MAROS Version 2, 2002, AFCEE
Wednesday, February 12, 2003
Page 2 of 5
-------
Project: Long Prairie site
Location: Todd County
Meng
State: Minnesota
Sampling
Event
Effective
Date
Well
Concentration
(mg/L)
Distance Down
Centerline (ft)
Coefficient
(1/ft)
Concentration
(mg/L)
Detection
Limit?
Used in
Analysis?
TETRACHLOROETHYLENE(PCE)
October 1 999
September 2000
September 2000
September 2000
September 2000
September 2000
September 2000
September 2000
September 2000
September 2000
September 2000
September 2000
September 2000
September 2000
September 2000
September 2000
September 2000
September 2000
September 2000
September 2000
September 2000
September 2000
September 2000
September 2000
September 2000
September 2000
September 2001
September 2001
September 2001
September 2001
September 2001
September 2001
September 2001
September 2001
September 2001
September 2001
10/26/1999
9/10/2000
9/10/2000
9/10/2000
9/10/2000
9/10/2000
9/10/2000
9/10/2000
9/10/2000
9/10/2000
9/10/2000
9/10/2000
9/10/2000
9/10/2000
9/10/2000
9/10/2000
9/10/2000
9/10/2000
9/10/2000
9/10/2000
9/10/2000
9/10/2000
9/10/2000
9/10/2000
9/10/2000
9/10/2000
9/15/2001
9/15/2001
9/15/2001
9/15/2001
9/15/2001
9/15/2001
9/15/2001
9/15/2001
9/15/2001
9/15/2001
RW-7
CW-3
MW-11B
MW-11C
MW-14B
MW-14C
MW-15A
MW-15B
MW-16B
MW-17B
MW-19B
MW-2A
MW-2B
MW-2C
MV\MB
MV\MC
MW-6A
MW-6B
MW-6C
RW-3
RW-4
RW-5
RW-6
RW-7
RW-8
RW-9
CW-3
MW-11B
MW-11C
MW-14B
MW-14C
MW-15A
MW-15B
MW-16B
MW-17B
MW-19B
3.600E-02
5.000E-04
5.000E-04
5.000E-04
1.400E-01
5.000E-04
5.000E-04
5.000E-04
1 .400E-02
1.000E-01
5.000E-04
1.100E-02
7.100E-02
5.000E-04
1.400E-01
1.150E-01
5.400E-02
4.700E-02
2.100E-02
5.100E-02
5.000E-04
7.400E-02
2.000E-02
3.100E-02
3.000E-02
1 .600E-02
5.000E-04
5.000E-04
5.000E-04
1.300E-01
5.000E-04
5.000E-04
5.000E-04
3.700E-03
9.200E-02
5.000E-04
1686.5
2938.7
1183.4
1183.4
1671.6
1671.6
10.0
10.0
520.6
1063.2
901.0
2897.1
2897.1
2897.1
1998.1
1998.1
2421.9
2421.9
2421.9
2922.2
2414.7
1730.8
2364.6
1686.5
1096.9
535.7
2938.7
1183.4
1183.4
1671.6
1671.6
10.0
10.0
520.6
1063.2
901.0
-5.38E-03
-1.50E-03
-1.50E-03
-1.50E-03
-1.50E-03
-1.50E-03
-1.50E-03
-1.50E-03
-1.50E-03
-1.50E-03
-1.50E-03
-1.50E-03
-1.50E-03
-1.50E-03
-1.50E-03
-1.50E-03
-1.50E-03
-1.50E-03
-1.50E-03
-1.50E-03
-1.50E-03
-1.50E-03
-1.50E-03
-1.50E-03
-1.50E-03
-1.50E-03
-1.51E-03
-1.51E-03
-1.51E-03
-1.51E-03
-1.51E-03
-1.51E-03
-1.51E-03
-1.51E-03
-1.51E-03
-1.51E-03
4.156E-06
6.121E-06
8.490E-05
8.490E-05
1.144E-02
4.086E-05
4.926E-04
4.926E-04
6.418E-03
2.033E-02
1 .296E-04
1 .433E-04
9.251 E-04
6.515E-06
7.015E-03
5.762E-03
1 .434E-03
1 .248E-03
5.576E-04
6.399E-04
1 .342E-05
5.534E-03
5.787E-04
2.477E-03
5.799E-03
7.170E-03
5.942E-06
8.390E-05
8.390E-05
1 .045E-02
4.017E-05
4.925E-04
4.925E-04
1 .687E-03
1.851E-02
1 .285E-04
Yes
Yes
Yes
Yes
No
Yes
Yes
Yes
No
No
Yes
Yes
No
Yes
No
No
No
No
No
No
Yes
No
No
No
No
No
Yes
Yes
Yes
No
Yes
Yes
Yes
No
No
Yes
No
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
No
No
No
No
No
No
No
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
MAROS Version 2, 2002, AFCEE
Wednesday, February 12, 2003
Page 3 of 5
-------
Project: Long Prairie site
Location: Todd County
Meng
State: Minnesota
Sampling
Event
Effective
Date
Well
Concentration
(mg/L)
Distance Down
Centerline (ft)
Coefficient
(1/ft)
Concentration
(mg/L)
Detection
Limit?
Used in
Analysis?
TETRACHLOROETHYLENE(PCE)
September 2001
September 2001
September 2001
September 2001
September 2001
September 2001
September 2001
September 2001
September 2001
September 2001
September 2001
September 2001
September 2001
September 2001
September 2001
October 2002
October 2002
October 2002
October 2002
October 2002
October 2002
October 2002
October 2002
October 2002
October 2002
October 2002
October 2002
October 2002
October 2002
October 2002
October 2002
October 2002
October 2002
October 2002
October 2002
October 2002
9/15/2001
9/15/2001
9/15/2001
9/15/2001
9/15/2001
9/15/2001
9/15/2001
9/15/2001
9/15/2001
9/15/2001
9/15/2001
9/15/2001
9/15/2001
9/15/2001
9/15/2001
10/14/2002
10/14/2002
10/14/2002
10/14/2002
10/14/2002
10/14/2002
10/14/2002
10/14/2002
10/14/2002
10/14/2002
10/14/2002
10/14/2002
10/14/2002
10/14/2002
10/14/2002
10/14/2002
10/14/2002
10/14/2002
10/14/2002
10/14/2002
10/14/2002
MW-2A
MW-2B
MW-2C
MW-4B
MW-4C
MW-6A
MW-6B
MW-6C
RW-3
RW-4
RW-5
RW-6
RW-7
RW-8
RW-9
BAL2B
BAL2C
CW-3
CW-6
MW-10
MW-11A
MW-11B
MW-11C
MW-13C
MW-14B
MW-14C
MW-15A
MW-15B
MW-16A
MW-16B
MW-17B
MW-18A
MW-18B
MW-19B
MW-1A
MW-1B
5.200E-03
3.800E-02
5.000E-04
1.400E-01
6.450E-02
5.400E-02
3.700E-02
1 .200E-02
2.550E-02
5.000E-04
1.025E-01
1 .250E-02
2.800E-02
1 .950E-02
1 .400E-02
5.000E-04
5.000E-04
5.000E-04
5.000E-04
3.800E-02
5.000E-04
5.000E-04
5.000E-04
5.000E-04
1.100E-01
5.000E-04
5.000E-04
5.000E-04
5.000E-04
2.000E-02
4.000E-02
5.000E-04
2.400E-03
5.000E-04
5.000E-04
5.000E-04
2897.1
2897.1
2897.1
1998.1
1998.1
2421.9
2421.9
2421.9
2922.2
2414.7
1730.8
2364.6
1686.5
1096.9
535.7
318.8
594.6
2938.7
430.4
3445.6
1183.4
1183.4
1183.4
1736.7
1671.6
1671.6
10.0
10.0
520.6
520.6
1063.2
357.3
357.3
901.0
2962.9
2962.9
-1.51E-03
-1.51E-03
-1.51E-03
-1.51E-03
-1.51E-03
-1.51E-03
-1.51E-03
-1.51E-03
-1.51E-03
-1.51E-03
-1.51E-03
-1.51E-03
-1.51E-03
-1.51E-03
-1.51E-03
-1.24E-03
-1.24E-03
-1.24E-03
-1.24E-03
-1.24E-03
-1.24E-03
-1.24E-03
-1.24E-03
-1.24E-03
-1.24E-03
-1.24E-03
-1.24E-03
-1.24E-03
-1.24E-03
-1.24E-03
-1.24E-03
-1.24E-03
-1.24E-03
-1.24E-03
-1.24E-03
-1.24E-03
6.580E-05
4.809E-04
6.327E-06
6.875E-03
3.167E-03
1 .399E-03
9.587E-04
3.109E-04
3.107E-04
1.310E-05
7.533E-03
3.531 E-04
2.200E-03
3.728E-03
6.240E-03
3.362E-04
2.385E-04
1 .288E-05
2.926E-04
5.210E-04
1.146E-04
1.146E-04
1.146E-04
5.754E-05
1 .373E-02
6.240E-05
4.938E-04
4.938E-04
2.615E-04
1 .046E-02
1 .065E-02
3.205E-04
1 .538E-03
1 .629E-04
1 .250E-05
1 .250E-05
Yes
Yes
Yes
No
No
No
No
Yes
Yes
Yes
No
Yes
No
No
No
Yes
Yes
Yes
Yes
No
Yes
Yes
Yes
Yes
No
Yes
Yes
Yes
Yes
No
No
Yes
No
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
No
No
No
No
No
No
No
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
MAROS Version 2, 2002, AFCEE
Wednesday, February 12, 2003
Page 4 of 5
-------
Project: Long Prairie site
Location: Todd County
Meng
State: Minnesota
Sampling
Event
Effective
Date
Well
Concentration
(mg/L)
Distance Down
Centerline (ft)
Coefficient
(1/ft)
Concentration
(mg/L)
Detection
Limit?
Used in
Analysis?
TETRACHLOROETHYLENE(PCE)
October 2002
October 2002
October 2002
October 2002
October 2002
October 2002
October 2002
October 2002
October 2002
October 2002
October 2002
October 2002
October 2002
October 2002
October 2002
October 2002
October 2002
October 2002
October 2002
October 2002
October 2002
October 2002
October 2002
10/14/2002
10/14/2002
10/14/2002
10/14/2002
10/14/2002
10/14/2002
10/14/2002
10/14/2002
10/14/2002
10/14/2002
10/14/2002
10/14/2002
10/14/2002
10/14/2002
10/14/2002
10/14/2002
10/14/2002
10/14/2002
10/14/2002
10/14/2002
10/14/2002
10/14/2002
10/14/2002
MW-2A
MW-2B
MW-2C
MW-3A
MW-3B
MW-4A
MW-4B
MW-4C
MW-5A
MW-5B
MW-6A
MW-6B
MW-6C
RW-1A
RW-1B
RW-1C
RW-3
RW-4
RW-5
RW-6
RW-7
RW-8
RW-9
9.700E-04
4.100E-02
6.900E-03
5.000E-04
5.000E-04
3.300E-02
7.700E-02
5.100E-02
5.000E-04
5.000E-04
5.000E-04
4.700E-03
2.500E-02
1 .OOOE-03
5.000E-04
5.000E-04
2.000E-02
5.000E-04
8.400E-02
9.500E-03
2.200E-02
1 .800E-02
1 .500E-02
2897.1
2897.1
2897.1
3151.4
3151.4
1998.1
1998.1
1998.1
1491.9
1491.9
2421.9
2421.9
2421.9
3468.7
3409.4
3412.5
2922.2
2414.7
1730.8
2364.6
1686.5
1096.9
535.7
-1.24E-03
-1.24E-03
-1.24E-03
-1.24E-03
-1.24E-03
-1.24E-03
-1.24E-03
-1.24E-03
-1.24E-03
-1.24E-03
-1.24E-03
-1.24E-03
-1.24E-03
-1.24E-03
-1.24E-03
-1.24E-03
-1.24E-03
-1.24E-03
-1.24E-03
-1.24E-03
-1.24E-03
-1.24E-03
-1.24E-03
2.633E-05
1.113E-03
1 .873E-04
9.887E-06
9.887E-06
2.743E-03
6.400E-03
4.239E-03
7.804E-05
7.804E-05
2.452E-05
2.305E-04
1 .226E-03
1 .332E-05
7.171E-06
7.144E-06
5.261 E-04
2.474E-05
9.739E-03
5.003E-04
2.695E-03
4.594E-03
7.699E-03
Yes
No
Yes
Yes
Yes
No
No
No
Yes
Yes
Yes
Yes
No
Yes
Yes
Yes
No
Yes
No
No
No
No
No
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
No
No
No
No
No
No
No
No
No
No
Note: Projected Concentrations that are below the user-specified detection limit are indicated by a check mark to its right; for sampling events
with less than 3 selected plume centerline wells, NO projected concentrations are calculated because no regression coefficient is available.
MAROS Version 2, 2002, AFCEE
Wednesday, February 12, 2003
Page 5 of 5
-------
MAROS Risk-Based Power Analysis for Site Cleanup
Project: Long Prairie site
Todd County
Meng
Minnesota
Parameters:
Groundwater Flow Direction: 90 degrees Distance to Receptor: 10 feet
From Period: May 1999 to October 2002
5/20/1999
Selected Plume
Centerline Wells:
10/14/2002
Well
MW-15B
MW-16B
MW-1 7B
MW-4B
MW-2B
The distance is
from the well to
Distance to Receptor (feet)
10.0
520.6
1063.2
1998.1
2897.1
measured in the Groundwater Flow Angle
the compliance boundary.
Normal Distribution Assumption Lognormal Distribution Assumption
Sample Event
Sample Sample
Szie Mean
Sample
Stdev.
TETRACHLOROETHYLENE(PCE)
May 1999
October 1999
September 2000
September 2001
October 2002
30
25
18
18
34
1 .34E-03
7.66E-05
3.15E-03
2.51 E-03
1 .66E-03
3.38E-03
1 .54E-04
5.40E-03
4.86E-03
3.44E-03
Cleanup
Status
Expected Celanup
Power Samp|e Size status
Expected
Power Sample Size
Alpha Expected
Level Power
Cleanup Goal = 0.005
Attained
Attained
Not Attained
Attained
Attained
1.000
1.000
0.414
0.685
1.000
7
<=3
54
25
8
Not Attained
Not Attained
Not Attained
Not Attained
Attained
S/E
S/E
S/E
S/E
0.728
S/E
S/E
S/E
S/E
42
0.05
0.05
0.05
0.05
0.05
0.8
0.8
0.8
0.8
0.8
Note: #N/C means "not conducted" due to a small sample size (N<4) or that the mean concentration is much greater than the cleanup
level; Sample Size is the number of sampling locations used in the power analysis; Expected Sample Size is the number of concentration
data needed to reach the Expected Power undercurrent sample variability.
MAROS Version 2, 2002, AFCEE
Wednesday, February 12, 2003
Page 1 of 1
-------
-------
Risk-Based Power Analysis — Projected Concentrations
Long Prairie site
Location: Todd County
From Period:
Sampling
Event
5/20/1999 to
Effective
Date
10/14/2002
Well
Meng
Minnesota
Distance from the most downgradient well to receptor: -100 feet
Observed
Concentration
(mg/L)
Distance Down
Centerline (ft)
Regression
Coefficient
(1/ft)
Projected
Concentration
(mg/L)
Below
Detection
Limit?
Used in
Analysis?
TETRACHLOROETHYLENE(PCE)
May 1999
May 1999
May 1999
May 1999
May 1999
May 1999
May 1999
May 1999
May 1999
May 1999
May 1999
May 1999
May 1999
May 1999
May 1999
May 1999
May 1999
May 1999
May 1999
May 1999
May 1999
May 1999
May 1999
May 1999
May 1999
May 1999
May 1999
May 1999
May 1999
May 1999
May 1999
May 1999
5/20/1999
5/20/1999
5/20/1999
5/20/1999
5/20/1999
5/20/1999
5/20/1999
5/20/1999
5/20/1999
5/20/1999
5/20/1999
5/20/1999
5/20/1999
5/20/1999
5/20/1999
5/20/1999
5/20/1999
5/20/1999
5/20/1999
5/20/1999
5/20/1999
5/20/1999
5/20/1999
5/20/1999
5/20/1999
5/20/1999
5/20/1999
5/20/1999
5/20/1999
5/20/1999
5/20/1999
5/20/1999
BAL2C
CW-3
CW-6
MW-10
MW-11C
MW-13C
MW-14B
MW-14C
MW-15A
MW-15B
MW-16A
MW-16B
MW-17B
MW-18A
MW-18B
MW-1A
MW-1B
MW-2A
MW-2B
MW-2C
MW-3A
MW-3B
MW-4A
MW-4B
MW-4C
MW-5A
MW-5B
MW-6A
MW-6B
MW-6C
RW-1A
RW-1B
5.000E-04
5.000E-04
5.000E-04
9.900E-02
5.000E-04
5.000E-04
3.500E-01
5.000E-04
5.000E-04
5.000E-04
5.000E-04
1 .450E-02
2.000E-01
5.000E-04
2.200E-03
5.000E-04
5.000E-04
7.600E-02
9.700E-01
5.100E-03
5.000E-04
5.000E-04
7.400E-03
2.300E-01
1.700E-01
5.000E-04
5.000E-04
3.750E-01
2.600E-01
2.100E-01
1 .500E-03
5.000E-04
484.6
2828.7
320.4
3335.6
1073.4
1626.7
1561.6
1561.6
-100.0
-100.0
410.6
410.6
953.2
247.3
247.3
2852.9
2852.9
2787.1
2787.1
2787.1
3041.4
3041.4
1888.1
1888.1
1888.1
1381.9
1381.9
2311.9
2311.9
2311.9
3358.7
3299.4
-2.31 E-03
-2.31 E-03
-2.31 E-03
-2.31 E-03
-2.31 E-03
-2.31 E-03
-2.31 E-03
-2.31 E-03
-2.31 E-03
-2.31 E-03
-2.31 E-03
-2.31 E-03
-2.31 E-03
-2.31 E-03
-2.31 E-03
-2.31 E-03
-2.31 E-03
-2.31 E-03
-2.31 E-03
-2.31 E-03
-2.31 E-03
-2.31 E-03
-2.31 E-03
-2.31 E-03
-2.31 E-03
-2.31 E-03
-2.31 E-03
-2.31 E-03
-2.31 E-03
-2.31 E-03
-2.31 E-03
-2.31 E-03
1 .635E-04
7.326E-07
2.387E-04
4.505E-05
4.202E-05
1.173E-05
9.538E-03
1 .363E-05
6.297E-04
6.297E-04
1 .939E-04
5.624E-03
2.218E-02
2.826E-04
1 .243E-03
6.927E-07
6.927E-07
1 .226E-04
1 .564E-03
8.226E-06
4.485E-07
4.485E-07
9.496E-05
2.952E-03
2.182E-03
2.062E-05
2.062E-05
1.810E-03
1 .255E-03
1.014E-03
6.471 E-07
2.473E-07
Yes
Yes
Yes
Yes
Yes
Yes
No
Yes
No
No
Yes
No
No
Yes
No
Yes
Yes
Yes
No
Yes
Yes
Yes
Yes
No
No
Yes
Yes
No
No
No
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
No
No
MAROS Version 2, 2002, AFCEE
Wednesday, February 12, 2003
Page 1 of 5
-------
Project: Long Prairie site
Location: Todd County
Meng
State: Minnesota
Sampling
Event
Effective
Date
Well
Concentration
(mg/L)
Distance Down
Centerline (ft)
Coefficient
(1/ft)
Concentration
(mg/L)
Detection
Limit?
Used in
Analysis?
TETRACHLOROETHYLENE(PCE)
May 1999
May 1999
May 1999
May 1999
May 1999
October 1 999
October 1 999
October 1 999
October 1 999
October 1 999
October 1 999
October 1 999
October 1 999
October 1 999
October 1 999
October 1 999
October 1 999
October 1 999
October 1 999
October 1 999
October 1 999
October 1 999
October 1 999
October 1 999
October 1 999
October 1 999
October 1 999
October 1 999
October 1 999
October 1 999
October 1 999
October 1 999
October 1 999
October 1 999
October 1 999
October 1 999
5/20/1999
5/20/1999
5/20/1999
5/20/1999
5/20/1999
10/26/1999
10/26/1999
10/26/1999
10/26/1999
10/26/1999
10/26/1999
10/26/1999
10/26/1999
10/26/1999
10/26/1999
10/26/1999
10/26/1999
10/26/1999
10/26/1999
10/26/1999
10/26/1999
10/26/1999
10/26/1999
10/26/1999
10/26/1999
10/26/1999
10/26/1999
10/26/1999
10/26/1999
10/26/1999
10/26/1999
10/26/1999
10/26/1999
10/26/1999
10/26/1999
10/26/1999
RW-3
RW-4
RW-5
RW-6
RW-7
BAL2C
CW-3
CW-6
MW-10
MW-11C
MW-13C
MW-14B
MW-14C
MW-15A
MW-15B
MW-16A
MW-16B
MW-17B
MW-18A
MW-18B
MW-19B
MW-1A
MW-1B
MW-2C
MW-3A
MW-3B
MV\MC
MW-5A
MW-5B
MW-6C
RW-1A
RW-1B
RW-3
RW-4
RW-5
RW-6
1.350E-01
5.000E-04
1.900E-01
4.100E-02
4.000E-02
5.000E-04
5.000E-04
5.000E-04
6.500E-02
5.000E-04
5.000E-04
2.100E-01
5.000E-04
5.000E-04
5.000E-04
5.000E-04
2.000E-03
1.400E-01
5.000E-04
1 .200E-03
5.000E-04
5.000E-04
5.000E-04
3.400E-03
5.000E-04
5.000E-04
1.400E-01
5.000E-04
5.000E-04
1.100E-01
1 .OOOE-03
5.000E-04
8.750E-02
5.000E-04
2.200E-01
4.400E-02
2812.2
2304.7
1620.8
2254.6
1576.5
484.6
2828.7
320.4
3335.6
1073.4
1626.7
1561.6
1561.6
-100.0
-100.0
410.6
410.6
953.2
247.3
247.3
791.0
2852.9
2852.9
2787.1
3041.4
3041.4
1888.1
1381.9
1381.9
2311.9
3358.7
3299.4
2812.2
2304.7
1620.8
2254.6
-2.31 E-03
-2.31 E-03
-2.31 E-03
-2.31 E-03
-2.31 E-03
-5.38E-03
-5.38E-03
-5.38E-03
-5.38E-03
-5.38E-03
-5.38E-03
-5.38E-03
-5.38E-03
-5.38E-03
-5.38E-03
-5.38E-03
-5.38E-03
-5.38E-03
-5.38E-03
-5.38E-03
-5.38E-03
-5.38E-03
-5.38E-03
-5.38E-03
-5.38E-03
-5.38E-03
-5.38E-03
-5.38E-03
-5.38E-03
-5.38E-03
-5.38E-03
-5.38E-03
-5.38E-03
-5.38E-03
-5.38E-03
-5.38E-03
2.055E-04
2.454E-06
4.518E-03
2.259E-04
1 .053E-03
3.693E-05
1.243E-10
8.929E-05
1 .059E-09
1 .559E-06
7.962E-08
4.744E-05
1.130E-07
8.560E-04
8.560E-04
5.500E-05
2.200E-04
8.329E-04
1 .323E-04
3.175E-04
7.116E-06
1.091E-10
1.091E-10
1 .057E-09
3.962E-1 1
3.962E-1 1
5.469E-06
2.968E-07
2.968E-07
4.401 E-07
1 .439E-1 1
9.897E-12
2.377E-08
2.080E-09
3.617E-05
2.396E-07
Yes
Yes
No
Yes
No
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
No
No
Yes
Yes
No
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
No
No
No
No
No
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
No
No
No
No
No
No
MAROS Version 2, 2002, AFCEE
Wednesday, February 12, 2003
Page 2 of 5
-------
Project: Long Prairie site
Location: Todd County
Meng
State: Minnesota
Sampling
Event
Effective
Date
Well
Concentration
(mg/L)
Distance Down
Centerline (ft)
Coefficient
(1/ft)
Concentration
(mg/L)
Detection
Limit?
Used in
Analysis?
TETRACHLOROETHYLENE(PCE)
October 1 999
September 2000
September 2000
September 2000
September 2000
September 2000
September 2000
September 2000
September 2000
September 2000
September 2000
September 2000
September 2000
September 2000
September 2000
September 2000
September 2000
September 2000
September 2000
September 2000
September 2000
September 2000
September 2000
September 2000
September 2000
September 2000
September 2001
September 2001
September 2001
September 2001
September 2001
September 2001
September 2001
September 2001
September 2001
September 2001
10/26/1999
9/10/2000
9/10/2000
9/10/2000
9/10/2000
9/10/2000
9/10/2000
9/10/2000
9/10/2000
9/10/2000
9/10/2000
9/10/2000
9/10/2000
9/10/2000
9/10/2000
9/10/2000
9/10/2000
9/10/2000
9/10/2000
9/10/2000
9/10/2000
9/10/2000
9/10/2000
9/10/2000
9/10/2000
9/10/2000
9/15/2001
9/15/2001
9/15/2001
9/15/2001
9/15/2001
9/15/2001
9/15/2001
9/15/2001
9/15/2001
9/15/2001
RW-7
CW-3
MW-11B
MW-11C
MW-14B
MW-14C
MW-15A
MW-15B
MW-16B
MW-17B
MW-19B
MW-2A
MW-2B
MW-2C
MV\MB
MV\MC
MW-6A
MW-6B
MW-6C
RW-3
RW-4
RW-5
RW-6
RW-7
RW-8
RW-9
CW-3
MW-11B
MW-11C
MW-14B
MW-14C
MW-15A
MW-15B
MW-16B
MW-17B
MW-19B
3.600E-02
5.000E-04
5.000E-04
5.000E-04
1.400E-01
5.000E-04
5.000E-04
5.000E-04
1 .400E-02
1.000E-01
5.000E-04
1.100E-02
7.100E-02
5.000E-04
1.400E-01
1.150E-01
5.400E-02
4.700E-02
2.100E-02
5.100E-02
5.000E-04
7.400E-02
2.000E-02
3.100E-02
3.000E-02
1 .600E-02
5.000E-04
5.000E-04
5.000E-04
1.300E-01
5.000E-04
5.000E-04
5.000E-04
3.700E-03
9.200E-02
5.000E-04
1576.5
2828.7
1073.4
1073.4
1561.6
1561.6
-100.0
-100.0
410.6
953.2
791.0
2787.1
2787.1
2787.1
1888.1
1888.1
2311.9
2311.9
2311.9
2812.2
2304.7
1620.8
2254.6
1576.5
986.9
425.7
2828.7
1073.4
1073.4
1561.6
1561.6
-100.0
-100.0
410.6
953.2
791.0
-5.38E-03
-1.50E-03
-1.50E-03
-1.50E-03
-1.50E-03
-1.50E-03
-1.50E-03
-1.50E-03
-1.50E-03
-1.50E-03
-1.50E-03
-1.50E-03
-1.50E-03
-1.50E-03
-1.50E-03
-1.50E-03
-1.50E-03
-1.50E-03
-1.50E-03
-1.50E-03
-1.50E-03
-1.50E-03
-1.50E-03
-1.50E-03
-1.50E-03
-1.50E-03
-1.51E-03
-1.51E-03
-1.51E-03
-1.51E-03
-1.51E-03
-1.51E-03
-1.51E-03
-1.51E-03
-1.51E-03
-1.51E-03
7.508E-06
7.217E-06
1.001E-04
1.001E-04
1 .349E-02
4.818E-05
5.808E-04
5.808E-04
7.568E-03
2.398E-02
1 .529E-04
1 .690E-04
1.091E-03
7.682E-06
8.272E-03
6.795E-03
1.691E-03
1.471E-03
6.575E-04
7.546E-04
1 .583E-05
6.526E-03
6.823E-04
2.921 E-03
6.838E-03
8.455E-03
7.014E-06
9.904E-05
9.904E-05
1 .233E-02
4.742E-05
5.814E-04
5.814E-04
1 .992E-03
2.185E-02
1.516E-04
Yes
Yes
Yes
Yes
No
Yes
No
No
No
No
Yes
Yes
No
Yes
No
No
No
No
No
No
Yes
No
No
No
No
No
Yes
Yes
Yes
No
Yes
No
No
No
No
Yes
No
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
No
No
No
No
No
No
No
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
MAROS Version 2, 2002, AFCEE
Wednesday, February 12, 2003
Page 3 of 5
-------
Project: Long Prairie site
Location: Todd County
Meng
State: Minnesota
Sampling
Event
Effective
Date
Well
Concentration
(mg/L)
Distance Down
Centerline (ft)
Coefficient
(1/ft)
Concentration
(mg/L)
Detection
Limit?
Used in
Analysis?
TETRACHLOROETHYLENE(PCE)
September 2001
September 2001
September 2001
September 2001
September 2001
September 2001
September 2001
September 2001
September 2001
September 2001
September 2001
September 2001
September 2001
September 2001
September 2001
October 2002
October 2002
October 2002
October 2002
October 2002
October 2002
October 2002
October 2002
October 2002
October 2002
October 2002
October 2002
October 2002
October 2002
October 2002
October 2002
October 2002
October 2002
October 2002
October 2002
October 2002
9/15/2001
9/15/2001
9/15/2001
9/15/2001
9/15/2001
9/15/2001
9/15/2001
9/15/2001
9/15/2001
9/15/2001
9/15/2001
9/15/2001
9/15/2001
9/15/2001
9/15/2001
10/14/2002
10/14/2002
10/14/2002
10/14/2002
10/14/2002
10/14/2002
10/14/2002
10/14/2002
10/14/2002
10/14/2002
10/14/2002
10/14/2002
10/14/2002
10/14/2002
10/14/2002
10/14/2002
10/14/2002
10/14/2002
10/14/2002
10/14/2002
10/14/2002
MW-2A
MW-2B
MW-2C
MW-4B
MW-4C
MW-6A
MW-6B
MW-6C
RW-3
RW-4
RW-5
RW-6
RW-7
RW-8
RW-9
BAL2B
BAL2C
CW-3
CW-6
MW-10
MW-11A
MW-11B
MW-11C
MW-13C
MW-14B
MW-14C
MW-15A
MW-15B
MW-16A
MW-16B
MW-17B
MW-18A
MW-18B
MW-19B
MW-1A
MW-1B
5.200E-03
3.800E-02
5.000E-04
1.400E-01
6.450E-02
5.400E-02
3.700E-02
1 .200E-02
2.550E-02
5.000E-04
1.025E-01
1 .250E-02
2.800E-02
1 .950E-02
1 .400E-02
5.000E-04
5.000E-04
5.000E-04
5.000E-04
3.800E-02
5.000E-04
5.000E-04
5.000E-04
5.000E-04
1.100E-01
5.000E-04
5.000E-04
5.000E-04
5.000E-04
2.000E-02
4.000E-02
5.000E-04
2.400E-03
5.000E-04
5.000E-04
5.000E-04
2787.1
2787.1
2787.1
1888.1
1888.1
2311.9
2311.9
2311.9
2812.2
2304.7
1620.8
2254.6
1576.5
986.9
425.7
208.8
484.6
2828.7
320.4
3335.6
1073.4
1073.4
1073.4
1626.7
1561.6
1561.6
-100.0
-100.0
410.6
410.6
953.2
247.3
247.3
791.0
2852.9
2852.9
-1.51E-03
-1.51E-03
-1.51E-03
-1.51E-03
-1.51E-03
-1.51E-03
-1.51E-03
-1.51E-03
-1.51E-03
-1.51E-03
-1.51E-03
-1.51E-03
-1.51E-03
-1.51E-03
-1.51E-03
-1.24E-03
-1.24E-03
-1.24E-03
-1.24E-03
-1.24E-03
-1.24E-03
-1.24E-03
-1.24E-03
-1.24E-03
-1.24E-03
-1.24E-03
-1.24E-03
-1.24E-03
-1.24E-03
-1.24E-03
-1.24E-03
-1.24E-03
-1.24E-03
-1.24E-03
-1.24E-03
-1.24E-03
7.768E-05
5.676E-04
7.469E-06
8.116E-03
3.739E-03
1 .652E-03
1.132E-03
3.670E-04
3.667E-04
1 .546E-05
8.892E-03
4.169E-04
2.597E-03
4.401 E-03
7.366E-03
3.855E-04
2.735E-04
1 .478E-05
3.355E-04
5.974E-04
1.314E-04
1.314E-04
1.314E-04
6.599E-05
1 .574E-02
7.155E-05
5.663E-04
5.663E-04
2.999E-04
1 .200E-02
1.221E-02
3.675E-04
1 .764E-03
1 .868E-04
1 .434E-05
1 .434E-05
Yes
No
Yes
No
No
No
No
Yes
Yes
Yes
No
Yes
No
No
No
Yes
Yes
Yes
Yes
No
Yes
Yes
Yes
Yes
No
Yes
No
No
Yes
No
No
Yes
No
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
No
No
No
No
No
No
No
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
MAROS Version 2, 2002, AFCEE
Wednesday, February 12, 2003
Page 4 of 5
-------
Project: Long Prairie site
Location: Todd County
Meng
State: Minnesota
Sampling
Event
Effective
Date
Well
Concentration
(mg/L)
Distance Down
Centerline (ft)
Coefficient
(1/ft)
Concentration
(mg/L)
Detection
Limit?
Used in
Analysis?
TETRACHLOROETHYLENE(PCE)
October 2002
October 2002
October 2002
October 2002
October 2002
October 2002
October 2002
October 2002
October 2002
October 2002
October 2002
October 2002
October 2002
October 2002
October 2002
October 2002
October 2002
October 2002
October 2002
October 2002
October 2002
October 2002
October 2002
10/14/2002
10/14/2002
10/14/2002
10/14/2002
10/14/2002
10/14/2002
10/14/2002
10/14/2002
10/14/2002
10/14/2002
10/14/2002
10/14/2002
10/14/2002
10/14/2002
10/14/2002
10/14/2002
10/14/2002
10/14/2002
10/14/2002
10/14/2002
10/14/2002
10/14/2002
10/14/2002
MW-2A
MW-2B
MW-2C
MW-3A
MW-3B
MW-4A
MW-4B
MW-4C
MW-5A
MW-5B
MW-6A
MW-6B
MW-6C
RW-1A
RW-1B
RW-1C
RW-3
RW-4
RW-5
RW-6
RW-7
RW-8
RW-9
9.700E-04
4.100E-02
6.900E-03
5.000E-04
5.000E-04
3.300E-02
7.700E-02
5.100E-02
5.000E-04
5.000E-04
5.000E-04
4.700E-03
2.500E-02
1 .OOOE-03
5.000E-04
5.000E-04
2.000E-02
5.000E-04
8.400E-02
9.500E-03
2.200E-02
1 .800E-02
1 .500E-02
2787.1
2787.1
2787.1
3041.4
3041.4
1888.1
1888.1
1888.1
1381.9
1381.9
2311.9
2311.9
2311.9
3358.7
3299.4
3302.5
2812.2
2304.7
1620.8
2254.6
1576.5
986.9
425.7
-1.24E-03
-1.24E-03
-1.24E-03
-1.24E-03
-1.24E-03
-1.24E-03
-1.24E-03
-1.24E-03
-1.24E-03
-1.24E-03
-1.24E-03
-1.24E-03
-1.24E-03
-1.24E-03
-1.24E-03
-1.24E-03
-1.24E-03
-1.24E-03
-1.24E-03
-1.24E-03
-1.24E-03
-1.24E-03
-1.24E-03
3.019E-05
1 .276E-03
2.148E-04
1.134E-05
1.134E-05
3.145E-03
7.339E-03
4.861 E-03
8.949E-05
8.949E-05
2.812E-05
2.643E-04
1 .406E-03
1 .528E-05
8.223E-06
8.192E-06
6.033E-04
2.837E-05
1.117E-02
5.737E-04
3.091 E-03
5.268E-03
8.829E-03
Yes
No
Yes
Yes
Yes
No
No
No
Yes
Yes
Yes
Yes
No
Yes
Yes
Yes
No
Yes
No
No
No
No
No
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
No
No
No
No
No
No
No
No
No
No
Note: Projected Concentrations that are below the user-specified detection limit are indicated by a check mark to its right; for sampling events
with less than 3 selected plume centerline wells, NO projected concentrations are calculated because no regression coefficient is available.
MAROS Version 2, 2002, AFCEE
Wednesday, February 12, 2003
Page 5 of 5
-------
MAROS Risk-Based Power Analysis for Site Cleanup
Project: Long Prairie site
Todd County
Meng
Minnesota
Parameters:
Groundwater Flow Direction: 90 degrees Distance to Receptor: -100 feet
From Period: May 1999 to October 2002
5/20/1999
Selected Plume
Centerline Wells:
10/14/2002
Well
MW-15B
MW-16B
MW-1 7B
MW-4B
MW-2B
The distance is
from the well to
Distance to Receptor (feet)
-100.0
410.6
953.2
1888.1
2787.1
measured in the Groundwater Flow Angle
the compliance boundary.
Normal Distribution Assumption Lognormal Distribution Assumption
Sample Event
Sample Sample
Szie Mean
Sample
Stdev.
TETRACHLOROETHYLENE(PCE)
May 1999
October 1999
September 2000
September 2001
October 2002
30
25
18
18
34
1 J3E-03
1 .38E-04
3.71 E-03
2.97E-03
1 .90E-03
4.35E-03
2.79E-04
6.36E-03
5.74E-03
3.95E-03
Cleanup
Status
Expected Celanup
Power Samp|e Size status
Expected
Power Sample Size
Alpha Expected
Level Power
Cleanup Goal = 0.005
Attained
Attained
Not Attained
Not Attained
Attained
0.992
1.000
0.211
0.432
0.998
12
<=3
>100
50
11
Not Attained
Not Attained
Not Attained
Not Attained
Attained
S/E
S/E
S/E
S/E
0.591
S/E
S/E
S/E
S/E
60
0.05
0.05
0.05
0.05
0.05
0.8
0.8
0.8
0.8
0.8
Note: #N/C means "not conducted" due to a small sample size (N<4) or that the mean concentration is much greater than the cleanup
level; Sample Size is the number of sampling locations used in the power analysis; Expected Sample Size is the number of concentration
data needed to reach the Expected Power undercurrent sample variability.
MAROS Version 2, 2002, AFCEE
Wednesday, February 12, 2003
Page 1 of 1
-------
DRAFT FINAL
THREE-TIERED
GROUND WATER MONITORING NETWORK
OPTIMIZATION EVALUATION
FOR
LONG PRAIRIE GROUND WATER CONTAMINATION
SUPERFUND SITE, MINNESOTA
Prepared for
US Environmental Protection Agency
MAY 2003
Denver, Colorado
-------
EXECUTIVE SUMMARY
This report presents a description and evaluation of the groundwater monitoring
program associated with the Long Prairie Ground Water Contamination Superfund Site in
Long Prairie, Minnesota. Groundwater at the site was contaminated by release of dry-
cleaning solvents into the primary drinking water aquifer. The monitoring program at
this site was evaluated to identify potential opportunities to streamline monitoring
activities while still maintaining an effective monitoring network. This evaluation is
being conducted as part of an independent assessment of monitoring network
optimization (MNO) methods by the US Environmental Protection Agency (USEPA) and
the Air Force Center for Environmental Excellence (AFCEE).
Objectives
Groundwater monitoring programs have two primary objectives (USEPA, 1994;
Gibbons, 1994):
1. Evaluate long-term temporal trends in contaminant concentrations (temporal
objective)', and
2. Evaluate the extent to which contaminant migration is occurring (spatial
objective).
The relative success of any remediation system (including the monitoring network) is
judged based on the degree to which it achieves the stated objectives of the system.
Designing an effective groundwater monitoring program involves locating monitoring
points and developing a site-specific strategy for groundwater sampling and analysis that
maximizes the amount of relevant information that can be obtained while minimizing
incremental costs. The effectiveness of a monitoring network in achieving the two
primary monitoring objectives can be evaluated quantitatively using statistical
techniques. Qualitative evaluation also is important to allow consideration of such
factors as hydrostratigraphy, locations of potential receptor exposure points with respect
to a dissolved contaminant plume, and the direction(s) and rate(s) of contaminant
migration.
The general objective of the project was to optimize the long-term groundwater
monitoring network at the Long Prairie site by applying a three-tiered MNO approach to
assess the degree to which the monitoring network addresses each of the two primary
objectives of monitoring listed above and other important considerations. The three-
ES-1
S:\ES\SHARED\CEN\MNO\EPA\LONGPRAIRIE\WRITEUP\LongPrairieMNODraftFinal.doc
-------
tiered MNO evaluation described in this report examines the 44 wells included in the
Long Prairie monitoring network. The specific objectives of the project were as follow:
« Apply a qualitative methodology that considers factors such as
hydrostratigraphy, locations of potential receptors with respect to the dissolved
plume, and the direction(s) and rate(s) of contaminant migration, to establish the
frequency at which monitoring should be conducted, and if each well should be
retained in or removed from the monitoring program.
« Conduct a Mann-Kendall statistical analysis to determine the temporal trends of
contaminants of concern (COCs) over time, and apply an algorithm to determine
the relevance of the trends within the monitoring network.
. Determine the relative amount of spatial information contributed by each
monitoring well by performing a spatial statistical analysis utilizing kriging error
predictions.
« Combine and evaluate the results of the three analyses to establish the frequency
at which monitoring should be conducted, as well as the optimal number and
locations of wells in the monitoring network.
Current Monitoring Program
The purposes of groundwater monitoring at Long Prairie are 1) to monitor progress
toward achieving the remedial action objectives set forth in the Record of Decision (as
amended), and 2) to gather adequate information to determine the status and effectiveness
of the groundwater extraction and treatment system. Wells are classified as recovery
(i.e., extraction) wells, monitoring wells, and municipal water-supply wells. The
purposes of the wells included in groundwater monitoring program are to:
« Evaluate the effectiveness of the groundwater recovery system on controlling the
plume and improving regional groundwater quality
. Confirm protection of the city of Long Prairie water supply wells (Barr, 2002).
The most recent monitoring event, conducted during October 2002, involved sampling
of 44 wells in the Long Prairie contaminant plume area, including 10 recovery wells, 2
municipal water-supply wells, and 32 monitoring wells. Several of the monitoring wells
are installed as clusters at a single location, but screened at different depths. The "A"
wells are screened across the water table, the "B" wells are screened at the based of the
upper glacial outwash deposits, and the "C" wells are screened in the lower outwash
deposits. Typically, about half of the wells sampled during the most recent monitoring
event are routinely sampled as part of the groundwater monitoring program; for example,
in the 2001 and 2000 sampling rounds, a subset of 26 of the 44 wells was sampled. This
subset included quarterly sampling of the 6 active extraction wells (RW3 through RW9,
ES-2
S:\ES\SHARED\CEN\MNO\EPA\LONGPRAIRIE\WRITEUP\LongPrairieMNODraftFinal.doc
-------
excluding RW4) and city well CW3, and annual sampling of 19 monitoring wells
(including RW4). In 2002, city well CW6 was added to the quarterly sampling schedule.
The "current" sampling plan includes the 25 extraction and monitoring wells sampled
during scheduled 2000 and 2001 monitoring events, plus city wells CW3 and CW6. The
three-tiered MNO evaluation in described in this report was used to examine the 44-well
network monitored during the October 2002 sampling event and to develop optimization
recommendations. The resulting optimized well network is then compared to the
currently monitored 27-well network.
Optimization Findings
The Long Prairie groundwater monitoring program was evaluated using results for
sampling events performed from May 1996 through October 2002. The analytical
database provided to Parsons contained from 1 to 29 sampling results for each constituent
for each of the 44 wells in the Long Prairie region. The primary COCs identified for the
Long Prairie plume are PCE, TCE, and cis-l,2-DCE. PCE is the primary COC because it
has been detected at the highest concentrations and has the broadest distribution in
groundwater at the site. PCE sampling results were used to conduct the spatial
component of the three-tiered MNO evaluation.
Results from the three-tiered MNO evaluation of the 2002 program for Long Prairie
indicate that 18 of the 44 wells could be removed from the groundwater monitoring
program with little loss of information. Based on these recommendations, the "current"
sampling plan (the monitoring and extraction wells included in the 2000 and 2001
sampling schedules, plus CW3 and CW6), could be optimized by removing 4 of the 27
active monitoring wells, and adding 3 additional area wells. A refined monitoring
program, consisting of 26 wells (2 to be sampled quarterly, 6 to be sampled semi-
annually, 14 to be sampled annually, and 4 to be sampled biennially) would be adequate
to address the two primary objectives of monitoring. This refined monitoring network
would result in an average 36 sampling events per year, compared to 51 events per year
under the current monitoring program. Implementing these recommendations for
optimizing the LTM monitoring program at Long Prairie could reduce current LTM
annual monitoring by over 29 percent. Additionally, based on a per well sampling cost
ranging from$l 00 to $280, these recommendations could reduce costs by an average of
$1500 to $4,200 per year.
ES-3
S:\ES\SHARED\CEN\MNO\EPA\LONGPRAIRIE\WRITEUP\LongPrairieMNODraftFinal.doc
-------
TABLE OF CONTENTS
Page
ACRONYMS AND ABBREVIATIONS iv
EXECUTIVE SUMMARY ES-1
SECTION 1 -INTRODUCTION 1-1
SECTION 2 - SITE BACKGROUND INFORMATION 2-1
2.1 Site Description 2-1
2.2 Geology and Hydrogeology 2-3
2.3 Nature and Extent of Contamination 2-6
2.4 Remedial Systems 2-7
SECTION 3 - LONG-TERM MONITORING PROGRAM AT LONG
PRAIRIE 3-1
3.1 Description of Monitoring Program 3-1
3.2 Summary of Analytical Data 3-2
SECTION 4 - QUALITATIVE MNO EVALUATION 4-1
4.1 Methodology for Qualitative Evaluation of Monitoring Network 4-2
4.2 Results of Qualitative MNO Evaluation 4-3
4.2.1 Monitoring Network and Sampling Frequency 4-3
4.2.1.1 Extraction Wells 4-4
4.2.1.2 Municipal Water-Supply Wells 4-6
4.2.1.3 Monitoring Wells 4-7
4.2.2 Laboratory Analytical Program 4-8
4.2.3 LTM Program Flexibility 4-8
-i-
S:\ES\SHARED\CEN\MNO\EPA\LONGPRAIRIE\WRITEUP\LongPrairieMNODraftFinal.doc
-------
TABLE OF CONTENTS (Continued)
Page
SECTION 5 - TEMPORAL STATISTICAL EVALUATION 5-1
5.1 Methodology for Temporal Trend Analysis of Contaminant
Concentrations 5-1
5.2 Temporal Evaluation Results 5-7
SECTION 6 - SPATIAL STATISTICAL EVALUATION 6-1
6.1 Geostatistical methods for evaluating Monitoring networks 6-1
6.2 Spatial Evaluation of Monitoring Network at Long Prairie 6-4
6.3 Spatial Statistical Evaluation Results 6-10
6.3.1 Kriging Ranking Results 6-10
SECTION 7 - SUMMARY OF THREE-TIERED MONITORING
NETWORK EVALUATION 7-1
SECTION 8 - REFERENCES 8-1
LIST OF TABLES
No. Title Page
3.1 Groundwater monitoring Program 3-3
3.2 Summary of Occurrence of Groundwater Contaminants of Concern 3-5
4.1 Monitoring Network Optimization Decision Logic 4-3
4.2 Monitoring Frequency Decision Logic 4-4
4.3 Qualitative Evaluation of Groundwater Monitoring Network 4-5
5.1 Results of Temporal Trend Analysis of COC Concentrations in Current
Groundwater Monitoring Network 5-8
6.1 Results of geostatistical Evaluation Ranking Wells by Relative Value of
Information 6-6
7.1 Summary of Evaluation of Current Groundwater Monitoring Program 7-4
-11-
S:\ES\SHARED\CEN\MNO\EPA\LONGPRAIRIE\WRITEUP\LongPrairieMNODraftFinal.doc
-------
TABLE OF CONTENTS (CONTINUED)
LIST OF FIGURES
No. Title Page
2.1 Current Monitoring Well Network and Sampling Frequencies 2-5
5.1 PCE Concentrations Through Time at Well MW6B 5-2
5.2 Conceptual Representation of Temporal Trends and Temporal Variation
in Concentrations 5-3
5.3 Conceptual Representation of Continued Monitoring at Location Where
No Temporal Trend in Concentration is Present 5-6
5.4 Mann-Kendall Temporal Trend Analysis for Concentrations of PCE 5-11
5.5 Temporal Chemical Concentration Trend Decision Rationale Flowchart 5-12
6.1 Idealized Semvariogram Model 6-3
6.2 Idealized Semivariogram Model 6-9
6.3 Results of Geostatistical Analysis Showing Relative Value of Spatial
Information of PCE Distribution in Select Wells 6-11
-111-
S:\ES\SHARED\CEN\MNO\EPA\LONGPRAIRIE\WRITEUP\LongPrairieMNODraftFinal.doc
-------
ACRONYMS AND ABBREVIATIONS
AFCEE
ASCE
bgs
CAH
BSD
coc
DCE
ESRI
GAC
Gpm
GIS
LTM
ug/L
MCL
MCPA
MDH
MNO
NPL
O&M
OU
PCE
PQL
RAO
ROD
SVE
TCE
USEPA
vc
voc
Air Force Center for Environmental Excellence
American Society of Chemical Engineers
below ground surface
chlorinated aliphatic hydrocarbon
explanation of significant difference
contaminant of concern
dichloroethene
Environmental Systems Research Institute, Inc.
granular activated carbon
gallons per minute
geographical information system
long-term monitoring
microgram(s) per liter
maximum contaminant level
Minnesota Pollution Control Agency
Minnesota Department of Health
monitoring network optimization
National Priorities List
operations and maintenance
operable unit
tetrachloroethene
practical quantitation limit
remedial action objective
Record of Decision
soil vapor extraction
trichloroethene
United States Environmental Protection Agency
vinyl chloride
volatile organic compound
-IV-
S:\ES\SHARED\CEN\MNO\EPA\LONGPRAIRIE\WRITEUP\LongPrairieMNODraftFinal.doc
-------
SECTION 1
INTRODUCTION
Groundwater monitoring programs have two primary objectives (U.S. Environmental
Protection Agency [USEPA], 1994; Gibbons, 1994):
1. Evaluate long-term temporal trends in contaminant concentrations at one or
more points within or outside of the remediation zone, as a means of
monitoring the performance of the remedial measure (temporal objective)', and
2. Evaluate the extent to which contaminant migration is occurring, particularly
if a potential exposure point for a susceptible receptor exists (spatial
objective).
The relative success of any remediation system and its components (including the
monitoring network) must be judged based on the degree to which it achieves the stated
objectives of the system. Designing an effective groundwater monitoring program
involves locating monitoring points and developing a site-specific strategy for
groundwater sampling and analysis so as to maximize the amount of relevant information
that can be obtained while minimizing incremental costs. Relevant information is that
required to effectively address the temporal and spatial objectives of monitoring. The
effectiveness of a monitoring network in achieving these two primary objectives can be
evaluated quantitatively using statistical techniques. In addition, there may be other
important considerations associated with a particular monitoring network that are most
appropriately addressed through a qualitative assessment of the network. The qualitative
evaluation may consider such factors as hydrostratigraphy, locations of potential receptor
exposure points with respect to a dissolved contaminant plume, and the direction(s) and
rate(s) of contaminant migration.
1-1
S:\ES\SHARED\CEN\MNO\EPA\LONGPRAIRIE\WRITEUP\LongPrairieMNODraftFinal.doc
-------
This report presents a description and evaluation of the groundwater monitoring
program associated with the Long Prairie Ground Water Contamination Superfund Site
(Long Prairie), Minnesota. A 44-well monitoring network was evaluated to identify
potential opportunities to streamline monitoring activities while still maintaining an
effective monitoring program. This evaluation is being conducted as part of an
independent assessment of monitoring network optimization (MNO) methods by the
USEPA and the Air Force Center for Environmental Excellence (AFCEE). A three-
tiered approach, consisting of a qualitative evaluation, an evaluation of temporal trends in
contaminant concentrations, and a statistical spatial analysis, was conducted to assess the
degree to which the monitoring network addresses each of the two primary objectives of
monitoring, and other important considerations. The results of the three evaluations were
combined and used to assess the optimal frequency of monitoring and the spatial
distribution of the components of the monitoring network. The results of the analysis
were then used to develop recommendations for optimizing the monitoring program at
Long Prairie.
1-2
S:\ES\SHARED\CEN\MNO\EPA\LONGPRAIRIE\WRITEUP\LongPrairieMNODraftFinal.doc
-------
SECTION 2
SITE BACKGROUND INFORMATION
The location, operational history, geology, and hydrogeology of Long Prairie are
briefly described in the following subsections.
2.1 SITE DESCRIPTION
The city of Long Prairie, Minnesota is a small farming community of fewer than 5,000
residents, and is located in Todd County in central Minnesota, about 120 miles northwest
of Minneapolis/St. Paul. The Long Prairie site comprises a 0.16-acre source area of
contaminated soil and an elongate plume of dissolved chlorinated aliphatic hydrocarbons
(CAHs) in the drinking-water aquifer underlying the north-central part of the city of Long
Prairie in central Minnesota. The source of the groundwater contamination was a dry-
cleaning establishment, formerly located in the city's commercial district at 243 Central
Street in the city of Long Prairie, which operated from 1949 through 1984. The
contamination resulted from the discharge of spent dry-cleaning solvents, primarily
tetrachloroethene (PCE), into the subsurface via a shallow, makeshift "french drain."
The contaminated soils served as a source of groundwater contamination, and a dissolved
CAH plume has migrated northward at least 3,600 feet from the source area, extending
beneath an older residential neighborhood and to within 500 feet of the Long Prairie
River.
The contamination was discovered in 1983, during a survey of municipal drinking-
water-supply wells for synthetic organic contaminants. PCE and other CAHs, including
trichloroethene (TCE) and czs-l,2-dichloroethene (DCE), were detected in two (CW4 and
CW5) of the five Long Prairie municipal water-supply wells, which are screened in the
lower unit of the Long Prairie Sand Plain aquifer. Eight of 21 residential wells sampled
2-1
S:\ES\SHARED\CEN\MNO\EPA\LONGPRAIRIE\WRITEUP\LongPrairieMNODraftFinal.doc
-------
also were contaminated. Because the detected concentrations of CAHs exceeded federal
maximum contaminant levels (MCLs) or other risk-based levels, the Minnesota
Department of Health (MDH) recommended that the two affected city wells be removed
from service, and issued a health advisory for a 15-block area in the northern part of the
city. The Minnesota Pollution Control Agency (MCPA) ordered that bottled water be
supplied for the 350 residents on city water or private wells in the advisory area. A new
municipal well (CW6) was installed in the deeper outwash deposits northeast of (outside)
the contaminant plume in 1984.
After enforcement activities failed to identify any viable potentially responsible parties
from among the three owners of the dry-cleaning property, a Multi-Site Cooperative
Agreement was signed on September 4, 1984 between MPCA and the US Environmental
Protection Agency (USEPA) to implement a remedial investigation and feasibility study.
Based on the results of the RI/FS, the Long Prairie Groundwater Contamination Site was
promulgated to the National Priorities List (NPL) in 1985, and a Record of Decision
(ROD) was signed in 1988.
The ROD and subsequent explanations of significant difference (ESDs) identify the
following remedial action objectives (RAOs) for groundwater:
• Provide a safe drinking-water supply for current and future users of the Long
Prairie San Plain aquifer by
- Restoring the aquifer by reducing the major contaminant (PCE) concentrations
to a health-based concentration of 5 micrograms per liter (ug/L) or less,
- Providing an alternate water supply to persons using the contaminated part of
the aquifer; and
- Reducing soil PCE concentrations to 1,200 micrograms per kilogram or less to
maintain an acceptable groundwater risk level (< 1x10"6) due to PCE leaching
from soils;
2-2
S:\ES\SHARED\CEN\MNO\EPA\LONGPRAIRIE\WRITEUP\LongPrairieMNODraftFinal.doc
-------
• Prevent the spread of contaminated groundwater to wells currently unaffected by
the CAH contamination, including municipal well CW6; and
• Prevent adverse effects on aquatic organisms in Long Prairie River due to
implementation of remedial actions by obtaining a PCE concentration of 5 ug/L or
less in treatment system effluent discharged to the river.
Pursuant to achieving these RAOs, the ROD identified three operable units (OUs).
OU1 addresses groundwater contamination through extraction of CAH-contaminated
groundwater via nine extraction wells (RWs 1 through 9), treatment of the extracted
water using granular activated carbon (GAC) filtration, and discharge of treated water to
the Long Prairie River. Installation of the pump-and-treat system, with a 250-gallon-per-
minute (gpm) treatment capacity, was completed in August 1997, and is intended to
restore aquifer quality to MCLs, and to prevent further migration and discharge of the
CAH plume to the Long Prairie River. The source area soils were addressed through soil
vapor extraction (SVE) under OU2, which operated from June 1997 through 1999, when
it was decommissioned. OU3 comprises an alternative water supply system, which
provided municipal water hookups to residents with private wells within the health-
advisory area; these hookups were completed in 1996.
The OU1 groundwater remedial system performance is monitored through quarterly to
annual sampling of a series of monitoring and city wells, and operation and maintenance
(O&M) monitoring of the extraction and treatment systems. The monitoring program is
fully described in Section 3.
2.2 GEOLOGY AND HYDROGEOLOGY
The city of Long Prairie is situated at an elevation of 1,300 feet above mean sea level
(amsl) on the eastern bank of the Long Prairie River. The sediments underlying the city
consist of a series of glacial till and outwash deposits nearly 700 feet thick, deposited in a
large valley along the Long Prairie River. Outwash sediments in the valley are composed
of coarse sands and gravels of deposited during two events, which are separated by finer
2-3
S:\ES\SHARED\CEN\MNO\EPA\LONGPRAIRIE\WRITEUP\LongPrairieMNODraftFinal.doc
-------
grained tills. The lower outwash is incised into the lower till, which forms the base of
that water-bearing unit. The lower outwash appears to extend to the east of the river, but
does not appear to be present west of the river. Above the lower outwash is a younger till
(the upper Wadena till). The surficial upper outwash unit incises into the lower outwash
unit through the Wadena till just east of the Long Prairie River, where the outwash
deposits form a single hydrogeologic unit. However, the till is intact along the eastern
side of the outwash valley, and where present, acts as an aquitard between the two
outwash units. The upper outwash unit pinches out at the eastern edge of the glacial
valley (Barr, 2002). The aquifer is recharged by inflow from upgradient lakes and
precipitation. Because of the high transmissivity of the outwash deposits, and the fact
that the river is "under-fit" in the much larger underlying glacial valley, the influence of
the Long Prairie River on groundwater flow in the site area may limited.
At the Long Prairie site, the solvent release occurred in an area in which the upper till
is present between the upper and lower outwash units. However, the saturated thickness
of the upper outwash at the source area (well MW10A) is only about 10 feet (Barr, 2002).
The till pinches out just north of the source area, and the CAH plume in groundwater is
present in both the shallow and deeper portions of the outwash deposits in the incised
channel west of the western edge of the upper till. However, because the upper and lower
outwash units are in direct hydraulic communication where the confining till is absent,
there is a pathway for contaminant migration to be drawn into the city wells screened in
the lower outwash unit to the east. The upper till is present east of the longitudinal axis
of the north/south-trending CAH plume; municipal wells CW3 and CW6 are screened in
the lower outwash deposits below the upper till.
Where the upper till unit is absent (along the incised upper outwash channel),
groundwater in the outwash aquifer occurs under water-table (unconfined) conditions.
Groundwater flow directions in the upper and lower outwash deposits generally parallel
the channel of the Long Prairie River; from the PCE source area, groundwater flows
northeast to the vicinity of extraction wells RW5 and RW7, then flows west-northwest
toward the Long Prairie River (Figure 2.1). In September 2001, the water table was
2-4
S:\ES\SHARED\CEN\MNO\EPA\LONGPRAIRIE\WRITEUP\LongPrairieMNODraftFinal.doc
-------
Q.
O
O c
O o
2
"o
CD
I
O)
g
'i_
2
'E
o
CD
O
O
CD
a:
O
§
o
O
O
O
O
"co
•5
LJJ o
O o
Q. CM
Q§
CM
o
>c::>
CM
O
^L
CM
O
s- O
a> CM
!i!
1 CD «
0)"^ 0
2§l
3 c C
•£ O o
SB °
T3 E^ ?
£=5-5.
jjj 5 ^
c ° 5
— !^ Q.
CD
g
-|— (
CO
•a
*^ *^
O) ~^
c "CD
CO ^
c o
o c
'ro O
O)
© e
o
O
CO
cxi
CO
~Z.
O
<
^^
5
CD o
J- 0
O
O
LO
O
CM
O
f
^.
"CD
o
o
LO
J/5
CO
cr
CD
o
CM
LJJ
tt
D
g
u_
_,
^J
LJJ
LJJ
o:
LJJ
PRAIRI
o
z
o
CM
o
o
CM
0
o
Q
^
o
0
o
~"
C/J
^y
o
1-
o:
1-
LJJ
o
z
o
o
o
o
o
o
"co
N
E
'•s.
o
o
"CD
z
O)
2
o
COS
oj
CDS
cs
cs
.921
'l!
QJ
il
o;
S
n
z
Q
Iff
a
i
c
5
e
-------
measured at elevations ranging from about 1,287 feet amsl near the source area at the
south end of the CAH plume, to 1,282 feet amsl at the northern end of the plume
(MW15A) (Barr, 2002). The average hydraulic gradient along the plume flowpath (i.e.,
between MW10A in the source area and MW16A near the plume toe) in the upper
outwash aquifer in September 2001 was approximately 0.0012 foot per foot. Using this
gradient, the calibrated hydraulic conductivity value of 426 feet per day (ft/day) used in
the MODFLOW model constructed for the site (Bangsund, 2003), and an estimated
effective porosity for sand and gravel of 0.30, the average advective groundwater flow
velocity in the upper outwash aquifer is estimated to be 1.7 ft/day. The vertical gradients
in the incised channel deposits are negligible, but appear to be slightly downward at the
northern extent of the CAH plume, suggesting that the plume may not directly threaten
wetlands in that area (Barr, 2002).
Where the upper till is present, groundwater in the lower outwash deposits is under
confined to semi-confined conditions, with a generally northwesterly groundwater flow
direction. Hydraulic properties of the lower glacial outwash deposits are inferred to be
comparable to those of the upper outwash deposits. Groundwater flow directions are
influenced locally by pumping of the city water-supply wells, as well as by operation of
the OU1 extraction wells.
2.3 NATURE AND EXTENT OF CONTAMINATION
The source of contamination at the Long Prairie site was discharge of dry-cleaning
solvents directly into glacial outwash deposits at the site of the former dry cleaning
establishment. Available records indicate that 2,200 gallons of PCE were used during the
period of operation from 1949 through 1984. While PCE was the primary solvent
disposed of, trace amounts of chlorinated ethanes also have been detected. The waste
solvents percolated through the coarse outwash soils at the source to the water table in the
Long Prairie Sand Plain aquifer, and subsequently migrated as dissolved constituents in
groundwater. PCE and its daughter products TCE and cis-1,2-DCE have been detected in
a plume about 1,000 feet wide and up to 3,600 feet long. In October 2002, the CAH
plume extended from the source area, near the inactive RW1A/1B/1C extraction well
2-6
S:\ES\SHARED\CEN\MNO\EPA\LONGPRAIRIE\WRITEUP\LongPrairieMNODraftFinal.doc
-------
cluster, approximately 3,200 feet downgradient to the northwest, to nested monitoring
well pair MW18A/B (Figure 2.1).
Contamination has been detected throughout the saturated thickness of the upper
glacial outwash deposits, and also historically has been detected in the lower outwash
deposits beneath the upper till at city well CW3 (Figure 2.1). Maximum historical
concentrations of PCE in groundwater were as high as 150,000 ug/L. Recent monitoring
data indicate that maximum PCE concentrations have decreased to around 100 ug/L in
the core of the plume, and that PCE is no longer present at detectable concentrations in
the lower outwash deposits east of the incised channel. However, contamination persists
throughout the saturated upper outwash deposits within the incised channel (along the
centerline of the plume), and the overall extent of the plume, as defined by the 5-ug/L
isolpleth for PCE, has not changed significantly since pumping began (Barr, 2002). In
October 2002, PCE concentrations in the plume ranged from 2.4 ug/L at the northern end
of the plume (well MW18B) to 110 ug/L near the center of the plume, at well MW14B
(Figure 2.1). In the shallow portion of the upper outwash deposits in the source area,
PCE was detected at 38 ug/L at well MW10A.
2.4 REMEDIAL SYSTEMS
As discussed in Section 2.1, an SVE system (OU2) was installed to remove PCE from
vadose-zone soils in the source area. The SVE system was pilot tested and completed in
June 1997, and operated continuously through the end of 1999, when it was disassembled
due to the low magnitude of extracted PCE concentrations.
The OU1 groundwater extraction system consists of 10 recovery wells, 6 of which are
currently operational. The RAOs for the OU1 groundwater extraction and treatment
system include restoring aquifer quality to MCLs, and preventing further migration and
discharge of the CAH plume to the Long Prairie River (Section 2.1). Initially, seven
recovery wells (RW1A/1B/1C/3/4/6/7) were installed, and closed municipal well CW5
was retrofitted to become RW5. This eight-well extraction system began operating in
May 1996. Two additional recovery wells (RW8 and 9) were installed in September
2-7
S:\ES\SHARED\CEN\MNO\EPA\LONGPRAIRIE\WRITEUP\LongPrairieMNODraftFinal.doc
-------
1999. The recovery wells are distributed along the axis of the plume, and the system is
designed to recover and treat up to 250 gpm of groundwater. According to the First Five-
Year Review Report (MPCA, 2002), operation of 4 of the 10 recovery wells (RW1A,
RW1B, RW1C, and RW4) was discontinued starting in 2000 to allow for higher pumping
rates at wells closer to the center of the plume. Recovered groundwater is discharged to
the Long Prairie River following treatment with GAC.
2-8
S:\ES\SHARED\CEN\MNO\EPA\LONGPRAIRIE\WRITEUP\LongPrairieMNODraftFinal.doc
-------
SECTION 3
LONG-TERM MONITORING PROGRAM AT LONG PRAIRIE
The current groundwater monitoring program at Long Prairie and the 44 wells
sampled during the comprehensive October 2002 monitoring event were examined to
identify potential opportunities for streamlining monitoring activities while still
maintaining an effective performance and compliance monitoring program. The 2002
and the current (2000/2001) monitoring programs at Long Prairie are reviewed in the
following subsections.
3.1 DESCRIPTION OF MONITORING PROGRAM
The purposes of monitoring at Long Prairie are to monitor progress toward achieving
the RAOs set forth in the ROD and ESDs, and to gather adequate information to
determine the effectiveness of the groundwater recovery and treatment system. Wells are
classified as recovery (extraction) wells, monitoring wells, and city wells. The
monitoring program is designed to:
« Evaluate the effectiveness of the groundwater extraction system on controlling the
plume and improving regional groundwater quality
. Confirm protection of the city of Long Prairie water-supply wells. (Barr, 2002)
The October 2002 sampling event included 44 wells in the Long Prairie region.
Several of the monitoring wells are installed as clusters at a single location, and screened
at different depths. The "A" wells are screened across the water table, the "B" wells are
screened at the based of the upper glacial outwash deposits, and the "C" wells are
screened in the lower outwash deposits. Typically, about half of the wells sampled
during the most recent monitoring event are routinely sampled as part of the groundwater
3-1
S:\ES\SHARED\CEN\MNO\EPA\LONGPRAIRIE\WRITEUP\LongPrairieMNODraftFinal.doc
-------
monitoring program; for example, in the 2001 and 2000 sampling rounds, a subset of 26
of the 44 wells was sampled. This subset included quarterly sampling of the 6 active
extraction wells (RW3 through RW9, excluding RW4) and city well CW3, and annual
sampling of 19 monitoring wells (including RW4). In 2002, city well CW6 was added to
the quarterly sampling schedule. In the second quarter of 2000, the suite of volatile
organic compounds (VOCs) for which groundwater samples were analyzed was reduced
from the MDH 465E list to the COCs: DCE, PCE, TCE, and vinyl chloride (VC).
Additionally, a gas chromatograph (GC) analytical method (assumed to be USEPA
Method 802IB) is now used instead of the gas chromatograph/mass spectrometer
(GC/MS) method (assumed to be USEPA Method SW8260B) formerly required.
The locations of 44 Long Prairie wells sampled in October 2002 are shown in relation
to the COC plume on Figure 2.1. Table 3.1 lists these wells with their screened intervals,
well type, and designation. The "current" sampling plan includes the 18 monitoring
wells and 7 recovery wells sampled during scheduled monitoring events in 2000 and
2001, as well as city wells CW3 and CW6. The three-tiered MNO evaluation in
described in this report examines the 44 wells sampled during October 2002, and
compared the recommended optimized well network to the 27-well network included in
the "current" monitoring program.
3.2 SUMMARY OF ANALYTICAL DATA
The Long Prairie groundwater monitoring program was evaluated using results for
sampling events performed from May 1996 through October 2002. These analytical data
were provided to Parsons by Ms. Jonelle Branca, the Data Management Coordinator for
Barr Engineering (MPCA's environmental contractor). The database was processed to
remove duplicate data by retaining the maximum result for each duplicate sample pair.
The analytical database provided to Parsons contained from 1 to 29 sampling results for
each constituent for each of the 44 wells in the Long Prairie site area. As discussed in
Section 2.3, the primary COCs identified for the Long Prairie plume are PCE, TCE, and
cis-l,2-DCE, therefore, the MNO evaluation focused on these constituents.
3-2
S:\ES\SHARED\CEN\MNO\EPA\LONGPRAIRIE\WRITEUP\LongPrairieMNODraftFinal.doc
-------
TABLE 3.1
GROUNDWATER MONITORING PROGRAM
THREE-TIERED MONITORING NETWORK OPTIMIZATION
LONG PRAIRIE, MINNESOTA
Well ID
Screened Interval
(ft bgs) *
Sampling
Frequency
Monitoring Wells in 2000/2001 Sampling Plan
MW2A
MW2B
MW2C
MW4B
MW4C
MW6A
MW6B
MW6C
MW10A
MW11B
MW11C
MW14B
MW14C
MW15A
MW15B
MW16B
MW17B
MW19B
15-20
31-35
48-53
31.5-35.5
42.0-46.0
12.0-17
31.0-37
46.0-59
16-21
50-55
21-26
50-55
12
33
25.3
33
26.5
City Water-Supply Wells
CW3
CW6
67-85
53-76
Recovery Wells
RW1A
RW1B
RW1C
RW3
RW4
RW5
RW6
RW7
RW8
RW9
15-30
13-45
12.5-23.5
17-52
10.0-50
41-56
10.0-55.5
15-45
30-40
25-35
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Quarterly
Quarterly
NA"'
NA
NA
Quarterly
Annual
Quarterly
Quarterly
Quarterly
Quarterly
Quarterly
Other Area Wells Sampled in October 2002
BAL2B
BAL2C
MW1A
MW1B
MW3A
MW3B
MW4A
MW5A
MW5B
MW11A
MW13C
MW16A
MW18A
MW18B
57-65
40-50
9.5-14.5
30-35
17.8-22.8
30-35
10.0-15
4.5-9.5
31.0-35
50-55
15.5
15.1
35
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
ft bgs = feet below ground surface
Reduced sampling frequency established prior to
022/742479/LongPraiTieTablesPinal.xls/Table3.1
3-3
-------
Table 3.2 presents a summary of the occurrence of the three primary COCs in Long
Prairie groundwater based on the data collected from 33 site and area monitoring wells
during the period from May 1996 through October 2002. The data summarized in Table
3.2 exclude results for the recovery wells (with the exception of inactive extraction well
RW4, which is sampled annually as a monitoring well) and city wells CW3 and CW6.
As indicated in this table and discussed in Section 2.3, PCE is the primary COC based on
its broad distribution at concentrations exceeding its MCL of 5 |u.g/L. PCE has been
detected in approximately 39 percent of groundwater samples, and has exceeded its MCL
in approximately 33 percent of the samples. PCE has been detected in 20 of the 33
monitoring wells in the Long Prairie site area, and has exceeded the MCL at 14 of these
wells. One of PCE's reductive-dechlorination daughter products, cis-l,2-DCE, is another
prevalent compound on site, and has been detected in 44 percent of the collected samples.
However, detected concentrations of cis-l,2-DCE have exceeded the MCL of 70 ug/L in
only about 1 percent of samples. TCE, another PCE daughter product, has been detected
at Long Prairie in approximately 35 percent of samples. Detected concentrations of TCE
have exceeded its MCLs of 5 |ug/L in approximately 21 percent of the samples.
PCE sampling results were the primary data used to conduct the qualitative and spatial
components of the three-tiered MNO evaluation due to the magnitude and spatial extent
of PCE concentrations in groundwater at Long Prairie compared to the other detected
compounds.
3-4
S:\ES\SHARED\CEN\MNO\EPA\LONGPRAIRIE\WRITEUP\LongPrairieMNODraftFinal.doc
-------
o
z
£3
o
H
Z
o
S
z
^^
u
z
o ^
u o
to H
O ^
H §
t-y P-H
5 g
Z o ^
TABLE 3.2
ROUNDWATER CONTAMI
WELLS
RED MONITORING NETWORK
LONG PRAIRIE, MINNESOT,
O w
fe *r
O w
W C£
U M
Z H
|
U
u
o
to
o
J*
5
S
S
^
cc
!/>
^5 JS °
o i3 u
« '? J §
A _K U -B
° "3 S "
S jj o
° ^ S
& *^
»« S<5 "~T
E S a>
5 !*• "
Z E? Q
~ «
° ? <*
« ^ 5
* — *
s 5 rt
Z ^
«g «,
0 4,
^ O
BJO j B
| U |
CJ CJ
aJ •*!
U "5fc
u "S 2s
bD Q> C^
B "g ^fc
^>^ ^-
rt ° ^
cs "S
"S w
§ "S
u Q
«s =
R ^
"S 21
cc
IH
"S
S
CS
IH
CS
"
2
O
y3 OJ
S-H
d
"S s
11
is] G
Ji* x
§ S
1 1
•1 ^
O ||
~^>
"§ o
< ^s
S
_OJ
I
-------
SECTION 4
QUALITATIVE MNO EVALUATION
An effective groundwater monitoring program will provide information regarding
contaminant plume migration and changes in chemical concentrations through time at
appropriate locations, enabling decision-makers to verify that contaminants are not
endangering potential receptors, and that remediation is occurring at rates sufficient to
achieve RAOs within a reasonable time frame. The design of the monitoring program
should therefore include consideration of existing receptor exposure pathways, as well as
exposure pathways arising from potential future use of the groundwater.
Performance monitoring wells located upgradient, within, and immediately
downgradient from a plume provide a means of evaluating the effectiveness of a
groundwater remedy relative to performance criteria. Long-term monitoring (LTM) of
these wells also provides information about migration of the plume and temporal trends
in chemical concentrations. Groundwater monitoring wells located downgradient from
the leading edge of a plume (i.e., sentry wells) are used to evaluate possible changes in
the extent of the plume and, if warranted, to trigger a contingency response action if
contaminants are detected.
Primary factors to consider when developing a groundwater monitoring program
include at a minimum:
• Aquifer heterogeneity,
• Types of contaminants,
• Distance to potential receptor exposure points,
4-1
S:\ES\SHARED\CEN\MNO\EPA\LONGPRAIRIE\WRITEUP\LongPrairieMNODraftFinal.doc
-------
• Groundwater seepage velocity and flow direction(s),
• Potential surface-water impacts, and
• The effects of the remediation system.
These factors will influence the locations and spacing of monitoring points and the
sampling frequency. Typically, the greater the seepage velocity and the shorter the
distance to receptor exposure points, the more frequently groundwater sampling should
be conducted.
One of the most important purposes of LTM is to confirm that the contaminant plume
is behaving as predicted. Graphical and statistical tests can be used to evaluate plume
stability. If a groundwater remediation system or strategy is effective, then over the long
term, groundwater-monitoring data should demonstrate a clear and meaningful
decreasing trend in concentrations at appropriate monitoring points. The current
groundwater monitoring program at Long Prairie was evaluated to identify potential
opportunities to LTM optimization.
4.1 METHODOLOGY FOR QUALITATIVE EVALUATION OF
MONITORING NETWORK
The three-tiered MNO evaluation of the Long Prairie groundwater LTM program
considered information for the 44 wells sampled during the October 2002 sampling
round. These wells, their respective screened intervals, and their current monitoring
frequencies are listed in Table 3.1, and their locations are depicted on Figure 2.1.
Multiple factors were considered in developing recommendations for continuation or
cessation of groundwater monitoring at each well. In some cases, a recommendation was
made to continue monitoring a particular well, but at a reduced frequency. A
recommendation to discontinue monitoring at a particular well based on the information
reviewed does not necessarily constitute a recommendation to physically abandon the
well. A change in site conditions might warrant resumption of monitoring at some time
in the future at wells that are not currently recommended for continued sampling.
4-2
S:\ES\SHARED\CEN\MNO\EPA\LONGPRAIRIE\WRITEUP\LongPrairieMNODraftFinal.doc
-------
Typical factors considered in developing recommendations to retain a well in, or remove
a well from, the monitoring program are summarized in Table 4.1. Typical factors
considered in developing recommendations for monitoring frequency are summarized in
Table 4.2.
TABLE 4.1
MONITORING NETWORK OPTIMIZATION DECISION LOGIC
THREE-TIERED MONITORING NETWORK OPTIMIZATION
LONG PRAIRIE, MINNESOTA
Reasons for Retaining a Well in
Monitoring Network
Well is needed to further characterize the
site or monitor changes in contaminant
concentrations through time
Well is important for defining the lateral or
vertical extent of contaminants
Well is needed to monitor water quality at
compliance point or receptor exposure
point (e.g., domestic well)
Well is important for defining background
water quality
Reasons for Removing a Well From
Monitoring Network
Well provides spatially redundant
information with a neighboring well (e.g.,
same constituents, and/or short distance
between wells)
Well has been dry for more than 2 yearsa/
Contaminant concentrations are
consistently below laboratory detection
limits or cleanup goals
Well is completed in same water-bearing
zone as nearby well(s)
a/ Water-level measurements in dry wells should continue,
becomes re-wetted.
and groundwater sampling should be resumed if the well
4.2 RESULTS OF QUALITATIVE MNO EVALUATION
The results of the qualitative evaluation of the 44 wells in the Long Prairie plume
vicinity are described in this subsection. Recommendations for optimizing the well
network are developed, and the basis for each recommendation is provided.
4.2.1 Monitoring Network and Sampling Frequency
The results of the qualitative evaluation of the 18 monitoring wells, 7 extraction wells,
and 2 municipal water-supply wells currently included in the LTM program at the Long
Prairie Superfund Site are included are summarized in Table 4.3, and described in the
following subsections. Other site wells that are not currently monitored on a regular
4-3
S:\ES\SHARED\CEN\MNO\EPA\LONGPRAIRIE\WRITEUP\LongPrairieMNODraftFinal.doc
-------
basis (i.e., 3 extraction wells and 14 monitoring wells) also are included in Table 4.3 for
completeness. The table includes recommendations for retaining or removing each well,
and for changing the sampling frequency, and lists the rationale for the recommendations.
TABLE 4.2
MONITORING FREQUENCY DECISION LOGIC
THREE-TIERED MONITORING NETWORK OPTIMIZATION
LONG PRAIRIE, MINNESOTA
Reasons for Increasing
Sampling Frequency
Groundwater velocity is high
Change in contaminant concentration
would significantly alter a decision or
course of action
Well is close to source area or operating
remedial system
Cannot predict if concentrations will
change significantly over time
Reasons for Decreasing
Sampling Frequency
Groundwater velocity is low
Change in contaminant concentration
would not significantly alter a decision or
course of action
Well is distal from source area or remedial
system
Concentrations are not expected to change
significantly over time, or contaminant
levels have been below groundwater
cleanup objectives for some prescribed
period of time
4.2.1.1
Extraction Wells
Six of the 10 groundwater extraction wells at the site (RW3 and RW5 through RW9)
are currently operating and are sampled quarterly. A seventh inactive extraction well
(RW4) is sampled annually, and the remaining three extraction wells (source-area wells
RW1A/1B/1C) are not sampled.
Continued sampling of the six active extraction wells is recommended to facilitate
periodic calculation of contaminant mass-removal rates and assessment of remedial
progress and system optimization needs. However, historical sampling data for these
wells indicate that temporal concentration trends could be adequately determined from
semiannual monitoring. Therefore, reduction in the sampling frequency for these wells
from quarterly to semiannually is recommended. This frequency also should be adequate
4-4
S:\ES\SHARED\CEN\MNO\EPA\LONGPRAIRIE\WRITEUP\LongPrairieMNODraftFinal.doc
-------
III
«ll
§ O a
«*%
11
P
1
•a
a
o
a
"i § a
.-g S, g
I 5 a
j Ifi 1
£ «
13
•I Is
3
>• "*
rt W
2
e* p.
f I
a 1
& aj
fa
"S -3 V
II 1
O H itj
g
1
a
a
01 Samp
1
a
1
[Monitoring
S
H
g
g
1
'3
pq
5
1
1
1
No temporal
S
<
o
M
g
3
'3
bC
g
1
p
O
.1
fc!
1
4
Monitors rela
1
s
-<
s
eq
S
H
g
g
1
'3
pq
5
id
1
1
1
No temporal
:
S
<
m
O
ition zones.
r relatively high-conce
11 MW4C monitc
S
id
g
a
1
S3
g
1
1
g
P
Monitors plui
1
s
<
m
1
ition zones.
r relatively high-conce
11 MW4B monitc
£
1
%
'S
1
S3
g
1
1
g
P
Monitors plui
1
S
<
42.0-46.0
1
"wo wells.
%
•B
P.
g
1 information tha
1
ith MW6B and MW6C; does not provide any
id
1
3
-<
O
1
S3
1
id
.23
a
p
fc!
§
orical COC cor
-S3
which has had highest
p
o
-S
1
ermec
p
g
3
'3
bC
a
p
.2
g
.S
1
i
:
S
-<
O
pq
P
g
3
'3
bC
a
p
.2
g
.S
1
i
1
S
•«
o
o
1
1
p
1
1
Monitors gro
1
s
-<
1
VOIAH|
d
1
a
5
S
en interval of
£
ward CW6. Note that
S
a
p
•a
1
•S
i
1
s
1
Provides con
1
s
<
m
wardCW6.
S
a
p
•a
1
a
i
a
s
1
Provides con
1
s
<
o
O
1
S3
g
1
1
g
P
Monitors plui
1
s
^
CN
|MW14B
on zone.
§
i
s
§
5
1
g
p
-fc!
I
§
p
.2
s
I
:
S
-<
o
|MW14C
S3
1
1
o
S
1
~a
s
Monitors dov
1
S
^
2
|
S3
1
1
o
S
1
~a
s
Monitors dov
1
S
^
m
pq
cis-l;2-DCE.
reasing trend for
.S
o
p
1
S3
g
1
1
g
P
Monitors plui
1
s
<
s
§
1
S3
g
1
1
g
P
Monitors plui
1
s
^
m
IMWIVB
01
£
"p.
8
•35
1
B
f MCL exceed
bx
.a
a
g
s
pfi
'§
ffi
0
1
f
P
1
I
a
a
43
Monitors con
'3
fi
m
S
<
S
|MW19B
-Supply Wells
| City Water
relations.
S
cal sensitivity/commur
i
1
p
i
litoring justified due to proximity of CAH pi
"S
1
a
S3
3
a
§
relations.
,-fi1
cal sensitivity/commul
i
1
p
i
litoring justified due to proximity of CAH pi
"S
1
a
S3
3
a
S
1
| Recovery V
C7
; area groundwat
1
1
o
59
o
s
bO
.S
0
Continued m
!
o
m
1
C7
; area groundwat
1
1
o
59
o
s
bO
.S
0
Continued m
!
"t
pq
1
1
; area groundwat
1
1
o
59
o
s
bO
.S
0
Continued m
!
m
O
1
bo
S
fi
a
annua
a
ely detected w
T
•a
1
1
£
bij
i
p
id
id
g
a
u
a
o
fc
13
4»
*
S3
a
K
1
8
6
f
1
°
8
g
•s
s
o
o
ng justified by la
S
a
1
1
-------
to achieve the extraction-system performance monitoring objectives stated above. In
addition, reduction in the sampling frequency for inoperative well RW4 from annual to
biennial (every other year) is recommended. This well was sampled 18 times from May
1996 through October 2002. Trace-level (i.e., almost exclusively < 3 ug/L)
concentrations of COCs were detected until late 1997; since then, COCs have for the
most part not been detected. Given this monitoring history, relatively infrequent
monitoring of RW4 is sufficient to define the southwestern boundary of the CAH plume
over time in this area.
Extraction wells RW1A/1B/1C are located in the source area and provide groundwater
quality data for the depth interval ranging from about 10 to 43 feet below the ground
surface. Wells RW1A and IB were sampled regularly from May 1996 through October
1999, and were sampled again in October 2002. Available data indicate that RW1C was
only sampled once (in October 2002). Since May 1996, COCs either have not been
detected in these wells, or have been present at trace concentrations (i.e., < 3 ug/L).
Given that monitoring well MW10A is located immediately adjacent to these extraction
wells, and is screened near the water table, where any fresh influx of contaminants from
the vadose zone would first be detected, continued monitoring of this well should be
sufficient to monitor groundwater quality in the source area. This well has historically
had higher COC concentrations than the adjacent extraction wells.
4.2.1.2 Municipal Water-Supply Wells
Two municipal water-supply wells, CW3 and CW6, currently are sampled on a
quarterly basis. Available data indicate that the CAH plume is not migrating eastward in
the lower outwash deposits toward these wells, and the only COCs detected in samples
from these wells have been very low-magnitude (< 1 ug/L) detections of cis-l,2-DCE in
CW3. This observation is supported by the prevailing groundwater flow direction to the
northwest, and by the lack of significant historical contaminant concentrations at other
lower-outwash wells (i.e., MW3B, MW11C, and BAL2C) along the eastern margin of the
COC plume (Figure 2.1). Therefore, from a purely technical standpoint, quarterly
sampling of CW3 and CW6 is probably not necessary. However, continuation of
4-6
S:\ES\SHARED\CEN\MNO\EPA\LONGPRAIRIE\WRITEUP\LongPrairieMNODraftFinal.doc
-------
"goodwill" monitoring at this frequency is recommended to reassure the public about the
safety of the municipal water supply (Table 4.3).
4.2.1.3 Monitoring Wells
Continued sampling of 14 of the 18 monitoring wells included in the current LTM
program is recommended. Wells MW2A, MW2C, MW6A, and MW14C are
recommended for removal from the LTM program. As indicated in Table 4.3, the aquifer
zones monitored by MW2A and MW2C have consistently contained lower COC
concentrations than the zone monitored by clustered well MW2B. Although this is
important information to obtain for initial plume characterization purposes, continued
monitoring of only the highest-concentration zone (well MW2B) is adequate to track the
magnitude of contaminant concentrations within the plume over time. Similarly,
continued monitoring of well MW14B should be adequate to track the plume magnitude
at the location of the 14B/14C well cluster over time, given the consistently low (to non-
detect) COC concentrations at MW14C.
During the first few sampling events for well cluster MW6A/6B/6C, the highest
concentrations of PCE, TCE, and cis-l,2-DCE were detected in groundwater from the
intermediate-depth well (MW6B). More recently, however, COC concentrations at each
of the three wells in this cluster have been similar, suggesting that continued sampling of
each well is not necessary. Continued sampling of MW6C is recommended because
recent samples from this well have contained the highest PCE concentrations. Continued
sampling of MW6B also is recommended because this well is screened in the vertical
interval that has historically contained the most elevated contaminant concentrations.
Removal of MW6A from the LTM program is recommended because the data provided
by sampling of this well does not provide any additional useful information that cannot
be obtained from other shallow wells downgradient from the MW6 cluster (Figure 2.1)..
Less-frequent monitoring is recommended for one monitoring well currently sampled
on an annual basis (MW19B). Retention of this well in the LTM program is
recommended due to its location in the wetland area adjacent to the Long Prairie River (a
4-7
S:\ES\SHARED\CEN\MNO\EPA\LONGPRAIRIE\WRITEUP\LongPrairieMNODraftFinal.doc
-------
potentially sensitive ecological receptor exposure area). However, the recommendation
to reduce the monitoring frequency is based on the low magnitude and stable nature of
COC concentrations detected to date. There is no reason to assume that COC
concentrations at this location could increase significantly unless the current pumping
regime is interrupted or altered.
Low-frequency (e.g., biennial) monitoring of two wells that currently are not included
in the LTM program is recommended. Samples from MW13C, which is located adjacent
to a potentially sensitive wetland area, have exhibited an increasing trend for cis-1,2-
DCE. If continued infrequent monitoring of this well does not indicate a continuation of
this trend, then monitoring should be discontinued. Sampling of well MW18B also
enables periodic evaluation of COC concentrations at the edge of the wetland area. In
addition, the location of this well near the downgradient plume toe facilitates continued
evaluation of plume dynamics (i.e., expanding, receding, steady-state).
4.2.2 Laboratory Analytical Program
The 2000/2001 Annual Report (Barr, 2002) indicates that the target analyte list (TAL)
was reduced to DCE, PCE, TCE, and VC in the second quarter of 2000, and that a
relatively inexpensive GC method that provides low detection limits is now used.
Parsons assumes that the presence of other VOCs at concentrations of potential concern
was ruled out based on previous analytical results obtained for a more extensive TAL
(MDH 465E list). If this is the case, then the current laboratory analytical program
appears to be reasonably optimized, and no further recommendations are provided. If
this is not the case, then the potential presence of other VOCs that could have been
present in the PCE as contaminants should be assessed.
4.2.3 LTM Program Flexibility
The LTM program recommendations summarized in Table 4.3 are based on available
data regarding current (and expected future) site conditions. Changing site conditions
(e.g., lengthy malfunction or significant adjustment of the groundwater extraction
system) could affect plume behavior. Therefore, the LTM program should be reviewed if
4-8
S:\ES\SHARED\CEN\MNO\EPA\LONGPRAIRIE\WRITEUP\LongPrairieMNODraftFinal.doc
-------
hydraulic conditions change significantly, and revised as necessary to adequately track
changes in plume magnitude and extent over time.
4-9
S:\ES\SHARED\CEN\MNO\EPA\LONGPRAIRIE\WRITEUP\LongPrairieMNODraftFinal.doc
-------
SECTION 5
TEMPORAL STATISTICAL EVALUATION
Chemical concentrations measured at different points in time (temporal data) can be
examined graphically, or using statistical tests, to evaluate dissolved-contaminant plume
stability. If removal of chemical mass is occurring in the subsurface as a consequence of
attenuation processes or operation of a remediation system, mass removal will be
apparent as a decrease in chemical concentrations through time at a particular sampling
location, as a decrease in chemical concentrations with increasing distance from chemical
source areas, and/or as a change in the suite of chemicals detected through time or with
increasing migration distance.
5.1 METHODOLOGY FOR TEMPORAL TREND ANALYSIS OF
CONTAMINANT CONCENTRATIONS
Temporal chemical-concentration data can be evaluated for trends by plotting
contaminant concentrations through time for individual monitoring wells (Figure 5.1), or
by plotting contaminant concentrations versus downgradient distance from the
contaminant source for several wells along the groundwater fiowpath, over several
monitoring events. Plotting temporal concentration data is recommended for any analysis
of plume stability (Wiedemeier and Haas, 2000); however, visual identification of trends
in plotted data may be a subjective process, particularly if (as is likely) the concentration
data do not exhibit a uniform trend, but are variable through time (Figure 5.2).
5-1
S:\ES\SHARED\CEN\MNO\EPA\LONGPRAIRIE\WRITEUP\LongPrairieMNODraftFinal.doc
-------
FIGURE 5.1
PCE CONCENTRATIONS THROUGH TIME
AT WELL MW6B
THREE-TIERED MONITORING NETWORK OPTIMIZATION
LONG PRAIRIE, MINNESOTA
PCE Concentration (^g/L)
•nnn -,
innn -
Qnn _
finn
DUU
/inn -
"inn -
A
0
* 0
Dec-96 Dec-97 Dec-98 Dec-99 Dec-00 Dec-01
Date
The possibility of arriving at incorrect conclusions regarding plume stability on the
basis of visual examination of temporal concentration data can be reduced by examining
temporal trends in chemical concentrations using various statistical procedures, including
regression analyses and the Mann-Kendall test for trends. The Mann-Kendall
nonparametric test (Gibbons, 1994) is well-suited for evaluation of environmental data
because the sample size can be small (as few as four data points), no assumptions are
made regarding the underlying statistical distribution of the data, and the test can be
adapted to account for seasonal variations in the data. The Mann-Kendall test statistic
can be calculated at a specified level of confidence to evaluate whether a statistically
significant temporal trend is exhibited by contaminant concentrations detected through
time in samples from an individual well. If a trend is identified, a nonparametric slope of
the trend line (change in concentration per unit time) also can be estimated using the test
5-2
S:\ES\SHARED\CEN\MNO\EPA\LONGPRAIRIE\WRITEUP\LongPrairieMNODraftFinal.doc
-------
Trend
Increasing Trend
No Trend
Confidence Factor
HIGH
Confidence Factor
LOW
Variation
LOW
Variation
HIGH
5.2
OF
IN
Monitoring Network Optimization
Long Prairie, Minnesota
draw\739732\diffusion\williamsA.cdr pg1 nap 4/3/02
-------
procedure. A negative slope (indicating decreasing contaminant concentrations through
time) or a positive slope (increasing concentrations through time) provides statistical
confirmation of temporal trends that may have been identified visually from plotted data
(Figure 5.2).
The relative value of information obtained from periodic monitoring at a particular
monitoring well can be evaluated by considering the location of the well with respect to
the dissolved contaminant plume and potential receptor exposure points, and the presence
or absence of temporal trends in contaminant concentrations in samples collected from
the well. The degree to which the amount and quality of information that can be obtained
at a particular monitoring point serve the two primary (i.e., temporal and spatial)
objectives of monitoring must be considered in this evaluation. For example, the
continued non-detection of a target contaminant in groundwater at a particular monitoring
location provides no information about temporal trends in contaminant concentrations at
that location, or about the extent to which contaminant migration is occurring, unless the
monitoring location lies along a groundwater fiowpath between a contaminant source and
a potential receptor exposure point. Therefore, a monitoring well having a history of
contaminant concentrations below detection limits may be providing little or no useful
information, depending on its location.
A trend of increasing contaminant concentrations in groundwater at a location between
a contaminant source and a potential receptor exposure point may represent information
critical in evaluating whether contaminants are migrating to the exposure point, thereby
completing an exposure pathway. Identification of a trend of decreasing contaminant
concentrations at the same location may be useful in evaluating decreases in the areal
extent of dissolved contaminants, but does not represent information that is critical to the
protection of a potential receptor. Similarly, a trend of decreasing contaminant
concentrations in groundwater near a contaminant source may represent important
information regarding the progress of remediation near, and downgradient from the
source, while identification of a trend of increasing contaminant concentrations at the
same location does not provide as much useful information regarding contaminant
5-4
S:\ES\SHARED\CEN\MNO\EPA\LONGPRAIRIE\WRITEUP\LongPrairieMNODraftFinal.doc
-------
conditions. By contrast, the absence of a temporal trend in contaminant concentrations at
a particular location within or downgradient from a plume indicates that virtually no
additional information can be obtained by continued monitoring of groundwater at that
location, in that the results of continued monitoring through time are likely to fall within
the historic range of concentrations that have already been detected (Figure 5.3).
Continued monitoring at locations where no temporal trend in contaminant
concentrations is present serves merely to confirm the results of previous monitoring
activities at that location. The relative amounts of information generated by the results of
temporal-trend evaluation at monitoring points near, upgradient from, and downgradient
from contaminant sources are presented schematically as follow:
Monitoring Point Near Contaminant Source
Relatively less information
Nondetect or no trend
Increasing trend in concentrations
Relatively more information
Decreasing trend in concentrations
Monitoring Point Upgradient from Contaminant Source
Relatively less information
Nondetect or no trend
Relatively more information
Decreasing trend in concentrations
Increasing trend in concentrations
5-5
S:\ES\SHARED\CEN\MNO\EPA\LONGPRAIRIE\WRITEUP\LongPrairieMNODraftFinal.doc
-------
c
_o
*•>
co
o
c
o
O
Likely Future
Results
• Historic Results-
Time
5.3
OF AT
NO
IN
IS
Monitoring Network Optimization
Long Prairie, Minnesota
draw\739732\diffusion\williamsA.cdr pg2 nap 4/3/02
-------
Monitoring Point Downgradient from Contaminant Source
Relatively less information Decreasing trend in concentrations
Nondetect or no trend
Relatively more information Increasing trend in concentrations
5.2 TEMPORAL EVALUATION RESULTS
The analytical data for groundwater samples collected from the 44 wells in Long
Prairie LTM program from May 1996 through October 2002 were examined for temporal
trends using the Mann-Kendall test. The objective of the evaluation was to identify those
wells having increasing or decreasing concentration trends for each COC, and to consider
the quality of information represented by the existence or absence of concentration trends
in terms of the location of each monitoring point.
Summary results of Mann-Kendall temporal trend analyses for COCs in groundwater
samples from wells in the PCE plume area are presented in Table 5.1. As implemented,
the algorithm used to evaluate concentration trends assigned a value of "ND" (not
detected) to those wells with sampling results that were consistently below analytical
detection limits through time, rather than assigning a surrogate value corresponding to the
detection limit - a procedure that could generate potentially misleading and anomalous
"trends" in concentrations. The color-coding of the Table 5.1 entries denotes the
presence/absence of temporal trends, and allows those monitoring points having
nondetectable concentrations, decreasing or increasing concentrations, or no discernible
trend in concentrations to be readily identified. The four wells that had fewer than four
analytical results for each of the COCs (wells RW1C, BAL2B, MW4A and MW11A)
5-7
S:\ES\SHARED\CEN\MNO\EPA\LONGPRAIRIE\WRITEUP\LongPrairieMNODraftFinal.doc
-------
TABLE 5.1
ORING RESULTS
H
Z
O
^
W
H
^
[^
Q
LYSIS OF GROUN
Z
O
z
td
H
-
O
H
N
S
H
PH
O
ITORING NETWORK
1
s
o
H
-
H
^
^
PRAIRIE, MINNESOT
O
O
hJ
a
td
H
u.
O
!/3
H
-
« -d
^3 «
I a
K ^
•8 -a
fi 5*
O
"2
O
V
V
w
o
Q
-------
«
o
u
TABLE 5.1
ORING RESULTS
H
Z
O
^
2:
td
H
^
[^
Q
GROUN
u.
O
ANALYSIS
O
Z
td
H
-
m
O
td
H
u.
0
H
-
O
H
SI
S
H
PH
O
ETWORK
Z
O
MONITORIN
O
K
-
H
^
a
H
^
[INNESOT
tJ"
,ONG PRAIRI]
—
^H
e
•|
tf
s ?
~5 Oi
5* ^
O _S
S <•>
Oi X
K td
U
U
PH
U
^
•8 -a v
§ C OJD
« « ,a
r* « fe
O
1
13
ca
^
| Addition
Oi
•a"
(U
£?
_g
§
•3
§
^
0?
(U
.1
co
8
1
w
0
PH
~f
I
Pf
|
GO
fN
fN
GO
^
|MW3A
x
o
•jd
00
o
monil
TJ
S
.g
"2
8
1
1
.minani
"3
o
o
~f
I
I
1
m
o
m
|MW3B
Oi
•a
GO"
^
since
o
1
w
0
PH
~f
M
No Trend
1
-«
_g
irf
-P3
a
(U
g
o
o
.g
•a
g
S
cd
8
T3
1
O
-------
could not be analyzed using the Mann-Kendall trend analysis, and have a "<4meas"
designation. Figure 5.4 displays the Mann-Kendall results thematically for PCE by well;
the analytical results for PCE in October 2002 are also presented.
The basis for the decision to remove or retain a well in the monitoring program based
on the value of its temporal information is described in the "Rationale" column of Table
5.1. In general, monitoring wells at which detected chemical concentrations display no
discernible temporal trends (e.g., wells MW2A, MW2C, MW4C, MW6A, MW18B)
represent points generating the least amount of useful information, and can be
recommended for removal from the monitoring network. Monitoring wells that are not
considered "sentry" wells (e.g., wells MW1A, MW1B, MW3B, MW5A, and MW18A) at
which concentrations of COCs consistently have been non-detected or
-------
s *
1 §>
m
•o
c
0
O)
3
C a>
0)Q
o -g =
'-~ .25 to
i 1=
-a£g)£§UJ-5
£ =5 a. jjj « > °
v—
o
CM ||
«<
^.
t— •
11
u
CD o
J- 0
Q}
Q}
o
o
CO
UO a;
o
o •—
CN *"
o
•*•
in
LU
tt
D
0
LJ_
_,
^
QL
O
0.
E
LU
1-
_l
<
O
Z
LJJ
^
i
Z
<
rv
ESULTS FOI
tt
Q
Z
LJJ
o:
i—
LJJ
O
Q.
LJ-
O
>
Z
o
1-
<
NCENTR
O
O
c
o
Network Optimizal
O)
c
'l_
2
'E
o
l§iii§j_MiDiis§2l§_
Q-!
il
o;
m
Z
Q
m
m
i
i
00 f)
O w
O
CM
Z ^
< S
5
O
03
o
• :>
CM
o
CO
Q a
-------
Increasing Trend?
Well in Source Area?
Recent ND
Concentrations?
Decreasing Trend? >—Yes-^—Yes
Cross or Downgradient
Sentry Well?
Cross or Downgradient
Sentry Well?
High Variation
Downgradient?
Cross or Downgradient
Sentry Well?
Remove
FIGURE 5.5
TEMPORAL CHEMICAL
CONCENTRATION TREND
DECISION RATIONALE FLOW
CHART
Monitoring Network Optimization
Long Prairie. Minnesota
PARSONS
Denver, Colorado
-------
SECTION 6
SPATIAL STATISTICAL EVALUATION
Spatial statistical techniques also can be applied to the design and evaluation of
groundwater monitoring programs to assess the quality of information generated during
monitoring, and to evaluate monitoring networks. Geostatistics, or the Theory of
Regionalized Variables (Clark, 1987; Rock 1988; American Society of Civil Engineers
[ASCE] Task Committee on Geostatistical Techniques in Hydrology, 1990a and 1990b),
is concerned with variables having values dependent on location, and which are
continuous in space, but which vary in a manner too complex for simple mathematical
description. Geostatistics is based on the premise that the differences in values of a
spatial variable depend only on the distances between sampling locations, and the relative
orientations of sampling locations — that is, the values of a variable (e.g., chemical
concentrations) measured at two locations that are spatially "close together" will be more
similar than values of that variable measured at two locations that are "far apart".
6.1 GEOSTATISTICAL METHODS FOR EVALUATING MONITORING
NETWORKS
Ideally, application of geostatistical methods to the results of the groundwater
monitoring program at Long Prairie could be used to estimate COC concentrations at
every point within the dissolved contaminant plume, and also could be used to generate
estimates of the "error," or uncertainty, associated with each estimated concentration
value. Thus, the monitoring program could be optimized by using available information
to identify those areas having the greatest uncertainty associated with the estimated
plume extent and configuration. Conversely, sampling points could be successively
eliminated from simulations, and the resulting uncertainty examined, to evaluate if
significant loss of information (represented by increasing error or uncertainty in
6-1
S:\ES\SHARED\CEN\MNO\EPA\LONGPRAIRIE\WRITEUP\LongPrairieMNODraftFinal.doc
-------
estimated chemical concentrations) occurs as the number of sampling locations is
reduced. Repeated application of geostatistical estimating techniques, using tentatively
identified sampling locations, then could be used to generate a sampling program that
would provide an acceptable level of uncertainty regarding the distribution of COCs with
the minimum possible number of samples collected. Furthermore, application of
geostatistical methods can provide unbiased representations of the distribution of COCs
at different locations in the subsurface, enabling the extent of COCs to be evaluated more
precisely.
Fundamental to geostatistics is the concept of semivariance [y(h)], which is a measure
of the spatial dependence between samples (e.g., chemical concentrations) in a specified
direction. Semivariance is defined for a constant spacing between samples (h) by:
1 2
y(h) = — zL[g(x) - g(x + h) ] Equation 6-1
2n
Where:
y(h) = semivariance calculated for all samples at a distance h from each other;
g(x) = value of the variable in sample at location x;
g(x + h) = value of the variable in sample at a distance h from sample at location x;
and
« = number of samples in which the variable has been determined.
Semivariograms (plots of y(h) versus h) are a means of depicting graphically the range
of distances over which, and the degree to which, sample values at a given point are
related to sample values at adjacent, or nearby, points, and conversely, indicate how close
together sample points must be for a value determined at one point to be useful in
predicting unknown values at other points. For h = 0, for example, a sample is being
compared with itself, so normally y(0) = 0 (the semivariance at a spacing of zero, is
6-2
S:\ES\SHARED\CEN\MNO\EPA\LONGPRAIRIE\WRITEUP\LongPrairieMNODraftFinal.doc
-------
FIGURE 6.1
IDEALIZED SEMIVARIOGRAM MODEL
THREE-TIERED MONITORING NETWORK OPTIMIZATION
LONG PRAIRIE, MINNESOTA
3500 -
3000
§ 2500 ~
I 2000 "
1
1 1500 -
g
o
1 1000
(U
sU
" 500 -
0
(
^ Ranae ^
i****"^ ^ w ~ *" B ®
Sill X"
/
/
/m Spherical model
^ Nugget: 1,500 mg 2/L 2
Range: 1,000 ft
Sill: 1,500 mg 2/L 2
f Nugget
j 1
' i i i i i ii
) 500 1000 1500 2000 2500 3000 3500
Distance (ft)
zero), except where a so-called nugget effect is present (Figure 6.1), which implies that
sample values are highly variable at distances less than the sampling interval. As the
distance between samples increases, sample values become less and less closely related,
and the semivariance, therefore, increases, until a "sill" is eventually reached, where y(h)
equals the overall variance (i.e., the variance around the average value). The sill is
reached at a sample spacing called the "range of influence," beyond which sample values
are not related. Only values between points at spacings less than the range of influence
can be predicted; but within that distance, the semivariogram provides the proper
weightings, which apply to sample values separated by different distances.
When a semivariogram is calculated for a variable over an area (e.g., concentrations of
PCE in the groundwater plume at Long Prairie), an irregular spread of points across the
6-3
S:\ES\SHARED\CEN\MNO\EPA\LONGPRAIRIE\WRITEUP\LongPrairieMNODraftFinal.doc
-------
semivariogram plot is the usual result (Rock, 1988). One of the most subjective tasks of
geostatistical analysis is to identify a continuous, theoretical semivariogram model that
most closely follows the real data. Fitting a theoretical model to calculated semivariance
points is accomplished by trial-and-error, rather than by a formal statistical procedure
(Davis, 1986; Clark, 1987; Rock, 1988). If a "good" model fit results, then tfh) (the
semivariance) can be confidently estimated for any value of h, and not only at the
sampled points.
6.2 SPATIAL EVALUATION OF MONITORING NETWORK AT LONG
PRAIRIE
PCE was used as the indicator chemical for the spatial evaluation of the groundwater
monitoring network at Long Prairie because this COC has the largest spatial distribution
of measurements that exceeded groundwater MCLs. The most recent (October 2002)
validated analytical data available at the start of this MNO evaluation were used in the
kriging evaluation because a spatial "snapshot" is required in order to conduct the
geospatial statistical analysis.
Of the 44 wells sampled in October 2002, 16 were included in the kriging evaluation.
Although the OU1 extraction wells have historically been used to define the plume
extent, data from extraction wells are not appropriate for use in a kriging analysis because
they represent COC concentrations averaged over the area within the well's capture zone,
and thus are not point specific, nor temporally discrete; the recovery wells are also
typically screened across a longer screening interval than the site monitoring wells.
Similarly, city wells CW3 and CW6 were excluded from the analysis because they also
are pumping wells. Kriging predicts concentrations over a two-dimensional surface and
thus including data from multiple co-located wells screened at different depths is not
appropriate. In this application, the well within each cluster of well with the highest
concentration of PCE was retained for use in the geostatistical evaluation; this
methodology is consistent with the values displayed on plume maps in the Long Prairie
annual reports (Barr, 2001; Barr, 2002). On the whole, of the clustered wells, the "B"
zone wells had the highest October 2002 PCE concentrations and were included in the
6-4
S:\ES\SHARED\CEN\MNO\EPA\LONGPRAIRIE\WRITEUP\LongPrairieMNODraftFinal.doc
-------
spatial analysis; however, the "C" zone well MW6C was included from the MW6 cluster.
The 16 wells analyzed are shown in Figure 6.3 and listed in Table 6.1.
The commercially available geostatistical software package Geostatistical Analyst™
(an extension to the Arc View® geographic information system [GIS] software package)
(Environmental Systems Research Institute, Inc. [ESRI], 2001) was used to develop a
semivariogram model depicting the spatial variation in PCE concentrations in
groundwater for the 16 Long Prairie area wells.
As semivariogram models were calculated for PCE (Equation 6-1), considerable
scatter of the data was apparent during fitting of the models. Several data
transformations (including a log transformation) were attempted to obtain a
representative semivariogram model. Ultimately, , the concentration data were
transformed to "rank statistics," in which the 16 wells were ranked from 1 to 16 (tie
values were assigned the median rank of the set) according to their October 2002 PCE
concentration. Transformations of this type can be less sensitive to outliers, skewed
distributions, or clustered data than semivariograms based on raw concentration values,
and thus may enable recognition and description of the underlying spatial structure of the
data in cases where ordinary data are too "noisy".
The PCE rank statistics were used to develop a semivariogram that most accurately
modeled the spatial distribution of the data. Anisotrophy was incorporated into the model
to adjust for the directional influence of groundwater to the northeast. The model was
unable to account for the "dog-leg" shift in groundwater flow direction to the northwest
in the northern half of the plume (Figure 2.1). Figure 6.2 shows the semivariogram
model in comparison to the site data. The large amount of scatter in the data due to the
small number of wells and shifting groundwater direction makes it difficult to develop a
representative semivariogram model. Thus, the geostatistical evaluation at this site can
not be as rigorous as at other sites with more wells and more consistent hydrology. The
best-fit semivariogram had the following parameters:
6-5
S:\ES\SHARED\CEN\MNO\EPA\LONGPRAIRIE\WRITEUP\LongPrairieMNODraftFinal.doc
-------
TABLE 6.1
RESULTS OF GEOSTATISTICAL EVALUATION RANKING OF WELLS BY
RELATIVE VALUE OF PCE INFORMATION
THREE-TIERED MONITORING NETWORK OPTIMIZATION
LONG PRAIRIE, MINNESOTA
Well ID ^
MW19B
MW13C
MW4B
MW16B
MW6C
MW2B
MW3B
BAL2C
MW18B
MW10A
MW11B
MW15B
MW14B
MW5B
MW17B
MW1B
Kriging Ranking
1
2
3.5
3.5
5.5
5.5
7
8
9.5
9.5
11
12
13
14.5
14.5
16
Remove/
Exclude
V
V
V
V
d/
-
-
—
-
-
-
—
Retain/ Add
—
-
-
—
-
-
-
—
V
V
V
V
Clustered wells included in the spatial analysis are those with the highest
relative October 2002 PCE concentration.
1= least relative amount of information; 16= most relative amount of information.
c Tie values receive the median ranking of the set.
Well in the "intermediate" range; received no recommendation for removal/exclusion or retention/addition
(see Section 6.2).
022/742479/LongPrairieTablesFinal.xls/Table 6.1
6-6
-------
Circular model
Range: 1200 feet
Sill: 17
Nugget: 10
Anisotrophy:
- Minor Range: 400 feet
- Direction: 20 degrees
FIGURE 6.2
LONG PRAIRIE SEMVARIOGRAM MODEL
THREE-TIERED MONITORING NETWORK OPTIMIZATION
LONG PRAIRIE, MINNESOTA
r - - !!
o
48 9-
eo : 9,
o
0
v
36 A °- -9 ' • e-» • I '
O g €P d (I QO
•3 if _ _ 0 O_ S- - ~ Q~ 3 q ? ?
-fefei' £J Q ^J " "IP" ft"
^ 2 ® 6 ^e e o **•e : =
Q O ft ^^P ^^O ;;;
6 fi Q fi :::
0 457 974 14B1 1948 2435 2922 34D9
-2
Distance, h-10
After this semivariogram model had been developed, it was used in the kriging system
implemented by the Geostatistical Analyst™ software package (ESRI, 2001) to develop
kriging realizations (estimates of the spatial distribution of PCE in groundwater at Long
Prairie), and to calculate the associated kriging prediction standard errors. The median
kriging standard deviation was obtained from the standard errors calculated using the
entire 16-well monitoring network for Long Prairie. Next, each of the 16 wells was
sequentially removed from the network, and for each resulting well network
6-7
S:\ES\SHARED\CEN\MNO\EPA\LONGPRAIRIE\WRITEUP\LongPrairieMNODraftFinal.doc
-------
configuration, a kriging realization was completed using the PCE concentration rankings
from the remaining 15 wells. The "missing-well" monitoring network realizations were
used to calculate prediction standard errors, and the median kriging standard deviations
were obtained for each "missing-well" realization and compared with the median kriging
standard deviation for the "base-case" realization (obtained using the complete 16-well
monitoring network), as a means of evaluating the amount of information loss (as
indicated by increases in kriging error) resulting from the use of fewer monitoring points.
Figure 6.3 illustrates the spatial-evaluation procedure by showing kriging prediction
standard-error maps for three kriging realizations. Each map shows the predicted
standard error associated with a given group of wells based on the semivariogram
parameters discussed above. Lighter colors represent areas with lower spatial
uncertainty, and darker colors represent areas with higher uncertainty; regions in the
vicinity of wells (i.e., data points) have the lowest associated uncertainty. Map A on
Figure 6.3 shows the predicted standard error map for the "base-case" realization in
which all 16 wells are included. Map B shows the realization in which well MW13C was
removed from the monitoring network, and Map C shows the realization in which well
MW17B was removed. Figure 6.3 shows that when a well is removed from the network,
the predicted standard error in the vicinity of the missing well increases (as indicated by a
darkening of the shading in the vicinity of that well). If a "removed" (missing) well is in
an area with several other wells (e.g., well MW13C; Map B on Figure 6.2), the predicted
standard error may not increase as much as if a well (e.g., MW17B; Map C) is removed
from an area with fewer surrounding wells. If removal of a particular well from the
monitoring network caused very little change in the resulting median kriging standard
deviation (less than about 1 percent), that well was regarded as contributing only a
limited amount of information to the LTM program. Likewise, if removal of a well from
the monitoring network produced larger increases in the kriging standard deviation, this
was regarded as an indication that the well contributes a relatively greater amount of
information, and is relatively more important to the monitoring network.
6-8
S:\ES\SHARED\CEN\MNO\EPA\LONGPRAIRIE\WRITEUP\LongPrairieMNODraftFinal.doc
-------
o
•>
i
s
6
gj
0
CO
03
CO 0
03 ^
OQ ±
- o
"0)00
co
CO
CD
Q.
CO
c
o
CD j^
i 2
0 S-
5.1"
•a^
o i
^^P ^^H . __ ^"^
C CD
HI CO c
> o
(0
w
c/j
E ^
= O
c
\D
CD
O
3
^
'-i—>
CD
Q.
(f)
0
CD
t
0
O
C
3
ID
'-i—>
CD
Q.
(f)
CD
CD
6
CD
CO
CO
O
-------
At the conclusion of the kriging realizations, each well was ranked from 1 (providing the
least information) to 16 (providing the most information), based on the amount of
information (as measured by changes in median kriging standard deviation) the well
contributed toward describing the spatial distribution of TCE, as shown in Table 6.1.
Wells providing the least amount of information represent possible candidates for
removal from the monitoring network at the Long Prairie.
6.3 SPATIAL STATISTICAL EVALUATION RESULTS
6.3.1 Kriging Ranking Results
Figure 6.4 and Table 6.1 present the ranking of the evaluated subset of monitoring
locations based on the relative value of recent PCE information provided by each well, as
calculated based on the kriging realizations. Examination of these results indicate that
monitoring wells in close proximity to several other monitoring wells (e.g., red color
coding on Figure 6.4) generally provide relatively lesser amounts of information than do
wells at greater distances from other wells, or wells located in areas having limited
numbers of monitoring points (e.g., blue color coding on Figure 6.4). This is intuitively
obvious, but the analysis allows the most valuable and least valuable wells to be
identified quantitatively. For example, Table 6.1 identifies the four wells ranked at or
below 3.5 (wells MW19B, MW13C, MW4B, MW16B) that provide the relative least
amount of information, and the four wells ranked at or above 13 (wells MW14B, MW5B,
MW17B, MW1B) that provide the greatest amount of relative information regarding the
occurrence and distribution of PCE in groundwater among those wells included in the
kriging analysis. The four lowest-ranked wells are potential candidates for removal from
the Long Prairie groundwater monitoring program, and the four highest-ranked wells are
candidates for retention in the monitoring program, intermediate ranked wells receive no
recommendation for removal or retention in the monitoring program based on the spatial
analysis.
6-10
S:\ES\SHARED\CEN\MNO\EPA\LONGPRAIRIE\WRITEUP\LongPrairieMNODraftFinal.doc
-------
CD
C
0
D)
0
CO
O) |
12
I *
O)
c
co
CN
O
O
CN
§«
o m
CD
,? °
LL o
3
O
3
CO
CO
o
CD
CN T-
CO
m -2 Q
•a a.
±£K
.? O O
O
o
LO
J/5
CO
cr i±
CD
o —
LO 1-
CN
FIGURE 6.4
^r
LTS OF GEOSTATISTIC/
3
W
LJJ
L±
LJJ
YSIS SHOWING RELATI\
^
^
^
<
O
OF SPATIAL INFORMAT
LJJ
i
^
IN PCE DISTRIBUTION
O
FOR SELECT WELLS
onitoring Network Optimization
^
CO:
1
.of
-t
m
z
o
a
?
o
CN
DQ
GO
'sT
GQ
GO
GO
CO
CQ
tP
QQ
O
<
O
-------
SECTION 7
SUMMARY OF THREE-TIERED MONITORING NETWORK
EVALUATION
The 44 wells sampled in October 2002 at Long Prairie were evaluated using
qualitative hydrogeologic and extraction-system information, temporal statistical
techniques, and spatial statistics. At each tier of the evaluation, monitoring points that
provide relatively greater amounts of information regarding the occurrence and
distribution of COCs in groundwater were identified, and were distinguished from those
monitoring points that provide relatively lesser amounts of information. In this section,
the results of the evaluations are combined to generate a refined monitoring program that
potentially could provide information sufficient to address the primary objectives of
monitoring, at reduced cost. Monitoring wells not retained in the refined monitoring
network could be removed from the monitoring program with relatively little loss of
information. The results of the evaluations were combined and summarized in
accordance with the following decision logic:
1. Each well retained in the monitoring network on the basis of the qualitative
hydrogeologic evaluation is recommended to be retained in the refined
monitoring program.
2. Those wells recommended for removal from the monitoring program on the
basis of all three evaluations, or on the basis of the qualitative and temporal
evaluations (with no recommendation resulting from the spatial evaluation)
should be removed from the monitoring program.
7-1
S:\ES\SHARED\CEN\MNO\EPA\LONGPRAIRIE\WRITEUP\LongPrairieMNODraftFinal.doc
-------
3. If a well is recommended for removal based on the qualitative evaluation and
recommended for retention based on the temporal or spatial evaluation, the final
recommendation is based on a case-by-case review of well information.
The results of the qualitative, temporal, and spatial evaluations are summarized in Table
7.1. These results indicate that 18 of the 44 wells sampled in October 2002 could be
excluded from the groundwater monitoring program with little loss of information.
These results further suggest that the "current" sampling plan (the 19 monitoring wells
and 6 active recovery wells included in the 2000 and 2001 sampling schedules, plus
water-supply wells CW3 and CW6) could be optimized by removing four of the 27 wells
now in the LTM program, and adding three wells not currently included in the program.
Two wells sampled in October 2002, but not included in the current monitoring program,
fall into case 3 of the decision logic (as listed above). Justifications for the
recommendation to continue to exclude these wells from routine sampling are as follow:
• Well MW1B was recommended for continued exclusion from the monitoring
network based on the qualitative and temporal evaluations, and for addition to the
program based on the spatial evaluation. This well should not be added to the
monitoring program because it is crossgradient from (west of) the plume, and well
RW4 provides adequate monitoring of the western boundary of the plume.
• Well MW5B also was recommended for continued exclusion from the monitoring
network based on the qualitative evaluation, and addition to the program based on
the temporal and spatial evaluations. Well MW5B should be added to the
scheduled monitoring program to determine if the recent low detection of cis-1,2-
DCE truly reflects an increasing concentration trend.
A refined monitoring program, consisting of 26 wells (2 to be sampled quarterly, 6 to be
sampled semi-annually, 14 to be sampled annually, and 4 to be sampled biennially)
would be adequate to address the two primary objectives of monitoring. This refined
monitoring network would result in an average of 36 sampling events per year, compared
to 51 events per year under the current (2000/2001) monitoring program. Implementing
7-2
S:\ES\SHARED\CEN\MNO\EPA\LONGPRAIRIE\WRITEUP\LongPrairieMNODraftFinal.doc
-------
these recommendations for optimizing the LTM monitoring program at Long Prairie
could reduce current LTM annual monitoring by more than 29 percent. Additionally,
based on a per well sampling cost ranging from$l 00 to $280, these recommendations
could reduce costs by an average of $1500 to $4,200 per year.
7-3
S:\ES\SHARED\CEN\MNO\EPA\LONGPRAIRIE\WRITEUP\LongPrairieMNODraftFinal.doc
-------
O
O
rt
o-
O
o
H
o 2
w
Z
Z
W
S
O
^ o
3s
(?
Summa
valuation
w
is
-*j
cs
a.
cc
"3
•—
o
^
O>
H
Evaluation
Qualitative
•I £?s>
Recommei
Monitor!
Frequen
5 «
« T5
(2 ^
Remove/
Exclude
5 «
« T5
(2 ^
Remove/
Exclude
5 «
« T5
(2 ^
Remove/
Exclude
5 «
•2 3
o> •
•9 °
t §
1 g-
cc £
u.
9 "3 ^-v
Ca >• ^
S C we
J- « -a
•» 5 *
Q
HH
"3
Bt
C
cs
PM
W3
e
"3.
S
cs
cc
^H
O
rl
o
o
o
ri
_9
!/5
"
£
we
_9
O
-t^
o
§
1
1
~r
Z
>•
•?•
C3
-^
O
CS
>n
n
m
m
PQ
^
s
!
~f
HH
Z
-?•
•?•
C3
-^
m
>n
oo
TT
O
^
S
C3
1
~f
~f
~f
~f
C3
-^
>n
>ri
m
>n
m
OQ
^
s
C3
1
>•
HH
£
>•
•?•
C3
-^
O
VO
Tt
o
CS
Tt
O
^
s
1
1
~r
Z
>•
•?•
C3
-^
t~-
O
CS
<
&
s
C3
1
~f
Z
>•
•?•
C3
1
f—
m
O
m
m
&
S
C3
1
~f
~f
~f
C3
1
ON
>n
O
VO
Tt
O
£
s
C3
1
>•
>•
•?•
C3
1
CS
^O
<
O
^
s
C3
1
>•
>•
~f
C3
1
«
^
s
C3
1
>•
Z
~f
~f
C3
1
>n
n
0
^
S
C3
1
-?•
>•
~f
~f
C3
-^
^O
CS
CN
OQ
•*
^
s
1
1
>•
z
~r
~r
C3
-^
>n
n
O
•*
£
S
C3
1
>•
HH
Z
>•
•?•
C3
-^
CN
<
>n
^
s
C3
1
>•
>•
~f
C3
-^
m
m
«
>n
^
s
C3
1
~r
~f
~f
~f
C3
1
m
>ri
CN
«
^O
^
s
C3
1
~r
~f
~f
~f
C3
1
m
m
OQ
r—
^
s
'3
«
~r
~f
~f
~f
C3
1
>n
VO
CS
OQ
ON
£
s
V5
"a!
£
jj»
a
cc
IH
U
-^^
CS
^
>%
-^— 1
U
>^
fe
cd
a
>•
z
>•
•?•
>^
s
1
a
>n
oo
r^
^o
m
^
O
>^
fe
cd
a
>•
z
>•
•?•
>^
s
1
a
^o
t~-
m
>n
^o
^
O
!*>
"a!
^
>%
IH
a>
>
o
u
tf
1
1
~f
Z
>•
~f
i
O
m
>n
n
Tt
m
OQ
1
1
1
~f
Z
'•a
u
s
>,
a
•<
+*
o
Z
•?•
i
>n
fi
CN
>n
CN
c;
1
"3
^
|
OJ
CO
•?•
z
•?•
•?•
>^
fe
1
a
CS
-------
o
O
rt
o-
O
o
H
o 2
5 UH S
3 5 H
^ n
1-- P^ H
•
H
Evaluation
Qualitative
nmended
litoring
quency
o o 2
l^ta
5 «
« T5
(2 ^
Remove/
Exclude
5 «
« T5
(2 ^
Remove/
Exclude
5 «
« T5
(2 ^
Remove/
Exclude
5 «
-2 3
o> •
.S °
t §
1 g-
cc 2
u.
B "3 ^-v
o> >. ^
S C we
b s -°
™ 5 *
Q
HH
"3
Bt
3
i
PQ
>•
Z
>•
•?•
C3
-^
O
>n
O
o
1
i-Annual |
OJ
CO
>•
Z
•?•
•?•
>1
fe
a
^o
Tl
*f
V~}
1
i-Annual |
OJ
CO
>•
Z
•?•
•?•
>1
fe
a
>n
>ri
>n
o
o
^o
1
i-Annual |
OJ
CO
-r
'^
-r
•?•
>^
fe
1
a
>n
Tt
•
>^
•?•
•?•
>^
fe
1
a
o
Tt
o
m
oo
1
i-Annual |
OJ
CO
-r
'^
~r
~r
>^
fe
1
a
>n
m
>n
"3
<
+*
o
Z
•?•
i
>n
^o
r^
>n
OQ
CN
J
<
«
1
1
>•
-?•
•?•
i
O
>n
O
Tt
0
CN
J
<
«
1
1
>•
(^
-r
•?•
i
>n
•*'
i
>n
o<
•
>•
-r
•?•
i
>n
m
O
m
OQ
^
S
1
1
•?•
(^
•?•
•?•
1
oo
CS
n
m
O
m
«
^
S
1
1
-r
1^
•e
u
s
j£»
"3
•<
-^^
o
Z
•?•
!
>n
o
o
<
&
s
1
1
-r
Z
>•
•?•
i
>n
ON
>n
Tt
<
&
S
"3
d
<
~r
~r
~r
~r
!
>n
m
O
^H
m
OQ
^
S
1
1
>•
z
•e
u
s
^>
"3
•<
-^^
o
Z
•?•
I
<
£
S
"3
'3
OQ
>•
>•
>•
•?•
i
n
0
m
^
S
1
1
-r
Z
>•
•?•
i
>n
>ri
<
^o
^
S
1
1
>•
HH
z
>•
•?•
1
>ri
<
OO
^
S
3
1
«
•?•
•?•
•?•
i
^
•I I
d §
fH -P
0> ^
H
OJ
T3 ^£
l> j3
a S .2
^ (ii -H
OJ
o
,d
(L> Cd
^ «->
S ^
M O
-rt d ^
1 s s
5 a1 --a
S «-> s.
^ d § -o
o .g o>
If ll
-M H *** o
1> TO ^ ^
•a 43
OJ -"^
SP-3
OJ
£
OJ
-------
SECTION 8
REFERENCES
American Society of Civil Engineers (ASCE) Task Committee on Geostatistical
Techniques in Hydrology. 1990a. Review of geostatistics in geohydrology - I.
Basic concepts. Journal of Hydraulic Engineering 116(5):612-632.
ASCE Task Committee on Geostatistical Techniques in Hydrology. 1990b. Review of
geostatistics in geohydrology - II. Applications. Journal of Hydraulic
Engineering 116(6): 63 3-65 8.
Barr Engineering (Barr). 2001. 1999/2000 Annual Report (September 1999 through
October 2000) Long Prairie Groundwater Remediation System, prepared for
Minnesota Pollution Control Agency, March.
Barr. 2002. 2000/2001 Annual Report (September 2000 through October 2001) Long
Prairie Groundwater Remediation System, prepared for Minnesota Pollution
Control Agency, March.
Clark, I. 1987. Practical Geostatistics. Elsevier Applied Science, Inc., London.
Environmental Systems Research Institute, Inc. (ESRI). 2001. ArcGIS Geostatistical
Analyst Extension to ArcGIS 8 Software, Redlands, CA.
Gibbons, R.D. 1994. Statistical Methods for Groundwater Monitoring. John Wiley &
Sons, Inc., New York.
Minnesota Pollution Control Agency (MPCA). 2002. First Five-Year Review Reprotfor
Long Prairie Ground Water Contamination Superfund Site, September.
8-1
S:\ES\SHARED\CEN\MNO\EPA\LONGPRAIRIE\WRITEUP\LongPrairieMNODraftFinal.doc
-------
Rock, N.M.S. 1988. Numerical Geology. Springer-Verlag. New York, New York
USEPA. 1994. Methods for Monitoring Pump-and-Treat Performance. Office of
Research and Development. EPA/600/R-94/123.
Wiedemeier, T.H., and P.E. Haas. 2000. Designing Monitoring Programs to Effectively
Evaluate the Performance of Natural Attenuation. Air Force Center for
Environmental Excellence (AFCEE). August.
8-2
S:\ES\SHARED\CEN\MNO\EPA\LONGPRAIRIE\WRITEUP\LongPrairieMNODraftFinal.doc
-------
APPENDIX D-3
OPTIMIZATION OF MONITORING PROGRAM
AT
OPERABLE UNIT D
McCLELLAN AIR FORCE BASE, CALIFORNIA
-------
G-2236-15
2.0
Air
Sacramento Valley, California
to
Air Force Center for Environmental
June 2, 2003
Groundwater Services, Inc.
2211 Norfolk, Suite 1000, Houston, Texas 77098-4044
-------
V
GROUNDWATER
SERVICES, INC.
MAROS 2.0 APPLICATION
ZONE A & B OU D MONITORING NETWORK OPTIMIZATION
MCCLELLAN AIR FORCE BASE
Sacramento Valley, California
Prepared
by
Groundwater Services, Inc.
2211 Norfolk, Suite 1000
Houston, Texas 77098
(713)522-6300
GSI Job No. G-2236
Revision No. DRAFT
Date: 6/02/03
-------
GSI Job No. G-2236-15 GROUNDWATER
January 15, 2003 SERVICES, INC.
MAROS 2.0 APPLICATION
ZONE A & B OU D MONITORING NETWORK
OPTIMIZATION, MCCLELLAN AIR FORCE BASE
Sacramento Valley, California
Table of Contents
Executive Summary 1
Project Objectives 1
Results 2
1.0 Introduction 4
1.1 Geology/Hydrogeology 4
1.2 Remedial Action 5
2.0 MAROS Methodology 7
2.1 MAROS Conceptual Model 7
2.2 Data Management 8
2.3 Site Details 8
2.4 Data Consolidation 9
2.5 Overview Statistics: Plume Trend Analysis 9
2.5.1 Mann-Kendall Analysis 10
2.5.2 Linear Regression Analysis 10
2.5.3 Overall Plume Analysis 11
2.5.4 Moment Analysis 12
2.6 Detailed Statistics: Optimization Analysis 13
2.6.1 Well Redundancy Analysis- Delaunay Method 14
2.6.2 Well Sufficiency Analysis - Delaunay Method 15
2.6.3 Sampling Frequency- Modified CES Method 15
2.6.4 Data Sufficiency - Power Analysis 16
3.0 Site Results 18
3.1 Data Consolidation 19
3.2 Overview Statistics: Plume Trend Analysis 19
3.2.1 Mann-Kendall/Linear Regression Analysis 19
3.2.2 Moment Analysis 21
3.2.3 Overall Plume Analysis 24
3.3 Detailed Statistics: Optimization Analysis 25
3.3.1 Well Redundancy Analysis 25
3.3.2 Well Sufficiency Analysis 27
3.3.3 Sampling Frequency Analysis 28
3.3.4 Data Sufficiency - Power Analysis 29
4.0 Summary and Recommendations 31
McClellanAFB Q MAROS 2.0 Application
Sacramento Valley, California Monitoring Network Optimization
-------
GSIJobNo. G-2236-15
January 15, 2003
GROUNDWATER
SERVICES, INC.
Tables
Table 1 Sampling Locations Used in the MAROS Analysis
Table 2 Mann-Kendall Analysis Decision Matrix
Table 3 Linear Regression Analysis Decision Matrix
Table 4 Zone A & B Aquifer Site-Specific Parameters
Table 5 Results of Zone A McClellan OU D Trend Analysis
Table 6 Results of Zone B McClellan OU D Trend Analysis
Table 7 Zone A Redundancy Analysis Results - Delaunay Method
Table 8 Zone B Redundancy Analysis Results - Delaunay Method
Table 9 Zone A Sampling Frequency Analysis Results - Modified CES
Table 10 Zone B Sampling Frequency Analysis Results- Modified CES
Table 11 Zone A Risk-Based Site Cleanup Evaluation Results - Power Analysis
Table 12 Zone A Selected Plume Centerline Wells for Risk-Based Site Cleanup
Evaluation - Power Analysis
Table 13 Zone A Plume Centerline Concentration Regression Results - Power
Analysis
Table 14 Summary of MAROS Sampling Optimization Results
Figures
Figure 1 Zone A McClellan OU D Groundwater Monitoring Network
Figure 2 Zone B McClellan OU D Groundwater Monitoring Network
Figure 3 MAROS Decision Support Tool Flow Chart
Figure 4 MAROS Overview Statistics Trend Analysis Methodology
Figure 5 Decision Matrix for Determining Provisional Frequency
Figure 6 Zone A McClellan OU D TCE Mann-Kendall Trend Results
Figure 7 Zone A McClellan OU D TCE Linear Regression Trend Results
Figure 8 Zone B McClellan OU D TCE Mann-Kendall Trend Results
Figure 9 Zone B McClellan OU D TCE Linear Regression Trend Results
Figure 10 Zone AB McClellan OU D TCE Mann-Kendall Trend Results, Extraction
Wells
Figure 11 Zone AB McClellan OU D TCE Linear Regression Trend Results,
Extraction Wells
Figure 12 Zone A McClellan OU D TCE First Moment (Center of Mass) Over Time
Figure 13 Zone B McClellan OU D TCE First Moment (Center of Mass) Over Time
Figure 14 Zone A Well Sufficiency Results
Appendices
Appendix A: Zone A and Zone B McClellan OU D Historical TCE Maps
Appendix B: Zone A and Zone B McClellan OU D MAROS 2.0 Reports
McClellan AFB
Sacramento Valley, California
MAROS 2.0 Application
Monitoring Network Optimization
-------
GROUNDWATER
June 2, 2003 SERVICES, INC.
MAROS 2.0 APPLICATION
ZONE A & B OU D MONITORING NETWORK OPTIMIZATION
MCCLELLAN AIR FORCE BASE
EXECUTIVE SUMMARY
Long-term monitoring programs, whether applied for process control, performance
measurement, or compliance purposes, require large scale data collection effort and
time commitment, making their cumulative costs very high. With the increasing use of
risk-based goals and natural attenuation in recent years as well as the move toward
long-term closure upon completion of cleanup activities, the need for better-designed
long-term monitoring plans that are cost-effective, efficient, and protective of human and
ecological health has greatly increased. The Monitoring and Remediation Optimization
System (MAROS) methodology provides an optimal monitoring network solution, given
the parameters within a complicated groundwater system which will increase its
effectiveness. By applying statistical techniques to existing historical and current site
analytical data, as well as considering hydrogeologic factors and the location of potential
receptors, the software suggests an optimal plan along with an analysis of individual
monitoring wells for the current monitoring system. This report summarizes the findings
of an application of the MAROS 2.0 software to Zone A and Zone B long-term monitoring
well networks in Operating Unit (OU) D at the McClellan Air Force Base in Sacramento
Valley, California.
The primary constituent of concern at the site is trichloroethylene (TCE) which is
analyzed at 32 and 14 monitoring wells respectively in the OU D Zone A and Zone B
well networks (Figures 1 and 2). Monitoring wells in both Zones A and B have been
sampled for TCE irregularly, ranging from quarterly to annually and in some cases,
biennially (every two years) since the implementation of the long-term monitoring plan in
1990. By December 2000, 40 sampling events had been carried out at the site,
however, many wells have only 5 analyses. The historical TCE data for all or in some
cases a subset of wells were analyzed using the MAROS 2.0 software in order to: 1)
gain an overall understanding of the plume stability, and 2) recommend changes in
sampling frequency and sampling locations without compromising the effectiveness of
the long-term monitoring network.
Project Objectives
The general objective of the project was to optimize the McClellan OU D long-term
monitoring network and sampling plan applying the MAROS 2.0 statistical and decision
support methodology. The key objectives of the project included:
Determining the overall plume stability through trend analysis and moment
analysis;
Evaluating individual well TCE concentration trends over time;
McClellan Air Force Base 1 MAROS 2.0 Application
Sacramento Valley, California Monitoring Network Optimization
-------
GROUNDWATER
June 2, 2003 SERVICES, INC.
Addressing adequate and effective sampling through reduction of redundant
wells without information loss and addition of new wells for future sampling;
• Assessing future sampling frequency recommendations while maintaining
sufficient plume stability information;
Evaluating risk-based site cleanup status using data sufficiency analysis.
Results
The MAROS 2.0 sampling optimization software/methodology has been applied to the
McClellan's existing monitoring program as of December 2000. Historical data from
2001 was not used in this analysis due to anomalous TCE concentrations from the
passive diffusion sampling technique utilized only in 2001. Results from the temporal
trend analysis, moment analysis, sampling location determination, sampling frequency
determination, and data sufficiency analysis indicate that:
• Site monitoring wells were divided into source wells and tail wells where source
wells are in the vicinity of NAPL or have historically elevated concentrations of
TCE.
• 9 out of 10 source wells and 11 out of 22 tail wells in Zone A have a Probably
Decreasing, Decreasing, or Stable trend. Both of the statistical methods used to
evaluate trends (Mann-Kendall and Linear Regression) gave similar trend
estimates for each well.
• 0 out of 1 source wells and 3 out of 13 tail wells in Zone B have a Probably
Decreasing, Decreasing, or Stable trend. The majority of the wells in Zone B
have no trend in the historical data. However, as of 2000, only one of the wells in
Zone B is actually above the MCL for TCE. Both of the statistical methods used
to evaluate trends (Mann-Kendall and Linear Regression) gave similar trend
estimates for each well.
• 5 out of 6 source area extraction wells in Zone AB have a Decreasing trend. Both
the Mann-Kendall and Linear Regression methods gave similar trend estimates
for each well.
• The dissolved mass shows stability over time, whereas the center of mass and
the plume spread show no trend over time. The results from the moment
analysis are not very strong due to the change in the wells sampled over the
sampling period analyzed.
• Overall plume stability results lead to the MAROS analysis system to indicate
that a monitoring system of "Moderate" intensity is appropriate for this plume
compared to "Limited" or "Extensive" systems due to a stable or decreasing
plume in both Zone A and Zone B.
McClellan Air Force Base 2 MAROS 2.0 Application
Sacramento Valley, California Monitoring Network Optimization
-------
GROUNDWATER
June 2, 2003 SERVICES, INC.
• The well redundancy optimization tool, using the Delaunay method, indicates that
3 existing monitoring wells may not be needed for Zone A plume monitoring and
can likely be eliminated from the existing Zone A monitoring network of 32 wells
without compromising the accuracy of the monitoring network. Similarly, two
existing monitoring wells may be eliminated from the existing Zone B monitoring
network of 14 wells.
• The well sufficiency optimization tool, using the Delaunay method, indicates that
there are no areas within the existing monitoring network that have high
uncertainty in the TCE concentration estimation. Therefore, no new monitoring
wells are recommended for Zones A or B.
• Application of the well sampling frequency determination tool, the Modified CES
method, leads to results that are in agreement with the sampling frequency
currently in use at the site. Therefore, no sampling frequency reduction is
recommended for the wells in the current monitoring network.
• The MAROS Data Sufficiency (Power Analysis) application indicates that the
monitoring record has sufficient statistical power to conclude that the site has
attained cleanup goal at (or farther than) the "hypothetical statistical compliance
boundary" located 100 feet downgradient of the most downgradient well at the
site. As more sampling records accumulate, this hypothetical statistical
compliance boundary will get closer and closer to the downgradient wells of the
monitoring system.
The recommended long-term monitoring strategy results in a moderate reduction in
sampling costs and allows site personnel to develop a better understanding of plume
behavior over time. The MAROS optimized plan results in a monitoring network of 47
wells: 17 sampled annually, and 30 sampled biennially. The MAROS optimized plan
would result in 32 samples per year, compared to 34 samples per year (17 annual and
34 biennial) in the current McClellan sampling program. Implementing these
recommendations could lead to a 6% reduction from the current monitoring plan in terms
of the samples to be collected per year. A reduction in the number of redundant wells is
expected to result in a moderate cost savings over the long-term at McClellan Air Force
Base. An approximate cost savings estimate of $300 per year is projected while still
maintaining adequate delineation of the plume as well as knowledge of the plume state
over time.
McClellan Air Force Base 3 MAROS 2.0 Application
Sacramento Valley, California Monitoring Network Optimization
-------
GROUNDWATER
June 2, 2003 SERVICES, INC.
1.0 INTRODUCTION
Long-term monitoring programs, whether applied for process control, performance
measurement, or compliance purposes, require large scale data collection effort and
time commitment, making their cumulative costs very high. With the increasing use of
risk-based goals and natural attenuation in recent years as well as the move toward
long-term closure upon completion of cleanup activities, the need for better-designed
long-term monitoring plans that are cost-effective, efficient, and protective of human and
ecological health has greatly increased. AFCEE's Monitoring and Remediation
Optimization System (MAROS) methodology provides an optimal monitoring network
solution, given the parameters within a complicated groundwater system which will
increase its effectiveness. By applying statistical techniques to existing historical and
current site analytical data, as well as considering hydrogeologic factors and the location
of potential receptors, the software suggests an optimal plan along with an analysis of
individual monitoring wells for the current monitoring system. This report summarizes
the findings of an application of the MAROS 2.0 software to the current Zone A and
Zone B OU D long-term monitoring well network at the McClellan Air Force Base,
Sacramento Valley, California.
1.1 Geology/Hydrogeology
McClellan Air Force Base (AFB) is located in the Sacramento Valley, approximately
seven miles northeast of Sacramento, California. The site has been divided into 8
operable units (OUs). Locations of these OUs at McClellan AFB are available in Figure
2-2 in the GMPF (Radian Corporation 1997). OU D is located in the northwest corner of
McClellan AFB.
The subsurface of McClellan AFB consists of alluvial and fluvial sediments, stretching
from the ground surface to a depth of 450 feet below ground surface (bgs). The ground
surface elevation in OU D is about 62 ft above mean sea level (msl). The subsurface
beneath McClellan AFB has been divided into the vadose zone and five monitoring
zones (A, B, C, D, and E, from shallowest to deepest) on the basis of lithologic, geologic,
and hydrologic characteristics. The monitoring zones are used to track the horizontal
migration of contaminants and to monitor local variations in hydraulic gradient. A
generalized geologic cross-section illustrating the designated monitoring zones is
presented in Figure 2-3 of the GMPF (Radian Corporation 1997). In OU D, the plume
has only impacted monitoring Zones A and B. Monitoring Zone A of OU D has an
average depth of 35 ft, ranging from 99 to 134 ft bgs; monitoring Zone B of OU D has an
average depth of 60 ft, ranging from 134 to 194 ft bgs (Table 2-3 in the GMPF, Radian
Corporation 1999).
Base-wide data collected during remedial investigations and groundwater sampling
efforts indicate that groundwater from 100 to 425 feet bgs beneath McClellan AFB is one
hydraulic system. Fine-grained deposits used to define the monitoring zones are not
continuous and allow groundwater movement and contaminant migration between
monitoring zones. The water elevation within the aquifer system has been declining
McClellan Air Force Base 4 MAROS 2.0 Application
Sacramento Valley, California Monitoring Network Optimization
-------
GROUNDWATER
June 2, 2003 SERVICES, INC.
continuously for approximately 50 years due to overdrawing by irrigation, pumping for
municipal water supply, and extraction wells. Within the last decade, water levels in
Zone A across the base have been declining at a rate of 1 to 2 feet per year.
Groundwater elevations rise and fall by an average of 5 feet due to seasonal
fluctuations. Groundwater elevations range from -38 to -40 ft msl at the northern and
southern edges of OU D, respectively, and the horizontal hydraulic gradient at OU D is
about 0.0006 ft/ft (1Q02 Monitoring Report, URS 2002).
Flow directions in the hydraulic system have varied over the past 8 decades, but have
persisted in a south to southwesterly direction over the past decade. The groundwater
flow direction for both monitoring Zone A and Zone B is predominantly toward the South-
Southwest and the groundwater seepage velocity is approximately 35 ft/yr. Base wells,
domestic production wells, extraction wells, and regional pumping affect local
groundwater flow directions. The vertical hydraulic gradients between monitoring zones
A and B are predominantly upward in the winter and downward during the rest of the
year. The horizontal hydraulic conductivity of layered sediments is about 5 to 15 times
the vertical hydraulic conductivity. For a detailed description of site geology and
hydrogeology refer to Radian Corporation (1997).
1.2 Remedial Action
McClellan Air Force Base was established in 1936 as an aircraft repair depot and supply
base. McClellan AFB and the OUs that have been established were put on the National
priorities List (NPL) in 1988. OU D located in the northwestern portion of the base is
approximately 192 acres and consists primarily of several disposal pits, sludge/oil pits,
fuel solvent pits, runway access, and an Industrial Wastewater Treatment Plant sludge
land farm. The Rl for OU D was complete in 1994 and indicated TCE is the main
constituent of concern for a single plume that extended approximately 1,750 feet off-
base to the northwest in Zones A and B.
A pump-and-treat system with 6 extraction wells was installed in OU D in 1987,
continuing to the present. The objective of the remediation is to restore Zones A and B
to drinking water standards by reducing the TCE concentration to less than the MCL (5
ppb). The groundwater long-term monitoring plan consists of 33 monitoring wells in Zone
A, 14 monitoring wells in Zone B and 6 extraction wells in Zone AB. It consists of
performance monitoring and compliance monitoring with the following goals: 1) plume
containment monitoring to confirm that the plume remains hydraulically controlled; and
2) plume reduction monitoring to verify progress toward achieving cleanup goals.
The number of monitoring wells that are currently sampled include 32 Zone A monitoring
wells, 14 Zone B monitoring wells, and 6 extraction wells screened in both Zone A and
Zone B (Figures 1 and 2). The sampling frequency for these wells has been very
irregular, ranging from quarterly or annual in 1990 to annual or biennial in 2000, although
the sampling was conducted on a quarterly basis. The frequency of sampling at the
base has been changed over time as part of the McClellan AFB Remediation Monitoring
Decision Logic implemented as part of the Installation Restoration Program (IRP)
McClellan Air Force Base 5 MAROS 2.0 Application
Sacramento Valley, California Monitoring Network Optimization
-------
June 2, 2003
If
GROUNDWATER
SERVICES, INC.
Groundwater Monitoring Plan (Radian Corporation, 1997). Some monitoring wells have
only 5 ~ 7 data records available during this period. This resulted in only a portion of the
wells being sampled on each quarterly sampling event. The MAROS 2.0 analysis
performed for this study utilizes the data from the current McClellan long-term monitoring
plan (1990 to 2000) and is summarized in Table 1. The 2001 data were not used
because a new sampling technique was being tested (passive diffusion sampling) in
sample collection and results from these sample events were not consistent with
previous results.
McClellan Air Force Base
Sacramento Valley, California
MAROS 2.0 Application
Monitoring Network Optimization
-------
GROUNDWATER
June 2, 2003 SERVICES, INC.
2.0 MAROS METHODOLOGY
The MAROS 2.0 software used to optimize the LTM network at the McClellan AFB is
explained in general terms in this section. MAROS is a collection of tools in one
software package that is used in an explanatory, non-linear fashion. The tool includes
models, statistics, heuristic rules, and empirical relationships to assist the user in
optimizing a groundwater monitoring network system while maintaining adequate
delineation of the plume as well as knowledge of the plume state over time. Different
users utilize the tool in different ways and interpret the results from a different viewpoint.
For a detailed description of the structure of the software and further utilities, refer to the
MAROS 2.0 Manual (Aziz et al. 2002).
2.1 MAROS Conceptual Model
In MAROS 2.0, two levels of analysis are used for optimizing long-term monitoring plans:
1) an overview statistical evaluation with interpretive trend analysis based on temporal
trend analysis and plume stability information; and 2) a more detailed statistical
optimization based on spatial and temporal redundancy reduction methods (see Figure 2
for further details). In general, the MAROS method applies to 2-D aquifers that have
relatively simple site hydrogeology. However, for a multi-aquifer (3-D) system, the user
could apply the statistical analysis layer-by-layer.
The overview statistics or interpretive trend analysis assesses the general monitoring
system category by considering individual well concentration trends, overall plume
stability, hydrogeologic factors (e.g., seepage velocity, and current plume length), and
the location of potential receptors (e.g., property boundaries or drinking water wells). The
analysis relies on temporal trend analysis to assess plume stability, which is then used
to determine the general monitoring system category. Since the temporal trend analysis
focuses on where the monitoring well is located, the site wells are divided into two
different zones: the source zone or the tail zone. The source zone includes areas with
non-aqueous phase liquids (NAPLs), contaminated vadose zone soils, and areas where
aqueous-phase releases have been introduced into ground water. The tail zone is
usually the area downgradient of the contaminant source zone. Although this
classification is a simplification of the well location, this broadness makes the user aware
on an individual well basis that the concentration trend results can have a different
interpretation depending on the well location in and around the plume. The location and
type of the individual wells allows further interpretation of the trend results, depending on
what type of well is being analyzed (e.g., remediation well, leading plume edge well, or
monitoring well). General recommendations for the monitoring network frequency and
density are suggested based on heuristic rules applied to the source and tail trend
results.
The detailed statistics level of analysis or sampling optimization, on the other hand,
consists of a well redundancy analysis and well sufficiency analysis using the Delaunay
method, a sampling frequency analysis using the Modified Cost Effective Sampling
(CES) method and a data sufficiency analysis using power analysis. The well
McClellan Air Force Base 7 MAROS 2.0 Application
Sacramento Valley, California Monitoring Network Optimization
-------
GROUNDWATER
June 2, 2003 SERVICES, INC.
redundancy analysis is designed to minimize monitoring locations and the Modified CES
method is designed to minimize the frequency of sampling. The data sufficiency
analysis uses power analysis to assess the sampling record to determine if the current
monitoring network and record is sufficient in terms of evaluating risk-based site target
level status.
2.2 Data Management
In MAROS, ground water monitoring data can be imported from simple database-format
Microsoft® Excel spreadsheets, Microsoft Access tables, previously created MAROS
database archive files, or entered manually. Compliance monitoring data interpretation in
MAROS is based on historical ground water monitoring data from a consistent set of
wells over a series of sampling events. Statistical validity of the concentration trend
analysis requires constraints on the minimum data input of at least four wells (ASTM
1998) in which COCs have been detected. Individual sampling locations need to include
data from at least the six most-recent sampling events. To ensure a meaningful
comparison of COC concentrations over time and space, both data quality and data
quantity need to be considered. Prior to statistical analysis, the user can consolidate
irregularly sampled data or smooth data that might result from seasonal fluctuations or a
change in site conditions.
Imported ground water monitoring data and the site-specific information entered in Site
Details can be archived and exported as MAROS archive files. These archive files can
be appended as new monitoring data becomes available, resulting in a dynamic long-
term monitoring database that reflects the changing conditions at the site (i.e.
biodegradation, compliance attainment, completion of remediation phase, etc.).
2.3 Site Details
Information needed for the MAROS analysis includes site-specific parameters such as
seepage velocity and current plume length. Part of the trend analysis methodology
applied in MAROS focuses on where the monitoring well is located, therefore the user
needs to divide site wells into two different zones: the source zone or the tail zone. The
source zone includes areas with non-aqueous phase liquids (NAPLs), contaminated
vadose zone soils, and areas where aqueous-phase releases have been introduced into
ground water. The source zone generally contains locations with historical high ground
water concentrations of the COCs. The tail zone is usually the area downgradient of the
contaminant source zone. It is up to the user to make further interpretation of the trend
results, depending on what type of well is being analyzed (e.g., remediation well, leading
plume edge well, or monitoring well).
MAROS allows the analysis of up to 5 COCs concurrently and users can pick COCs
from a list of compounds existing in the monitoring data, or select COCs based on
recommendations provided in MAROS based on toxicity, prevalence, and mobility of
compounds.
McClellan Air Force Base 8 MAROS 2.0 Application
Sacramento Valley, California Monitoring Network Optimization
-------
GROUNDWATER
June 2, 2003 SERVICES, INC.
2.4 Data Consolidation
Typically long-term monitoring raw data have been measured irregularly in time or
contain many non-detects, trace level results, and duplicates. Therefore, before the data
can be further analyzed, raw data are filtered, consolidated, transformed, and possibly
smoothed to allow for a consistent dataset meeting the minimum data requirements for
statistical analysis mentioned previously.
MAROS allows users to specify the period of interest in which data will be consolidated
(i.e., monthly, bi-monthly, quarterly, semi-annual, yearly, or a biennial basis). In
computing the representative value when consolidating, one of four statistics can be
used: median, geometric mean, mean, and maximum. Non-detects can be transformed
to one half the reporting or method detection limit (DL), the DL, or a fraction of the DL.
Trace level results can be represented by their actual values, one half of the DL, the DL,
or a fraction of their actual values. Duplicates are reduced in MAROS by one of three
ways: assigning the average, maximum, or first value. The reduced data for each COC
and each well can be viewed as a time series in a graphical form on a linear or semi-log
plot generated by the software.
2.5 Overview Statistics: Plume Trend Analysis
Within the MAROS software there are historical data analyses that support a conclusion
about plume stability (e.g., increasing plume, etc.) through statistical trend analysis of
historical monitoring data. Plume stability results are assessed from time-series
concentration data with the application of three statistical tools: Mann-Kendall Trend
analysis, linear regression trend analysis and moment analysis. The two trend methods
are used to estimate the concentration trend for each well and each COC based on a
statistical trend analysis of concentrations versus time at each well (Figure 2). These
trend analyses are then consolidated to give the user a general plume stability and
general monitoring frequency and density recommendations (see Figure 3 for further
step-by-step details). Both qualitative and quantitative plume information can be gained
by these evaluations of monitoring network historical data trends both spatially and
temporally. The MAROS Overview Statistics are the foundation the user needs to make
informed optimization decisions at the site. The Overview Statistics are designed to
allow site personnel to develop a better understanding of the plume behavior over time
and understand how the individual well concentration trends are spatially distributed
within the plume. This step allows the user to gain information that will support a more
informed decision to be made in the next level or detailed statistics optimization analysis
(Figure 2).
2.5.1 Mann-Kendall Analysis
The Mann-Kendall test is a non-parametric statistical procedure that is well suited for
analyzing trends in data over time. The Mann-Kendall test can be viewed as a
McClellan Air Force Base 9 MAROS 2.0 Application
Sacramento Valley, California Monitoring Network Optimization
-------
GROUNDWATER
June 2, 2003 SERVICES, INC.
nonparametric test for zero slope of the first-order regression of time-ordered
concentration data versus time. The Mann-Kendall test does not require any
assumptions as to the statistical distribution of the data (e.g. normal, lognormal, etc.)
and can be used with data sets which include irregular sampling intervals and missing
data. The Mann-Kendall test is designed for analyzing a single groundwater constituent,
multiple constituents are analyzed separately. The Mann-Kendall S statistic measures
the trend in the data: positive values indicate an increase in concentrations over time
and negative values indicate a decrease in concentrations over time. The strength of the
trend is proportional to the magnitude of the Mann-Kendall statistic (i.e., a large value
indicates a strong trend). The confidence in the trend is determined by consulting the S
statistic and the sample size n in a Kendall probability table such as the one reported in
Hollander and Wolfe (1973).
The concentration trend is determined for each well and each COC based on results of
the S statistic, the confidence in the trend, and the Coefficient of Variation (COV). The
decision matrix for this evaluation is shown in Table 2. A Mann-Kendall statistic that is
greater than 0 combined with a confidence of greater than 95% is categorized as an
Increasing trend while a Mann-Kendall statistic of less than 0 with a confidence between
90% and 95% is defined as a Probably Increasing trend, and so on.
Depending on statistical indicators, the concentration trend is classified into six
categories:
• Decreasing (D),
• Probably Decreasing (PD),
• Stable (S),
• No Trend (NT),
• Probably Increasing (PI)
• Increasing (I).
These trend estimates are then analyzed to identify the source and tail region overall
stability category (see Figure 2 for further details).
2.5.2 Linear Regression Analysis
Linear Regression is a parametric statistical procedure that is typically used for
analyzing trends in data over time. Using this type of analysis, a higher degree of
scatter simply corresponds to a wider confidence interval about the average log-slope.
Assuming the sign (i.e., positive or negative) of the estimated log-slope is correct, a level
of confidence that the slope is not zero can be easily determined. Thus, despite a poor
goodness of fit, the overall trend in the data may still be ascertained, where low levels of
confidence correspond to "Stable" or "No Trend" conditions (depending on the degree of
scatter) and higher levels of confidence indicate the stronger likelihood of a trend. The
linear regression analysis is based on the first-order linear regression of the log-
transformed concentration data versus time. The slope obtained from this log-
transformed regression, the confidence level for this log-slope, and the COV of the
untransformed data are used to determine the concentration trend. The decision matrix
for this evaluation is shown in Table 3. To estimate the confidence in the log-slope, the
McClellan Air Force Base 10 MAROS 2.0 Application
Sacramento Valley, California Monitoring Network Optimization
-------
GROUNDWATER
June 2, 2003 SERVICES, INC.
standard error of the log-slope is calculated. The coefficient of variation, defined as the
standard deviation divided by the average, is used as a secondary measure of scatter to
distinguish between "Stable" or "No Trend" conditions for negative slopes. The Linear
Regression Analysis is designed for analyzing a single groundwater constituent; multiple
constituents are analyzed separately, (up to five COCs simultaneously). For this
evaluation, a decision matrix developed by Groundwater Services, Inc. is also used to
determine the "Concentration Trend" category (plume stability) for each well.
Depending on statistical indicators, the concentration trend is classified into six
categories:
Decreasing (D),
Probably Decreasing (PD),
Stable (S),
No Trend (NT),
Probably Increasing (PI)
Increasing (I).
The resulting confidence in the trend, together with the log-slope and the COV of the
untransformed data, are used in the linear regression analysis decision matrix to
determine the concentration trend. For example, a positive log-slope with a confidence
of less than 90% is categorized as having No Trend whereas a negative log-slope is
considered Stable if the COV is less than 1 and categorized as No Trend if the COV is
greater than 1.
2.5.3 Overall Plume Analysis
General recommendations for the monitoring network frequency and density are
suggested based on heuristic rules applied to the source and tail trend results.
Individual well trend results are consolidated and weighted by the MAROS software
according to user input, and the direction and strength of contaminant concentration
trends in the source zone and tail zone for each COC are determined. Based on
i) the consolidated trend analysis,
ii) hydrogeologic factors (e.g., seepage velocity), and
iii) location of potential receptors (e.g., wells, discharge points, or property
boundaries),
the software suggests an general optimization plan for the current monitoring system in
order to efficiently effectively monitor in the future. A flow chart of the MAROS
methodology utilizing trend analysis results and other site-specific parameters to form a
general sampling frequency and well density recommendation is outlined in Figure 3. For
example, a generic plan for a shrinking petroleum hydrocarbon plume (BTEX) in a slow
hydrogeologic environment (silt) with no nearby receptors would entail minimal, low
frequency sampling of just a few indicators. On the other hand, the generic plan for a
chlorinated solvent plume in a fast hydrogeologic environment that is expanding but has
very erratic concentrations over time would entail more extensive, higher frequency
sampling. The generic plan is based on a heuristically derived algorithm for assessing
McClellan Air Force Base 11 MAROS 2.0 Application
Sacramento Valley, California Monitoring Network Optimization
-------
GROUNDWATER
June 2, 2003 SERVICES, INC.
future sampling duration, location and density that takes into consideration plume
stability. For a detailed description of the heuristic rules used in the MAROS software,
refer to the MAROS 2.0 Manual (Aziz et al. 2002).
2.5.3 Moment Analysis
An analysis of moments can help resolve plume trends, where the zeroth moment shows
change in dissolved mass vs. time, the first moment shows the center of mass location
vs. time, and the second moment shows the spread of the plume vs. time. Moment
calculations can predict how the plume will change in the future if further statistical
analysis is applied to the moments to identify a trend (in this case, Mann Kendall Trend
Analysis is applied). The trend analysis of moments can be summarized as:
• Zeroth Moment: Change in dissolved mass over time
• First Moment: Change in the center of mass location over time
• Second Moment: Spread of the plume over time
The role of moment analysis in MAROS is to provide a relative measure of plume
stability and condition. Plume stability may vary by constituent, therefore the MAROS
moment analysis can be used to evaluate multiple COCs simultaneously in order to
provide used to provide a quick way of comparing individual plume parameters to
determine the size and movement of constituents relative to one another. Moment
analysis in the MAROS software can also be used to assist the user in evaluating the
impact on plume delineation in future sampling events by removing identified
"redundant" wells from a long-term monitoring program (this analysis was not performed
as part of this study, for more details on this application of moment analysis refer to the
MAROS 2.0 Manual (Aziz et al. 2002).
The zeroth moment is a mass estimate. The zeroth moment calculation can show high
variability over time, largely due to the fluctuating concentrations at the most
contaminated wells as well as varying monitoring well network. Plume analysis and
delineation based exclusively on concentration can exhibit a fluctuating degree of
temporal and spatial variability. The mass estimate is also sensitive to the extent of the
site monitoring well network over time. The zeroth moment trend over time is determined
by using the Mann-Kendall Trend Methodology. The zeroth Moment trend test allows
the user to understand how the plume mass has changed over time. Results for the
trend include: Increasing, Probably Increasing, No Trend, Stable, Probably Decreasing,
Decreasing or Not Applicable (Insufficient Data). When considering the results of the
Zeroth moment trend, the following factors should be considered which could effect the
calculation and interpretation of the plume mass over time: 1) Change in the spatial
distribution of the wells sampled historically 2) Different wells sampled within the well
network over time (addition and subtraction of well within the network). 3) Adequate
versus inadequate delineation of the plume over time.
The first moment estimates the center of mass, coordinates (Xc and Yc) for each
sample event and COC. The changing center of mass locations indicate the movement
McClellan Air Force Base 12 MAROS 2.0 Application
Sacramento Valley, California Monitoring Network Optimization
-------
GROUNDWATER
June 2, 2003 SERVICES, INC.
of the center of mass over time. Whereas, the distance from the original source location
to the center of mass locations indicate the movement of the center of mass over time
relative to the original source. Calculation of the first moment normalizes the spread by
the concentration indicating the center of mass. The first moment trend of the distance to
the center of mass over time shows movement of the plume in relation to the original
source location over time. Analysis of the movement of mass should be viewed as it
relates to 1) the original source location of contamination 2) the direction of groundwater
flow and/or 3) source removal or remediation. Spatial and temporal trends in the center
of mass can indicate spreading or shrinking or transient movement based on seasonal
variation in rainfall or other hydraulic considerations. No appreciable movement or a
neutral trend in the center of mass would indicate plume stability. However, changes in
the first moment over time do not necessarily completely characterize the changes in the
concentration distribution (and the mass) over time. Therefore, in order to fully
characterize the plume the First Moment trend should be compared to the Zeroth
moment trend (mass change over time).
The second moment indicates the spread of the contaminant about the center of mass
(Sxx and Syy), or the distance of contamination from the center of mass for a particular
COC and sample event. The Second Moment represents the spread of the plume over
time in both the x and y directions. The Second Moment trend indicates the spread of
the plume about the center of mass. Analysis of the spread of the plume should be
viewed as it relates to the direction of groundwater flow. An increasing trend in the
second moment indicates an expanding plume, whereas a declining trend in the plume
indicates a shrinking plume. No appreciable movement or a neutral trend in the center of
mass would indicate plume stability. The second moment provides a measure of the
spread of the concentration distribution about the plume's center of mass. However,
changes in the second moment over time do not necessarily completely characterize the
changes in the concentration distribution (and the mass) over time. Therefore, in order to
fully characterize the plume the Second Moment trend should be compared to the zeroth
moment trend (mass change over time).
2.6 Detailed Statistics: Optimization Analysis
Although the overall plume analysis shows a general recommendation regarding
sampling frequency reduction and general sampling density, a more detailed analysis is
also available with the MAROS 2.0 software in order to allow for further reductions on a
well-by-well basis for frequency, well redundancy, well sufficiency and sampling
sufficiency. The MAROS Detailed Statistics allows for a quantitative analysis for spatial
and temporal optimization of the well network on a well-by-well basis. The results from
the Overview Statistics should be considered along with the MAROS optimization
recommendations gained from the Detailed Statistical Analysis described previously.
The MAROS Detailed Statistics results should be reassessed in view of site knowledge
and regulatory requirements as well as in consideration of the Overview Statistics
(Figure 2).
McClellan Air Force Base 13 MAROS 2.0 Application
Sacramento Valley, California Monitoring Network Optimization
-------
GROUNDWATER
June 2, 2003 SERVICES, INC.
The Detailed Statistics or Sampling Optimization MAROS module can be used to
determine the minimal number of sampling locations and the lowest frequency of
sampling that can still meet the requirements of sampling spatially and temporally for an
existing monitoring program. It also provides an analysis of the sufficiency of data for
the monitoring program.
Sampling optimization in MAROS consists of four parts:
• Well redundancy analysis using the Delaunay method
• Well sufficiency analysis using the Delaunay method
• Sampling frequency determination using the Modified CES method
• Data sufficiency analysis using statistical power analysis.
The well redundancy analysis using the Delaunay method identifies and eliminates
redundant locations from the monitoring network. The well sufficiency analysis can
determine the areas where new sampling locations might be needed. The Modified CES
method determines the optimal sampling frequency for a sampling location based on the
direction, magnitude, and uncertainty in its concentration trend. The data sufficiency
analysis examines the risk-based site cleanup status and power and expected sample
size associated with the cleanup status evaluation.
2.6.1 Well Redundancy Analysis - Delaunav Method
The well redundancy analysis using the Delaunay method is designed to select the
minimum number of sampling locations based on the spatial analysis of the relative
importance of each sampling location in the monitoring network. The approach allows
elimination of sampling locations that have little impact on the historical characterization
of a contaminant plume. The delaunay methodology application assumes that the
current sampling network adequately delineates the plume (bounding wells have non-
detect values) and that if a hydraulic containment system is currently in operation, this
will continue. An extended method or wells sufficiency analysis, based on the Delaunay
method, can also be used for recommending new sampling locations. Details about the
Delaunay method can be found in Appendix A.2 of the MAROS Manual (AFCEE 2002).
Well redundancy analysis uses the Delaunay triangulation method to determine the
significance of the current sampling locations relative to the overall monitoring network.
The Delaunay method calculates the network Area and Average concentration of the
plume using data from multiple monitoring wells. A slope factor (SF) is calculated for
each well to indicate the significance of this well in the system (i.e. how removing a well
changes the average concentration.)
The well redundancy optimization process is performed in a stepwise fashion. Step one
involves assessing the significance of the well in the system, if a well has a small SF
(little significance to the network), the well may be removed from the monitoring network.
Step two involves evaluating the information loss of removing a well from the network. If
one well has a small SF, it may or may not be eliminated depending on whether the
McClellan Air Force Base 14 MAROS 2.0 Application
Sacramento Valley, California Monitoring Network Optimization
-------
GROUNDWATER
June 2, 2003 SERVICES, INC.
information loss is significant. If the information loss is not significant, the well can be
eliminated from the monitoring network and the process of optimization continues with
fewer wells. However if the well information loss is significant then the optimization
terminates. This sampling optimization process allows the user to assess "redundant"
wells that will not incur significant information loss on a constituent-by-constituent basis
for individual sampling events.
Before applying the Delaunay method for spatial redundancy analysis, it is important to
select the appropriate set of wells for analysis, i.e., only the wells that contribute to the
spatial delineation of the plume. For example, if wells are far from the plume and
contribute little or nothing to the delineation of the plume (e.g., some sentry wells or
background wells far from the plume), they should be excluded from the analysis. One
reason not to use these wells is that these wells usually are on the boundary of the
triangulation and are hard to be eliminated since the Delaunay method protects
boundary wells from being easily removed. The elimination status of these wells, in fact,
should be determined from the regulatory standpoint. Another well type that could be
excluded from analysis is one of a clustered well set because the Delaunay method is a
two-dimensional method. Generally, only one well is picked from the clustered well set to
represent the concentration at this point. This well can be the one that has the highest
concentration or is screened in the representative aquifer interval with the geologic unit.
Data from clustered wells can also be averaged to form a single sample and then used
in the Delaunay method.
2.6.2 Well Sufficiency Analysis - Delaunav Method
The well sufficiency analysis, using the Delaunay method, is designed to recommend
new sampling locations in areas within the existing monitoring network where there is a
high level uncertainty in plume concentration. Details about the well sufficiency analysis
can be found in Appendix A.2 of the MAROS Manual (AFCEE 2002).
In many cases, new sampling locations need to be added to the existing network to
enhance the spatial plume characterization. In MAROS, the method for determining new
sampling locations recommends the area for a possible new sampling location where
there is a high level of uncertainty in concentration estimation. The Slope Factor (SF)
values obtained from the redundancy reduction described above are used to calculate
the concentration estimation error at each triangle area formed in the Delaunay
triangulation. The estimated SF value at each triangle area is then classified into four
levels: Small, Moderate, Large, or Extremely large because the larger the estimated SF
value, the higher the estimation error at this area. Therefore, the triangle areas with the
estimated SF value at the Extremely large or Large level are candidate regions for new
sampling locations.
The results from the Delaunay method and the method for determining new sampling
locations are derived solely from the spatial configuration of the monitoring network and
the spatial pattern of the contaminant plume. No parameters such as the hydrogeologic
McClellan Air Force Base 15 MAROS 2.0 Application
Sacramento Valley, California Monitoring Network Optimization
-------
GROUNDWATER
June 2, 2003 SERVICES, INC.
conditions are considered in the analysis. Therefore, professional judgement and
regulatory considerations must be used to make final decisions.
2.6.3 Sampling Frequency Determination - Modified CES Method
The Modified Cost Effective Sampling (MCES) method optimizes sampling frequency for
each sampling location based on the magnitude, direction, and uncertainty of its
concentration trend derived from its recent and historical monitoring records. The MCES
estimates the lowest-frequency sampling schedule for a given groundwater monitoring
location yet still provide needed information for regulatory and remedial decision-making.
The Modified CES method was developed on the basis of the Cost Effective Sampling
(Ridley et al. 1995). Details about the Modified CES method can be found in Appendix
A.3 of the MAROS Manual (AFCEE 2002).
In order to estimate the least frequent sampling schedule for a monitoring location that
still provides enough information for regulatory and remedial decision-making, MCES
employs three steps to determine the sampling frequency. The first step involves
analyzing frequency based on recent trends (Figure 4). A preliminary location sampling
frequency (PLSF) is determined based on the trends determined by rates of change
from linear regression and Mann-Kendall analysis of the most recent monitoring data.
The variability of the sequential sampling data is accounted for by the Mann-Kendall
analysis. The PLSF is then adjusted based on overall trends. If the long-term history of
change is significantly greater than the recent trend, the frequency may be reduced by
one level. Otherwise, no change could be made. The final step in the analysis involves
reducing frequency based on risk. Since not all compounds in the target being
assessed are equally harmful, frequency is reduced by one level if recent maximum
concentration for compound of high risk is less than 1/2 of the Maximum Concentration
Limit (MCL). The result of applying this method is a suggested sampling frequency
based on recent sampling data trends and overall sampling data trends.
The finally determined sampling frequency from the Modified CES method can be
Quarterly, Semiannual, Annual, and Biennial. Users can further reduce the sampling
frequency to, for example, once every three years, if the trend estimated from Biennial
data (i.e., data drawn once every two years from the original data) is the same as that
estimated from the original data.
2.6.4 Data Sufficiency Analysis - Power Analysis
Statistical power analysis is a technique for interpreting the results of statistical tests. It
provides additional information about a statistical test: 1) the power of the statistical test,
i.e., the probability of finding a difference in the variable of interest when a difference
truly exists; and 2) the expected sample size of a future sampling plan given the
minimum detectable difference it is supposed to detect. For example, if the mean
McClellan Air Force Base 16 MAROS 2.0 Application
Sacramento Valley, California Monitoring Network Optimization
-------
June 2, 2003
If
GROUNDWATER
SERVICES, INC.
concentration is lower than the cleanup goal but a statistical test cannot prove this, the
power and expected sample size can tell the reason and how many more samples are
needed to result in a significant test. The additional samples can be obtained by a
longer period of sampling or an increased sampling frequency. Details about the data
sufficiency analysis can be found in Appendix A.6 of the MAROS Manual (AFCEE 2002).
When applying the MAROS power analysis method, a hypothetical statistical compliance
boundary (HSCB) is assigned to be a line perpendicular to the groundwater flow
direction (see figure below). Monitoring well concentrations are projected onto the
HSCB using the distance from each well to the compliance boundary along with a decay
coefficient. The projected concentrations from each well and each sampling event are
then used in the risk-based power analysis. Since there may be more than one sampling
event selected by the user, the risk-based power analysis results are given on an event-
by-event basis. This power analysis can then indicate if target are statistically achieved
at the HSCB. For instance, at a site where the historical monitoring record is short with
few wells, the HSCB would be distant; whereas, at a site with longer duration of
sampling with many wells, the HSCB would be close. Ultimately, at a site the goal would
be to have the HSCB coincide with or be within the actual compliance boundary
(typically the site property line).
®-
L
c>
Groundwater flow direction
Concentrations
projected to this
•* line
The nearest
downgradient
receptor
In order to perform a risk-based cleanup status evaluation for the whole site, a strategy
was developed as follows.
• Estimate concentration versus distance decay coefficient from plume centerline
wells.
• Extrapolate concentration versus distance for each well using this decay
coefficient.
• Comparing the extrapolated concentrations with the compliance concentration
using power analysis.
McClellan Air Force Base
Sacramento Valley, California
17
MAROS 2.0 Application
Monitoring Network Optimization
-------
GROUNDWATER
June 2, 2003 SERVICES, INC.
Results from this analysis can be Attained or Not Attained, providing a statistical
interpretation of whether the cleanup goal has been met on the site-scale from the risk-
based point of view. The results as a function of time can be used to evaluate if the
monitoring system has enough power at each step in the sampling record to indicate
certainty of compliance by the plume location and condition relative to the compliance
boundary. For example, if results are Not Attained at early sampling events but are
Attained in recent sampling events, it indicates that the recent sampling record provides
a powerful enough result to indicate compliance of the plume relative to the location of
the receptor or compliance boundary.
McClellan Air Force Base IS MAROS 2.0 Application
Sacramento Valley, California Monitoring Network Optimization
-------
GROUNDWATER
June 2, 2003 SERVICES, INC.
3.0 SITE RESULTS
The groundwater long-term monitoring plan for McClellan AFB was started in 1990. The
monitoring plan consisted of performance monitoring and compliance monitoring with the
following goals:
1) plume containment monitoring to confirm that the TCE plume remains
hydraulically controlled; and
2) plume reduction monitoring to verify progress toward achieving cleanup goals.
32 monitoring wells in Zone A were included in the long-term monitoring network as of
2000 along with 14 monitoring wells in Zone B, and 6 extraction wells screened in both
Zone A and Zone B (Figures 1 and 2). The sampling frequency for these wells has been
irregular, ranging from quarterly or annual in 1990 to annual or biennial in 2000, although
the sampling was conducted on a quarterly basis. Some monitoring wells have only 5 ~
7 data records available during this period. This resulted in only a portion of the wells
being sampled on each quarterly sampling event.
Monitoring data from 1990 to 2000 were used for the detailed optimization analysis, with
a subset of this data used in some of the analyses. The 2001 data were not used in the
overview analysis or the detailed analysis because a new sampling technique was being
tested (passive diffusion sampling) in sample collection and results from these sample
events were not consistent with previous results.
In applying the MAROS methodology to develop a revised monitoring strategy for the
McClellan AFB Zones A and B, many site and dataset parameters were applied. General
site assumptions include:
• All wells that were part of the network in between 1990 and 2000 were
considered in the temporal concentration trend analysis.
• Five chemicals of concern (COCs) that have been historically present at the site:
tetrachloroethene (PCE), trichloroethene (TCE), cis-1,2-dichloroethene (cis-1,2-
DCE), and 1,2-dichloroethane (DCA), however, TCE is the only constituent that
is currently above the MCL in the OU D plume.
• All source/tail assignments were made based on the TCE Plume. Source wells
were selected based on historically elevated concentrations of TCE.
• Site-specific hydrogeologic parameters related to Zones A and B including
groundwater flow direction, seepage velocity, saturated thickness, porosity,
receptor locations, can be found in the Table 4.
• Monitoring data from 1990 to 2000 were used for the "overall" trend analysis in
the sampling frequency optimization, and data from January, 1995 to December,
2000 "recent" trend analysis and other analyses in the MAROS detailed
optimization analysis.
McClellan Air Force Base 19 MAROS 2.0 Application
Sacramento Valley, California Monitoring Network Optimization
-------
GROUNDWATER
June 2, 2003 SERVICES, INC.
• There will be a continuation of the current hydraulic containment system in place
for the near future of the monitoring network.
• The current monitoring network adequately delineates the plume at the site for
the constituent of concern, TCE.
3.1 Data Consolidation
In MAROS, ground water monitoring data can be imported from simple database-format
Microsoft® Excel spreadsheets, Microsoft® Access tables, previously created MAROS
database archive files, or entered manually. The historical monitoring data from
McClellan were received in Excel database format. The URS, 2002 report defined the
wells into Zone A, Zone B and Zone AB (URS 2002, Tables 3.2 and 3.3). The columns
in the file where formatted to the MAROS Access file import format and then imported
into the MAROS software using the import tool. The long-term monitoring raw data
contained many non-detects, trace level results, and duplicates. Therefore, in the
MAROS software the raw data are filtered, consolidated, and the period of interest was
specified (i.e. monitoring data from 1990 to 2000) as well as the wells of interest for the
zone of interest. The MAROS analysis was applied separately for Zone A and Zone B
monitoring networks. For statistical evaluation of the data, a representative value for
each sample point in time is needed. MAROS has many automated options to choose
how these values are assigned. For the McClellan data, non-detects values were
chosen to be set to the minimum detection limit, allowing for uniform detection limits over
time. Trace level results were chosen to be represented by their actual values and
duplicates samples were chosen to be assigned the average of the two samples. The
reduced data for each well were viewed as a time series in a graphical form on a linear
or semi-log plot generated by the software.
3.2 Overview Statistics: Plume Trend Analysis
3.2.1 Mann-Kendall/Linear Regression Analysis
The goal of the Mann-Kendall and Linear Regression temporal trend analysis is to
assess the historical trend in the concentrations over time. These trend estimates are
then analyzed to identify the source and tail region overall stability category as well as
gaining an understanding of the individual well concentrations overtime (see Figure 3 for
further details). The TCE historical data for monitoring wells in both Zones A and B as
well as the extraction wells in Zone AB were assessed for trends. No data consolidation
was performed to condense the sampling into regular sample intervals.
Zone A
All 32 monitoring wells in Zone A had sufficient data within the time period of 1990 to
2000 (greater than 6 sample events) to assess the trends in the wells. Trend results
from the Mann-Kendall and Linear Regression temporal trend analysis for Zone A
monitoring wells are given in Table 5. The monitoring well trend results for Zone A show
McClellan Air Force Base 20 MAROS 2.0 Application
Sacramento Valley, California Monitoring Network Optimization
-------
June 2, 2003
If
GROUNDWATER
SERVICES, INC.
that 9 out of 10 source wells and 9 out of 22 tail wells have a Probably Decreasing,
Decreasing, or Stable trend. Both methods gave similar trend estimates for each well.
When considering the spatial distribution of the trend results (Figures 6 and 7 - maps
created in ArcGIS from MAROS results), the majority of the decreasing or stable trend
results are located near in the source area, indicating a decreasing source region
concentration. Areas with no trend tended to be in the tail or edge of the plume where
the wells have been sampled less frequently.
Well Type
Source
Tail
Extraction (Zone AB)
Zone A MAROS Trend Analysis
PD, D, S
9 of 10 (90%)
1 1 of 22 (50%)
5 of 6 (83%)
I, PI
Oof 10(0%)
1 of 22 (5%)
0 of 6 (0%)
Note: Decreasing (D), Probably Decreasing (PD), Stable (S), Probably Increasing (PI), and Increasing (I)
Zone B
All 14 monitoring wells in Zone B had sufficient data within the time period of 1990 to
2000 (greater than 6 sample events) to assess the trends in the wells. Trend results
from the Mann-Kendall and Linear Regression temporal trend analysis for Zone B
monitoring wells are given in Table 6. The majority of the wells in Zone B have no trend
in the historical data. However, as of 2000, only one of the wells in Zone B is actually
above the MCL for TCE. Both of the statistical methods used to evaluate trends (Mann-
Kendall and Linear Regression) gave similar trend estimates for each well. When
considering the spatial distribution of the trend results (Figures 8 and 9 - maps created
in ArcGIS from MAROS results), the majority of the decreasing or stable trend results
are located near in the source area, indicating a decreasing source region concentration.
Areas with no trend tended to be in the tail or edge of the plume where the wells have
been sampled less frequently.
Well Type
Source
Tail
Zone B MAROS Trend Analysis
PD, D, S
0 of 1 (0%)
6 of 13 (50%)
1, PI
1 of 1 (100%)
1 of 13(8%)
Note: Decreasing (D), Probably Decreasing (PD), Stable (S), Probably Increasing (PI), and Increasing (I)
Zone AB
All 6 extraction wells had sufficient data within the time period of 1990 to 2000 (greater
than 6 sample events) to assess the trends in the wells. Trend results from the Mann-
Kendall and Linear Regression temporal trend analysis for Zone AB extraction wells are
given in Table 5. The extraction well trend results show that 5 out of 6 wells have a
McClellan Air Force Base
Sacramento Valley, California
21
MAROS 2.0 Application
Monitoring Network Optimization
-------
GROUNDWATER
June 2, 2003 SERVICES, INC.
Probably Decreasing, Decreasing, or Stable trend. Both methods gave similar trend
estimates for each well. The extraction wells in the source mostly show decreasing or
probably decreasing trends (Figures 10 and 11 - maps created in ArcGIS from MAROS
results).
Although monitoring wells and extraction wells are present in the well network, these
well trend results need to be treated differently for the purpose of individual trend
analysis interpretation primarily due to the different course of action possible for the two
types of wells. For monitoring wells, strongly decreasing concentration trends may lead
the site manager to decrease their monitoring frequency, as well look at the well as
possibly attaining its remediation goal. Conversely, strongly decreasing concentration
trends in extraction wells may indicate ineffective or near-asymptotic contamination
extraction, which may in turn lead to either the shutting down of the well or a drastic
change in the extraction scheme. Other reasons favoring the separation of these two
types of wells in the trend analysis interpretation is the fact that they produce very
different types of samples. Typically extraction wells possess screens that are much
larger than those of the average monitoring well. Therefore, the potential for the dilution
of extraction well samples is far greater than monitoring well samples.
3.2.2 Moment Analysis
The moment analysis in the MAROS software was applied at the McClellan site in order
to gain a better understanding of the overall plume stability in both Zones A and B.
Monitoring well data from 1990 to 2000 were used for the moment analysis. Sampling
frequency for these wells was very irregular, ranging from quarterly or annual in 1990 to
annual or biennial in 2000. Therefore, all Zone A spatial moment analyses were based
on sampling events redefined on a yearly basis, that is, data collected between January
1st and December 31st of a year were treated as if from the same sampling event
performed on July 1st of that year, with the geometric mean result utilized for each
location. Whereas, all Zone B spatial moment analyses were based on sampling events
redefined on a biennial basis, that is, data collected between January 1st and December
31st of two years were treated as is from the same sampling event performed on July
1st of the first year, with the geometric mean result utilized for each location.
Zone A
Moment trend results from the Zeroth, First, and Second Moment analyses for the Zone
A monitoring well network were varied. Moment Trend results from the moment trend
analysis for selected Zone A well dataset are given in the Moment Analysis Report,
Appendix B. Approximately 32 wells were used in the Zone A moment analysis. Wells
with redundant spatial concentration information were not utilized in the moment analysis
(i.e. MW-1041).
The zeroth moment analysis showed a stable trend (no change in dissolved mass) over
time (Appendix B). The zeroth moment or mass estimate can show high variability over
time, largely due to the fluctuating concentrations at the most contaminated wells as well
McClellan Air Force Base 22 MAROS 2.0 Application
Sacramento Valley, California Monitoring Network Optimization
-------
June 2, 2003
If
GROUNDWATER
SERVICES, INC.
as a varying monitoring well network. In order to reduce the fluctuating factors that could
influence a mass trend, the data were consolidated to annual sampling and the zeroth
moment trend evaluated. Another factor to consider when interpreting the mass
increase over time is the change in the spatial distribution of the wells sampled
historically. At the McClellan OU D site there were changes in the well distribution over
time, due to addition and subtraction of wells from the well network as well as changes in
sampling frequency. Therefore, the results from the MAROS mass trend over tim at the
site should be evaluated along with the trend analysis results. The trend in mass is more
likely decreasing over time, in accordance with the decreasing trend results (decreasing
concentrations) seen in the majority of wells in the source area.
Moment
Type
Zeroth
First
Second
Zone A Mann-Kendall Trend Analysis
Trend
Zone A
Stable to Decreasing
No Trend
No Trend
Comment
The amount of dissolved mass has decreased over time.
The center of mass remained in relatively the same location through
time, with slight movement forward or backward along the direction of
groundwater flow.
Stable to no trend, indicating that wells representing very large areas
both on the tip and the sides of the plume show little conclusive change
in concentrations.
The first moment, or center of mass, for each sample event in Zone A remained
relatively stable to no trend in distance relative to the approximate source location, see
Figure 12, as well as the MAROS First Moment Reports in Appendix B. The center of
mass remained in relatively the same location through time, with slight movement
forward or backward along the direction of groundwater flow. These spatial and
temporal trends in the center of mass distance from the source location can indicate
transient movement based on season variation in rainfall or other hydraulic
considerations. With no appreciable movement or a neutral trend in center of mass as is
the case at McClellan there is additional confirmation separate from the individual well
trend analysis, that the plume is relatively stable to decreasing. This stable center of the
mass indicates that both the mass and mass movement over time, the plume itself is
stable.
Zone B
Moment trend results from the Zeroth, First, and Second Moment analyses for the Both
the Zone B monitoring well network were similar. Moment Trend results from the
moment trend analysis for selected Zone B monitoring well dataset are given in the
Moment Analysis Report, Appendix B. Approximately 12 wells were used in the Zone B
moment analysis. Wells with redundant spatial concentration information were not
utilized in the moment analysis (i.e. MW-1003 and MW-1028).
The zeroth moment analysis showed a stable trend (no change in dissolved mass) over
time (Appendix B). The zeroth moment or mass estimate can show high variability over
McClellan Air Force Base
Sacramento Valley, California
23
MAROS 2.0 Application
Monitoring Network Optimization
-------
June 2, 2003
If
GROUNDWATER
SERVICES, INC.
time, largely due to the fluctuating concentrations at the most contaminated wells as well
as a varying monitoring well network. In order to reduce the fluctuating factors that could
influence a mass trend, the data were consolidated to biennial sampling and the zeroth
moment trend evaluated. Another factor to consider when interpreting the mass
increase over time is the change in the spatial distribution of the wells sampled
historically. At the McClellan OU D site there were changes in the well distribution over
time, due to addition and subtraction of wells from the well network as well as changes in
sampling frequency. . Therefore, the results from the MAROS mass trend over time at
the site should be evaluated along with the trend analysis results. The trend in mass is
more likely decreasing over time, in accordance with the decreasing trend results
(decreasing concentrations) seen in the majority of wells.
The first moment, or center of mass, for each sample event in Zone B remained
relatively stable in distance relative to the approximate source location, see Figure 13,
as well as the MAROS First Moment Reports in Appendix B. The center of mass
remained in relatively the same location through time, with slight movement forward or
backward along the direction of groundwater flow. Similar to Zone A, these spatial and
temporal trends in the center of mass distance from the source location can indicate
transient movement based on season variation in rainfall or other hydraulic
considerations. With no appreciable movement or a neutral trend in center of mass as is
the case at McClellan there is additional confirmation separate from the individual well
trend analysis, that the plume is relatively stable to decreasing. This stable center of the
mass indicates that both the mass and mass movement over time, the plume itself is
stable.
Moment
Type
Zeroth
First
Second
Zone B Mann-Kendall Trend Analysis
Trend
Zone B
Stable to Decreasing
Stable
No Trend
Comment
The amount of dissolved mass has decreased over time.
The center of mass remained in relatively the same location through
time, with slight movement forward or backward along the direction of
groundwater flow.
Stable to no trend, indicating that wells representing very large areas
both on the tip and the sides of the plume show little conclusive change
in concentrations.
As was the case with the Zone A moment analysis, the second moment, or spread of the
plume over time in both the x and y directions for the sample events in Zone B, showed
no trend over time, Appendix B. The second moment provides a measure of the spread
of the concentration distribution about the plume's center of mass. Analysis of the
spread of the plume indicates no trend in the plume, indicating that wells representing
very large areas both on the tip and the sides of the plume show highly variable
concentrations over time or that the wells have not been sampled consistently enough to
show a clear trend. The no trend results indicate that at the McClellan OU D site the
highly variable well network from changes in the well distribution over time, due to
McClellan Air Force Base
Sacramento Valley, California
24
MAROS 2.0 Application
Monitoring Network Optimization
-------
GROUNDWATER
June 2, 2003 SERVICES, INC.
addition and subtraction of wells from the well network as well as changes in sampling
frequency, results in non-specific trend results.
McClellan Air Force Base 25 MAROS 2.0 Application
Sacramento Valley, California Monitoring Network Optimization
-------
GROUNDWATER
June 2, 2003 SERVICES, INC.
3.2.3 Overview Statistics: Plume Analysis
Overview Statistics Results:
• Overall trend for Source region: Stable,
• Overall trend for Tail region: near Stable (No Trend),
• Overall results from moment analysis indicate a stable to decreasing plume,
• Overall monitoring intensity needed: Moderate.
In evaluating overall plume stability, the trend analysis results and all monitoring wells
were assigned "Medium" weights within the MAROS software (as described in Figure 4),
assuming equal importance for each well and each trend result in the overall analysis.
These results matched with the judgment based on the visual comparison of TCE
plumes over time, as well as the Moment Analysis. The TCE concentrations observed
over the history of monitoring at the site are plotted in Appendix A. The Zone A TCE
plume observed in 1995 was very similar to that of 2000, indicating that the TCE plume
is relatively stable over time, even when the individual well concentration trends in the
MAROS analysis indicate a near stable overall plume trend.
For a generic plume, the MAROS software indicates:
• No recommendation for sampling frequency
• Zone A may need 25 wells for sampling network
• Zone B may need 25 wells for sampling network
These MAROS results are for a generic site, and are based on knowledge gained from
applying the MAROS Overview Statistics. There is no recommendation for frequency of
sampling for the whole monitoring network due to some uncertainty in the trends and the
presence of an active remediation system. Also, the recommended the number of wells
seems high when applied to each zone individually. So, although the overall plume
trend analysis shows a near stable plume, no general sampling frequency
recommendation was assessed by the MAROS software. Therefore, a more detailed
analysis was performed using the MAROS 2.0 software in order to allow for possible
reductions on a well-by-well basis, frequency and well redundancy analysis were
conducted. These overview statistics were also used when evaluating a final
recommendation for each well after the detailed statistical analysis was applied.
McClellan Air Force Base 26 MAROS 2.0 Application
Sacramento Valley, California Monitoring Network Optimization
-------
GROUNDWATER
June 2, 2003 SERVICES, INC.
3.3 Detailed Statistics: Optimization Analysis
Monitoring data from 1990 to 2000 were used for the optimization analysis. 2001 data
were not used because passive diffusion sampling was used in sample collection, with
anomalous results. Wells used in the analysis include 32 Zone A monitoring wells, 14
Zone B monitoring wells and 6 extraction wells screened in Zone AB (Table 1).
Due to the variable and irregular sampling frequency as described in Section 1.2, some
monitoring wells have only 5-7 data records available from 1990 to 2000. This
resulted in only a portion of the wells being sampled on each quarterly sampling event.
Therefore, all spatial analyses (sampling location determination and risk-based site
cleanup evaluation) were based on sampling events redefined on a yearly basis, that is,
data collected between January 1st and December 31st of a year were treated as coming
from the same sampling event performed on July 1st of that year.
In the well redundancy and well sufficiency analyses, only the monitoring wells were
used. For the sampling frequency analysis, both the monitoring wells and the extraction
wells were analyzed. Only Zone A was analyzed with the data sufficiency analysis for
evaluating the risk-based site cleanup. The data sufficiency or power analysis for Zone
B was not performed because Zone B has only one monitoring well with concentrations
above MCL. Results for well redundancy, well sufficiency, sampling frequency, and data
sufficiency analyses are detailed in the following sub-sections.
3.3.1 Well Redundancy Analysis - Delaunav Method
The goal of the well redundancy analysis is to identify wells that are spatially redundant
within monitoring network as candidates for removal from the sampling plan. The
approach allows elimination of sampling locations that have little impact on the historical
characterization of a contaminant plume.
Zone A
Among the 32 Zone A monitoring wells, 31 were used in the well redundancy analysis
(Table 1). MW-1041 was excluded from the analysis because it duplicates MW-1042
both spatially and in concentration levels over time. The Delaunay analysis was
conducted with yearly averages from the latest 6 years of data (1995 to 2000). The
MAROS results show that 3 monitoring wells (MW-14, MW-241, and MW-72) are
candidates for elimination from the existing long-term monitoring network (Table 7).
These wells are overall most redundant in the past 6 years, from the standpoint of their
contribution to the spatial definition of the plume.
After consideration of the MAROS recommendations and the need for plume and site
characterization, 3 wells were recommended for elimination from the existing 32-well
Zone A monitoring network, resulting in a reduction of 9%.
McClellan Air Force Base 27 MAROS 2.0 Application
Sacramento Valley, California Monitoring Network Optimization
-------
GROUNDWATER
June 2, 2003 SERVICES, INC.
Well removal candidates include (Table 7):
MW-1041
MW-14
MW-241
A qualitative confirmation was performed to assess that eliminating these 3 wells from
the 32-well monitoring network would not cause inadequate plume delineation and
spatial concentration representation. The Zone A TCE plume observed in 2000 was
hand contoured before and after removal of the 3 wells resulting in no significant plume
size or concentration changes, indicating that the information loss in by eliminating these
wells would be negligible. Also, the TCE plume shown in Figure 14 generated based on
the existing and optimized networks using 1999 data agree with each other quite well,
indicating that eliminating these wells from the monitoring network does not show any
significant loss of information. Therefore, information gained from the MAROS trend
analysis and a qualitative assessment of the concentration history of the wells, indicated
these wells could be removed from the Zone A monitoring network without significant
loss of information.
The MAROS software suggested eliminating MW-72 from the monitoring network,
however, there were site-specific reasons to keep the well within the monitoring network
(Table 7). Well MW-72 currently (as of the 2000 sampling) has concentration greater
than the MCL for TCE and the well is located on the plume centerline and is the basis for
risk-based power analysis for attainment at the compliance boundary. Similarly, there
was a well that was not used in the Delaunay analysis that is clustered with an
equivalent duplicate, MW-1041 is very close to MW-1042 with similar concentrations and
concentration trends over time (Table 1 and Table 7). The clustered well MW-1041 with
similar screen intervals, concentration trends, and concentration ranges to the nearby
well, MW-1042, is suggested for elimination without having used them in the MAROS
well redundancy analysis.
Zone B
Among the 14 Zone B monitoring wells, 12 were used in the Delaunay analysis (Table
1). MW-1003 and MW-1028 were excluded from the analysis because they duplicate
MW-1001 and MW-1027, respectively. The Delaunay analysis was conducted with
yearly averages from the latest 6 years (1995 to 2000). The MAROS results show that
no monitoring wells that could be eliminated from the existing long-term monitoring
network (Table 8).
After consideration of the MAROS recommendations and the need for plume and site
characterization, 2 wells were recommended for elimination from the existing 14-well
monitoring network, resulting in a reduction of 14%.
McClellan Air Force Base 28 MAROS 2.0 Application
Sacramento Valley, California Monitoring Network Optimization
-------
GROUNDWATER
June 2, 2003 SERVICES, INC.
Well removal candidates include (Table 8):
MW-1003
MW-1028
A qualitative confirmation was performed to assess that eliminating these 2 wells from
the 14-well monitoring network would not cause inadequate plume delineation and
spatial concentration representation. The Zone B TCE plume observed in 2000 was
hand contoured before and after removal of the 2 wells resulting in no significant plume
size or concentration changes, indicating that the information loss by eliminating these
wells would be negligible. Also, the TCE plume shown in Figure 15 generated based on
the existing and optimized networks using 1999 data agree with each other quite well,
indicating that eliminating these wells from the monitoring network does not show any
significant loss of information. Therefore, information gained from the MAROS trend
analysis and a qualitative assessment of the concentration history of the wells, indicated
these wells could be removed from the Zone B monitoring network without significant
loss of information.
Although these wells (MW-1003 and MW-1028) were not used in the well redundancy
analysis they were clustered wells with equivalent duplicates very close to a well that
had similar concentrations (lower than the MCL or DL) and concentration trends over
time, which indicated these wells could be eliminated without significant loss of
information. These clustered wells with similar screen intervals, concentration trends,
and concentration ranges to a nearby well were suggested for elimination without having
used them in the MAROS well redundancy analysis. However, information gained from
the MAROS trend analysis and a qualitative assessment of the concentration history of
the well, indicated these wells could be eliminated without significant loss of information.
Zones A & B
An additional redundancy analysis was performed on the McClellan well network that
excluded the wells on the extreme periphery of the network (Zone A: far up-gradient
wells: MW-1041, 1042, MW-1064; far cross-gradient wells: MW-237, MW-1026; and far
down-gradient well: MW-350; Zone B: far up-gradient wells MW-1043 and MW-1010;
far cross-gradient wells MW-1027 and MW-1028). The above analysis included these
wells in the MAROS analysis, and these wells were not recommended for removal from
the network as they are required to define the boundary of the Zone A and Zone B
plumes. However, a monitoring network that includes the periphery wells significantly
"over-captures" the plume as these wells are located far from the likely plume boundary.
The MAROS analysis that excluded these periphery wells indicated that additional wells
would be required at the down-gradient edge of the plume to define the boundary of the
plumes in Zone A and B. For example in Zone A, the area East of MW-12 is a likely
location for a new down-gradient boundary well. In the Figure 9 from the well sufficiency
analysis in the next section, almost all triangles outside the plume region are "Medium"
McClellan Air Force Base 29 MAROS 2.0 Application
Sacramento Valley, California Monitoring Network Optimization
-------
GROUNDWATER
June 2, 2003 SERVICES, INC.
in estimation errors, those areas that are close to the plume could be potential new
locations if the periphery network wells were removed.
Although the MAROS analysis indicates that new wells might be able to replace the
periphery wells, the decision to stop sampling the periphery wells should be made with
consideration to non-statistical considerations, such as regulatory, community, and/or
public health issues. Non-statistical considerations may indicate that continued
sampling of the periphery wells may be warranted.
3.3.2 Well Sufficiency Analysis - Delaunav Method
The goal of the sampling location determination was to identify wells that are redundant
within the monitoring network as candidates for removal from the sampling plan. The
approach allows elimination of sampling locations that have little impact on the historical
characterization of a contaminant plume. An extended method based on the Delaunay
method can also be used for recommending new sampling locations in areas where
additional plume information is needed.
Zone A
The well sufficiency analysis SF values obtained from the aforementioned analysis were
used to generate Figure 16, which recommends the triangular regions for placing new
sampling locations. It is seen that almost all triangular regions (except one) have S
(small) or M (medium) estimation errors. Also, considering the relatively small size of
the TCE plume, which is adequately delineated by the current monitoring system, the
current sampling locations are sufficient. Zone A well sufficiency analysis was
performed and no new locations are recommended.
Zone B
The Zone B well sufficiency analysis was not performed because only one well (MW-54)
out of the 14 monitoring wells has concentrations above the MCL for TCE. Therefore,
considering the relatively small size of the TCE plume, which is adequately delineated by
the current monitoring system, the current sampling locations are sufficient. No new
locations are recommended for the Zone B monitoring network.
3.3.3 Sampling Frequency Analysis - Modified CES Method
The sampling frequency analysis, using the Modified CES method, was applied to
optimize the sampling frequency for each sampling location based on the magnitude,
direction, and uncertainty of its concentration trend of its recent and historical TCE data.
The Modified Cost Effective Sampling is a temporal analysis that estimates the lowest-
frequency sampling schedule for a given groundwater monitoring location yet still
provide needed information for regulatory and remedial decision-making. In the
McClellan Air Force Base 30 MAROS 2.0 Application
Sacramento Valley, California Monitoring Network Optimization
-------
GROUNDWATER
June 2, 2003 SERVICES, INC.
sampling frequency analysis, sampling events were defined on a quarterly basis
(corresponding to the actual sampling schedules) so that all the data records could be
used in the temporal analysis.
Zone A
For the monitoring well system, considering all the wells prior to the well redundancy
analysis, 20 wells are recommended to be sampled biennially, and 12 annually. All 6
extraction wells are recommended to be sampled annually. Because the pre-
optimization sampling frequencies for the Zone A wells were already very low (mostly
annual and biennial), the sampling frequency optimization results in no significant
reduction and in some cases there is an increase in sampling recommended. Results
from the sampling frequency analysis for the 32 Zone A monitoring wells and the 6
extraction wells are given in Table 9. Most of the annual or lower-than-annual sampling
frequency recommendations (Table 9) were due to insufficient recent data (i.e., less than
6 data records), which prevented the MAROS estimation of concentration trend using
recent monitoring data. After considering the MAROS results and the historic and recent
concentration levels at these wells, final sampling frequency recommendations are
provided in Table 9.
In most cases, the frequency recommendations from the MAROS software were not
adopted due to data inadequacy. Sampling frequencies for all the wells was irregular,
ranging between quarterly and biennial from 1990 to 2000 (Table 1). Many monitoring
wells have only 5-7 data records available during an 11-year period (from 1990 to
2000). Because the minimum data requirement for the sampling frequency trend
analysis is 6 sampling events, the recent trends for many wells were not able to be
estimated. This resulted in frequency results that were solely estimated from the overall
data trend, which was not very reliable given the data inadequacy. For instance, well
MW-74 has only two concentration records between 1994 and 2000, making the
estimation of recent trend impossible. In cases of data inadequacy, the MAROS
frequency analysis will always assign conservative results, i.e., quarterly or semiannual
instead of annual or biennial.
However, a qualitative assessment of the concentration levels and concentration history
for these wells resulted in more reasonable sampling frequency recommendations
(Table 9). For example, well MW-1026 was suggested for a biennial sampling because
its concentrations have been below the MCL or DL since 1993. Also, the relative size of
the plume as well as the plume stability which remains stable according to the overview
statistical analysis, it is unlikely that the plume will show rapid changes over the long-
term. Therefore, keeping the frequency of most wells at annual and biennial level will
continue to allow for adequate plume delineation.
Zone B
For the Zone B monitoring well system, assessing all the wells prior to the well
redundancy analysis, 12 wells can be sampled biennially and 2 annually. In all cases
McClellan Air Force Base 31 MAROS 2.0 Application
Sacramento Valley, California Monitoring Network Optimization
-------
GROUNDWATER
June 2, 2003 SERVICES, INC.
except for well MW-1028, these sampling recommendations agree with the sampling
frequency currently performed at the site. Results from the Modified CES method for the
14 Zone B monitoring wells are given in Table 10. Similar to the MAROS Zone A
analysis, some Zone B annual sampling frequency recommendations needed to be
adjusted when taking into consideration the historic and recent concentration levels at
these wells.
3.3.4 Data Sufficiency - Power Analysis
The MAROS data sufficiency analysis indicates the current monitoring network is
sufficient in terms of evaluating risk-based site target level status, if the pump-and-treat
remedial system contains the plume and keeps reducing the TCE concentration in the
aquifer. Table 11 shows the risk-based site cleanup status at selected sampling events
for both analyses (i.e., HSCB at 1000 ft and HSCB at 100 ft downgradient of the
monitoring system) assuming normality of the projected data. The cleanup standard has
been "attained" for both HSCBs for all the years in the assessment. The high power
indicates the site-wide TCE concentration level at the HSCB is much lower than the
cleanup goal and the number of sampling points is more than sufficient. This analysis
indicates that the monitoring system is working because it is powerful enough to
accurately reflect the location of the plume relative to the compliance point.
In the MAROS data sufficiency analysis, statistical power analysis was used to assess
the sufficiency of monitoring plans for detecting the difference between the mean
concentration and cleanup goal. Results from the analysis indicate plume location from
the risk-based standpoint at a hypothetical statistical compliance boundary (HSCB). The
power and expected sample size associated with the target level evaluation may indicate
the need for expansion or redundancy reduction of future sampling plans. This analysis
was performed based on sampling events defined on a yearly basis for the Zone A
monitoring well system only.
In the risk-based site cleanup evaluation, two analyses were performed. In the first
analysis, the distance from the most downgradient well to the nearest downgradient
receptor (HSCB) was assumed to be 1000 ft. The general groundwater flow angle is to
the Southwest. Selected plume centerline wells are MW-11, MW-72, MW-91, and MW-
92 (Table 12). The analysis was conducted with yearly averages of the TCE data from
the latest 6 years (1995 to 2000). The second analysis used the same parameters
except that the distance from the HSCB was assumed to be 100 ft. Table 13 shows
plume centerline concentration regression results for each selected sampling event,
which range from 3.8 x 10"5 to 6.8 x 10"4 per ft.
McClellan Air Force Base 32 MAROS 2.0 Application
Sacramento Valley, California Monitoring Network Optimization
-------
GROUNDWATER
June 2, 2003 SERVICES, INC.
4.0 SUMMARY AND RECOMMENDATIONS
In recent years, the high cost of long-term monitoring as part of active or passive
remediation of affected ground water has made the design of efficient and effective
ground water monitoring plans a pressing concern. Periodically updating and revising
long-term monitoring programs with changing conditions at the site can mean
considerable savings in site monitoring costs. The MAROS decision-support software
presented in this report assists in revising existing long-term monitoring plans based on
the historical and current monitoring data and plume behavior over time.
The MAROS 2.0 sampling optimization software/methodology has been applied to the
McClellan existing OU D long-term monitoring program as of December 2000. The
optimization results and subsequent recommendations allow for optimization of the
spatial and temporal groundwater monitoring system in place at the McClellan site. The
current long-term monitoring network could be optimized through reduction in sampling
locations (Results are summarized in Table 14).
Overview Statistics
Both the Mann-Kendall and Linear Regression temporal trend methods gave similar
trend estimates for each well. Results from the temporal trend analysis indicate that
90% of the plume source area monitoring wells in Zone A indicate a Probably
Decreasing, Decreasing, or Stable TCE concentration trend, whereas only about half of
the wells in the tail and edges of the plume have similar trends. The majority of the wells
in Zone B have no trend in the historical data. However, as of 2000, only one of the
wells in Zone B is actually above the MCL for TCE. The trend results for the extraction
wells in the source area indicate most wells have Probably Decreasing, or Decreasing
concentrations over time.
Results from the moment trend analysis give some evidence of a stable plume, with the
dissolved mass showing a decrease over time, whereas the center of mass and the
plume spread shows no trend over time, probably due to the change in sample locations
and frequency over time. Overall plume stability temporal results recommend a
moderate monitoring strategy due to the near stable to decreasing OU D TCE plume.
The overview results are relatively generic and not well-by-well specific, therefore, a
detailed statistical analysis with a well-by-well analysis was performed.
Detailed Statistics
Further analysis from the well redundancy spatial analysis using the Delaunay method
optimization indicate that
• 3 monitoring wells could be eliminated from the existing Zone A monitoring
network of 32 wells and
• 2 wells could be eliminated from the existing Zone B monitoring network of
14 wells (Table 14)
without compromising the reliability of the monitoring system.
McClellan Air Force Base 33 MAROS 2.0 Application
Sacramento Valley, California Monitoring Network Optimization
-------
GROUNDWATER
June 2, 2003 SERVICES, INC.
In addition, the well sufficiency spatial analysis indicated there are no areas within the
network where there are high uncertainties in the predicted TCE concentration.
Therefore there are no recommended new monitoring wells for Zones A or B. The
sampling frequency optimization analysis using the temporal MCES method, resulted in
sampling frequency that is relatively consistent with the sampling schedule currently in
use at the site for both Zone A and Zone B well networks (Table 14).
Data sufficiency analysis using power analysis methods, shows that the site has
achieved target levels at (or further than) the compliance boundary 100 ft downgradient
from the most downgradient well. This analysis indicates that the monitoring system is
working because it is powerful enough to accurately reflect the location of the plume
relative to the compliance boundary. This shows the sufficiency of the monitoring system
in terms of evaluating risk-based site target level status if the pump-and-treat remedial
system continues to contain the plume and keeps reducing the TCE concentration in the
OU-D plume.
The recommended long-term monitoring strategy results in small reduction in sampling
costs and allows site personnel to develop a better understanding of plume behavior
over time. A reduction in the number of redundant wells is expected to result in a
moderate cost savings over the long-term at McClellan AFB. The MAROS optimized
plan consists of 47 wells: 17 sampled annually, and 30 sampled biennially. The MAROS
optimized plan would result in 32 samples per year, compared to 34 samples per year
(17 annual and 34 biennial) in the current sampling program. Implementing these
recommendations could lead to a 6% reduction from the current monitoring plan in terms
of the samples to be collected per year. An approximate cost savings estimate of $300
per year is projected while still maintaining adequate delineation of the plume as well as
knowledge of the plume state over time.
McClellan Air Force Base 34 MAROS 2.0 Application
Sacramento Valley, California Monitoring Network Optimization
-------
GROUNDWATER
June 2, 2003 SERVICES, INC.
CITED REFERENCES
Air Force Center for Environmental Excellence (AFCEE), 2002. Monitoring and
Remediation Optimization System (MAROS) 2.0 Software Users Guide.
Ridley, M.N. et al., 1995. Cost-Effective Sampling of Groundwater Monitoring Wells, the
Regents of UC/LLNL, Lawrence Livermore National Laboratory.
Air Force Center for Environmental Excellence (AFCEE), 1997, AFCEE Long-Term
Monitoring Optimization Guide, http://www.afcee.brooks.af.mil.
Aziz, J, Ling, M., Newell, C., Rifai, H., and Gonzales, J., 2002, MAROS: a Decision
Support System for Optimizing Monitoring Plans, Ground Water, In Press.
Gilbert, R. O., 1987, Statistical Methods for Environmental Pollution Monitoring, Van
Nostrand Reinhold, New York, NY, ISBN 0-442-23050-8.
Radian Corporation, 1997. Installation Restoration Program, Groundwater Monitoring
Plan - Final.
Radian Corporation, 1999. Five Year Review Report, McClellan Air Force Base - Final.
URS, 2002. Former McClellan Air Force Base Installation Restoration Program,
Groundwater Monitoring Program - Quarterly Report - 1st Quarter 2002.
McClellan Air Force Base 35 MAROS 2.0 Application
Sacramento Valley, California Monitoring Network Optimization
-------
January 15, 2003
GSI Job No. G-2236-15
V
GROUNDWATER
SERVICES, INC.
TABLES
MAROS 2.0 APPLICATION
ZONE A & B OU D MONITORING NETWORK OPTIMIZATION
McClellan AFB
Sacramento Valley, California
Table 1 Sampling Locations Used in the MAROS Analysis
Table 2 Mann-Kendall Analysis Decision Matrix
Table 3 Linear Regression Analysis Decision Matrix
Table 4 Zone A & B Aquifer Site-Specific Parameters
Table 5 Results of Zone A McClellan OU D Trend Analysis
Table 6 Results of Zone B McClellan OU D Trend Analysis
Table 7 Zone A Redundancy Analysis Results - Delaunay Method
Table 8 Zone B Redundancy Analysis Results - Delaunay Method
Table 9 Zone A Sampling Frequency Analysis Results - Modified CES
Table 10 Zone B Sampling Frequency Analysis Results - Modified CES
Table 11 Zone A Selected Plume Centerline Wells for Risk-Based Site Cleanup
Evaluation - Power Analysis
Table 12 Zone A Plume Centerline Concentration Regression Results - Power
Analysis
Table 13 Zone A Risk-Based Site Cleanup Evaluation Results - Power Analysis
Table 14 Summary of MAROS Sampling Optimization Results
-------
GSI Job No. G-2236-15
Issued: 1/15/03
Page 1 of 2
If
GROUNDWATER
SERVICES, INC.
Table 1
Sampling Locations Used in the MAROS Analysis
McClellan AFB OU-D
Sacramento Valley, California
Welt
Name
MW-10
MW-11
MW-12
MW-14
MW-15
MW-38D
MW-52
MW-53
MW-55
MW-70
MW-72
MW-74
MW-76
MW-88
MW-89
Monitoring
Zone 1
A
A
A
A
A
lABorA
lABorA
lABorA
lABorA
lABorA
lABorA
lABorA
lABorA
A
A
Used in Delaunay
Analysis?
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Used in Modified
CES Analysis?
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Summary of Sampling History (sampling data
available since 1 990)
Sampled annually; recent data were <= TL but
>=MCL
Sampled annually; recent data were <= TL but
>=MCL
Sampled annually; it is an interior plume well
Sampled annually since 90, biennially since 99
because data fell below MCL
Sampled annually; it is an interior plume well
Sampled annually; recent data were <= TL but
>=MCL
Sampled annually, then biennially since 99
because data fell below MCL
Sampled annually, then biennially since 99
because data fell below MCL
Sampled annually on average, then biennially
since 01 because data fell below MCL
Sampled annually or less frequently since 91 ,
then biennially since 01 because data fell below
DL
Sampled annually; recent data were <= TL but
>=MCL
Sampled quarterly the first 6 quarters, then
annually on average, then biennially since 01
because data fall below MCL
Sampled quarterly the first 6 quarters, then
biennially on average, then annually since 99
because data were <= TL but >= MCL
Sampled annually, then biennially since 99
because data fell below MCL
Sampled annually, then biennially since 99
because data fell below MCL
Note: 1) Monitoring Zones assigned as shown in URS, 2002, 1Q02 report (Tables 3.2 and 3.3).
2) TL = the upper 90% tolerance limit from Groundwater Monitoring Plan Final (Radian Corporation 1997)
3) MCL = the maximum contaminant level of TCE, DL = detection limit, IAB = intermediate zone between zone
A and Zone B, AB = both zone A and zone B
-------
GSI Job No. G-2236-15
Issued: 1/15/03
Page 2 of 2
If
GROUNDWATER
SERVICES, INC.
Table 1
Sampling Locations Used in the MAROS Analysis
McClellan AFB OU-D
Sacramento Valley, California
Welt
Name
MW-90
MW-91
MW-92
MW-237
MW-240
MW-241
MW-242
MW-350
MW-351
MW-412
MW-458
MW-1004
MW-1026
MW-1041
MW-1042
Monitoring
Zone
A
A
A
A
A
A
A
A
A
A
A
A
A
A
A
Used in Delaunay
Analysis?
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
No, duplicates MW-
1042
Yes
Used in Modified
CES Analysis?
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Summary of Sampling History (sampling
started in 1990)
Sampled annually, then biennially since 95
because data fell below MCL
Sampled annually, then biennially since 95
because data fell below MCL
Sampled quarterly the first year, then annually on
average, then biennially since 99 because data
fall below MCL
Sampled annually since 93, then biennially since
96 because data fell below MCL
Sampled annually since 93, then biennially since
99 because data fell below MCL
Sampled annually since 93; recent data were <=
TL but >= MCL
Sampled annually since 93; recent data were <=
TL but >= MCL
Sampled at least annually since 95, then
biennially since 99 because data fell below MCL
Sampled at least annually since 95, then
biennially since 98 because data were <= MCL
Sampled quarterly since 97, then biennially since
99 because data fell below MCL
Sampled quarterly since 99, then biennially since
00 because data fell below MCL
Sampled quarterly in the first 2 years, then
biennially because data fell below MCL
Sampled biennially since data were <= MCL
Sampled biennially since data were <= MCL
Sampled annually in the first 4 years, then
biennially since 95 because data fell below MCL
Note: 1) Monitoring Zones assigned as shown in URS, 2002, 1Q02 report (Tables 3.2 and 3.3).
2) TL = the upper 90% tolerance limit from Groundwater Monitoring Plan Final (Radian Corporation 1997)
3) MCL = the maximum contaminant level of TCE, DL = detection limit, IAB = intermediate zone between zone
A and Zone B, AB = both zone A and zone B
-------
GSI Job No. G-2236-15
Issued: 1/15/03
Page 3 of 2
If
GROUNDWATER
SERVICES, INC.
Table 1
Sampling Locations Used in the MAROS Analysis
McClellan AFB OU-D
Sacramento Valley, California
Welt
Name
MW-1064
MW-1073
EW-73
EW-83
EW-84
EW-85
EW-86
EW-87
MW-19D
MW-51
MW-54
MW-57
MW-58
MW-59
MW-104
Monitoring
Zone
A
A
AB
AB
AB
AB
AB
AB
B
B
B
B
B
B
B
Used in Delaunay
Analysis?
Yes
Yes
No, screened in
zones A and B
No, screened in
zones A and B
No, screened in
zones A and B
No, screened in
zones A and B
No, screened in
zones A and B
No, screened in
zones A and B
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Used in Modified
CES Analysis?
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Summary of Sampling History (sampling
started in 1990)
Sampled quarterly in the first 6 quarters, then
biennially since 93 because data fell below MCL
Sampled biennially since 93 because data were
<=MCL
Sampled annually except between 94 and 96 it
was sampled quarterly
Sampled annually except between 94 and 96 it
was sampled quarterly
Sampled annually except between 94 and 96 it
was sampled quarterly
Sampled annually except between 94 and 96 it
was sampled quarterly
Sampled annually except between 94 and 96 it
was sampled quarterly
Sampled annually except between 94 and 96 it
was sampled quarterly
Sampled annually on average, then biennially
since 95 because data fell below MCL
Sampled biennially because data were below
MCL or DL
Sampled annually on average, then annually
since 98; recent data were <= TL but >= MCL
Sample quarterly in the first 6 quarters, then
annually, then biennially since 99 because data
fell below MCL
Sample quarterly in the first 6 quarters, then
annually, then biennially since 01 because data
fell below MCL
Sample quarterly in the first 5 quarters, then
annually, then biennially since 97 because data
fell below MCL
Sampled annually, then biennially since 95
because data fell below MCL
Note: 1) Monitoring Zones assigned as shown in URS, 2002, 1Q02 report (Tables 3.2 and 3.3).
2) TL = the upper 90% tolerance limit from Groundwater Monitoring Plan Final (Radian Corporation 1997)
3) MCL = the maximum contaminant level of TCE, DL = detection limit, IAB = intermediate zone between zone
A and Zone B, AB = both zone A and zone B
-------
GSI Job No. G-2236-15
Issued: 1/15/03
Page 4 of 2
If
GROUNDWATER
SERVICES, INC.
Table 1
Sampling Locations Used in the MAROS Analysis
McClellan AFB OU-D
Sacramento Valley, California
Well
Name
MW-105
MW-1001
MW-1003
MW-1010
MW-1027
MW-1028
MW-1043
Monitoring
Zone
B
B
B
B
B
B
B
Used in Delaunay
Analysis?
Yes
Yes
No, duplicates MW-
1001
Yes
Yes
No, duplicates MW-
1027
Yes
Used in Modified
CES Analysis?
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Summary of Sampling History (sampling
started in 1990)
Sampled annually on average, then biennially
since 01 because data fell below MCL
Sampled annually on average, then biennially
since 96 because data fell below MCL or DL
Sampled annually on average, then biennially
since 01 because data fell below MCL
Sampled biennially on average, then biennially
since 01 because data fell below MCL or DL
Sampled biennially on average, then biennially
since 99 because data fell below MCL or DL
Sampled annually on average, then biennially
since 99 because data fell below MCL or DL
Sampled annually, then biennially since 95
because data fell below MCL or DL
Note: 1) Monitoring Zones assigned as shown in URS, 2002, 1Q02 report (Tables 3.2 and 3.3).
2) TL = the upper 90% tolerance limit from Groundwater Monitoring Plan Final (Radian Corporation 1997)
3) MCL = the maximum contaminant level of TCE, DL = detection limit, IAB = intermediate zone between zone
A and Zone B, AB = both zone A and zone B
-------
GSI Job No. G-2236-15
Issued 1/15/03
Page 1 of 1
If
GROUNDWATER
SERVICES, INC.
Mann-Kendall
Mann-Kendall
Statistic
S>0
S>0
S>0
S<0
S<0
S<0
S<0
Table 2
Analysis Decision Matrix
Confidence in the
Trend
> 95%
90 - 95%
< 90%
< 90% and COV > 1
< 90% and COV < 1
90 - 95%
> 95%
(Aziz, et. al., 2002)
Concentration Trend
Increasing
Probably Increasing
No Trend
No Trend
Stable
Probably Decreasing
Decreasing
Table 3
Linear Regression Analysis Decision Matrix (Aziz, et. al., 2002)
Confidence in the
Trend
Log-slope
Positive
Negative
< 90%
90 - 95%
> 95%
No Trend
Probably Increasing
Increasing
COV < 1 Stable
COV > 1 No Trend
Probably Decreasing
Decreasing
-------
£
c
o
o
c
o
o
LU
LJ
CQ
L
^^
0>
"S
E
re
re
Q.
_0
'o
0>
Q.
«
*-•
CO
CO
c
re
o>
c
o
N
Q
^
O
0
co
CD
CQ
0
0
LL
.t
E
re
re
Q.
^_
>>
LO
CO
^
'o
o
CD
CD
D5
CD
CL
CD
CD
CO
CO
CM CN
^5 S
O -5
CO ^
"CD
e Location
o
t_
3
o
CO
<
CD
c~
o
N
CD
"CD
^ E
±± x
CO O
2 Q.
O CL
CL <
CO
?
^
LU
"CD
e Location
o
t_
3
o
CO
CQ
CD
c~
o
N
CD
"CD
E
'x
o
CL
CL
^
t^
O
CO
73
22
!_
p3
CD
CO
<
CD
c~
o
N
CD
"ro in
E co
x £
o 5
0.0
CL i
< h-
^
t^
o
CD
73
03
!_
p3
CD
CO
CQ
CD
c~
o
N
CD
| CO
E w
x ^
o 5
0.0
Q- i
< H
>
CO
Direction
Z
o
i j
j_
S
"CD
73
o
O
"ro
CD
CD
O
in
03
O
c
ro
in
c
ro
"fi)
2
O)
f~
in
ro
g
"3
0
^ro
"ro
E
^
'^
CT
ro
'o
03
CL
•*"*
c
o
-o
03
in
ro
JD
c
1^ O
O3 -p^
O3 CO
c" 'x
ro g
ro CL
l± <
c
ro
c
.0
1
c
03
O
8
LU
o
1—
1
s
^
O)
E
S
-o
03
C
-Q
0
ro
03
ro
03
D
0
in
'o
c
0
"ro
o
o
~5 CU CU
55 « §>
O J2 CL
-------
GSIJobNo. G-2236-15
Issued 1/15/03
Page 1 of 2
TABLE 5
Results of Zone A Trend Analysis
McClellan Air Force Base OU D
Sacramento Valley, California
Notes:
1. Consolidation of data included non-detect values set to the minium detection limit (0.001 mg/L)
and duplicate data for the quarter were averaged.
2. All wells that were part of the network in between 1990 and 2000 were analyzed.
If HV1CI-S. tXC-
Well
MW-10
MW-11
MW-12
MW-14
MW-15
MW-38D
MW-52
MW-53
MW-55
MW-70
MW-72
MW-74
MW-76
MW-88
MW-89
MW-90
MW-91
MW-92
MW-237
MW-240
MW-241
MW-242
MW-350
MW-351
MW-412
MW-458
MW-1004
MW-1026
MW-1041
MW-1042
MW-1064
MW-1073
EW-73
EW-86
EW-87
EW-83
EW-84
EW-85
Well
Type3
MW
MW
MW
MW
MW
MW
MW
MW
MW
MW
MW
MW
MW
MW
MW
MW
MW
MW
MW
MW
MW
MW
MW
MW
MW
MW
MW
MW
MW
MW
MW
MW
EW
EW
EW
EW
EW
EW
Well
Category '"
S
S
S
S
S
S
T
T
T
T
S
T
T
T
T
T
T
T
T
T
S
S
T
S
T
T
T
T
T
T
T
T
S
S
S
S
S
S
Mann-Kendall
Trend 4
D
D
D
D
D
D
NT
I
S
S
D
D
I
NT
NT
NT
D
NT
NT
NT
D
D
D
NT
S
S
PD
NT
NT
S
NT
NT
D
D
I
NT
D
D
Linear
Regression
Trend 4
D
D
D
D
D
D
PD
I
S
S
D
D
I
NT
NT
NT
D
NT
NT
NT
D
D
D
NT
S
NT
PD
NT
NT
S
PI
NT
D
D
I
NT
D
D
Overall
Trend 6
D
D
D
D
D
D
S
I
S
S
D
D
I
NT
NT
NT
D
NT
NT
NT
D
D
D
NT
S
S
PD
NT
NT
S
PI
NT
D
D
I
NT
D
D
Number
of
Samples
11
11
12
12
11
8
9
10
11
6
10
10
10
11
10
8
10
12
7
9
12
12
8
9
6
4
12
7
7
7
8
6
15
15
15
16
16
16
Number
of
Detects
11
11
12
12
11
8
3
9
11
0
10
10
4
4
2
6
9
9
5
5
12
12
5
9
4
3
6
5
1
1
4
5
15
15
15
16
16
16
-------
GSIJobNo. G-2236-15
Issued 1/15/03
Page 1 of 1
GSQUNDWATE!
SERVICES, INC.
TABLE 6
Results of Zone B Trend Analysis
McClellan Air Force Base OU D
Sacramento Valley, California
Well
MW-19D
MW-51
MW-54
MW-57
MW-58
MW-59
MW-104
MW-105
MW-1001
MW-1003
MW-1010
MW-1027
MW-1028
MW-1043
Well
Type3
T
T
S
T
T
T
T
T
T
T
T
T
T
T
Well
Category !
T
T
S
T
T
T
T
T
T
T
T
T
T
T
Mann-Kendall
Trend 4
NT
S
I
NT
NT
S
NT
NT
S
PD
S
NT
S
NT
Linear
Regression
Trend 4
NT
PD
I
NT
NT
S
NT
I
D
D
S
NT
S
NT
Overall
Trend 6
NT
S
I
NT
NT
S
NT
PI
PD
D
S
NT
S
NT
Number
of
Samples
10
8
10
12
15
12
8
8
10
11
5
6
10
6
Number
of
Detects
6
1
8
5
7
2
2
2
1
3
0
2
2
1
Notes:
1. Consolidation of data included non-detect values set to the minium detection limit (0.001 mg/L)
and duplicate data for the quarter were averaged.
2. All wells that were part of the network in between 1990 and 2000 were analyzed.
3. EW = Extraction Well; MW = Monitoring Well
4. Decreasing (D), Probably Decreasing (PD), Stable (S), No Trend (NT), Probably Increasing (PI), and Increasing (I)
5.3 = Source Zone Well; T = Tail Zone Well
6. Overall Trend is calculated from a weighted average of the Linear Regression and Mann-Kendall Trends.
For further details on this methodolgy refer to the MAROS Manual Appendix A.8.
-------
GSI Job No. G-2236-15
Issued: 1/15/03
Page 1 of 2
If
GROUNDWATER
SERVICES, INC.
Table 7
Well Redundancy Analysis Results - Delaunay Method
McClellan AFB OU-D Zone A
Sacramento Valley, California
Well Name
MW-10
MW-11
MW-12
MW-14
MW-15
MW-38D
MW-52
MW-53
MW-55
MW-70
MW-72
MW-74
MW-76
MW-88
MW-89
MW-90
MW-91
MW-92
MW-237
MW-240
Well Used in Analysis?
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
MAROS Well
Redundancy
Analysis Result
Keep
Keep
Keep
Eliminate
Keep
Keep
Keep
Keep
Keep
Keep
Eliminate
Keep
Keep
Keep
Keep
Keep
Keep
Keep
Keep
Keep
MAROS
Interpreted
Well
Redundancy
Keep
Keep
Keep
Eliminate
Keep
Keep
Keep
Keep
Keep
Keep
Keep
Keep
Keep
Keep
Keep
Keep
Keep
Keep
Keep
Keep
Comments
Spatially redundant
On plume centerline and used in
MAROS data sufficiency
analysis
Notes: 1) Yearly averages from 6 sampling events (1995 ~ 2000) were used in the above analysis
2) InsideSF = 0.20, HullSF = 0.01, AR = CR = 0.95
3)"-" = Not Applicable.
-------
GSI Job No. G-2236-15
Issued: 1/15/03
Page 2 of 2
If
GROUNDWATER
SERVICES, INC.
Table 7
Well Redundancy Analysis Results - Delaunay Method
McClellan AFB OU-D Zone A
Sacramento, California
Well Name
MW-241
MW-242
MW-350
MW-351
MW-412
MW-458
MW-1004
MW-1026
MW-1041
MW-1042
MW-1064
MW-1073
Well Used in Analysis?
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
No: duplicates MW-1042
Yes
Yes
Yes
MAROS Weil
Redundancy
Analysis Result
Eliminate
Keep
Keep
Keep
Keep
Keep
Keep
Keep
-
Keep
Keep
Keep
MAROS
Interpreted
Well
Redundancy
Eliminate
Keep
Keep
Keep
Keep
Keep
Keep
Keep
Eliminate
Keep
Keep
Keep
Comments
Spatially redundant
Duplicates MW-1042
Notes: Yearly averages from 6 sampling events (1995 ~ 2000) were used in the above analysis
InsideSF = 0.20, HullSF = 0.01, AR = CR = 0.95
-------
GSI Job No. G-2236-15
Issued: 1/15/03
Page 1 of 1
If
GROUNDWATER
SERVICES, INC.
Table 8
Well Redundancy Analysis Results - Delaunay Method
McClellan AFB OU-D Zone B
Sacramento Valley, California
Well Name
MW-19D
MW-51
MW-54
MW-57
MW-58
MW-59
MW-104
MW-105
MW-1001
MW-1003
MW-1010
MW-1027
MW-1028
MW-1043
Well Used in Analysis?
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
No: duplicates MW-1001
Yes
Yes
No: duplicates MW-1027
Yes
MAROS Well
Redundancy
Analysis Result
Keep
Keep
Keep
Keep
Keep
Keep
Keep
Keep
Keep
-
Keep
Keep
-
Keep
MAROS
Interpreted
Well
Redundancy
Keep
Keep
Keep
Keep
Keep
Keep
Keep
Keep
Keep
Eliminate
Keep
Keep
Eliminate
Keep
Comments
Duplicates MW-1003
Duplicates MW-1027
Notes: 1) Yearly averages from 6 sampling events (1995 ~ 2000) were used in the above analysis
2) InsideSF = 0.05 or 0.20, HullSF = 0.01 or 0.05, AR = CR = 0.95
3)"-" = Not Applicable.
-------
GSI Job No. G-2236-15
Issued: 1/15/03
Page 1 of 2
If
GROUNDWATER
SERVICES, INC.
Table 9
Sampling Frequency Analysis Results - Modified CES
McClellan AFB OU-D Zone A
Sacramento Valley, California
Well
Name
MW-10
MW-11
MW-12
MW-14
MW-15
MW-38D
MW-52
MW-53
MW-55
MW-70
MW-72
MW-74
MW-76
MW-88
MW-89
MW-90
MW-91
MW-92
MW-237
MW-240
MAROS
Frequency
Based on
Recent
Trend"1
Annual
Annual
Annual
Annual
Semiannual
Annual
Annual
Annual
Annual
Annual
Annual
Semiannual
Quarterly
Annual
Annual
Annual
Annual
Annual
Annual
Annual
MAROS
Frequency
Based on
Overall
Trend'21
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Semiannual
Quarterly
Annual
Annual
Annual
Annual
Annual
Annual
Annual
MAROS
Recommended
Frequency'31
Annual
Annual
Annual
Annual
Semiannual
Annual
Biennial
Annual
Biennial
Biennial
Annual
Semiannual
Quarterly
Annual
Biennial
Annual
Biennial
Annual
Biennial
Annual
Frequency for
Optimized
Network
Annual
Annual
Annual
-
Annual
Annual
Biennial
Biennial
Biennial
Biennial
Annual
Annual
Annual
Biennial
Biennial
Biennial
Biennial
Biennial
Biennial
Biennial
Comments
Decreasing trend but still higher than MCL
Decreasing trend but still higher than MCL
Decreasing trend but still higher than MCL
-
Inside-plume well although with slightly
increasing trend
Concentrations higher than MCL
All historical concentrations below MCL or
DL
All historical concentrations (except one)
below MCL or DL
All historical concentrations below MCL or
DL
All historical concentrations below MCL or
DL
Recent concentrations higher than MCL
Recent concentrations higher than MCL
(the MAROS result was due to insufficient
recent data)
Recent concentrations higher than MCL
(the MAROS result was due to insufficient
recent data)
Recent concentrations below MCL or DL
All historical concentrations below MCL or
DL
Recent concentrations (except one) below
MCLorDL
Recent concentrations below MCL or DL
Recent concentrations (except one) below
MCLorDL
All historical concentrations below MCL or
DL
All historical concentrations below MCL or
DL
Notes: 1) The frequency determined by MAROS based on the analysis of recent data (data between 1994 and 2000)
2) The frequency determined by MAROS based on the analysis of overall data (data between 1990 and 2000)
3) The frequency finally recommended by MAROS after considering recent and overall frequency results as well
as the rates of change in these trends Rate parameters used are 0.5MCL/year, 1.0MCL/year, and 2.0MCL/year
for Low, Medium, and High rates, respectively; the MCL of TCE is 0.005 mg/L
-------
GSI Job No. G-2236-15
Issued: 1/15/03
Page 2 of 2
If
GROUNDWATER
SERVICES, INC.
Table 9
Sampling Frequency Analysis Results - Modified CES
McClellan AFB OU-D Zone A
Sacramento Valley, California
Well
Name
MW-241
MW-242
MW-350
MW-351
MW-412
MW-458
MW-1004
MW-1026
MW-1041
MW-1042
MW-1064
MW-1073
MAROS
Frequency
Based on
Recent
Trend11
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Semiannual
Annual
Annual
Quarterly
Annual
MAROS
Frequency
Based on
Overall
Trend'2'
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Semiannual
Annual
Annual
Quarterly
Annual
MAROS
Recommended
Frequency13'
Annual
Annual
Annual
Annual
Biennial
Biennial
Annual
Semiannual
Biennial
Biennial
Quarterly
Annual
MAROS
Interpreted
Sampling
Frequency
Results
-
Annual
Biennial
Annual
Biennial
Biennial
Biennial
Biennial
-
Biennial
Biennial
Biennial
Comments
-
Decreasing trend but still higher than
MCL
Recent concentrations below MCL or
DL
Recent concentrations higher than
MCL
All historical concentrations below MCL
orDL
All historical concentrations below MCL
orDL
All historical concentrations below MCL
or DL (the MAORS result was due to
insufficient recent data)
Concentrations since 93 below MCL or
DL (the MAORS result was due to
insufficient recent data)
-
Upgradient wells with all historical
concentrations below MCL or DL
Upgradient well with recent
concentrations below MCL or DL (the
MAORS result was due to insufficient
recent data)
Recent concentrations below MCL or
DL (the MAORS result was due to
insufficient data)
Extraction wells below:
EW-73
EW-83
EW-84
EW-85
EW-86
EW-87
Annual
Annual
Annual
Annual
Annual
Quarterly
Annual
Annual
Annual
Annual
Annual
Quarterly
Annual
Annual
Annual
Annual
Annual
Quarterly
Annual
Annual
Annual
Annual
Annual
Annual
Performance monitoring
Performance monitoring
Performance monitoring
Performance monitoring
Performance monitoring
Performance monitoring
Notes: 1) The frequency determined by MAROS based on the analysis of recent data (data between 1994 and 2000)
2) The frequency determined by MAROS based on the analysis of overall data (data between 1990 and 2000)
3) The frequency finally recommended by MAROS after considering recent and overall frequency results as well
as the rates of change in these trends Rate parameters used are 0.5MCL/year, 1.0MCL/year, and 2.0MCL/year
for Low, Medium, and High rates, respectively; the MCL of TCE is 0.005 mg/L
-------
GSI Job No. G-2236-15
Issued: 1/15/03
Page 1 of 1
If
GROUNDWATER
SERVICES, INC.
Table 10
Sampling Frequency Analysis Results - Modified CES
McClellan AFB OU-D Zone B
Sacramento Valley, California
Well
Name
MW-19D
MW-51
MW-54
MW-57
MW-58
MW-59
MW-104
MW-105
MW-1001
MW-1003
MW-1010
MW-1027
MW-1028
MW-1043
MAROS
Frequency
Based on
Recent
Trend111
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
MAROS
Frequency
Based on
Overall
Trend'21
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
MAROS
Recommended
Frequency'31
Annual
Biennial
Annual
Biennial
Biennial
Biennial
Annual
Annual
Biennial
Biennial
Annual
Annual
Annual
Annual
MAROS
Interpreted
Sampling
Frequency
Results
Biennial
Biennial
Annual
Biennial
Biennial
Biennial
Biennial
Biennial
Biennial
-
Biennial
Biennial
-
Biennial
Comments
All historical concentrations below
MCL or DL (the MAORS result
was due to insufficient recent data)
All historical concentrations below
MCL or DL
Recent concentrations above MCL
All historical concentrations below
MCL or DL
All historical concentrations below
MCL or DL
All historical concentrations below
MCL or DL
All historical concentrations below
MCL or DL (the MAORS result
was due to insufficient recent data)
All historical concentrations below
MCL or DL
All historical concentrations below
MCL or DL
-
All historical concentrations below
MCL or DL (the MAORS result
was due to insufficient data)
All historical concentrations below
MCL or DL (the MAORS result
was due to insufficient recent data)
-
All historical concentrations below
MCL or DL (the MAORS result
was due to insufficient recent data)
Notes: 1) The frequency determined by MAROS based on the analysis of recent data (data between 1995 and 2000)
2) The frequency determined by MAROS based on the analysis of overall data (data between 1990 and 2000)
3) The frequency finally recommended by MAROS after considering recent and overall frequency results as well
as the rates of change in these trends Rate parameters used are 0.5MCL/year, 1.0MCL/year, and 2.0MCL/year
for Low, Medium, and High rates, respectively; the MCL of TCE is 0.005 mg/L
-------
GSI Job No. G-2236-15
Issued: 1/15/03
Page 1 of 1
If
GROUNDWATER
SERVICES, INC.
Table 11
Risk-Based Site Cleanup Evaluation Results - Power Analysis
McClellan AFB OU-D Zone A
Sacramento Valley, California
Sampling Event
(Yearly Averaged)
1995
1997
1998
1999
Sample
Size
29
20
19
29
Distance to HSCB = 1000 ft
Cleanup Status
Attained
Attained
Attained
Attained
Power
1.0
1.0
1.0
1.0
Distance to HSCB = 100 ft
Cleanup Status
Attained
Attained
Attained
Attained
Power
1.0
1.0
1.0
1.0
Note: The power analysis used for this application assumes normality of data. Distance to the Hypothetical
Statistical Compliance Boundary (HSCB) is the distance from the most downgradient well to the HSCB; S/E
= extrapolated result significantly exceeds the target level (0.005 mg/L).
-------
GSI Job No. G-2236-15
Issued: 1/15/03
Page 1 of 1
If
GROUNDWATER
SERVICES, INC.
Table 12
Selected Plume Centerline Wells
Risk-Based Site Cleanup Evaluation - Power Analysis
McClellan AFB OU-D Zone A
Sacramento Valley, California
Well Name
MW-92
MW-91
MW-72
MW-11
Distance from Well to Receptor (feet)
1866.2
1996.1
2547.8
3115.4
Note: Groundwater flow angle is to the Southeast; the distance from the most
downgradient well to the nearest downgradient receptor is assumed to be 1000
feet.
-------
GSI Job No. G-2236-15
Issued: 1/15/03
Page 1 of 1
If
GROUNDWATER
SERVICES, INC.
Table 13
Plume Centerline Concentration
Regression Results - Power Analysis
McClellan AFB OU-D Zone A
Sacramento Valley, California
Sampling Event
(Yearly Averaged)
1995
1996
1997
1998
1999
2000
Number of
Centerline Wells
4
2
3
3
4
2
Regression
Coefficient (1/ft)
-6.77E-03
_
-4.74E-03
-3.84E-03
-4.75E-03
-
Confidence in
Coefficient
99.2%
_
83.5%
88.9%
96.2%
-
Note: Regression is on natural log concentration of TCE versus distance from source
centerline wells shown in Table 12; no regression was performed for sampling event with less
than 3 centerline wells.
-------
a>
a:
c
_0
^
re
N
•e Q
a.
O
.2 a
II
CQ >^
LL tu
< ro
c >
_
O
o
o
ro
CO
re
E
E
CO
CN
CN
e>8
CO
•5
-o
cu
i/)
£
>
^ w £ 0" a)
slw|D1
1 c o
> ro c
oil:l|
s 1"! o01
£ u. 5
|£ J
*?>
£M C
a a)
3 £ ^
O ro 5
it
W 2*8
0=0,31
s 5 «
O)
O C
S 0
c N
O
s
Q)
__,
§i
source well
08
o
ncentrations >
8
"c
8
g
-o
c
IT
D)
C
'c/3
CC
§
03
Q
1
C
^
I
Q
to
C
c
*t
W
^
0
§
source well
08
o
ncentrations >
8
"c
8
03
g
-o
C
IT
D)
.C
CO
§
03
Q
1
C
^
|
Q
to
C
c
*t
W
^
•*-
§
source well
08
o
ncentrations >
8
"c
8
03
g
-o
C
^T
D)
C
'c/3
CC
§
03
Q
To
1
^
1
Q
to
C
c
*t
W
^
CN
§
undant
-o
03
>^
|ro
"cc
Q.
C/3
08
-o
C
^T
D)
.C
CO
§
03
Q
1
^
'iminate
Ul
Q
to
"c
c
03
in
w
^
^t
§
1
>
1
o
C/3
08
-o
C
03
i
D)
.C
CO
§
03
Q
To
1
^
1
Q
to
C
c
*t
W
^
LO
§
source well
08
o
ncentrations >
8
"c
8
03
g
-o
C
^T
D)
.C
CC
§
03
Q
1
C
^
I
Q
to
C
c
*t
W
<
m"
<^
Q
ro
CO
§
f source area
o
c
.03
CO
D)
Q.
08
Q
0
O
^
V
C/3
C
O
i
03
O
C
o
o
to
o
"en
±
3
£
c
0)
CQ
|
«
—
"E
c
03
in
H
<
m"
<£
CN
§
f source area
o
c
.03
D)
Q.
08
Q
0
O
^
V
C/3
C
O
1
c
03
O
c
o
o
to
o
"c/3
±
3
£
c
0)
CQ
|
1-
z
|ro
"c
c
03
in
H
<
OQ
<^
CO
IO
§
8
"c
8
03
g
-o
C
^T
D)
C
"c/3
CC
§
03
Q
To
1
^
1
Q
to
C
c
*t
W
<
OQ"
<^
CN
§
CO
03
CO
o
D
O
C/3
03
C/3
_g
08
O
ncentrations >
8
"c
8
03
g
-o
C
^T
D)
C
"c/3
CO
§
03
Q
1
^
|
Q
!2
"c
c
03
in
H
<
OQ
<^
j*
§
CO
CO
o
o
tn
jpgradient of !
08
o
•^
centrations > l>
c
o
o
recen
_c
'i
c
03
D)
C
CO
03
O
c
To
1
"*
1
-
to
c
c
*t
1-
<
OQ
<^
to
§
cc
CO
8
o
C/3
M—
oss gradient c
o
08
_i
Q
ons < MCL or
1
"c
conce
c
8
2?
f—
'i
-o
c
0
z
3
£
c
0)
CQ
|
1-
z
to
"c
c
03
in
i-
^
ro
§
CO
03
it of source ai
.03
CO
D)
C/3
C/3
§
08
Q
0
O
^
V
W
C
o
1
03
O
C
o
o
to
o
"c/3
±
3
£
c
0)
CQ
|
-
—
'E
c
03
in
i-
^
§3
§
CO
03
CO
8
o
C/3
M—
oss gradient c
o
08
_i
Q
O
O
V
C/3
C
g
"E
conce
c
8
2?
f—
'i
-o
c
0
z
3
£
c
0)
CQ
|
1-
z
—
'E
c
03
in
i-
^
8
§
'o
"E
.03
CO
D)
C
O
-o
08
Q
o
O
ncentrations <
8
"c
8
03
g
-o
C
03
i
"c/3 CO
CO m
03 °j
03 g
a 8
3
£
c
0)
CQ
|
Q
—
"E
c
03
in
\-
^
S
§
'o
"E
.03
CO
D)
C
o
-o
08
Q
o
o
ncentrations <
8
"c
8
03
g
-o
C
03
i
"c/3 CO
CO m
03 °j
03 g
a 8
3
£
c
0)
CQ
|
Q
—
"E
c
03
in
\-
^
CN
§
m
03
CO
<
03
C
03
03
03
S
03
CO
I
03
C
-—. 03
CN 03
§ P
D)
CN
C/3
L DO
D)
C
'c/3
CO
03
O
c
s
Q-
&= ?
Q^o P
< S
55 « c?
O J2 D.
a- ±= 03 c/3
£ O
O 03
cn Q
S.
o
-------
o>
Qi
C
_0
^
re
N
•e Q
Q.
O
« Q.
E
re
re
CQ >^
U_ CD
< ro
c >
O
o
o
ro
V)
re
E
E
LO
5
CO
CM
CM
e>8
-b ™
cu cu
Comments
»1 -S.
ufn
= ^5 202
"ai c o
>
C
'in
CO
2^
o
CD
Q
.^
3
C
C
*
S
re
c
|
ul
Q
"co
c
c
CO
^
5
§
5
<:
^
sntrations > MCL & inside plume
o
C
o
o
"E
CD
O
'i
-^
C
O)
C
"tn
ro
2?
o
CD
Q
.^
3
C
C
*
Q
1
Q
"to
C
CO
^
CM
§
^
"c
CD
W
-l_l
C
_CD
"m
i_
sntrations < MCL or DL & downg
t conc<
I
'g
-^
1
D)
C
'in
CO
2?
CD CD
Q g
3
CD
CO
Q
1
Q
H
'c
c
CD
in
i-
^
o
un
CO
§
5
"CD
s > MCL & downgradient sentry '
c
g
2
§
c
o
o
"E
CD
0
C~
'1
c
2?
o
z
^^
3
C
C
*
Q
1
H
"(0
c
CO
<(•
CO
§
5
DL & downgradient sentry well
o
_i
o
V
in
c
0
•^
s
c
8
C
8
8
0
in
±
're
.CD
CO
Q
1
Q
Q_
_
'c
c
CD
in
i-
^
CM
§
5
CO
p
DL & downgradient of source ar
o
_i
o
V
in
c
.g
"(0
c
8
C
8
8
0
in
±
're
1
CO
Q_
CD
1-
—
'E
c
P
in
i-
^
CO
un
§
5
DL & cross gradient sentry well
0
_i
o
V
in
c
.g
j=
c
8
C
8
8
0
in
±
're
.CD
CO
Q
1
Q
_
'c
c
P
in
i-
^
^
o
o
§
5
P
DL & up/cross gradient sentry w
0
O
V
CO
P
o
c
'in
in
c
o
"E
o
o
O
're
1
CO
Q
1
1-
_
'c
c
P
in
i-
^
CD
CM
O
§
5
duplicates MW-1042
_i
o
V
in
c
.g
j=
c
8
8
8
0
in
±
're
1
CO
S
re
c
|
Uj
1-
_
'c
c
P
in
i-
^
i
§
5
P
1
DL & upgradient well far from pi
o
_i
o
V
in
c
.g
j=
c
8
C
8
8
0
in
±
're
.CD
CO
Q
1
CO
—
'E
c
P
in
i-
^
CM
S
§
5
(11
C
—
^-
or DL & upgradient well far from
O
_g
c
o
•-£=
CO
5
o
C
0
o
CD
CD
OL
're
.CD
CO
Q
1
\-
—
'E
c
CD
in
i-
<£
^
CD
O
§
5
CD
5
lions < MCL or DL & upgradient '
CO
§
C
o
o
§
CD
"1
C
CD
"o
s
're
.CD
CO
Q
1
CO
—
'E
C
CD
in
i-
^
CO
o
§
5
D)
C
5
'c
xtraction well & performance mo
ume e
Q.
CD
-o
'in
08
-^
c
CD
D>
C
'in
CO
CD
O
CD
Q
"re
3
C
C
*
Q
1
Q
"to
C
C
CO
CD
^
CO
§
HI
well & performance monitoring
C
g
CO
1
CD
D
Q.
CD
-o
'in
08
C
CD
O
z
"re
3
C
C
*
Q
1
H
"co
C
C
CO
CO
O J2 D.
S
o
-------
O 53
o>
Qi
C
_0
^
re
N
•S Q
Q.
O
re
E
re
O
o:
re
E
E
CQ >^
LL tu
< ro
c >
= &
® c
O i
o E
o
ro
CO
CO
CM
CM
e>8
O
—3
o
CO
cu
42
« c
f /» >» ® (A
|f|fl
S|f =5*
W _ .H
i = =
E£
"P^ (0 o
2 » S
30 =
I £
LL
w £*g
O = g, 3 «
Sw "
O w
D)
C
•c <»
£5
C N
O
_ 01
3Z
D)
0
'E
erformance me
u.
"CD
o
•fi
CO
•g
CD
-Q_
CD
r^
08
C
CD
D)
'to
CO
CD
Q
3
C
^
I
Q
03
C
C
**•
CO
CD
un
LJJ
D)
p
'E
erformance me
u.
08
"CD
o
Ti
CO
•g
CD
"o.
CD
r^
08
C
CD
i
D)
'to
CO
CD
Q
3
C
^
1
Q
03
C
C
**•
CO
CD
CD
111
CD
O
C
CO
1
CD
Q.
08
1
C
.0
o
£
X
^
"o.
CD
W
C
08
c_
co
c
03
§
>,.£
-° o
S o
Q- E
3
c
H
I
Q_
03
C
c
^
CO
CD
LJJ
CD
^
P
O
ent of zone B s
CO
D)
1
TJ
08
1
Q
O
O
^
V
in
r
0
CO
.b
c~
conce
8
0
X
iennial
CQ
!
ro
c
r
03
m
i-
m
Q
o>
§
CD
^
P
O
ent of zone B s
CO
D)
1
TJ
08
1
Q
O
O
^
V
to
c
0
CO
.b
c
conce
8
0
X
iennial
CQ
!
ro
c
c
03
m
i-
m
§
o
A
to
c
.0
"2
s
8
• '• —
&
CD
g
LL)
0
N
-^
CD
rtf
1
C
"w
03
O
To
3
C
^
1
03
C
C
**•
CO
CD
^
§
CD
^
P
O
ent of zone B s
CO
D)
C
TJ
08
1
Q
O
O
^
V
in
r
0
CO
i
c~
conce
8
0
X
iennial
CQ
!
ro
c
r
03
m
i-
CD
^
§
—
CD
S
8
of zone B sou
c
CD
CO
2
^
oa
i
Q
O
O
^
V
to
c
0
CO
i
c
conce
8
0
X
'S
CO
I
ro
c
c
03
CD
1-
CD
CO
§
CD
^
P
O
ent of zone B s
CO
D)
1
TJ
08
1
Q
O
O
^
V
in
r
0
ro
c~
conce
8
0
X
iennial
CQ
!
ro
c
r
03
m
i-
CD
«
§
—
c
CD
2
^
oa
i
Q
O
O
^
V
to
c
0
ro
.b
c
conce
8
0
X
iennial
CQ
!
ro
c
c
03
m
i-
CD
S
§
CD
1
"E
CD
CO
U)
to
to
p
O
08
1
Q
O
O
„>
V
to
c
0
CO
c
conce
8
0
X
iennial
CQ
1
ro
c
c
03
m
i-
CD
un
o
§
CD
1
"E
CD
CO
U)
to
to
p
O
08
1
Q
O
O
„>
V
in
r
0
CO
c~
conce
8
0
X
iennial
CQ
!
ro
c
r
03
CD
\-
CD
O
O
§
m
O
0
§
in
_CD
"ro
=
Q
iennial
CQ
01
Z
|
UI
Q
ro
c
r
CD
m
H
m
CO
o
o
§
m
background w
c
CD
2
^
oa
i
Q
O
O
^
V
in
r
0
ro
c~
conce
8
0
X
iennial
CQ
!
ro
c
r
03
CD
\-
CD
0
O
§
£»
radient sentry v
D)
to
g
^
08
i
Q
O
O
^
V
to
c
0
ro
c
conce
8
0
X
'S
CO
I
ro
c
c
03
CD
1-
CD
CN
O
§
CN
0
§
to
_CD
"ro
=
^
Q
3
C
^
a
imina
ui
ro
c
c
CD
m
i-
m
CO
CN
o
§
CD
background w
c
CD
ro
2
^
oa
i
Q
O
O
^
V
to
c
0
ro
c
conce
8
p
to
X
'S
CO
I
ro
c
c
03
CD
1-
CD
CO
S
§
ro
<
CD
CD
CD
S
CD
.Q
CD
a
03
ro
I -
"CD ™
£ -I
W
g
Q-
OT co
O J2 D.
•C CD ^3 O
o "S -Q z
i-s.1 p
S1! ^ §
§«~ I5
Q-£ 03 'to
CD M c ro
£ " 03 g
.C O 0 03
I— S to Q
?C\T CO~
-------
January 15 2003 GROUNDWATER
GSI Job No. G-2236-15 SERVICES, INC.
MAROS 2.0 APPLICATION
ZONE A & B OU D MONITORING NETWORK OPTIMIZATION
McClellan AFB
Sacramento Valley, California
FIGURES
Figure 1 Zone A McClellan OU D Groundwater Monitoring Network
Figure 2 Zone B McClellan OU D Groundwater Monitoring Network
Figure 3 MAROS Decision Support Tool Flow Chart
Figure 4 MAROS Overview Statistics Trend Analysis Methodology
Figure 5 Decision Matrix for Determining Provisional Frequency
Figure 6 Zone A McClellan OU D TCE Mann-Kendall Trend Results
Figure 7 Zone A McClellan OU D TCE Linear Regression Trend Results
Figure 8 Zone B McClellan OU D TCE Mann-Kendall Trend Results
Figure 9 Zone B McClellan OU D TCE Linear Regression Trend Results
Figure 10 Zone AB McClellan OU D TCE Mann-Kendall Trend Results, Extraction
Wells
Figure 11 Zone AB McClellan OU D TCE Linear Regression Trend Results,
Extraction Wells
Figure 12 Zone A McClellan OU D TCE First Moment (Center of Mass) Over Time
Figure 13 Zone B McClellan OU D TCE First Moment (Center of Mass) Over Time
Figure 14 Zone A Well Sufficiency Results
-------
-Q v>
o >
S§
-------
-I—•
8
D)
c
o
-------
GSI Job No. G-2236-15
Issued: 1/15/03 GROUNDWATER
Page 1 of 1 SERVICES, INC.
MAROS: Decision Support Tool
MAROS is a collection of tools in one software package that is used in an explanatory, non-linear fashion. The tool
includes models, geostatistics, heuristic rules, and empirical relationships to assist the user in optimizing a
groundwater monitoring network system while maintaining adequate delineation of the plume as well as knowledge
of the plume state over time. Different users utilize the tool in different ways and interpret the results from a different
viewpoint.
Overview Statistics
What it is: Simple, qualitative and quantitative plume information can be gained through evaluation of monitoring
network historical data trends both spatially and temporally. The MAROS Overview Statistics are the foundation the
user needs to make informed optimization decisions at the site.
What it does: The Overview Statistics are designed to allow site personnel to develop a better understanding of the
plume behavior over time and understand how the individual well concentration trends are spatially distributed within
the plume. This step allows the user to gain information that will support a more informed decision to be made in the
next level of optimization analysis.
What are the tools: Overview Statistics includes two analytical tools:
1) Trend Analysis: includes Mann-Kendall and Linear Regression statistics for individual wells and results in
general heuristically-derived monitoring categories with a suggested sampling density and monitoring
frequency.
2) Moment Analysis: includes dissolved mass estimation (0th Moment), center of mass (1st Moment), and
plume spread (2nd Moment) over time. Trends of these moments show the user another piece of
information about the plume stability over time.
What is the product: A first-cut blueprint for a future long-term monitoring program that is intended to be a
foundation for more detailed statistical analysis.
T
Detailed Statistics
What it is: The MAROS Detailed Statistics allows for a quantitative analysis for spatial and temporal optimization of
the well network on a well-by-well basis.
What it does: The results from the Overview Statistics should be considered along side the MAROS optimization
recommendations gained from the Detailed Statistical Analysis. The MAROS Detailed Statistics results should be
reassessed in view of site knowledge and regulatory requirements as well as the Overview Statistics.
What are the tools: Detailed Statistics includes four analytical tools:
1) Sampling Frequency Optimization: uses the Modified CES method to establish a recommended future
sampling frequency.
2) Well Redundancy Analysis: uses the Delaunay Method to evaluate if any wells within the monitoring
network are redundant and can be eliminated without any significant loss of plume information.
3) Well Sufficiency Analysis: uses the Delaunay Method to evaluate areas where new wells are
recommended within the monitoring network due to high levels of concentration uncertainty.
4) Data Sufficiency Analysis: uses Power Analysis to assess if the historical monitoring data record has
sufficient power to accurately reflect the location of the plume relative to the nearest receptor or
compliance point.
What is the product: List of wells to remove from the monitoring program, locations where monitoring wells may
need to be added, recommended frequency of sampling for each well, analysis if the overall system is statistically
powerful to monitor the plume.
Figure 3. MAROS Decision Support Tool Flow Chart
-------
GSI Job No. G-2236-15
Issued: 1/15/03
Page 1 of 1
If
GROUNDWATER
SERVICES, INC.
In Ml
i
(21 S>* of
V By
of
tor lip wtll ma by
C IK
• fi '
« (P1J ——
» _____________
|S|' -
(DJ
for Will On All
I.Oi: SmiPCt W.ljMf WIIGHT
tor
Tall
(I
IPI
(MI)
PI
Pi
«nl
^1
faj
w
Oil»pny
tut By
I tan
*«
•ml
pnnwlm.
•
»
•
Figure 4:
MAROS Overview Statistics Trend Analysis Methodology
CtmsHJtttlon
Fvei SsSwent
-------
GSI Job No. G-2236-15
Issued: 1/15/03
Page 1 of 1
If
GROUNDWATER
SERVICES, INC.
i i
i i
i I E
; : 2
S: : h-
i JL; i !§
• : C
Q>
S^
c
c
to
S
I
PI
NT
S
PD
D
Rate of Change (Linear Regression)
High MH Medium LM Low
HI =::::::
;EEEEEEEEEEE;;EEEEEEEEE=EEEEEEEE;
Figure 5. Decision Matrix for Determining Provisional Frequency (Figure A.3.1 of the
MAROS Manual (AFCEE 2001))
-------
O
CD CO
CO O
Si £
6 ^
6 -a
~
.a
o
-------
£
O) I-
0) £
Q
Q Q.
Q. _
T3
CD
O
a:
LU
_l
<
O
O
CO
CO
UJ
ZJ
O
o ^F
Q
3
O
;O
"Z>
O
CD CO
CO O
Si £
6 ^
6 -a
~
.a
o
-------
C Q
0) Q CL
CO
CL _
T3
03
O
LU
o
o
o
LO
CO
c/)
o ^F
UJ
Zl
O
O
CO CO
CO O
CN rs
CN !£
6 ^
6 -a
Z
%s
-------
o 8
-------
-------
£
O) I-
0) £
Q
Q Q.
CO
Q. _
T3
CD
O
a:
LU
CO
O
00
CD
O
"3"
co
O
"o
X
LU (0
UJ
ZJ
O
o ^^
j
Q
3
O
'Zl
O
CD CO
CO O
Si £
6 ^
6 -a
.a
o
-------
o
o
0
p
CO
"ro
E
'x
•o
C
(D
(D
o
00
1
s
o
W
o ^F
CD CO
co o
Si £
6 ^
CD -a
^ 1
7T
J2
O
Q
3
O
o
&_
B
CD
O
X CD
E E
a. ^
a. o
< W
yj
-------
(D
O
(D
o
o
0
p
CO
"ro
E
'x
o
CO
o
3
o
a: tn
O ~
OT
a
s
00
o
W
o ^F
Q
3
O
CD CO
CO O
CN 5>
6 ^
CD -a
"Z. ^
I 1
7T
-------
GSI JobNo. G-2236-15
Issued: 1/15/03
Page 1 of 1
If
GROUNDWATER
SERVICES, INC.
Potential areas for
new locations are
indicated by triangles
with a high SF level, t
Estimated SF Level;
S - S mall
M - i
L^ Large I
E ~ Extremely large }
'-1073
\j
\\
4i»,vMi :^:8
\Kfi " '• ' ' '
"
/ I / '.
/' s'i/ s
lf MW-458\
MlW-1026
MW-35D
Figure 14. Well Sufficiency Analysis for possible new sampling locations in Zone A. Areas with L
symbols are candidate regions for placing new wells. No new wells need to be recommended
since the current network has enough sampling points.
-------
January 15 2003 GROUNDWATER
GSI Job No. G-2236-15 SERVICES, INC.
MAROS 2.0 APPLICATION
ZONE A & B OU D MONITORING NETWORK OPTIMIZATION
McClellan AFB
Sacramento Valley, California
APPENDICES
Appendix A: Zone A and Zone B McClellan OU D Historical TCE Maps
Appendix B: Zone A and Zone B McClellan OU D MAROS 2.0 Reports
-------
January 15 2003 GROUNDWATER
GSI Job No. G-2236-15 SERVICES, INC.
MAROS 2.0 APPLICATION
ZONE A & B OU D MONITORING NETWORK OPTIMIZATION
McClellan AFB
Sacramento Valley, California
APPENDIX A: Zone A and B McClellan AFB Historical TCE Maps
Zone A and Zone B McClellan OU D Historical TCE Maps (1990 - 2001)
-------
5
05 ~^>
E ?
m
o •<-
O o
O
m
o
<5 o
o m
0) O
in
A A
© © ©
D)
O "£> C CD
T-^ o '"* W
A o:
g,
LU
<
o
w
o
00
CD
o ^r
Q
o
- .:- ©
x ©
--.©© © ©
X ©
© ©
0
0
CN
CD O
CN y
6 ^
6 -a
^
O (/)
7T
-------
c
_g
2
+•«
c
o
o
uj
D)
ID
O
O
D)
ID
O
9 d TJ 1
T £ & &
V A A A A Qi ^ OQ
© © ©
O
00
O ^^
uj
©
Q
O
X
X
x x
© ©
X © ©
5U©=
X ©
0
CN
CD O
CN y
6 2
6 -a
^
O j/5
7T CD
-------
c
_g
2
+•«
c
o
o
uj
D)
ID
O
O
0
D)
ID
O
O
CD
A
O
A
O
A
a:
© © ©
<
o
O
00
CD
O
00
uj
o ^r
©
Q
O
©
0
CM
CD O
CNI y
6 2
6 -a
^
O j/5
7T
-------
c
_g
2
+•«
c
o
o
uj
D)
ID
O
O
D)
ID
O
V A A A
A a:
© © ©
O
00
O ^^
uj
©
Q
O
X
© ©
© ©
x;=x=
X ©
©
CN
CD O
CN y
6 ™
6 -a
^
O j/5
7T CD
-------
c
_g
2
+•«
c
o
o
uj
D)
ID
O
O
D)
ID
O
a:
© © ©
O
00
O ^^
uj
©
Q
O
X
X
© ©
© ©
1®
CN
CD O
CN y
6 ™
6 -a
^
O j/5
7T CD
»c
©
-------
c
_g
2
+•«
c
o
o
uj
D)
ID
O
O
D)
ID
O
V A A A
A a:
© © ©
O
00
O ^^
uj
Q
O
©
©
i©
©
X © ©
3U©=
© ©
©
©
1®
CN
CD O
CN 5
6 2
6 -a
^
O V)
7T CD
GX
-------
C
_g
2
05 ~^>
E ?
m
o •<-
O o
O
m
o
<5 o
o m
0) O
in
A A
© © ©
D)
O "£> C CD
T-^ o '"* W
A o:
g,
LU
<
o
w
o
00
CD
o ^r
©
Q
o
©
-.0®
©
,.,0,::^
© ©
© ©
,©=
©
CN
CD O
CN y
6 ^
6 -a
^
O (/)
7T
-------
c
_g
2
+•«
c
o
o
uj
D)
ID
O
O
D)
ID
O
a:
© © ©
O
00
O ^^
uj
©
©
: ©
"-. © ©
...0© © ©
0
©
0
1®
CN
CD O
CN y
6 ™
6 -a
^
O j/5
7T CD
-------
c
_g
2
+•«
c
o
o
ill
D)
ID
O
O
D)
ID
O
a:
© © ©
O
00
O ^^
uj
Q
O
©
X
©
©
CN
CD O
CN y
6 ™
6 -a
^
O j/5
7T CD
-------
c
_g
2
+•«
c
o
o
uj
D) -g^
10 "3)
o ^- E
O o _
m
o
9 d TJ 1
T £ & &
V A A A A Qi ^ OQ
© © ©
O
00
O ^^
uj
0
Q
O
x-
iX
X
© ©
X © ©
© ©
©
CN
CD O
CN y
6 2
6 -a
^
O j/5
7T CD
-------
c
_g
2
+•«
c
o
o
uj
D)
ID
O
O
D)
ID
O
a:
© © ©
O
00
O ^^
uj
Q
O
©
.-.0® ©
X ©
©
© ,..
-------
CM
CD O
CN 55
CM 5
6 2
6 -a
Z CD
3
O (/)
— 32
^^^w ^^^r
^^H
^
c
O)
OJ
1
_l
.11
o
o ^r
X
X
XX
c
o
nt
5
+•«
c
U
£ c
O ;
HI
0 '
H
l,v_
••.-.-
x
^
^5> ~
P c
C c
m
f") 'V
o c
O^ i
i JC
0 c
D S c
Z V A
X « ©
I
J
/
J/
/•
p/r
. r~ : •• =-
D _,
*^^
O)
m o :
ci -^
1 1
T- m <
d> d>
A A
© © (
, •-"-'-
^^^^^e^^^, ..
"
-
" \ 3
;X .-©
X X
ii f£\ -„
'• •: ,~J£,r ~™~~^~
©
X
_J
•^~,
O
E
o
c —
A
s
<
x
^**
CO
•Q
"O C
CO 3
o o
a: oa
"° C CD
0 TO |S
a: S tn
: 1 ' - - . '
—
*~*
3
O
X
X 6
© :
___:^.@':: ?
Q'i-"
©
'
-J
:
i !l
i UJ :-
, 0 P-^=^"y .
|| !' _v:::v5i
"!«"X"; '
"' ' i: ••':
-- -. --
1: •! .
.;'
- -?^'' i '•''•.' :. •
©
._ - :::::
:' /" ' ' ^
:-.===:::: - T=-~: -T* =
*"' ^^
i;
X ii
i i]
i : j
X :i
O
00
CD
O
CO
o
LU
O
1-
a3
^
^
O
<
(U
c
d
=>
o
^~
o
o
CM
a
CT
C
C
(-
(/
(/
^
c
"c
-1-
c
a
c
£
c
C
LI
c
§
-1—
^
c
CI
0
£
c
a
a
C]
z
CO
1
o"
1 •
(D
- s—
o
(D
m
o
o
1 1
l_l_
<
c
(D
-(D
O
O
J CD
: CD
> *-
i .1
s "co
3 ^
| |
> O
i Q
cc
Jo
! ^—.
1 ¥
c
CD
! O
L ^
1
1
-------
c
U
o
c E == ^
in c en Jo
| 5 | . j |
17 g ° 7 E a: m
V A A A A Qi ^ OQ
<
o
O LJJ
0 Q X « © © ©
O ^^
O
o
©
© © ©
•©
CN
CD O
CN ^
CN 5
6 2
6 -a
~7 CD
3
O (/)
~3 —
7T -i
2 "S3
c E
X o
CD ^
a: £
CD
•5
-------
c
U
o
c E == ^
in c en Jo
| 5 | . j |
17 g ° 7 E a: m
V A A A A Qi ^ OQ
<
o
O LJJ
0 Q X « © © ©
O ^^
O
o
© © ©
IT*
CN
CD O
CN ^
CN 5
6 2
6 -a
~7 CD
3
O (/)
~3 —
7T -i
2 "S3
c E
X o
CD ^
a: £
CD
•5
-------
CN
CO CD
CN 5
6 2
6 -a
^
O l/)
-3 —
— 32
^f
1~~ '
_ _ ,_, ^ ~ ~ ~ . w .. ro
C ^ Z V A A A A Oi S CD
II ' ' II:
- ' • . I I: - - '"
;; M 1: ,.
--=- ; ---= .,.. jjj ':. -
?: ' 3 :'; .-• •••_
11 'I ' — '
'1 ! '' ., - "
g ir .. .
o . i "".;'••...
1 I :-. ' " "" ; ..r ' v
"X _J
i
.
j
_X : s; "•-
^^ . - -"- ' :
O •""•'"
'i •::::. .. '',
•; ':" . ,j:-so
;.'; --,. ! ^ ':'-:--'' O
1 ,-;=- !; ill; !• •!•
'",--:.. 1 . - — ----••.••••-' , ! i: ill
'. '"" .- .::::-: :.- " ~~ " :., !i :|:
i i
l|
: i : :
" ". i
: 1 ;j
i i i:
i i ' •
.^^l j
— |
i s:
| |i
o
o
\ —
to
ro
V J
o
LU
O
1-
^
'3
a
<
CQ
0)
c
o
N
Q
^
O
CN
CO
^~
L.
cc
c
c
j:
(/
(/
c
c
1
"c
a
c
^
c
C
LI
(__
4
H—
p
-i-
c
Cj
(/
a
L.
C
a
a
•
Cj
Z
CD
|
CD
^
o"
•*-•
CD
E
CD
C/)
CD"
CO
CD
CQ
8
o
^
<
c
"CD
O
o
CD
! CD
: >^
: CD
>S
is
: c
i o
jl
| |
> o
o
CO
! o
'i=
5 1
O
L a.
i
•
-------
10
<5 o
o m
§ , , E a: DQ
Z V A A A A Oi ^ OQ
o
o
r~~
o
LO
CO
CN
CD O
CN ^O
CN 5
6 ™
6 -a
~7 CD
3
O (/)
~3 —
7T
© ©
0
(0
LU $
O U
I- O
I |
o- S
CQ C/3
c (0
o co
N m
o-g
^ O
O t
CO <
CD T_
CD b
CD £-
CO CD
.0 O
-HJ '*-!
£ 2
CD CD
O O
3§
2
c E
CD o
|o
g-o
as
-------
_g
to
0)
u
O) LU
fl) Q X
E |>
10
o ^-
q d
O i
E
LO
,
0)
E
V A A A A Qi
• ©00©
UJ
_l
<
o
O
O
CM
CD O
CN ^O
CNI 5
6 2
6 -a
~7 CD
3
O (ft
7T
o ^r
Q
O
©
©
(0
O
| I
—) CO
°" b
^ (0
CQ CO
(U (U
C
o
N
(0
co
s
^ °
o t
^r <
S c
^
O
o
*= CO
CO CD
§ CD
.0 O
-HJ '-^J
2 2
CD CD
O O
3§
2
c E
CD o
|o
CD fD
a: £
CD
•5
-------
10
<5 o
o m
§ , , E a: DQ
Z V A A A A Oi ^ OQ
o
o
r~~
o
LO
CO
CN
CD O
CN ^
6 2
6 -a
~7 CD
3
O (/)
~3 —
7T
©
®
®
©
(0
LU ^
O U
I- O
I |
§ g
< (0
CQ C/)
C (0
o co
N CO
O o
—\ o
o t
10 <
CD T_
CD b
CD ^~
CO CD
.0 O
-HJ '*-!
CO CO
e^
8 8
3§
^ o
g-o
as
CD
•5
-------
10
<5 o
o uj
§ , , E a: DQ
Z V A A A A Oi ^ OQ
o
o
r~~
o
LO
CO
CN
CD O
CN ^
6 2
6 -a
Z CD
O (/)
—} —
7T
ZJ
O
©
©
©
(0
-3
I- o
m v}
c w
O ^5
N CQ
Q" ^
^ £
O .<_
CD <
CD c
C^ (0
CD £-
CO CD
.0 O
-HJ '*-!
£ 2
CD CD
O O
C§§
^ o
g-o
as
CD
•5
-------
_g
to
0)
u
O) UJ
fl) Q X
E |>
10
o •<-
q d
O i
E
LO
~
0)
E
V A A A A Qi
• ©00©
UJ
_l
<
o
O
O
CN
« §
SI °
6 ™
6 -a
~7 CD
^
O j/5
7T
o ^r
ZJ
O
Q
O
©
©
0
O
=5 (0
< (0
CQ w
cu aT
c <2
08
^ £
o .•_
f-^ <^
CO .-
O
O
P 5s
5s ro
CO CD
5 CD
.0 O
-HJ '*-!
£ 2
CD CD
O O
3§
o
CD ^
a: £
CD
•5
-------
m
o
o
O)
c
o
C O ^
O) LU
0) O
b T3
d - LO q ^ |§
17 g ° 7 E a: oa
V A A A A Qi ^ OQ
« © © © © II
O ^^
ZJ
O
Q
O
© © ©
•©
CN
CD O
CN
O
O ™
6 -a
~7 CD
^
O (/)
~3 —
7T
-------
_g
to
LO
o
o
O)
,
0)
E
O) LU
fl) Q X
V A A A A Qi
• ©00©
UJ
_l
<
o
O
O
CM
« §
SI °
6 ™
6 -a
~7 CD
^
O j/5
7T
O ^^
©
Q
O
©
©
©
'©
•O
O
LU
O
=5 (0
< (0
CQ w
(U
-------
m
o
o
O)
c
o
C O ^
O) LU
0) O
i
V A A A
E o: tn
^D "^ C fl)
CD •— (/)
A on S m
© © ©
O ^^
O
O
©
©
CN
CD O
CN
O
O ™
6 -a
~7 CD
^
O (/)
~3 —
7T
-------
in
0)
U
©
O
o
o
LO
CO
(0
-3
I- 2
LJ c
0) £
t E
13 (0
°" b
< (0
CQ W
cu a)
O ^5
N CQ
Q" ^
^ £
O .•_
o c
8 J
^
o
o
a> s-
^ ro
CO CO
c ><
g O)
O £
« o
.0 O
-HJ '-^J
CO CO
-I—' ^.J
a> CD
o o
o o
O o
LLJ c
O a>
S|
m ^
^
CD ^
a: £
ai
•5
-------
January 15 2003 GROUNDWATER
GSI Job No. G-2236-15 SERVICES, INC.
MAROS 2.0 APPLICATION
ZONE A & B OU D MONITORING NETWORK OPTIMIZATION
McClellan AFB
Sacramento Valley, California
APPENDIX B: Zone A and B McClellan AFB MAROS 2.0 Reports
Linear Regression Statistics Summary
Mann-Kendall Statistics Summary
Spatial Moment Analysis Summary
Zeroth, First, and Second Moment Reports
Plume Analysis Summary
Site Results Summary
Sampling Location Optimization Results
Sampling Frequency Optimization Results
Risk-Based Power Analysis - Plume Centerline Concentrations
Risk-Based Power Analysis - Site Cleanup Status
-------
MAROS Linear Regression Statistics Summary
Project: McClellan Zone A OU D
Location: McClellan AFB
Julia Aziz
California
Time Period: 5/1/1990 to 12/31/2000
Consolidation Period: No Time Consolidation
Consolidation Type: Median
Duplicate Consolidation: Average
ND Values: Specified Detection Limit
J Flag Values : Actual Value
Well
Average
Source/ Cone
Tail (mg/L)
Median
Cone
(mg/L)
Standard
Deviation
All
Samples
"ND" ?
Coefficient
Ln Slope of Variation
Confidence Concentration
in Trend Trend
TRICHLOROETHYLENE (TCE)
MW-12
EW-73
MW-72
MW-38D
MW-351
MW-242
MW-241
MW-14
MW-11
MW-10
EW-87
EW-86
EW-85
EW-84
EW-83
MW-15
MW-1042
MW-1041
MW-1004
MW-52
MW-91
MW-90
MW-89
MW-88
MW-76
MW-74
MW-70
MW-1026
MW-53
MW-92
MW-458
MW-412
MW-350
MW-240
MW-237
MW-1073
MW-1064
MW-55
s
s
s
s
s
s
s
s
s
s
s
s
s
s
s
s
T
T
T
T
T
T
T
T
T
T
T
T
T
T
T
T
T
T
T
T
T
T
5.8E-01
2.6E-01
1.3E-01
1.5E-01
5.2E-03
2.2E-02
4.1E-02
1 .4E+00
9.5E-01
3.4E-01
1.2E-01
1 .2E-02
1.9E-01
4.0E-01
1.0E-01
2.4E-01
1.1E-04
1 .OE-04
2.0E-04
1 .8E-04
2.3E-03
3.1E-03
1 .7E-04
8.6E-03
1 .4E-03
2.5E-03
1 .OE-04
2.7E-03
1 .2E-03
1 .OE-03
2.0E-04
4.2E-04
2.7E-03
5.9E-04
4.2E-04
1 .3E-03
3.7E-03
1 .5E-03
5.8E-01
1 .7E-01
6.7E-02
1 .5E-01
2.6E-03
1 .4E-02
2.1E-02
1.0E+00
4.4E-01
3.3E-01
8.4E-02
7.7E-03
1 .2E-01
2.5E-01
9.6E-02
8.3E-02
1 .OE-04
1 .OE-04
1 .3E-04
1 .OE-04
1 .5E-03
2.4E-04
1 .OE-04
1 .OE-04
1 .OE-04
2.8E-03
1 .OE-04
3.4E-04
4.8E-04
6.3E-04
1 .9E-04
2.7E-04
7.1E-04
1 .OE-04
1 .OE-04
4.9E-04
3.8E-04
1 .7E-03
3.5E-01
2.4E-01
1 .2E-01
6.4E-02
4.9E-03
2.1E-02
5.7E-02
1.5E+00
1.5E+00
2.8E-01
8.2E-02
1 .5E-02
1 .5E-01
3.3E-01
3.1E-02
3.9E-01
2.5E-05
8.7E-06
1 .4E-04
2.8E-04
2.8E-03
4.2E-03
2.0E-04
2.8E-02
2.1E-03
6.7E-04
O.OE+00
3.3E-03
1 .8E-03
1 .4E-03
8.3E-05
3.7E-04
5.0E-03
1 .2E-03
5.7E-04
1 .8E-03
7.1 E-03
9.9E-04
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
Yes
No
No
No
No
No
No
No
No
No
No
No
-6.9E-04
-7.2E-04
-7.8E-04
-3.7E-04
4.9E-04
-8.8E-04
-1.1 E-03
-2.5E-03
-1 .4E-03
-7.5E-04
4.6E-04
-7.6E-04
-6.6E-04
-1 .1 E-03
2.3E-05
-6.5E-04
^.OE-05
2.9E-05
-3.0E-04
-4.0E-04
-8.9E-04
5.5E-04
3.4E-05
4.9E-04
1.1 E-03
-1 .7E-04
O.OE+00
2.4E-04
7.4E-04
-3.3E-04
1 .2E-03
-2.7E-03
-2.3E-03
-6.9E-05
4.7E-04
-8.2E-04
1 .2E-03
-1 .4E-04
0.60
0.92
0.92
0.42
0.94
0.92
1.41
1.12
1.53
0.83
0.68
1.23
0.79
0.83
0.30
1.64
0.23
0.08
0.69
1.52
1.19
1.37
1.14
3.28
1.53
0.26
0.00
1.26
1.45
1.39
0.42
0.88
1.88
2.07
1.36
1.35
1.92
0.65
100.0%
100.0%
100.0%
98.9%
82.3%
100.0%
99.8%
100.0%
100.0%
100.0%
99.9%
99.6%
99.9%
100.0%
60.7%
99.3%
63.0%
72.6%
94.4%
91 .6%
98.6%
75.7%
56.1%
77.4%
100.0%
97.0%
100.0%
60.2%
99.2%
80.5%
64.9%
85.0%
97.4%
53.8%
72.2%
78.6%
91 .8%
72.5%
D
D
D
D
NT
D
D
D
D
D
I
D
D
D
NT
D
S
NT
PD
PD
D
NT
NT
NT
I
D
S
NT
I
NT
NT
S
D
NT
NT
NT
PI
S
MAROS Version 2, 2002, AFCEE
Monday, January 13, 2003
Page 1 of 2
-------
project; McClellan Zone A OU D
Julia Aziz
McClellan AFB
California
Well
Average Median
Source/ Cone Cone
Tail (mg/L) (mg/L)
All
Standard Samples Coefficient Confidence Concentration
Deviation "ND" ? Ln Slope of Variation in Trend Trend
TRICHLOROETHYLENE (TCE)
Note: Increasing (I); Probably Increasing (PI); Stable (S); Probably Decreasing (PD); Decreasing (D); No Trend (NT); Not Applicable (N/A) -
Due to insufficient Data (< 4 sampling events); COV = Coefficient of Variation
MAROS Version 2, 2002, AFCEE
Monday, January 13, 2003
Page 2 of 2
-------
MAROS Mann-Kendall Statistics Summary
McClellan Zone A OU D
Location: McClellan AFB
Julia Aziz
California
Time Period: 5/1/1990 to 12/31/2000
Consolidation Period: No Time Consolidation
Consolidation Type: Median
Duplicate Consolidation: Average
ND Values: Specified Detection Limit
J Flag Values : Actual Value
Source/ Number of
Well Tail Samples
TRICHLOROETHYLENE
MW-72
MW-351
MW-38D
MW-242
MW-15
MW-12
MW-241
MW-11
EW-73
MW-10
EW-87
EW-86
EW-85
EW-84
EW-83
MW-14
MW-1041
MW-1073
MW-1042
MW-1026
MW-1004
MW-1064
MW-55
MW-91
MW-90
MW-89
MW-88
MW-76
MW-237
MW-70
MW-240
MW-53
MW-52
MW-458
MW-412
MW-350
MW-92
MW-74
(TCE)
S
s
S
s
s
s
s
s
s
s
s
s
s
s
s
s
T
T
T
T
T
T
T
T
T
T
T
T
T
T
T
T
T
T
T
T
T
T
10
9
8
12
11
12
12
11
15
11
15
15
16
16
16
12
7
6
7
7
12
8
11
10
8
10
11
10
7
6
9
10
9
4
6
8
12
10
Number of
Detects
10
9
8
12
11
12
12
11
15
11
15
15
16
16
16
12
1
5
1
5
6
4
11
9
6
2
4
4
5
0
5
9
3
3
4
5
9
10
Coefficient
of Variation
0.92
0.94
0.42
0.92
1.64
0.60
1.41
1.53
0.92
0.83
0.68
1.23
0.79
0.83
0.30
1.12
0.08
1.35
0.23
1.26
0.69
1.92
0.65
1.19
1.37
1.14
3.28
1.53
1.36
0.00
2.07
1.45
1.52
0.42
0.88
1.88
1.39
0.26
Mann-Kendall
Statistic
-37
7
-18
-46
-23
-52
-42
-49
-83
-53
60
-53
-84
-105
22
-58
4
-3
-2
-4
-23
6
-11
-31
7
3
-4
20
2
0
0
23
-13
0
-4
-15
-18
-28
Confidence
in Trend
100.0%
72.8%
98.4%
100.0%
95.7%
100.0%
99.8%
100.0%
100.0%
100.0%
99.9%
99.6%
100.0%
100.0%
82.5%
100.0%
66.7%
64.0%
55.7%
66.7%
93.3%
72.6%
77.7%
99.8%
76.4%
56.9%
59.0%
95.5%
55.7%
42.3%
46.0%
97.7%
89.0%
37.5%
70.3%
95.8%
87.5%
99.4%
All
Samples Concentration
"ND" ? Trend
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
Yes
No
No
No
No
No
No
No
No
D
NT
D
D
D
D
D
D
D
D
I
D
D
D
NT
D
NT
NT
S
NT
PD
NT
S
D
NT
NT
NT
I
NT
S
NT
I
NT
S
S
D
NT
D
MAROS Version 2, 2002, AFCEE
Monday, January 13, 2003
Page 1 of 2
-------
McClellan Zone A OU D Julia Aziz
Location: McClellan AFB California
All
Source/ Number of Number of Coefficient Mann-Kendall Confidence Samples Concentration
We|| Tai| Samples Detects of Variation Statistic in Trend "ND" ? Trend
TRICHLOROETHYLENE (TCE)
Note: Increasing (I); Probably Increasing (PI); Stable (S); Probably Decreasing (PD); Decreasing (D); No Trend (NT); Not Applicable (N/A)-
Due to insufficient Data (< 4 sampling events); Source/Tail (S/T)
The Number of Samples and Number of Detects shown above are post-consolidation values.
MAROS Version 2, 2002, AFCEE Monday, January 13, 2003 Page 2 of 2
-------
MAROS Statistical Trend Analysis Summary
Project: McClellan Zone A OU D
Location: McClellan AFB
Julia Aziz
California
Time Period: 5/1/1990 to 12/31/2000
Consolidation Period: No Time Consolidation
Consolidation Type: Median
Duplicate Consolidation: Average
ND Values: Specified Detection Limit
J Flag Values : Actual Value
Well
TRICHLOROETHYLENE
EW-73
EW-83
EW-84
EW-85
EW-86
EW-87
MW-10
MW-1004
MW-1026
MW-1041
MW-1042
MW-1064
MW-1073
MW-11
MW-12
MW-14
MW-15
MW-237
MW-240
MW-241
MW-242
MW-350
MW-351
MW-38D
MW-412
MW-458
MW-52
MW-53
MW-55
MW-70
MW-72
MW-74
MW-76
MW-88
MW-89
Source/
Tail
(TCE)
s
s
s
s
s
s
s
T
T
T
T
T
T
S
S
s
s
T
T
S
s
T
S
s
T
T
T
T
T
T
S
T
T
T
T
Number Number Average Median
of of Cone. Cone.
Samples Detects (mg/L) (mg/L)
15
16
16
16
15
15
11
12
7
7
7
8
6
11
12
12
11
7
9
12
12
8
9
8
6
4
9
10
11
6
10
10
10
11
10
15
16
16
16
15
15
11
6
5
1
1
4
5
11
12
12
11
5
5
12
12
5
9
8
4
3
3
9
11
0
10
10
4
4
2
2.6E-01
1 .OE-01
4.0E-01
1 .9E-01
1 .2E-02
1 .2E-01
3.4E-01
2.0E-04
2.7E-03
1 .OE-04
1.1E-04
3.7E-03
1 .3E-03
9.5E-01
5.8E-01
1 .4E+00
2.4E-01
4.2E-04
5.9E-04
4.1E-02
2.2E-02
2.7E-03
5.2E-03
1 .5E-01
4.2E-04
2.0E-04
1 .8E-04
1 .2E-03
1 .5E-03
1 .OE-04
1 .3E-01
2.5E-03
1 .4E-03
8.6E-03
1 .7E-04
1 .7E-01
9.6E-02
2.5E-01
1 .2E-01
7.7E-03
8.4E-02
3.3E-01
1 .3E-04
3.4E-04
1 .OE-04
1 .OE-04
3.8E-04
4.9E-04
4.4E-01
5.8E-01
1.0E+00
8.3E-02
1 .OE-04
1 .OE-04
2.1E-02
1 .4E-02
7.1E-04
2.6E-03
1 .5E-01
2.7E-04
1 .9E-04
1 .OE-04
4.8E-04
1 .7E-03
1 .OE-04
6.7E-02
2.8E-03
1 .OE-04
1 .OE-04
1 .OE-04
All
Samples
"ND" ?
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
Yes
No
No
No
No
No
Mann-
Kendall
Trend
D
NT
D
D
D
I
D
PD
NT
NT
S
NT
NT
D
D
D
D
NT
NT
D
D
D
NT
D
S
S
NT
I
S
S
D
D
I
NT
NT
Linear
Regression
Trend
D
NT
D
D
D
I
D
PD
NT
NT
S
PI
NT
D
D
D
D
NT
NT
D
D
D
NT
D
S
NT
PD
I
S
S
D
D
I
NT
NT
MAROS Version 2, 2002, AFCEE
Monday, January 13, 2003
Page 1 of 2
-------
MAROS Statistical Trend Analysis Summary
Source/
Well Tai|
Number
of
Samples
Number
of
Detects
Average
Cone.
(mg/L)
Median
Cone.
(mg/L)
All
Samples
"ND" ?
Mann-
Kendall
Trend
Linear
Regression
Trend
TRICHLOROETHYLENE (TCE)
MW-90
MW-91
MW-92
T
T
T
8
10
12
6
9
9
3.1E-03
2.3E-03
1 .OE-03
2.4E-04
1 .5E-03
6.3E-04
No
No
No
NT
D
NT
NT
D
NT
Note: Increasing (I); Probably Increasing (PI); Stable (S); Probably Decreasing (PD); Decreasing (D); No Trend (NT); Not Applicable
(N/A); Not Applicable (N/A) - Due to insufficient Data (< 4 sampling events); No Detectable Concentration (NDC)
The Number of Samples and Number of Detects shown above are post-consolidation values.
MAROS Version 2, 2002, AFCEE
Monday, January 13, 2003
Page 2 of 2
-------
MAROS Site Results
Project: McClellan Zone A OU D
Location: McClellan AFB
Julia Aziz
California
User Defined Site and Data Assumptions:
Hydrogeology and Plume Information:
Groundwater
Seepage Velocity: 35 ft/yr
Current Plume Length: 1000 ft
Current Plume Width 600 ft
Number of Tail Wells: 22
Number of Source Wells: 16
Source Information:
Source Treatment: Pump and Treat
NAPL is not at this site.
Down-gradient Information:
Distance from Edge of Tail to Nearest:
Down-gradient receptor: 1000 ft
Down-gradient property: 10ft
Distance from Source to Nearest:
Down-gradient receptor: 6000ft
Down-gradient property: 10 ft
Consolidation Assumptions:
Time Period: 5/1/1990 to 12/31/2000
Consolidation Period: No Time Consolidation
Consolidation Type: Median
Duplicate Consolidation: Average
ND Values: Specified Detection Limit
J Flag Values : Actual Value
Plume Information Weighting Assumptions:
Consolidation Step 1. Weight Plume Information by Chemical
Summary Weighting: Weighting Applied to All Chemicals Equally
Consolidation Step 2. Weight Well Information by Chemical
Well Weighting: No Weighting of Wells was Applied.
Chemical Weighting: No Weighting of Chemicals was Applied.
Note: These assumptions made when consolidating the historical mentoring lumping the Wells and COCs.
1.
Preliminary Monitoring System Optimization Results: Based on site classification, source treatment and Monitoring System
Category the following suggestions are made for site Sampling Frequency, Duration of Sampling, and Well Density. These
criteria take into consideration: Plume Stability, Type of Plume, and Groundwater Velocity.
coc
Tail Source Level of
Stability Stability Effort
Sampling
Duration
Sampling
Frequency
Sampling
Density
TRICHLOROETHYLENE (TCE)
PD
M
25
Remove treatment No Recommendation
system if previously
reducing concentation
Note:
Plume Status: (I) Increasing; (Pl)Probably Increasing; (S) Stable; (NT) No Trend; (PD) Probably Decreasing; (D) Decreasing
Design Categories: (E) Extensive; (M) Moderate; (L) Limited (N/A) Not Applicable, Insufficient Data Available
Level of Monitoring Effort Indicated by Analysi I Moderate
2,
MAROS Version 2, 2002, AFCEE
Tuesday, January 14, 2003
Page 1 of 2
-------
Moment Type Consituent
Zeroth Moment: Mass
TRICHLOROETHYLENE (TCE)
1st Moment: Distance to Source
TRICHLOROETHYLENE (TCE)
2nd Moment: Sigma XX
TRICHLOROETHYLENE (TCE)
2nd Moment: Sigma YY
TRICHLOROETHYLENE (TCE)
Coefficient
of Variation
1.44
1.04
1.26
1.72
Mann-Kendall
S Statistic
-116
48
-18
-18
Confidence
in Trend
91 .8%
87.7%
66.2%
66.2%
Moment
Trend
PD
NT
NT
NT
Note: The following assumptions were applied for the calculation of the Zeroth Moment:
Porosity: 0.30
Saturated Thickness: Uniform: 30 ft
Mann-Kendall Trend test performed on all sample events for each constituent. Increasing (I); Probably Increasing (PI); Stable (S);
Probably Decreasing (PD); Decreasing (D); No Trend (NT); Not Applicable (N/A)-Due to insufficient Data (< 4 sampling events).
MAROS Version 2, 2002, AFCEE
Tuesday, January 14, 2003
Page 2 of 2
-------
MAROS Zeroth Moment Analysis
McClellan AFB Zone A OU D
Location: McClellan AFB
Julia Aziz
California
COC: TRICHLOROETHYLENE (TCE)
Change in Dissolved Mass Over Time
Date
S
1 9F+ni -
1.0E+01 -
8.0E+00 -
6.0E+00 -
4.0E+00 -
2.0E+00 -
n nF+nn -
*
*
• *
* * * *
Porosity: 0.30
Saturated Thickness:
Uniform: 30 ft
Mann Kendall S Statistic:
-15
Confidence in
Trend:
I 85.9%
Coefficient of Variation:
I 0.96
Zeroth Moment
Trend:
Data Table:
Effective Date
7/1/1990
7/1/1991
7/1/1992
7/1/1993
7/1/1994
7/1/1995
7/1/1996
7/1/1997
7/1/1998
7/1/1999
7/1/2000
Constituent
TRICHLOROETHYLENE (TCE)
TRICHLOROETHYLENE (TCE)
TRICHLOROETHYLENE (TCE)
TRICHLOROETHYLENE (TCE)
TRICHLOROETHYLENE (TCE)
TRICHLOROETHYLENE (TCE)
TRICHLOROETHYLENE (TCE)
TRICHLOROETHYLENE (TCE)
TRICHLOROETHYLENE (TCE)
TRICHLOROETHYLENE (TCE)
TRICHLOROETHYLENE (TCE)
Estimated
Mass (Kg)
1.3E+01
5.8E+00
8.8E-01
2.1E+00
1.2E+01
2.2E+00
1.3E+01
7.1E+00
7.9E-01
9.4E-01
9.0E-01
Number of Wells
20
21
11
25
18
28
15
20
19
29
13
Note: Increasing (I); Probably Increasing (PI); Stable (S); Probably Decreasing (PD); Decreasing (D); No Trend (NT); Not Applicable (N/A) -
Due to insufficient Data (< 4 sampling events); ND = Non-detect. Moments are not calculated for sample events with less than 6 wells.
MAROS Version 2, 2002, AFCEE
1/2/2003
Page 1 of 1
-------
MAROS First Moment Analysis
McClellan AFB Zone A OU D
Location: McClellan AFB
Julia Aziz
California
COC: TRICHLOROETHYLENE (TCE)
Distance from Source to Center of Mass
2.5E+03
£, 2.0E+03
o 1.5E+03
CO
o
* 1.0E+03
8
w 5.0E+02 -
Q
O.OE+00
Mann Kendall S Statistic:
Date
I 17
Confidence in
Trend:
* * *
*
* *
I 89.1%
Coefficient of Variation:
First Moment Trend:
NT
Data Table:
Effective Date
7/1/1990
7/1/1991
7/1/1992
7/1/1993
7/1/1994
7/1/1995
7/1/1996
7/1/1997
7/1/1998
7/1/1999
7/1/2000
Constituent
TRICHLOROETHYLENE (TCE)
TRICHLOROETHYLENE (TCE)
TRICHLOROETHYLENE (TCE)
TRICHLOROETHYLENE (TCE)
TRICHLOROETHYLENE (TCE)
TRICHLOROETHYLENE (TCE)
TRICHLOROETHYLENE (TCE)
TRICHLOROETHYLENE (TCE)
TRICHLOROETHYLENE (TCE)
TRICHLOROETHYLENE (TCE)
TRICHLOROETHYLENE (TCE)
Xc (ft)
2,166,613
2,166,683
2,166,819
2,166,669
2,166,500
2,166,698
2,168,405
2,167,678
2,166,569
2,166,817
2,167,148
Yc (ft)
366,307
366,147
366,337
366,336
366,370
366,444
366,904
367,043
366,068
366,152
366,103
Distance from Source (ft)
220
56
287
242
322
351
1,918
1,387
100
162
482
Number of Wells
20
21
11
25
18
28
15
20
19
29
13
Note: Increasing (I); Probably Increasing (PI); Stable (S); Probably Decreasing (PD); Decreasing (D); No Trend (NT); Not Applicable (N/A) -
Due to insufficient Data (< 4 sampling events). Moments are not calculated for sample events with less than 6 wells.
MAROS Version 2, 2002, AFCEE
1/2/2003
Page 1 of 1
-------
MAROS First Moment Analysis
McClellan AFB Zone A OU D
Location: McClellan AFB
Julia Aziz
California
COC: TRICHLOROETHYLENE(TCE)
Change in Location of Center of Mass Over Time
367200
367000 •
366800 •
366600 •
366400 •
366200 •
366000
07/96
'/95
• 07/94
* 07AW/9i
«• 07/97
07/99
Groundwater
Flow Direction:
Source
Coordinate:
X: J 2,166,666
Y: I 366,094
2166000 2166500 2167000 2167500 2168000 2168500
Xc (ft)
Effective Date
Constituent
Xc (ft)
Yc (ft) Distance from Source (ft) Number of Wells
7/1/1990
7/1/1991
7/1/1992
7/1/1993
7/1/1994
7/1/1995
7/1/1996
7/1/1997
7/1/1998
7/1/1999
7/1/2000
TRICHLOROETHYLENE (TCE)
TRICHLOROETHYLENE (TCE)
TRICHLOROETHYLENE (TCE)
TRICHLOROETHYLENE (TCE)
TRICHLOROETHYLENE (TCE)
TRICHLOROETHYLENE (TCE)
TRICHLOROETHYLENE (TCE)
TRICHLOROETHYLENE (TCE)
TRICHLOROETHYLENE (TCE)
TRICHLOROETHYLENE (TCE)
TRICHLOROETHYLENE (TCE)
2,166,613
2,166,683
2,166,819
2,166,669
2,166,500
2,166,698
2,168,405
2,167,678
2,166,569
2,166,817
2,167,148
366,307
366,147
366,337
366,336
366,370
366,444
366,904
367,043
366,068
366,152
366,103
220
56
287
242
322
351
1,918
1,387
100
162
482
20
21
11
25
18
28
15
20
19
29
13
Note: Increasing (I); Probably Increasing (PI); Stable (S); Probably Decreasing (PD); Decreasing (D); No Trend (NT); Not Applicable (N/A) •
Due to insufficient Data (< 4 sampling events). Moments are not calculated for sample events with less than 6 wells.
MAROS Version 2, 2002, AFCEE
1/2/2003
Page 1 of 1
-------
MAROS Second Moment Analysis
McClellan AFB Zone A OU D
McClellan AFB
COC: TRICHLOROETHYLENE (TCE)
Change in Pin
10000000
1000000
£- 100000
Jl 10000
CM
V 1000
w
100
10
1
mnnnnnn _i
1000000 -
£- 100000 -
tr
~ 10000 -
CM
g 1000 -
«
100-
10-
1 .
\me Spread Over Time
Date
* » *
* • * * * * *
*
Date
* * * *
* * * *
• *
Data Table:
Effective Date Constituent Sigma XX (sq ft) Si
7/1/1990 TRICHLOROETHYLENE (TCE) 103,011
7/1/1991 TRICHLOROETHYLENE (TCE) 107,861
7/1/1992 TRICHLOROETHYLENE (TCE) 282,555
7/1/1993 TRICHLOROETHYLENE (TCE) 209,993
7/1/1994 TRICHLOROETHYLENE (TCE) 432,219
7/1/1995 TRICHLOROETHYLENE (TCE) 617,888
7/1/1996 TRICHLOROETHYLENE (TCE) 1,645,598
7/1/1997 TRICHLOROETHYLENE (TCE) 2,180,836
7/1/1998 TRICHLOROETHYLENE (TCE) 509,750
7/1/1999 TRICHLOROETHYLENE (TCE) 983,522
7/1/2000 TRICHLOROETHYLENE (TCE) 1,997,612
Julia Aziz
California
Mann Kendall S Statistic:
J?> I15
Confidence in
Trend:
| 85.9%
Coefficient of Variation:
| 0.91
Second Moment
Trend:
l| NT
5§>
Mann Kendall S Statistic:
j 41
Confidence in
Trend:
| 100.0%
Coefficient of Variation:
I 0.93
Second Moment
Trend:
I '
gmaYY(sqft) Number of Wells
478,732 20
321,324 21
565,080 1 1
382,557 25
101,139 18
803,729 28
2,578,566 15
2,062,276 20
338,208 19
1,859,400 29
596,906 13
MAROS Version 2, 2002, AFCEE
1/2/2003
Page 1 of 2
-------
MAROS Second Moment Analysis
Effective Date
Constituent
Sigma XX (sq ft) Sigma YY (sq ft)
Number of Wells
Note: Increasing (I); Probably Increasing (PI); Stable (S); Probably Decreasing (PD); Decreasing (D); No Trend (NT); Not Applicable (N/A) -
Due to insufficient Data (< 4 sampling events)
The Sigma XX and Sigma YY components are estimated using the given field coordinate system and then rotated to align with the
estimated groundwater flow direction. Moments are not calculated for sample events with less than 6 wells.
MAROS Version 2, 2002, AFCEE
1/2/2003
Page 2 of 2
-------
MAROS Spatial Moment Analysis Summary
McClellan AFB Zone A OU D
Location: McClellan AFB
Julia Aziz
California
Oth Moment
1st (Center of
2nd Moment
Effective Date
TRICHLOROETHYLENE
7/1/1990
7/1/1991
7/1/1992
7/1/1993
7/1/1994
7/1/1995
7/1/1996
7/1/1997
7/1/1998
7/1/1999
7/1/2000
Estimated
Mass (Kg) Xc (ft)
(TCE)
1.3E+01
5.8E+00
8.8E-01
2.1E+00
1.2E+01
2.2E+00
1.3E+01
7.1E+00
7.9E-01
9.4E-01
9.0E-01
2,166,613
2,166,683
2,166,819
2,166,669
2,166,500
2,166,698
2,168,405
2,167,678
2,166,569
2,166,817
2,167,148
Source
Yc (ft) Distance (ft)
366,307
366,147
366,337
366,336
366,370
366,444
366,904
367,043
366,068
366,152
366,103
220
56
287
242
322
351
1,918
1,387
100
162
482
Sigma XX
(sq ft)
103,011
107,861
282,555
209,993
432,219
617,888
1,645,598
2,180,836
509,750
983,522
1,997,612
Sigma YY
(sq ft)
478,732
321,324
565,080
382,557
101,139
803,729
2,578,566
2,062,276
338,208
1,859,400
596,906
Number of
Wells
20
21
11
25
18
28
15
20
19
29
13
MAROS Version 2, 2002, AFCEE
Thursday, January 02, 2003
Page 1 of 2
-------
McClellan AFB Zone A OU D
Location: McClellan AFB
Julia Aziz
California
Moment Type Consituent
Zeroth Moment: Mass
TRICHLOROETHYLENE (TCE)
1st Moment: Distance to Source
TRICHLOROETHYLENE (TCE)
2nd Moment: Sigma XX
TRICHLOROETHYLENE (TCE)
2nd Moment: Sigma YY
TRICHLOROETHYLENE (TCE)
Coefficient
of Variation
0.96
1.18
0.93
0.91
Mann-Kendall
S Statistic
-15
17
41
15
Confidence
in Trend
85.9%
89.1%
100.0%
85.9%
Moment
Trend
S
NT
I
NT
Note: The following assumptions were applied for the calculation of the Zeroth Moment:
Porosity: 0.30 Saturated Thickness: Uniform: 30 ft
Mann-Kendall Trend test performed on all sample events for each constituent. Increasing (I); Probably Increasing (PI); Stable (S);
Probably Decreasing (PD); Decreasing (D); No Trend (NT); Not Applicable (N/A)-Due to insufficient Data (< 4 sampling events).
Note: The Sigma XX and Sigma YY components are estimated using the given field coordinate system and then rotated to align with the
estimated groundwater flow direction. Moments are not calculated for sample events with less than 6 wells.
MAROS Version 2, 2002, AFCEE
Thursday, January 02, 2003
Page 2 of 2
-------
1 MAROS Sampling Location Optimization Re suits 1
Project: McClellan OUD ZoneA
McClellan AFB
Sampling Events Analyzed: From 1995
7/1/1995
Parameters used: Constituent
TRICHLOROETHYLENE (TCE)
Meng
California
to 2000
7/1/2000
Inside SF Hull SF Area Ratio Cone. Ratio
0.2 0.01 0.95 0.95
Average Minimum Maximum
Well X (feet) Y (feet) Removable? Slope Factor* Slope Factor* Slope Factor* Eliminated?
TRICHLOROETHYLENE (TCE)
MW-10 2166666.50 366103.81 0
MW-1004 2165862.25 366561.97 0
MW-1026 2168527.25 367760.69 0
MW-1042 2165134.00 368116.88 0
MW-1064 2166045.50 368281.13 0
MW-1073 2165967.75 367415.38 0
MW-11 2166621.50 366694.28 0
MW-12 2167022.25 366593.50 0
MW-14 2166783.25 365890.03 0
MW-15 2166639.25 365844.72 0
MW-237 2164664.25 365534.06 0
MW-240 2165901.00 365548.66 0
MW-241 2166666.00 366093.94 0
MW-242 2166642.50 365855.91 0
MW-350 2169199.50 365231.88 0
MW-351 2167722.25 364740.31 0
MW-38D 2166617.50 366507.66 0
MW-412 2167940.50 364881.28 0
MW-458 2167267.75 365669.84 0
MW-52 2166983.75 366826.34 0
MW-53 2166865.50 366619.34 0
MW-55 2166860.25 366290.91 0
MW-70 2166867.50 366745.16 0
MW-72 2166654.25 366123.69 0
MW-74 2166534.25 366146.22 0
MW-76 2166529.50 366444.34 0
MW-88 2167425.50 366186.38 0
MW-89 2167204.00 366317.66 0
0.208 0.156 0.232 D
0.677 0.529 0.892 D
0.506 0.474 0.538 D
0.533 0.398 0.606 D
0.431 0.258 0.555 D
0.373 0.096 0.908 D
0.299 0.080 0.542 D
0.601 0.387 0.778 D
0.135 0.018 0.404 0
0.254 0.190 0.323 D
0.357 0.261 0.499 D
0.591 0.217 0.703 D
0.169 0.081 0.241 0
0.203 0.081 0.292 D
0.392 0.209 0.661 D
0.416 0.068 0.675 D
0.294 0.162 0.374 D
0.477 0.104 0.721 D
0.300 0.187 0.553 D
0.663 0.538 0.779 D
0.374 0.172 0.458 D
0.430 0.212 0.649 D
0.664 0.534 0.896 D
0.032 0.000 0.070 0
0.264 0.262 0.266 D
0.235 0.136 0.336 D
0.457 0.291 0.549 D
0.651 0.481 0.898 D
MAROS Version 2, 2002, AFCEE
Friday, December 13, 2002
Page 1 of2
-------
McClellan OUD ZoneA
McClellan AFB
Meng
California
Well
MW-90
MW-91
MW-92
X (feet)
2167220.00
2166909.75
2166911.50
Y (feet)
365966.75
365608.53
365476.94
Removable?
0
0
0
Average
Slope Factor*
0.249
0.331
0.364
Minimum
Slope Factor*
0.021
0.095
0.139
Maximum
Slope Factor*
0.347
0.736
0.505
Eliminated?
D
D
D
Note: The Slope Factor indicates the relative importance of a well in the monitoring network at a given sampling event; the larger the SF
value of a well, the more important the well is and vice versa; the Average Slope Factor measures the overall well importance in the
selected time period; the state coordinates system (i.e., X and Y refer to Easting and Northing respectively) or local coordinates systems
may be used; wells that are NOT selected for analysis are not shown above.
* When the report is generated after running the Excel module, SF values will NOT be shown above.
MAROS Version 2, 2002, AFCEE
Friday, December 13, 2002
Page 2 of2
-------
MAROS Sampling Frequency Optimization Results
McClellan OUD ZoneA
McClellan AFB
The Overall Number of Sampling Events: 42
"Recent Period" defined by events: From 1994 1st Quarter
Meng
California
To 2000 3rd Quarter
"Rate of Change" f
Well
1/15/1994
larameters used:
8/15/2000
Constituent Cleanup Goal Low Rate Medium Rate High Rate
TRICHLOROETHYLENE (TCE) 0.005
Units: Cleanup Goal is in mg/L; all rate parameters are
Recommended
Sampling Frequency
0.0025 0.005 0.01
in mg/L/year.
Frequency Based Frequency Based
on Recent Data on Overall Data
TRICHLOROETHYLENE (TCE)
EW-73
EW-83
EW-84
EW-85
EW-86
EW-87
MW-10
MW-1004
MW-1026
MW-1041
MW-1042
MW-1064
MW-1073
MW-11
MW-12
MW-14
MW-15
MW-237
MW-240
MW-241
MW-242
MW-350
MW-351
MW-38D
Annual
Annual
Annual
Annual
Annual
Quarterly
Annual
Annual
SemiAnnual
Biennial
Biennial
Quarterly
Annual
Annual
Annual
Annual
SemiAnnual
Biennial
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Quarterly Quarterly
Annual
Annual
Annual
Annual
SemiAnnual SemiAnnual
Annual
Annual
Annual
Annual
Quarterly Quarterly
Annual
Annual
Annual
Annual
SemiAnnual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
MAROS Version 2, 2002, AFCEE
Monday, December 16, 2002
Page 1 of2
-------
McClellan OUD ZoneB
Location: McClellanAFB
Meng
California
Well
MW-412
MW-458
MW-52
MW-53
MW-55
MW-70
MW-72
MW-74
MW-76
MW-88
MW-89
MW-90
MW-91
MW-92
Recommended
Sampling Frequency
Biennial
Biennial
Biennial
Annual
Biennial
Biennial
Annual
SemiAnnual
Quarterly
Annual
Biennial
Annual
Biennial
Annual
Frequency Based
on Recent Data
Annual
Annual
Annual
Annual
Annual
Annual
Annual
SemiAnnual
Quarterly
Annual
Annual
Annual
Annual
Annual
Frequency Based
on Overall Data
Annual
Annual
Annual
Annual
Annual
Annual
Annual
SemiAnnual
Quarterly
Annual
Annual
Annual
Annual
Annual
Note: Sampling frequency is determined considering both recent and overall concentration trends. Sampling Frequency is the
final recommendation; Frequency Based on Recent Data is the frequency determined using recent (short) period of monitoring
data; Frequency Based on Overall Data is the frequency determined using overall (long) period of monitoring data. If the "recent
period" is defined using a different series of sampling events, the results could be different.
MAROS Version 2, 2002, AFCEE
Monday, December 16, 2002
Page 2 of2
-------
MAROS Risk-Based Power Analysis for Site Cleanup
McClellan OUD ZoneA
McClellan AFB
Meng
California
Parameters:
Groundwater Flow Direction: 280 degrees Distance to Receptor: 1000 feet
From Period: 1995 to 2000
7/1/1995 7/1/2000
Sample Event
Selected Plume
Centerline Wells:
N
Well
MW-92
MW-91
MW-72
MW-11
The distance
from the well
Distance
to Receptor (feet)
1866.2
1996.1
2547.8
3115.4
is measured in the Groundwater Flow Angle
to the compliance boundary.
ormal Distribution Assumption
Sample Sample Sample Cleanup Expected
Szie Mean Stdev. Status Power Samp|e size
TRICHLOROETHYLENE (TCE) Cleanup Goal =
1995
1997
1998
1999
29 2.05E-07 9.05E-07 Attained 1
20 1.60E-06 4.97E-06 Attained 1
19 4.06E-06 9.41 E-06 Attained 1
29 4.23E-06 2.21 E-05 Attained 1
= 0.005
000 <=3
000 <=3
000 <=3
000 <=3
Lognormal Distribution Assumption
Celanup Expected Alpha Expected
Status Power Sample Size Level Power
Not Attained S/E S/E 0.05 0.8
Not Attained S/E S/E 0.05 0.8
Attained 1.000 4 0.05 0.8
Attained 1.000 4 0.05 0.8
Note: #N/C means "not conducted" due to a small sample size (N<4) or that the mean concentration is much greater than the cleanup
level; Sample Size is the number of sampling locations used in the power analysis; Expected Sample Size is the number of concentration
data needed to reach the Expected Power undercurrent sample variability.
MAROS Version 2, 2002, AFCEE
Friday, December 13, 2002
Page 1 of 1
-------
Risk-Based Power Analysis — Projected Concentrations
McClellan OUD ZoneA
Location : McClellan AFB
From Period:
Sampling
Event
7/1/1995 to
Effective
Date
7/1/2000
Well
Meng
California
Distance from the most downgradient well to receptor: 1000 feet
Observed
Concentration
(mg/L)
Distance Down
Centerline (ft)
Regression
Coefficient
(1/ft)
Projected
Concentration
(mg/L)
Below
Detection
Limit?
Used in
Analysis?
TRICHLOROETHYLENE (TCE)
1995
1995
1995
1995
1995
1995
1995
1995
1995
1995
1995
1995
1995
1995
1995
1995
1995
1995
1995
1995
1995
1995
1995
1995
1995
1995
1995
1995
1995
1995
1995
1995
7/1/1995
7/1/1995
7/1/1995
7/1/1995
7/1/1995
7/1/1995
7/1/1995
7/1/1995
7/1/1995
7/1/1995
7/1/1995
7/1/1995
7/1/1995
7/1/1995
7/1/1995
7/1/1995
7/1/1995
7/1/1995
7/1/1995
7/1/1995
7/1/1995
7/1/1995
7/1/1995
7/1/1995
7/1/1995
7/1/1995
7/1/1995
7/1/1995
7/1/1995
7/1/1995
7/1/1995
7/1/1995
EW-73
EW-83
EW-84
EW-85
EW-86
EW-87
MW-10
MW-1026
MW-1041
MW-1042
MW-1064
MW-1073
MW-11
MW-12
MW-14
MW-15
MW-237
MW-240
MW-241
MW-242
MW-350
MW-351
MW-38D
MW-52
MW-53
MW-55
MW-70
MW-72
MW-74
MW-76
MW-88
MW-89
3.007E-01
1.057E-01
2.618E-01
2.160E-01
5.140E-03
8.076E-02
2.848E-01
6.150E-05
1 .230E-04
6.150E-05
2.063E-02
3.052E-03
7.106E-01
6.823E-01
1.021E+00
7.601 E-02
6.945E-05
7.484E-05
2.257E-02
1 .425E-02
3.657E-03
4.216E-03
1.937E-01
5.514E-05
4.805E-04
1 .507E-04
6.150E-05
6.046E-02
2.583E-03
2.093E-03
8.494E-05
4.948E-05
2900.5
2852.5
2602.5
2307.1
2261.4
2550.4
2526.1
3834.7
4817.5
4774.7
4778.2
3939.1
3115.4
2946.6
2295.3
2275.7
2312.7
2112.3
2516.5
2286.1
1227.6
1000.0
2932.3
3182.6
2999.3
2676.7
3122.8
2547.8
2590.8
2885.3
2475.6
2643.4
-6.77E-03
-6.77E-03
-6.77E-03
-6.77E-03
-6.77E-03
-6.77E-03
-6.77E-03
-6.77E-03
-6.77E-03
-6.77E-03
-6.77E-03
-6.77E-03
-6.77E-03
-6.77E-03
-6.77E-03
-6.77E-03
-6.77E-03
-6.77E-03
-6.77E-03
-6.77E-03
-6.77E-03
-6.77E-03
-6.77E-03
-6.77E-03
-6.77E-03
-6.77E-03
-6.77E-03
-6.77E-03
-6.77E-03
-6.77E-03
-6.77E-03
-6.77E-03
8.875E-10
4.316E-10
5.811E-09
3.544E-08
1.150E-09
2.551 E-09
1.061E-08
3.247E-16
8.361 E-1 9
5.585E-19
1.830E-16
7.947E-15
4.893E-10
1 .474E-09
1.814E-07
1 .543E-08
1.097E-11
4.592E-1 1
8.972E-10
2.695E-09
8.974E-07
4.831 E-06
4.607E-10
2.409E-14
7.265E-13
2.023E-12
4.028E-14
1 .944E-09
6.206E-1 1
6.849E-12
4.452E-12
8.328E-13
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
No
No
No
No
No
No
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
MAROS Version 2, 2002, AFCEE
Friday, December 13, 2002
Page 1 of4
-------
McClellan OUD ZoneA
: McClellan AFB
Meng
California
Sampling
Event
Effective
Date
Well
Concentration
(mg/L)
Distance Down
Centerline (ft)
Coefficient
(1/ft)
Concentration
(mg/L)
Detection
Limit?
Used in
Analysis?
TRICHLOROETHYLENE (TCE)
1995
1995
1995
1997
1997
1997
1997
1997
1997
1997
1997
1997
1997
1997
1997
1997
1997
1997
1997
1997
1997
1997
1997
1997
1997
1997
1997
1997
1997
1998
1998
1998
1998
1998
1998
1998
7/1/1995
7/1/1995
7/1/1995
7/1/1997
7/1/1997
7/1/1997
7/1/1997
7/1/1997
7/1/1997
7/1/1997
7/1/1997
7/1/1997
7/1/1997
7/1/1997
7/1/1997
7/1/1997
7/1/1997
7/1/1997
7/1/1997
7/1/1997
7/1/1997
7/1/1997
7/1/1997
7/1/1997
7/1/1997
7/1/1997
7/1/1997
7/1/1997
7/1/1997
7/1/1998
7/1/1998
7/1/1998
7/1/1998
7/1/1998
7/1/1998
7/1/1998
MW-90
MW-91
MW-92
EW-73
EW-83
EW-84
EW-85
EW-86
EW-87
MW-10
MW-1004
MW-1064
MW-1073
MW-11
MW-12
MW-14
MW-15
MW-240
MW-241
MW-242
MW-350
MW-351
MW-38D
MWM12
MW-72
MW-88
MW-89
MW-90
MW-91
EW-73
EW-83
EW-84
EW-85
EW-86
EW-87
MW-10
5.136E-03
7.550E-04
1 .360E-04
9.730E-02
1.770E-01
1.600E-01
5.880E-02
7.750E-03
2.200E-01
1.430E-01
1 .720E-04
6.520E-04
1 .300E-04
7.190E-02
3.280E-01
2.520E-02
5.175E-02
3.820E-03
4.580E-02
1.150E-02
1 .090E-03
2.570E-03
1.620E-01
2.563E-04
6.660E-02
9.310E-02
1 .940E-04
5.400E-03
3.480E-04
1.210E-01
1.480E-01
5.120E-02
4.080E-02
6.700E-03
2.850E-01
9.990E-02
2295.0
1996.1
1866.2
2900.5
2852.5
2602.5
2307.1
2261.4
2550.4
2526.1
3117.0
4778.2
3939.1
3115.4
2946.6
2295.3
2275.7
2112.3
2516.5
2286.1
1227.6
1000.0
2932.3
1100.9
2547.8
2475.6
2643.4
2295.0
1996.1
2900.5
2852.5
2602.5
2307.1
2261.4
2550.4
2526.1
-6.77E-03
-6.77E-03
-6.77E-03
-4.74E-03
-4.74E-03
-4.74E-03
-4.74E-03
-4.74E-03
-4.74E-03
-4.74E-03
-4.74E-03
-4.74E-03
-4.74E-03
-4.74E-03
-4.74E-03
-4.74E-03
-4.74E-03
-4.74E-03
-4.74E-03
-4.74E-03
-4.74E-03
-4.74E-03
-4.74E-03
-4.74E-03
-4.74E-03
-4.74E-03
-4.74E-03
-4.74E-03
-4.74E-03
-3.84E-03
-3.84E-03
-3.84E-03
-3.84E-03
-3.84E-03
-3.84E-03
-3.84E-03
9.147E-10
1.018E-09
4.418E-10
1 .040E-07
2.374E-07
7.020E-07
1 .046E-06
1.713E-07
1 .236E-06
9.011E-07
6.585E-1 1
9.492E-14
1.010E-12
2.773E-08
2.816E-07
4.743E-07
1 .069E-06
1.712E-07
3.021 E-07
2.260E-07
3.237E-06
2.245E-05
1 .488E-07
1 .388E-06
3.787E-07
7.453E-07
7.012E-10
1.018E-07
2.705E-08
1 J54E-06
2.580E-06
2.331 E-06
5.778E-06
1.1 31 E-06
1 .585E-05
6.100E-06
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
No
No
No
No
No
No
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
No
No
No
No
No
No
Yes
MAROS Version 2, 2002, AFCEE
Friday, December 13, 2002
Page 2 of4
-------
McClellan OUD ZoneA
: McClellan AFB
Meng
California
Sampling
Event
Effective
Date
Well
Concentration
(mg/L)
Distance Down
Centerline (ft)
Coefficient
(1/ft)
Concentration
(mg/L)
Detection
Limit?
Used in
Analysis?
TRICHLOROETHYLENE (TCE)
1998
1998
1998
1998
1998
1998
1998
1998
1998
1998
1998
1998
1998
1998
1998
1998
1998
1998
1999
1999
1999
1999
1999
1999
1999
1999
1999
1999
1999
1999
1999
1999
1999
1999
1999
1999
7/1/1998
7/1/1998
7/1/1998
7/1/1998
7/1/1998
7/1/1998
7/1/1998
7/1/1998
7/1/1998
7/1/1998
7/1/1998
7/1/1998
7/1/1998
7/1/1998
7/1/1998
7/1/1998
7/1/1998
7/1/1998
7/1/1999
7/1/1999
7/1/1999
7/1/1999
7/1/1999
7/1/1999
7/1/1999
7/1/1999
7/1/1999
7/1/1999
7/1/1999
7/1/1999
7/1/1999
7/1/1999
7/1/1999
7/1/1999
7/1/1999
7/1/1999
MW-11
MW-12
MW-14
MW-15
MW-240
MW-241
MW-242
MW-350
MW-351
MW-38D
MWM12
MW-52
MW-53
MW-55
MW-72
MW-88
MW-89
MW-92
EW-73
EW-83
EW-84
EW-85
EW-86
EW-87
MW-10
MW-1004
MW-1026
MW-1073
MW-11
MW-12
MW-14
MW-15
MW-237
MW-240
MW-241
MW-242
8.695E-02
2.405E-01
3.905E-03
6.195E-02
9.050E-05
7.670E-03
1 .290E-02
1 .OOOE-04
1.915E-03
1.340E-01
5.224E-04
5.400E-05
3.845E-03
3.580E-04
4.630E-02
9.050E-05
9.050E-05
7.853E-04
6.040E-02
9.975E-02
9.923E-03
4.850E-02
6.630E-03
2.653E-01
7.840E-02
1 .095E-04
3.910E-03
4.330E-04
3.475E-02
8.290E-02
4.990E-03
7.890E-02
4.840E-04
4.645E-05
1 .030E-02
5.680E-03
3115.4
2946.6
2295.3
2275.7
2112.3
2516.5
2286.1
1227.6
1000.0
2932.3
1100.9
3182.6
2999.3
2676.7
2547.8
2475.6
2643.4
1866.2
2900.5
2852.5
2602.5
2307.1
2261.4
2550.4
2526.1
3117.0
3834.7
3939.1
3115.4
2946.6
2295.3
2275.7
2312.7
2112.3
2516.5
2286.1
-3.84E-03
-3.84E-03
-3.84E-03
-3.84E-03
-3.84E-03
-3.84E-03
-3.84E-03
-3.84E-03
-3.84E-03
-3.84E-03
-3.84E-03
-3.84E-03
-3.84E-03
-3.84E-03
-3.84E-03
-3.84E-03
-3.84E-03
-3.84E-03
-4.75E-03
-4.75E-03
-4.75E-03
-4.75E-03
-4.75E-03
-4.75E-03
-4.75E-03
-4.75E-03
-4.75E-03
-4.75E-03
-4.75E-03
-4.75E-03
-4.75E-03
-4.75E-03
-4.75E-03
-4.75E-03
-4.75E-03
-4.75E-03
5.520E-07
2.920E-06
5.787E-07
9.899E-06
2.709E-08
4.860E-07
1 .980E-06
8.955E-07
4.111E-05
1.719E-06
7.610E-06
2.649E-10
3.814E-08
1 .226E-08
2.601 E-06
6.709E-09
3.522E-09
6.049E-07
6.351 E-08
1.317E-07
4.291 E-08
8.521 E-07
1 .447E-07
1 .469E-06
4.872E-07
4.120E-11
4.879E-1 1
3.292E-12
1.317E-08
7.002E-08
9.272E-08
1 .609E-06
8.280E-09
2.057E-09
6.700E-08
1.102E-07
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
No
No
No
No
No
No
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
MAROS Version 2, 2002, AFCEE
Friday, December 13, 2002
Page 3 of4
-------
McClellan OUD ZoneA
: McClellan AFB
Meng
California
Sampling
Event
Effective
Date
Well
Concentration
(mg/L)
Distance Down
Centerline (ft)
Coefficient
(1/ft)
Concentration
(mg/L)
Detection
Limit?
Used in
Analysis?
TRICHLOROETHYLENE (TCE)
1999
1999
1999
1999
1999
1999
1999
1999
1999
1999
1999
1999
1999
1999
1999
1999
1999
7/1/1999
7/1/1999
7/1/1999
7/1/1999
7/1/1999
7/1/1999
7/1/1999
7/1/1999
7/1/1999
7/1/1999
7/1/1999
7/1/1999
7/1/1999
7/1/1999
7/1/1999
7/1/1999
7/1/1999
MW-350
MW-351
MW-38D
MW-412
MW-458
MW-52
MW-53
MW-55
MW-70
MW-72
MW-74
MW-76
MW-88
MW-89
MW-90
MW-91
MW-92
1 .095E-04
1 .370E-02
7.697E-02
1 .095E-04
1 .202E-04
1 .095E-04
4.420E-04
1 .730E-03
1 .095E-04
2.770E-02
1.610E-03
5.430E-03
1 .095E-04
1 .095E-04
1 .095E-04
3.970E-04
1 .095E-04
1227.6
1000.0
2932.3
1100.9
1994.3
3182.6
2999.3
2676.7
3122.8
2547.8
2590.8
2885.3
2475.6
2643.4
2295.0
1996.1
1866.2
-4.75E-03
-4.75E-03
-4.75E-03
-4.75E-03
-4.75E-03
-4.75E-03
-4.75E-03
-4.75E-03
-4.75E-03
-4.75E-03
-4.75E-03
-4.75E-03
-4.75E-03
-4.75E-03
-4.75E-03
-4.75E-03
-4.75E-03
3.230E-07
1.190E-04
6.956E-08
5.892E-07
9.320E-09
3.018E-11
2.908E-10
5.260E-09
4.008E-1 1
1 .553E-07
7.358E-09
6.136E-09
8.646E-10
3.900E-10
2.037E-09
3.052E-08
1 .559E-08
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Note: Projected Concentrations that are below the user-specified detection limit are indicated by a check mark to its right; for sampling events
with less than 3 selected plume centerline wells, NO projected concentrations are calculated because no regression coefficient is available.
MAROS Version 2, 2002, AFCEE
Friday, December 13, 2002
Page 4 of4
-------
Regression of Plume Centerline Concentrations
McClellan OUD ZoneA
Location: McClellanAFB
Meng
California
Groundwater Flow Direction: 280 degrees Distance to Receptor: 1000 feet
From Period: 7/1/1995 to 7/1/2000
Selected Plume
Centerline Wells:
Sample Even
Well Distance to Receptor (feet)
MW-92
MW-91
MW-72
MW-11
1866.2
1996.1
2547.8
3115.4
The distance is measured in the Groundwater Flow Angle
from the well to the compliance boundary.
Number of Regression Confidence in
Effective Date Centerline Wells Coefficient (1 /ft) Coefficient
TRICHLOROETHYLENE (TCE)
1995 7/1/1995 4
1996 7/1/1996 2
1997 7/1/1997 3
1998 7/1/1998 3
1999 7/1/1999 4
2000 7/1/2000 2
-6.77E-03 99.2%
O.OOE+00 0.0%
-4.74E-03 83.5%
-3.84E-03 88.9%
-4.75E-03 96.2%
O.OOE+00 0.0%
Note: when the number of plume centerline wells is less than 3, no analysis is performed and all related values
are set to ZERO; Confidence in Coefficient is the statistical confidence that the estimated coefficient is
different from ZERO (for details, please refer to "Conference in Trend" in Linear Regression Analysis).
MAROS Version 2, 2002, AFCEE
Friday, December 13, 2002
Page 1 of 1
-------
MAROS Risk-Based Power Analysis for Site Cleanup
McClellan OUD ZoneA
McClellan AFB
Meng
California
Parameters:
Groundwater Flow Direction: 280 degrees Distance to Receptor: 100 feet
From Period: 1995 to 2000
7/1/1995 7/1/2000
Sample Event
Selected Plume
Centerline Wells:
N
Well
MW-92
MW-91
MW-72
MW-11
The distance
from the well
Distance
to Receptor (feet)
966.2
1096.1
1647.8
2215.4
is measured in the Groundwater Flow Angle
to the compliance boundary.
ormal Distribution Assumption
Sample Sample Sample Cleanup Expected
Szie Mean Stdev. Status Power Samp|e size
TRICHLOROETHYLENE (TCE) Cleanup Goal =
1995
1997
1998
1999
29 9.09E-05 4.01E-04 Attained 1
20 1.14E-04 3.54E-04 Attained 1
19 1.29E-04 2.99E-04 Attained 1
29 3.03E-04 1.58E-03 Attained 1
= 0.005
000 <=3
000 <=3
000 <=3
000 <=3
Lognormal Distribution Assumption
Celanup Expected Alpha Expected
Status Power Sample Size Level Power
Not Attained S/E S/E 0.05 0.8
Not Attained S/E S/E 0.05 0.8
Not Attained 0.457 49 0.05 0.8
Attained 0.740 35 0.05 0.8
Note: #N/C means "not conducted" due to a small sample size (N<4) or that the mean concentration is much greater than the cleanup
level; Sample Size is the number of sampling locations used in the power analysis; Expected Sample Size is the number of concentration
data needed to reach the Expected Power undercurrent sample variability.
MAROS Version 2, 2002, AFCEE
Monday, December 16, 2002
Page 1 of 1
-------
Risk-Based Power Analysis — Projected Concentrations
McClellan OUD ZoneA
Location : McClellan AFB
From Period:
Sampling
Event
7/1/1995 to
Effective
Date
7/1/2000
Well
Meng
California
Distance from the most downgradient well to receptor: 100 feet
Observed
Concentration
(mg/L)
Distance Down
Centerline (ft)
Regression
Coefficient
(1/ft)
Projected
Concentration
(mg/L)
Below
Detection
Limit?
Used in
Analysis?
TRICHLOROETHYLENE (TCE)
1995
1995
1995
1995
1995
1995
1995
1995
1995
1995
1995
1995
1995
1995
1995
1995
1995
1995
1995
1995
1995
1995
1995
1995
1995
1995
1995
1995
1995
1995
1995
1995
7/1/1995
7/1/1995
7/1/1995
7/1/1995
7/1/1995
7/1/1995
7/1/1995
7/1/1995
7/1/1995
7/1/1995
7/1/1995
7/1/1995
7/1/1995
7/1/1995
7/1/1995
7/1/1995
7/1/1995
7/1/1995
7/1/1995
7/1/1995
7/1/1995
7/1/1995
7/1/1995
7/1/1995
7/1/1995
7/1/1995
7/1/1995
7/1/1995
7/1/1995
7/1/1995
7/1/1995
7/1/1995
EW-73
EW-83
EW-84
EW-85
EW-86
EW-87
MW-10
MW-1026
MW-1041
MW-1042
MW-1064
MW-1073
MW-11
MW-12
MW-14
MW-15
MW-237
MW-240
MW-241
MW-242
MW-350
MW-351
MW-38D
MW-52
MW-53
MW-55
MW-70
MW-72
MW-74
MW-76
MW-88
MW-89
3.007E-01
1.057E-01
2.618E-01
2.160E-01
5.140E-03
8.076E-02
2.848E-01
6.150E-05
1 .230E-04
6.150E-05
2.063E-02
3.052E-03
7.106E-01
6.823E-01
1.021E+00
7.601 E-02
6.945E-05
7.484E-05
2.257E-02
1 .425E-02
3.657E-03
4.216E-03
1.937E-01
5.514E-05
4.805E-04
1 .507E-04
6.150E-05
6.046E-02
2.583E-03
2.093E-03
8.494E-05
4.948E-05
2000.5
1952.5
1702.5
1407.1
1361.4
1650.4
1626.1
2934.7
3917.5
3874.7
3878.2
3039.1
2215.4
2046.6
1395.3
1375.7
1412.7
1212.3
1616.5
1386.1
327.6
100.0
2032.3
2282.6
2099.3
1776.7
2222.8
1647.8
1690.8
1985.3
1575.6
1743.4
-6.77E-03
-6.77E-03
-6.77E-03
-6.77E-03
-6.77E-03
-6.77E-03
-6.77E-03
-6.77E-03
-6.77E-03
-6.77E-03
-6.77E-03
-6.77E-03
-6.77E-03
-6.77E-03
-6.77E-03
-6.77E-03
-6.77E-03
-6.77E-03
-6.77E-03
-6.77E-03
-6.77E-03
-6.77E-03
-6.77E-03
-6.77E-03
-6.77E-03
-6.77E-03
-6.77E-03
-6.77E-03
-6.77E-03
-6.77E-03
-6.77E-03
-6.77E-03
3.935E-07
1.914E-07
2.577E-06
1 .572E-05
5.098E-07
1.131E-06
4.703E-06
1.440E-13
3.707E-16
2.476E-16
8.115E-14
3.524E-12
2.169E-07
6.535E-07
8.043E-05
6.841 E-06
4.865E-09
2.036E-08
3.978E-07
1.195E-06
3.979E-04
2.142E-03
2.043E-07
1.068E-11
3.221 E-10
8.972E-10
1.786E-11
8.618E-07
2.752E-08
3.037E-09
1 .974E-09
3.693E-10
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
No
No
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
No
No
No
No
No
No
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
MAROS Version 2, 2002, AFCEE
Monday, December 16, 2002
Page 1 of4
-------
McClellan OUD ZoneA
: McClellan AFB
Meng
California
Sampling
Event
Effective
Date
Well
Concentration
(mg/L)
Distance Down
Centerline (ft)
Coefficient
(1/ft)
Concentration
(mg/L)
Detection
Limit?
Used in
Analysis?
TRICHLOROETHYLENE (TCE)
1995
1995
1995
1997
1997
1997
1997
1997
1997
1997
1997
1997
1997
1997
1997
1997
1997
1997
1997
1997
1997
1997
1997
1997
1997
1997
1997
1997
1997
1998
1998
1998
1998
1998
1998
1998
7/1/1995
7/1/1995
7/1/1995
7/1/1997
7/1/1997
7/1/1997
7/1/1997
7/1/1997
7/1/1997
7/1/1997
7/1/1997
7/1/1997
7/1/1997
7/1/1997
7/1/1997
7/1/1997
7/1/1997
7/1/1997
7/1/1997
7/1/1997
7/1/1997
7/1/1997
7/1/1997
7/1/1997
7/1/1997
7/1/1997
7/1/1997
7/1/1997
7/1/1997
7/1/1998
7/1/1998
7/1/1998
7/1/1998
7/1/1998
7/1/1998
7/1/1998
MW-90
MW-91
MW-92
EW-73
EW-83
EW-84
EW-85
EW-86
EW-87
MW-10
MW-1004
MW-1064
MW-1073
MW-11
MW-12
MW-14
MW-15
MW-240
MW-241
MW-242
MW-350
MW-351
MW-38D
MWM12
MW-72
MW-88
MW-89
MW-90
MW-91
EW-73
EW-83
EW-84
EW-85
EW-86
EW-87
MW-10
5.136E-03
7.550E-04
1 .360E-04
9.730E-02
1.770E-01
1.600E-01
5.880E-02
7.750E-03
2.200E-01
1.430E-01
1 .720E-04
6.520E-04
1 .300E-04
7.190E-02
3.280E-01
2.520E-02
5.175E-02
3.820E-03
4.580E-02
1.150E-02
1 .090E-03
2.570E-03
1.620E-01
2.563E-04
6.660E-02
9.310E-02
1 .940E-04
5.400E-03
3.480E-04
1.210E-01
1.480E-01
5.120E-02
4.080E-02
6.700E-03
2.850E-01
9.990E-02
1395.0
1096.1
966.2
2000.5
1952.5
1702.5
1407.1
1361.4
1650.4
1626.1
2217.0
3878.2
3039.1
2215.4
2046.6
1395.3
1375.7
1212.3
1616.5
1386.1
327.6
100.0
2032.3
200.9
1647.8
1575.6
1743.4
1395.0
1096.1
2000.5
1952.5
1702.5
1407.1
1361.4
1650.4
1626.1
-6.77E-03
-6.77E-03
-6.77E-03
-4.74E-03
-4.74E-03
-4.74E-03
-4.74E-03
-4.74E-03
-4.74E-03
-4.74E-03
-4.74E-03
-4.74E-03
-4.74E-03
-4.74E-03
-4.74E-03
-4.74E-03
-4.74E-03
-4.74E-03
-4.74E-03
-4.74E-03
-4.74E-03
-4.74E-03
-4.74E-03
-4.74E-03
-4.74E-03
-4.74E-03
-4.74E-03
-4.74E-03
-4.74E-03
-3.84E-03
-3.84E-03
-3.84E-03
-3.84E-03
-3.84E-03
-3.84E-03
-3.84E-03
4.056E-07
4.512E-07
1 .959E-07
7.409E-06
1 .692E-05
5.002E-05
7.457E-05
1.221E-05
8.804E-05
6.421 E-05
4.693E-09
6.764E-12
7.200E-1 1
1 .976E-06
2.007E-05
3.380E-05
7.617E-05
1 .220E-05
2.153E-05
1.6 11 E-05
2.307E-04
1 .600E-03
1.061 E-05
9.889E-05
2.698E-05
5.311 E-05
4.997E-08
7.252E-06
1 .927E-06
5.565E-05
8.185E-05
7.397E-05
1 .833E-04
3.589E-05
5.030E-04
1 .935E-04
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
No
No
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
No
Yes
Yes
Yes
Yes
No
No
No
No
No
No
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
No
No
No
No
No
No
Yes
MAROS Version 2, 2002, AFCEE
Monday, December 16, 2002
Page 2 of4
-------
McClellan OUD ZoneA
: McClellan AFB
Meng
California
Sampling
Event
Effective
Date
Well
Concentration
(mg/L)
Distance Down
Centerline (ft)
Coefficient
(1/ft)
Concentration
(mg/L)
Detection
Limit?
Used in
Analysis?
TRICHLOROETHYLENE (TCE)
1998
1998
1998
1998
1998
1998
1998
1998
1998
1998
1998
1998
1998
1998
1998
1998
1998
1998
1999
1999
1999
1999
1999
1999
1999
1999
1999
1999
1999
1999
1999
1999
1999
1999
1999
1999
7/1/1998
7/1/1998
7/1/1998
7/1/1998
7/1/1998
7/1/1998
7/1/1998
7/1/1998
7/1/1998
7/1/1998
7/1/1998
7/1/1998
7/1/1998
7/1/1998
7/1/1998
7/1/1998
7/1/1998
7/1/1998
7/1/1999
7/1/1999
7/1/1999
7/1/1999
7/1/1999
7/1/1999
7/1/1999
7/1/1999
7/1/1999
7/1/1999
7/1/1999
7/1/1999
7/1/1999
7/1/1999
7/1/1999
7/1/1999
7/1/1999
7/1/1999
MW-11
MW-12
MW-14
MW-15
MW-240
MW-241
MW-242
MW-350
MW-351
MW-38D
MWM12
MW-52
MW-53
MW-55
MW-72
MW-88
MW-89
MW-92
EW-73
EW-83
EW-84
EW-85
EW-86
EW-87
MW-10
MW-1004
MW-1026
MW-1073
MW-11
MW-12
MW-14
MW-15
MW-237
MW-240
MW-241
MW-242
8.695E-02
2.405E-01
3.905E-03
6.195E-02
9.050E-05
7.670E-03
1 .290E-02
1 .OOOE-04
1.915E-03
1.340E-01
5.224E-04
5.400E-05
3.845E-03
3.580E-04
4.630E-02
9.050E-05
9.050E-05
7.853E-04
6.040E-02
9.975E-02
9.923E-03
4.850E-02
6.630E-03
2.653E-01
7.840E-02
1 .095E-04
3.910E-03
4.330E-04
3.475E-02
8.290E-02
4.990E-03
7.890E-02
4.840E-04
4.645E-05
1 .030E-02
5.680E-03
2215.4
2046.6
1395.3
1375.7
1212.3
1616.5
1386.1
327.6
100.0
2032.3
200.9
2282.6
2099.3
1776.7
1647.8
1575.6
1743.4
966.2
2000.5
1952.5
1702.5
1407.1
1361.4
1650.4
1626.1
2217.0
2934.7
3039.1
2215.4
2046.6
1395.3
1375.7
1412.7
1212.3
1616.5
1386.1
-3.84E-03
-3.84E-03
-3.84E-03
-3.84E-03
-3.84E-03
-3.84E-03
-3.84E-03
-3.84E-03
-3.84E-03
-3.84E-03
-3.84E-03
-3.84E-03
-3.84E-03
-3.84E-03
-3.84E-03
-3.84E-03
-3.84E-03
-3.84E-03
-4.75E-03
-4.75E-03
-4.75E-03
-4.75E-03
-4.75E-03
-4.75E-03
-4.75E-03
-4.75E-03
-4.75E-03
-4.75E-03
-4.75E-03
-4.75E-03
-4.75E-03
-4.75E-03
-4.75E-03
-4.75E-03
-4.75E-03
-4.75E-03
1.751E-05
9.266E-05
1 .836E-05
3.141E-04
8.594E-07
1 .542E-05
6.283E-05
2.841 E-05
1 .304E-03
5.453E-05
2.414E-04
8.403E-09
1.210E-06
3.889E-07
8.253E-05
2.129E-07
1.117E-07
1.919E-05
4.548E-06
9.432E-06
3.073E-06
6.103E-05
1 .037E-05
1 .052E-04
3.489E-05
2.951 E-09
3.494E-09
2.358E-10
9.433E-07
5.015E-06
6.640E-06
1.152E-04
5.930E-07
1 .473E-07
4.798E-06
7.895E-06
Yes
Yes
Yes
No
Yes
Yes
Yes
Yes
No
Yes
No
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
No
No
No
No
No
No
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
MAROS Version 2, 2002, AFCEE
Monday, December 16, 2002
Page 3 of4
-------
McClellan OUD ZoneA
: McClellan AFB
Meng
California
Sampling
Event
Effective
Date
Well
Concentration
(mg/L)
Distance Down
Centerline (ft)
Coefficient
(1/ft)
Concentration
(mg/L)
Detection
Limit?
Used in
Analysis?
TRICHLOROETHYLENE (TCE)
1999
1999
1999
1999
1999
1999
1999
1999
1999
1999
1999
1999
1999
1999
1999
1999
1999
7/1/1999
7/1/1999
7/1/1999
7/1/1999
7/1/1999
7/1/1999
7/1/1999
7/1/1999
7/1/1999
7/1/1999
7/1/1999
7/1/1999
7/1/1999
7/1/1999
7/1/1999
7/1/1999
7/1/1999
MW-350
MW-351
MW-38D
MW-412
MW-458
MW-52
MW-53
MW-55
MW-70
MW-72
MW-74
MW-76
MW-88
MW-89
MW-90
MW-91
MW-92
1 .095E-04
1 .370E-02
7.697E-02
1 .095E-04
1 .202E-04
1 .095E-04
4.420E-04
1 .730E-03
1 .095E-04
2.770E-02
1.610E-03
5.430E-03
1 .095E-04
1 .095E-04
1 .095E-04
3.970E-04
1 .095E-04
327.6
100.0
2032.3
200.9
1094.3
2282.6
2099.3
1776.7
2222.8
1647.8
1690.8
1985.3
1575.6
1743.4
1395.0
1096.1
966.2
-4.75E-03
-4.75E-03
-4.75E-03
-4.75E-03
-4.75E-03
-4.75E-03
-4.75E-03
-4.75E-03
-4.75E-03
-4.75E-03
-4.75E-03
-4.75E-03
-4.75E-03
-4.75E-03
-4.75E-03
-4.75E-03
-4.75E-03
2.313E-05
8.523E-03
4.982E-06
4.220E-05
6.675E-07
2.161E-09
2.082E-08
3.767E-07
2.870E-09
1.112E-05
5.270E-07
4.395E-07
6.192E-08
2.793E-08
1 .459E-07
2.185E-06
1.117E-06
Yes
No
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Note: Projected Concentrations that are below the user-specified detection limit are indicated by a check mark to its right; for sampling events
with less than 3 selected plume centerline wells, NO projected concentrations are calculated because no regression coefficient is available.
MAROS Version 2, 2002, AFCEE
Monday, December 16, 2002
Page 4 of4
-------
Regression of Plume Centerline Concentrations
McClellan OUD ZoneA
Location: McClellanAFB
Meng
California
Groundwater Flow Direction: 280 degrees
From Period: 7/1/1995 to 7/1/2000
Distance to Receptor: 100 feet
Selected Plume
Centerline Wells:
Sample Even
Well Distance to Receptor (feet)
MW-92
MW-91
MW-72
MW-11
966.2
1096.1
1647.8
2215.4
The distance is measured in the Groundwater Flow Angle
from the well to the compliance boundary.
Number of Regression Confidence in
Effective Date Centerline Wells Coefficient (1 /ft) Coefficient
TRICHLOROETHYLENE (TCE)
1995 7/1/1995 4
1996 7/1/1996 2
1997 7/1/1997 3
1998 7/1/1998 3
1999 7/1/1999 4
2000 7/1/2000 2
-6.77E-03 99.2%
O.OOE+00 0.0%
-4.74E-03 83.5%
-3.84E-03 88.9%
-4.75E-03 96.2%
O.OOE+00 0.0%
Note: when the number of plume centerline wells is less than 3, no analysis is performed and all related values
are set to ZERO; Confidence in Coefficient is the statistical confidence that the estimated coefficient is
different from ZERO (for details, please refer to "Conference in Trend" in Linear Regression Analysis).
MAROS Version 2, 2002, AFCEE
Monday, December 16, 2002
Page 1 of 1
-------
MAROS Mann-Kendall Statistics Summary
McClellan Zone B OU D
Location: McClellan AFB
Julia Aziz
California
Time Period: 5/1/1990 to 12/31/2000
Consolidation Period: No Time Consolidation
Consolidation Type: Median
Duplicate Consolidation: Average
ND Values: Specified Detection Limit
J Flag Values : Actual Value
Source/ Number of
Well Tail Samples
TRICHLOROETHYLENE
EW-83
EW-84
MW-54
EW-85
EW-86
EW-87
EW-73
MW-1028
MW-1001
MW-1003
MW-1027
MW-59
MW-104
MW-1043
MW-105
MW-19D
MW-51
MW-57
MW-58
MW-1010
(TCE)
S
s
S
s
s
s
s
T
T
T
T
T
T
T
T
T
T
T
T
T
16
16
10
16
15
15
15
10
10
11
6
12
8
6
8
10
8
12
15
5
Number of
Detects
16
16
8
16
15
15
15
2
1
3
2
2
2
1
2
6
1
5
7
0
Coefficient
of Variation
0.30
0.83
1.46
0.79
1.23
0.68
0.92
0.61
0.03
0.30
0.71
0.31
1.60
0.09
1.94
1.67
0.18
1.44
1.75
0.00
Mann-Kendall
Statistic
22
-105
28
-84
-53
60
-83
-1
-3
-19
3
-1
-7
5
11
7
-7
-3
13
0
Confidence
in Trend
82.5%
100.0%
99.4%
100.0%
99.6%
99.9%
100.0%
50.0%
56.9%
91 .8%
64.0%
50.0%
76.4%
76.5%
88.7%
70.0%
76.4%
55.4%
72.1%
40.8%
All
Samples Concentration
"ND" ? Trend
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
Yes
NT
D
I
D
D
I
D
S
S
PD
NT
S
NT
NT
NT
NT
S
NT
NT
S
Note: Increasing (I); Probably Increasing (PI); Stable (S); Probably Decreasing (PD); Decreasing (D); No Trend (NT); Not Applicable (N/A)-
Due to insufficient Data (< 4 sampling events); Source/Tail (S/T)
The Number of Samples and Number of Detects shown above are post-consolidation values.
MAROS Version 2, 2002, AFCEE
Monday, January 13, 2003
Page 1 of 1
-------
MAROS Linear Regression Statistics Summary
Project: McClellan Zone B OU D
Location: McClellan AFB
Julia Aziz
California
Time Period: 5/1/1990 to 12/31/2000
Consolidation Period: No Time Consolidation
Consolidation Type: Median
Duplicate Consolidation: Average
ND Values: Specified Detection Limit
J Flag Values : Actual Value
Well
Average
Source/ Cone
Tail (mg/L)
Median
Cone
(mg/L)
All
Standard Samples
Deviation "ND" ?
Coefficient
Ln Slope of Variation
Confidence Concentration
in Trend Trend
TRICHLOROETHYLENE (TCE)
EW-83
EW-84
EW-85
EW-86
EW-87
MW-54
EW-73
MW-1027
MW-1001
MW-1010
MW-59
MW-1028
MW-104
MW-1043
MW-105
MW-19D
MW-51
MW-57
MW-58
MW-1003
s
s
s
s
s
s
s
T
T
T
T
T
T
T
T
T
T
T
T
T
1.0E-01
4.0E-01
1.9E-01
1 .2E-02
1.2E-01
1 .7E-03
2.6E-01
1 .5E-04
9.9E-05
1 .OE-04
1 .OE-04
1 .3E-04
2.1E-04
1 .OE-04
4.5E-04
5.1E-04
1.1E-04
2.6E-04
3.0E-04
1.1E-04
Note: Increasing (I); Probably Increasing (PI);
Due to insufficient Data (< 4 sampling events)
9.6E-02
2.5E-01
1 .2E-01
7.7E-03
8.4E-02
2.4E-04
1 .7E-01
1 .OE-04
1 .OE-04
1 .OE-04
1 .OE-04
1 .OE-04
1 .OE-04
1 .OE-04
1 .OE-04
1 .7E-04
1 .OE-04
1 .OE-04
1 .OE-04
1 .OE-04
3.1E-02
3.3E-01
1 .5E-01
1 .5E-02
8.2E-02
2.5E-03
2.4E-01
1 .OE-04
2.6E-06
O.OE+00
3.2E-05
7.8E-05
3.3E-04
9.4E-06
8.7E-04
8.5E-04
1 .9E-05
3.8E-04
5.3E-04
3.3E-05
No
No
No
No
No
No
No
No
No
Yes
No
No
No
No
No
No
No
No
No
No
Stable (S); Probably Decreasing (PD);
; COV = Coefficient of Variation
2.3E-05
-1.1E-03
-6.6E-04
-7.6E-04
4.6E-04
1 .2E-03
-7.2E-04
6.6E-05
-3.9E-06
O.OE+00
-2.9E-05
-1 .1 E-04
-3.5E-04
7.1E-05
6.3E-04
1 .8E-04
-8.1 E-05
3.4E-05
6.7E-05
-1.1 E-04
Decreasing (D);
0.30
0.83
0.79
1.23
0.68
1.46
0.92
0.71
0.03
0.00
0.31
0.61
1.60
0.09
1.94
1.67
0.18
1.44
1.75
0.30
No Trend
60.7%
100.0%
99.9%
99.6%
99.9%
99.9%
100.0%
60.4%
100.0%
100.0%
60.9%
75.7%
83.6%
85.4%
98.6%
67.4%
92.4%
55.7%
62.3%
96.6%
(NT); Not Applicable
NT
D
D
D
I
I
D
NT
D
S
S
S
NT
NT
I
NT
PD
NT
NT
D
(N/A) -
MAROS Version 2, 2002, AFCEE
Monday, January 13, 2003
Page 1 of 1
-------
MAROS Statistical Trend Analysis Summary
Project: McClellan Zone B OU D
Location: McClellan AFB
Julia Aziz
California
Time Period: 5/1/1990 to 12/31/2000
Consolidation Period: No Time Consolidation
Consolidation Type: Median
Duplicate Consolidation: Average
ND Values: Specified Detection Limit
J Flag Values : Actual Value
Well
TRICHLOROETHYLENE
EW-73
EW-83
EW-84
EW-85
EW-86
EW-87
MW-1001
MW-1003
MW-1010
MW-1027
MW-1028
MW-104
MW-1043
MW-105
MW-19D
MW-51
MW-54
MW-57
MW-58
MW-59
Source/
Tail
(TCE)
s
s
s
s
s
s
T
T
T
T
T
T
T
T
T
T
S
T
T
T
Number Number Average Median
of of Cone. Cone.
Samples Detects (mg/L) (mg/L)
15
16
16
16
15
15
10
11
5
6
10
8
6
8
10
8
10
12
15
12
15
16
16
16
15
15
1
3
0
2
2
2
1
2
6
1
8
5
7
2
2.6E-01
1 .OE-01
4.0E-01
1 .9E-01
1 .2E-02
1 .2E-01
9.9E-05
1.1E-04
1 .OE-04
1 .5E-04
1 .3E-04
2.1E-04
1 .OE-04
4.5E-04
5.1E-04
1.1E-04
1 .7E-03
2.6E-04
3.0E-04
1 .OE-04
1 .7E-01
9.6E-02
2.5E-01
1 .2E-01
7.7E-03
8.4E-02
1 .OE-04
1 .OE-04
1 .OE-04
1 .OE-04
1 .OE-04
1 .OE-04
1 .OE-04
1 .OE-04
1 .7E-04
1 .OE-04
2.4E-04
1 .OE-04
1 .OE-04
1 .OE-04
All
Samples
"ND" ?
No
No
No
No
No
No
No
No
Yes
No
No
No
No
No
No
No
No
No
No
No
Mann-
Kendall
Trend
D
NT
D
D
D
I
S
PD
S
NT
S
NT
NT
NT
NT
S
I
NT
NT
S
Linear
Regression
Trend
D
NT
D
D
D
I
D
D
S
NT
S
NT
NT
I
NT
PD
I
NT
NT
S
Note: Increasing (I); Probably Increasing (PI); Stable (S); Probably Decreasing (PD); Decreasing (D); No Trend (NT); Not Applicable
(N/A); Not Applicable (N/A) - Due to insufficient Data (< 4 sampling events); No Detectable Concentration (NDC)
The Number of Samples and Number of Detects shown above are post-consolidation values.
MAROS Version 2, 2002, AFCEE
Monday, January 13, 2003
Page 1 of 1
-------
MAROS Site Results
McClellan Zone B OU D
Location: McClellan AFB
Julia Aziz
California
User Defined Site and Data Assumptions:
Hydrogeology and Plume Information:
Groundwater
Seepage Velocity: 35 ft/yr
Current Plume Length: 1000 ft
Current Plume Width 600 ft
Number of Tail Wells: 13
Number of Source Wells: 7
Source Information:
Source Treatment: Pump and Treat
NAPL is not at this site.
Down-gradient Information:
Distance from Edge of Tail to Nearest:
Down-gradient receptor: 1000 ft
Down-gradient property: 10ft
Distance from Source to Nearest:
Down-gradient receptor: 6000ft
Down-gradient property: 10 ft
Consolidation Assumptions:
Time Period: 5/1/1990 to 12/31/2000
Consolidation Period: No Time Consolidation
Consolidation Type: Median
Duplicate Consolidation: Average
ND Values: Specified Detection Limit
J Flag Values : Actual Value
Plume Information Weighting Assumptions:
Consolidation Step 1. Weight Plume Information by Chemical
Summary Weighting: Weighting Applied to All Chemicals Equally
Consolidation Step 2. Weight Well Information by Chemical
Well Weighting: No Weighting of Wells was Applied.
Chemical Weighting: No Weighting of Chemicals was Applied.
Note: These assumptions made when consolidating the historical mentoring lumping the Wells and COCs.
1.
Preliminary Monitoring System Optimization Results: Based on site classification, source treatment and Monitoring System
Category the following suggestions are made for site Sampling Frequency, Duration of Sampling, and Well Density. These
criteria take into consideration: Plume Stability, Type of Plume, and Groundwater Velocity.
coc
Tail Source Level of
Stability Stability Effort
Sampling
Duration
Sampling
Frequency
Sampling
Density
TRICHLOROETHYLENE (TCE)
M
25
Remove treatment No Recommendation
system if previously
reducing concentation
Note:
Plume Status: (I) Increasing; (Pl)Probably Increasing; (S) Stable; (NT) No Trend; (PD) Probably Decreasing; (D) Decreasing
Design Categories: (E) Extensive; (M) Moderate; (L) Limited (N/A) Not Applicable, Insufficient Data Available
Level of Monitoring Effort Indicated by Analysi I Moderate
2,
MAROS Version 2, 2002, AFCEE
Tuesday, January 14, 2003
Page 1 of 2
-------
Moment Type Consituent
Zeroth Moment: Mass
TRICHLOROETHYLENE (TCE)
1st Moment: Distance to Source
TRICHLOROETHYLENE (TCE)
2nd Moment: Sigma XX
TRICHLOROETHYLENE (TCE)
2nd Moment: Sigma YY
TRICHLOROETHYLENE (TCE)
Coefficient
of Variation
1.32
0.76
1.68
1.06
Mann-Kendall
S Statistic
-13
-50
-50
-56
Confidence
in Trend
57.7%
93.0%
93.0%
95.1%
Moment
Trend
NT
PD
PD
D
Note: The following assumptions were applied for the calculation of the Zeroth Moment:
Porosity: 0.30
Saturated Thickness: Uniform: 60 ft
Mann-Kendall Trend test performed on all sample events for each constituent. Increasing (I); Probably Increasing (PI); Stable (S);
Probably Decreasing (PD); Decreasing (D); No Trend (NT); Not Applicable (N/A)-Due to insufficient Data (< 4 sampling events).
MAROS Version 2, 2002, AFCEE
Tuesday, January 14, 2003
Page 2 of 2
-------
MAROS Zeroth Moment Analysis
McClellan AFB Zone B OU D
Location: McClellan AFB
Julia Aziz
California
COC: TRICHLOROETHYLENE (TCE)
Change in Dissolved Mass Over Time
Date
H-.UC-U 1 •
3.5E-01 •
3.0E-01 •
2.5E-01 •
2.0E-01 •
1.5E-01 -
1 nF.ni
i .\ic \i i
5.0E-02 •
n np+nn -
*
*
* *
*
Porosity: 0.30
Saturated Thickness:
Uniform: 60 ft
Mann Kendall S Statistic:
-5
Confidence in
Trend:
I 76.5%
Coefficient of Variation:
I 0.70
Zeroth Moment
Trend:
Data Table:
Effective Date
1/1/1991
1/1/1993
1/1/1995
1/1/1997
1/1/1999
1/1/2001
Constituent
TRICHLOROETHYLENE (TCE)
TRICHLOROETHYLENE (TCE)
TRICHLOROETHYLENE (TCE)
TRICHLOROETHYLENE (TCE)
TRICHLOROETHYLENE (TCE)
TRICHLOROETHYLENE (TCE)
Estimated
Mass (Kg)
2.0E-01
1 .5E-01
1 .7E-01
1 .2E-01
3.6E-01
O.OE+00
Number of Wells
12
12
12
7
10
3
Note: Increasing (I); Probably Increasing (PI); Stable (S); Probably Decreasing (PD); Decreasing (D); No Trend (NT); Not Applicable (N/A) -
Due to insufficient Data (< 4 sampling events); ND = Non-detect. Moments are not calculated for sample events with less than 6 wells.
MAROS Version 2, 2002, AFCEE
1/2/2003
Page 1 of 1
-------
MAROS First Moment Analysis
McClellan AFB Zone B OU D
Location: McClellan AFB
COC: TRICHLOROETHYLENE (TCE)
Julia Aziz
California
Distance from Source to Center of Mass
Date
Mann Kendall S Statistic:
Confidence in
Trend:
o
o
3
CO
E
o
o
o
c
to
1.6E+03 -
1.4E+03 -
1.2E+03 -
1.0E+03 -
8.0E+02 -
6.0E+02 -
4.0E+02 -
2.0E+02 -
n np+nn .
*
» *
*
40.8%
Coefficient of Variation:
I 029
First Moment Trend:
Data Table:
Effective Date
1/1/1991
1/1/1993
1/1/1995
1/1/1997
1/1/1999
1/1/2001
Constituent
TRICHLOROETHYLENE (TCE)
TRICHLOROETHYLENE (TCE)
TRICHLOROETHYLENE (TCE)
TRICHLOROETHYLENE (TCE)
TRICHLOROETHYLENE (TCE)
TRICHLOROETHYLENE (TCE)
Xc (ft)
2,167,260
2,167,260
2,167,280
2,167,915
2,167,287
Yc (ft)
367,204
367,030
367,265
367,431
366,773
Distance from Source (ft)
1,138
986
1,201
1,709
799
Number of Wells
12
12
12
7
10
3
Note: Increasing (I); Probably Increasing (PI); Stable (S); Probably Decreasing (PD); Decreasing (D); No Trend (NT); Not Applicable (N/A) -
Due to insufficient Data (< 4 sampling events). Moments are not calculated for sample events with less than 6 wells.
MAROS Version 2, 2002, AFCEE
1/2/2003
Page 1 of 1
-------
MAROS First Moment Analysis
Project: McClellan AFB Zone B OU D
Location: McClellan AFB
Julia Aziz
California
COC: TRICHLOROETHYLENE (TCE)
Change in Location of Center of Mass Over Time
367400
367300
367200
367000
366900
366800
01/9!
Xc (ft)
Groundwater
Flow Direction:
Source
Coordinate:
X: | 2,166,737
Y: I 366,194
Effective Date
Constituent
Xc (ft)
Yc (ft) Distance from Source (ft) Number of Wells
1/1/1991
1/1/1993
1/1/1995
1/1/1997
1/1/1999
1/1/2001
TRICHLOROETHYLENE (TCE)
TRICHLOROETHYLENE (TCE)
TRICHLOROETHYLENE (TCE)
TRICHLOROETHYLENE (TCE)
TRICHLOROETHYLENE (TCE)
TRICHLOROETHYLENE (TCE)
2,167,260
2,167,260
2,167,280
2,167,915
2,167,287
367,204
367,030
367,265
367,431
366,773
1,138
986
1,201
1,709
799
12
12
12
7
10
3
Note: Increasing (I); Probably Increasing (PI); Stable (S); Probably Decreasing (PD); Decreasing (D); No Trend (NT); Not Applicable (N/A) -
Due to insufficient Data (< 4 sampling events). Moments are not calculated for sample events with less than 6 wells.
MAROS Version 2, 2002, AFCEE
1/2/2003
Page 1 of 1
-------
MAROS Second Moment Analysis
McClellan AFB Zone B OU D
McClellan AFB
COC: TRICHLOROETHYLENE (TCE)
Julia Aziz
California
Change in Plume Spread Over Time
Date
Mann Kendall S Statistic:
Confidence in
1000000 -
£- 100000 -
-£- 10000 -
CM
£ 1000 -
CO
10-
1 .
4(1(1(1(1(1(1(1(1
10000000 -
_ 1000000 -
=• 100000 -
5* 10000 -
$ 1000 -
100 -
10-
•1 .
Data Table:
Effective Date
1/1/1991
1/1/1993
1/1/1995
1/1/1997
1/1/1999
1/1/2001
* • * I 59'2%
Coefficient of Variation:
l| 1 -23
Second Moment
Trend:
| NT
Date
/& $> & $> &
•$" •$" •$" •$" •$" Mann Kendall S Statistic:
' * ' I4
Confidence in
Trend:
» » » * | 75.8%
Coefficient of Variation:
Second Moment
Trend:
I NT
1
Constituent Sigma XX (sq ft) Sigma YY (sq ft) Number of Wells
TRICHLOROETHYLENE (TCE) 237,221 766,710 12
TRICHLOROETHYLENE (TCE) 240,845 790,322 12
TRICHLOROETHYLENE (TCE) 231,216 837,265 12
TRICHLOROETHYLENE (TCE) 21,689,115 4,613,693 7
TRICHLOROETHYLENE (TCE) 255,472 251,790 10
TRICHLOROETHYLENE (TCE) 3
Note: Increasing (I); Probably Increasing (PI); Stable (S); Probably Decreasing (PD); Decreasing (D); No Trend (NT); Not Applicable (N/A) -
Due to insufficient Data (< 4 sampling events)
The Sigma XX and Sigma YY components are estimated using the given field coordinate system and then rotated to align with the
estimated groundwater flow direction. Moments are not calculated for sample events with less than 6 wells.
MAROS Version 2, 2002, AFCEE
1/2/2003
Page 1 of 1
-------
MAROS Spatial Moment Analysis Summary
McClellan AFB Zone B OU D
Location: McClellan AFB
Julia Aziz
California
Effective Date
TRICHLOROETHYLENE
1/1/1991
1/1/1993
1/1/1995
1/1/1997
1/1/1999
1/1/2001
Oth Moment
Estimated
Mass (Kg)
(TCE)
2.0E-01
1.5E-01
1.7E-01
1.2E-01
3.6E-01
O.OE+00
1st (Center of
Xc (ft)
2,167,260
2,167,260
2,167,280
2,167,915
2,167,287
Yc (ft)
367,204
367,030
367,265
367,431
366,773
Source
Distance (ft)
1,138
986
1,201
1,709
799
2nd Moment
Sigma XX
(sq ft)
237,221
240,845
231,216
21,689,115
255,472
Sigma YY
(sq ft)
766,710
790,322
837,265
4,613,693
251,790
Number of
Wells
12
12
12
7
10
3
MAROS Version 2, 2002, AFCEE
Thursday, January 02, 2003
Page 1 of 2
-------
McClellan AFB Zone B OU D
Location: McClellan AFB
Julia Aziz
California
Moment Type Consituent
Zeroth Moment: Mass
TRICHLOROETHYLENE (TCE)
1st Moment: Distance to Source
TRICHLOROETHYLENE (TCE)
2nd Moment: Sigma XX
TRICHLOROETHYLENE (TCE)
2nd Moment: Sigma YY
TRICHLOROETHYLENE (TCE)
Coefficient
of Variation
0.70
0.29
2.12
1.23
Mann-Kendall
S Statistic
-5
0
4
2
Confidence
in Trend
76.5%
40.8%
75.8%
59.2%
Moment
Trend
S
S
NT
NT
Note: The following assumptions were applied for the calculation of the Zeroth Moment:
Porosity: 0.30 Saturated Thickness: Uniform: 60 ft
Mann-Kendall Trend test performed on all sample events for each constituent. Increasing (I); Probably Increasing (PI); Stable (S);
Probably Decreasing (PD); Decreasing (D); No Trend (NT); Not Applicable (N/A)-Due to insufficient Data (< 4 sampling events).
Note: The Sigma XX and Sigma YY components are estimated using the given field coordinate system and then rotated to align with the
estimated groundwater flow direction. Moments are not calculated for sample events with less than 6 wells.
MAROS Version 2, 2002, AFCEE
Thursday, January 02, 2003
Page 2 of 2
-------
MAROS Sampling Location Optimization Results
Project: McClellan OUD ZoneB
McClellan AFB
Meng
California
Sampling Events Analyzed:
From 1995
7/1/1995
to 2000
7/1/2000
Parameters used:
Constituent
Inside SF Hull SF Area Ratio Cone. Ratio
TRICHLOROETHYLENE (TCE)
0.1
0.01
0.95
0.95
Well
Average Minimum Maximum
X (feet) Y (feet) Removable? Slope Factor* Slope Factor* Slope Factor* Eliminated?
TRICHLOROETHYLENE (TCE)
MW-1001
MW-1010
MW-1027
MW-104
MW-1043
MW-105
MW-19D
MW-51
MW-54
MW-57
MW-58
MW-59
2165839.75
2165998.00
2168527.25
2166810.00
2165134.50
2168172.25
2167162.75
2166767.25
2166658.50
2166663.50
2166656.25
2166608.50
366560.94 0
368273.56 0
367775.44 0
367362.16 0
368138.75 0
366814.81 0
365619.84 0
366264.69 0
366509.63 0
365863.28 0
366634.56 0
365816.88 0
0.574
0.003
0.262
0.712
0.324
0.571
0.156
0.642
0.556
0.257
0.329
0.377
0.262
0.003
0.000
0.534
0.324
0.279
0.084
0.390
0.292
0.098
0.048
0.025
0.949
0.003
0.481
0.925
0.324
0.921
0.242
0.944
0.873
0.447
0.665
0.799
D
D
D
D
D
D
D
D
D
D
D
D
Note: The Slope Factor indicates the relative importance of a well in the monitoring network at a given sampling event; the larger the SF
value of a well, the more important the well is and vice versa; the Average Slope Factor measures the overall well importance in the
selected time period; the state coordinates system (i.e., X and Y refer to Easting and Northing respectively) or local coordinates systems
may be used; wells that are NOT selected for analysis are not shown above.
* When the report is generated after running the Excel module, SF values will NOT be shown above.
MAROS Version 2, 2002, AFCEE
Friday, December 13, 2002
Page 1 of 1
-------
MAROS Sampling Frequency Optimization Results
McClellan OUD ZoneB
McClellan AFB
Meng
California
The Overall Number of Sampling Events: 34
"Recent Period" defined by events: From 1995 1st Quarter
To 2000 3rd Quarter
"Rate of Change" f
Well
1/15/1995
larameters used:
8/15/2000
Constituent Cleanup Goal Low Rate Medium Rate High Rate
TRICHLOROETHYLENE (TCE) 0.005
Units: Cleanup Goal is in mg/L; all rate parameters are
Recommended
Sampling Frequency
0.0025 0.005 0.01
in mg/L/year.
Frequency Based Frequency Based
on Recent Data on Overall Data
TRICHLOROETHYLENE (TCE)
EW-73
EW-83
EW-84
EW-85
EW-86
EW-87
MW-1001
MW-1003
MW-1010
MW-1027
MW-1028
MW-104
MW-1043
MW-105
MW-19D
MW-51
MW-54
MW-57
MW-58
MW-59
Annual
Annual
Annual
Annual
Annual
Quarterly
Biennial
Biennial
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Biennial
Annual
Biennial
Biennial
Biennial
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Quarterly Quarterly
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Note: Sampling frequency is determined considering both recent and overall concentration trends. Sampling Frequency is the
final recommendation; Frequency Based on Recent Data is the frequency determined using recent (short) period of monitoring
data; Frequency Based on Overall Data is the frequency determined using overall (long) period of monitoring data. If the "recent
period" is defined using a different series of sampling events, the results could be different.
MAROS Version 2, 2002, AFCEE
Monday, December 16, 2002
Page 1 of 1
-------
DRAFT FINAL
THREE-TIERED
GROUND WATER MONITORING NETWORK
OPTIMIZATION EVALUATION
FOR
OPERABLE UNIT D
MCCLELLAN AIR FORCE BASE, CALIFORNIA
Prepared for
US Environmental Protection Agency
May 2003
Denver, Colorado
-------
EXECUTIVE SUMMARY
This report presents a description and evaluation of the groundwater monitoring
program associated with Operable Unit D (OU D) at McClellan Air Force Base (AFB),
California. The monitoring program at this site was evaluated to identify potential
opportunities to streamline monitoring activities while still maintaining an effective
monitoring network. This evaluation is being conducted as part of an independent
assessment of monitoring network optimization (MNO) methods by the US
Environmental Protection Agency (USEPA) and the Air Force Center for Environmental
Excellence (AFCEE).
Objectives
Groundwater monitoring programs have two primary objectives (USEPA, 1994b;
Gibbons, 1994):
1. Evaluate long-term temporal trends in contaminant concentrations (temporal
objective)', and
2. Evaluate the extent to which contaminant migration is occurring (spatial
objective).
The relative success of any remediation system (including the monitoring network) is
judged based on the degree to which it achieves the stated objectives of the system.
Designing an effective groundwater monitoring program involves locating monitoring
points and developing a site-specific strategy for groundwater sampling and analysis that
maximizes the amount of relevant information that can be obtained while minimizing
incremental costs. The effectiveness of a monitoring network in achieving the two
primary monitoring objectives can be evaluated quantitatively using statistical
techniques. Qualitative evaluation also is important to allow consideration of such
factors as hydrostratigraphy, locations of potential receptor exposure points with respect
to a dissolved contaminant plume, and the direction(s) and rate(s) of contaminant
migration.
The general objective of the project was to optimize the OU D LTM network by
applying a three-tiered MNO approach to assess the degree to which the monitoring
network addresses each of the two primary objectives of monitoring listed above and
other important considerations. The three-tiered MNO evaluation described in this report
examines the 51 wells (6 extraction wells and 45 monitoring wells) included in OU D
monitoring network. The specific objectives of the project included:
ES-1
S:\ES\shared\cen\MNO\EPA\McClellan\Writeup\McClellanMNODraftFinal.doc
-------
« Apply a qualitative methodology that considers factors such as
hydrostratigraphy, locations of potential receptors with respect to the dissolved
plume, and the direction(s) and rate(s) of contaminant migration to establish the
frequency at which monitoring should be conducted, and if each well should be
retained in or removed from the monitoring program.
« Conduct a Mann-Kendall statistical analysis to determine the temporal trends of
COCs over time, and apply an algorithm to determine the relevance of the trends
within the monitoring network.
. Determine the relative amount of spatial information contributed by each
monitoring well by performing a spatial statistical analysis utilizing kriging error
predictions.
« Combine and evaluate the results of the three analyses to establish the frequency
at which monitoring should be conducted, as well as the optimal number and
locations of wells in the monitoring network.
Current Monitoring Program
Quarterly sampling of off-Base wells begin in 1984 as part of the Installation
Restoration Program at McClellan AFB. In 1986, the Groundwater Sampling and
Analysis Program was established to support ongoing remedial investigation/feasibility
study activities. In 1996, the Groundwater Monitoring Plan (GWMP) for on- and off-
Base wells was established under the Long-Term Monitoring Program to update the
GSAP and to support groundwater operable unit (GWOU) activities associated with the
Interim Record of Decision (IROD). Under the GWMP, groundwater samples are
collected quarterly from selected wells across the entire Base and analyzed for
contaminants of concern (COCs) associated with the respective plumes. Groundwater
sampling data are used to monitor how plume sizes and shapes are changing in response
to extraction well (EW) pumping to determine if remedial action objectives are being
met.
In the OU D area, groundwater sampling is conducted primarily to monitor areas
where dissolved VOC concentrations exceed MCLs (termed VOC target areas) in
monitoring zones A and B (monitoring of deeper zones in the OU D area is not
performed). In addition, selected wells that are more distant from the VOC target areas
are monitored to confirm and demonstrate that significant concentrations of COCs are not
present. The field sampling plan identifies the wells to be sampled in the area based on
the rationale and decision logic presented in the GWMP (Radian, 1997). Based on the
logic presented in the GWMP, the monitoring frequency and sampling rationale for each
well are continually evaluated, and can change as new sampling data are obtained.
Because the OU D plume is contained by the extraction system and the plume is well
defined, all of the wells associated with the OU D plume are sampled relatively
infrequently (annually or biennially). The Monitoring and Extraction Well Baseline
ES-2
S:\ES\shared\cen\MNO\EPA\McClellan\Writeup\McClellanMNODraftFinal.doc
-------
Sampling Frequency and Rationale (URS, 2002) identifies 6 extraction wells and 45
monitoring wells to be monitored in the vicinity of the OU D plume, based on
groundwater-quality data collected through the first quarter of 2002.
Optimization Findings
The OU D groundwater monitoring program was evaluated using results for sampling
events performed from April 1990 through August 2001. The analytical database
provided to Parsons contained from 5 to 18 sampling results for each of the 51 wells in
the OU D monitoring program. The primary COCs identified for the OU D plume in the
GWOU IROD are tricholoroethene (TCE), tetrachlroethene (PCE) , cis-1,2-
dicholorethene, (DCE), and 1,2-dichloroethane (DCA); therefore, the MNO evaluation
focused on these constituents. TCE is the COC with the highest concentrations and
largest spatial extent in groundwater at McClellan AFB OU D; therefore TCE sampling
results were the primary data used to conduct the three-tiered MNO.
Results from the three-tiered monitoring network optimization for OU D at McClellan
AFB indicate that 30 of the 51 OU D area wells could be removed from the groundwater
LTM program with little loss of information. A refined monitoring program consisting of
21 wells (13 to be sampled annually, and 8 to be sampled biennially) would be adequate
to address the two primary objectives of monitoring. This refined monitoring network
would result in an average of 17 sampling events per year, compared to 34 events per
year in the current monitoring program. Implementing these recommendations for
optimizing the LTM monitoring program at OU D at McClellan AFB could reduce
current LTM annual monitoring events by 50 percent. Based on analytical costs alone,
implementing these recommendations could save $2550 per year based on an estimate
of $150 per sample analysis. The recommendations provided in this report for removal
of off-Base monitoring wells from the LTM program were based largely on technical
considerations, and should be evaluated in the light of relevant community relations
concerns.
ES-3
S:\ES\shared\cen\MNO\EPA\McClellan\Writeup\McClellanMNODraftFinal.doc
-------
TABLE OF CONTENTS
Page
ACRONYMS AND ABBREVIATIONS iv
EXECUTIVE SUMMARY ES-1
SECTION 1 INTRODUCTION 1-1
SECTION 2 - SITE BACKGROUND INFORMATION 2-1
2.1 Site Description 2-1
2.2 Geology and Hydrogeology 2-2
2.2.1 Geology 2-2
2.2.2 Local Hydrogeology 2-3
2.3 Nature and Extent of Contamination 2-5
2.4 Remedial Systems 2-8
SECTION 3 - LONG-TERM MONITORING PROGRAM AT OU D 3-1
3.1 Description of Monitoring Program 3-1
3.2 Summary of Analytical Data 3-5
SECTION 4 - QUALITATIVE MNO EVALUATION 4-1
4.1 Methodology for Qualitative Evaluation of Monitoring Network 4-2
4.2 Results of Qualitative MNO Evaluation 4-3
4.2.1 Monitoring Network and Sampling Frequency 4-4
4.2.1.1 Extraction Wells 4-8
4.2.1.2 A-Zone Monitoring Wells 4-9
4.2.1.3 B-Zone Monitoring Wells 4-12
4.2.2 Laboratory Analytical Program 4-12
4.2.3 LTM Program Flexibility 4-13
SECTION 5 - TEMPORAL STATISTICAL EVALUATION 5-1
5.1 Methodology for Temporal Trend Analysis of Contaminant
Concentrations 5-1
-i-
S:\ES\shared\cen\MNO\EPA\McClellan\Writeup\McClellanMNODraftFinal.doc
-------
TABLE OF CONTENTS (Continued)
Page
5.2 Temporal Evaluation Results 5-7
SECTION 6 - SPATIAL STATISTICAL EVALUATION 6-1
6.1 Geostatistical methods for evaluating Monitoring networks 6-1
6.2 Spatial Evaluation of Monitoring Network at OU D 6-4
6.3 Spatial Statistical Evaluation Results 6-8
6.3.1 Kriging Ranking Results 6-8
SECTION 7 - SUMMARY OF THREE-TIERED MONITORING
NETWORK EVALUATION 7-1
SECTION 8 - REFERENCES 8-1
LIST OF TABLES
No. Title Page
3.1 Current Groundwater Monitoring Program 3-4
3.2 Summary of Occurrence of Groundwater Contaminants of Concern 3-7
4.1 Monitoring Network Optimization Decision Logic 4-3
4.2 Monitoring Frequency Decision Logic 4-4
4.3 Qualitative Evaluation of Groundwater Monitoring Network 4-5
5.1 Results of Temporal Trend Analysis of Groundwater Monitoring Results 5-10
6.1 Results of Geostatistical Evaluation Ranking of Zone A Wells by
Relative Value of TCE Information 6-10
7.1 Summary of Evaluation of Current Groundwater Monitoring Program 7-3
-11-
S:\ES\shared\cen\MNO\EPA\McClellan\Writeup\McClellanMNODraftFinal.doc
-------
TABLE OF CONTENTS (CONTINUED)
LIST OF FIGURES
No. Title Page
3.1 Current Monitoring Well Network and Sampling Procedures 3-3
4.1 Recommended Sampling Frequencies Based on Qualitative Evaluation 4-7
5.1 TCE Concentrations Through Time at Well MW-38D 5-2
5.2 Conceptual Representation of Temporal Trends and Temporal Variations
in Concentrations 5-3
5.3 Conceptual Representation of Continued Monitoring at Location Where
no Temporal Trend in Concentrations is Present 5-6
5.4 Mann-Kendall Temporal Trend Analysis for Concentrations of TCE 5-12
5.5 Temporal Trend Decision Rationale Flow Chart 5-13
6.1 Idealized Semvariogram Model 6-3
6.2 Impact of Missing Wells on Predicted Standard Error 6-7
6.3 Results of Geostatistical Analysis Showing Relative Value of
Information on TCE Distribution in Zone A Wells 6-9
6.4 Results of Geostatistical Analysis Showing Relative Value of Spatial
Information on TCE in Select Zone A Wells 6-12
-111-
S:\ES\shared\cen\MNO\EPA\McClellan\Writeup\McClellanMNODraftFinal.doc
-------
ACRONYMS AND ABBREVIATIONS
1,2-DCA
AFB
AFCEE
ASCE
bgs
bmsl
coc
DNAPL
ESRI
ft/day
ft/ft
GIS
GSAP
GWMP
GWOU
IROD
IRP
LNAPL
LTM
LTMP
ug/L
MCL
MNO
OUD
PCE
PQL
RAO
TCA
TCE
USEPA
voc
dichloroethane
Air Force Base
Air Force Center for Environmental Excellence
American Society of Chemical Engineers
below ground surface
below mean sea level
contaminant of concern
dense nonaqueous-phase liquid
Environmental Systems Research Institute, Inc.
foot (feet) per day
foot per foot
geographical information system
Groundwater Sampling and Analysis Program
Groundwater Monitoring Plan
Groundwater Operable Unit
Interim Record of Decision
Installation Restoration Program
light nonaqueous-phase liquid
long-term monitoring
Long-Term Monitoring Program
microgram(s) per liter
maximum contaminant level
monitoring network optimization
Operable Unit D
tetrachloroethene
practical quantitation limit
remedial action objective
trichloroethene
trichloroethene
United States Environmental Protection Agency
volatile organic compound
-IV-
S:\ES\shared\cen\MNO\EPA\McClellan\Writeup\McClellanMNODraftFinal.doc
-------
SECTION 1
INTRODUCTION
Groundwater monitoring programs have two primary objectives (U.S. Environmental
Protection Agency [USEPA], 1994b; Gibbons, 1994):
1. Evaluate long-term temporal trends in contaminant concentrations at one or
more points within or outside of the remediation zone, as a means of
monitoring the performance of the remedial measure (temporal objective); and
2. Evaluate the extent to which contaminant migration is occurring, particularly if
a potential exposure point for a susceptible receptor exists (spatial objective).
The relative success of any remediation system and its components (including the
monitoring network) must be judged based on the degree to which it achieves the stated
objectives of the system. Designing an effective groundwater monitoring program
involves locating monitoring points and developing a site-specific strategy for
groundwater sampling and analysis so as to maximize the amount of relevant information
that can be obtained while minimizing incremental costs. Relevant information is that
required to effectively address the temporal and spatial objectives of monitoring. The
effectiveness of a monitoring network in achieving these two primary objectives can be
evaluated quantitatively using statistical techniques. In addition, there may be other
important considerations associated with a particular monitoring network that are most
appropriately addressed through a qualitative assessment of the network. The qualitative
evaluation may consider such factors as hydrostratigraphy, locations of potential receptor
exposure points with respect to a dissolved contaminant plume, and the direction(s) and
rate(s) of contaminant migration.
This report presents a description and evaluation of the groundwater monitoring
program associated with Operable Unit D (OU D) McClellan Air Force Base (AFB),
1-1
S:\ES\shared\cen\MNO\EPA\McClellan\Writeup\McClellanMNODraftFinal.doc
-------
California. The current monitoring program at this site was evaluated to identify
potential opportunities to streamline monitoring activities while still maintaining an
effective monitoring network. This evaluation is being conducted as part of an
independent assessment of monitoring network optimization (MNO) methods by the
USEPA and the Air Force Center for Environmental Excellence (AFCEE). A three-
tiered approach, consisting of a qualitative evaluation, an evaluation of temporal trends in
contaminant concentrations, and a statistical spatial analysis, was conducted to assess the
degree to which the monitoring network addresses each of the two primary objectives of
monitoring, and other important considerations. The results of the three evaluations were
combined and used to assess the optimal frequency of monitoring and the spatial
distribution of the components of the monitoring network. The results of the analysis
were then used to develop recommendations for optimizing the monitoring program at
OUD.
1-2
S:\ES\shared\cen\MNO\EPA\McClellan\Writeup\McClellanMNODraftFinal.doc
-------
SECTION 2
SITE BACKGROUND INFORMATION
The location, operational history, geology, and hydrogeology of OU D at McClellan
AFB are briefly described in the following subsections.
2.1 SITE DESCRIPTION
McClellan AFB is located approximately 7 miles northeast of downtown Sacramento,
California. The installation comprises approximately 3,000 acres and is bounded by the
city of Sacramento on the west and southwest, the unincorporated areas of Antelope on
the north, Rio Linda on the northwest, and North Highlands on the east. OU D is located
in the northwestern part of McClellan AFB, northwest of Building 1069 along Patrol
Road, and occupies approximately 192 acres.
Through most of its operational history, McClellan AFB was engaged in a wide
variety of military/industrial operations involving the use, storage, and disposal of
hazardous materials, including industrial solvents, caustic cleaners, electroplating
chemicals, metals, polychlorinated biphenyls, low-level radioactive wastes, and a variety
of fuel oils and lubricants. Historical waste-disposal practices included the use of burial
pits for the disposal and/or burning of these materials.
Fifteen sites that were formerly operated as waste pits are located at OU D. These
waste pits were operational from the mid-1950s through the 1970s (CH2M Hill, 1994a).
In 1956, the first burial pit was created for disposal of sodium valves from aircraft
engines. Additional burial and burn pits were constructed throughout the 1960s and
1970s, and were used for the disposal of refuse, other solid waste, oil, various chemicals,
and industrial sludges. From the late 1970s into the early 1980s, many of the burial and
2-1
S:\ES\shared\cen\MNO\EPA\McClellan\Writeup\McClellanMNODraftFinal.doc
-------
burn pits were closed and covered with soil. In 1985, the "Area D" cap was constructed
over waste pits CS-1, CS-2, CS-3, CS-4, CS-5, CS-S, CS-A, and CS-T to reduce the
infiltration of precipitation through the waste pits, thereby also reducing the formation
and migration of leachate to groundwater at the site. Prior to 1985, waste pit CS-4 was
excavated to remove the sludge waste, and pits CS-6 and CS-26 also were excavated and
backfilled. The waste-disposal pits at OU D are no longer used for disposal of waste
products (CH2M Hill, 1992).
Based on evidence of environmental contamination, McClellan AFB was included on
the National Priorities List of Superfund sites in 1987. A single OU was established for
groundwater at the Base, and an Interim Record of Decision (IROD) was signed for the
Base-wide Groundwater OU (GWOU) in 1995 (CH2M Hill, 1995). The IROD specifies
groundwater extraction and treatment as the interim remedy for groundwater at OU D,
thereby endorsing the extraction well system that was installed in 1987 (see Section 2.5).
In 1995, McClellan AFB was recommended for closure under the Base Realignment
and Closure Act (BRAC). The recommendation became effective on September 28,
1995, and the installation was closed in July, 2001. Ongoing environmental restoration
activities are being directed by the Air Force Real Property Agency (formerly the Air
Force Base Conversion Agency).
2.2 GEOLOGY AND HYDROGEOLOGY
2.2.1 Geology
The Central Valley is an elongate basin, bounded on the east and west by nearly-
continuous mountain ranges. Sediments derived from the bordering mountain ranges
have accumulated in the basin for millions of years, and now comprise a sequence of
unconsolidated to partly consolidated deposits that, in places along the western side of the
Valley, may be as much as 30,000 feet thick (Norris and Webb, 1990). The sediments
that fill the Valley thin to the east, but the sequence is probably several thousand feet
thick beneath the city of Sacramento.
2-2
S:\ES\shared\cen\MNO\EPA\McClellan\Writeup\McClellanMNODraftFinal.doc
-------
The sediments in the upper few hundred feet of the subsurface beneath the Base
consist of coalescing deposits laid down by fluvial systems of various sizes and
competence that flowed generally from northeast to southwest or west. Soils are
primarily sand, silt, and clay, generally poorly sorted, with localized occurrences of
gravel, generally in the southern part of the Base. The nature of fluvial deposition,
including stream meandering and abandonment/reoccupation of channels, produced
morphologically irregular lenses and strata that are laterally and vertically discontinuous.
The coalescing and intercalating nature of the sediments makes distinction among units,
or stratigraphic correlation over distances greater than a few tens of feet, difficult (CH2M
Hill, 1994a).
2.2.2 Local Hydrogeology
Although the stratigraphy of the sediments beneath McClellan AFB is complex, the
juxtaposed and intercalated strata of sand, silt, clay, and gravel comprise a single
groundwater system. The geologic and hydraulic properties of the upper water-bearing
unit vary over short distances, and the more permeable intervals serve to establish
hydraulic interconnection vertically and horizontally, so that in general, groundwater
movement (and associated advective migration of contaminants) may occur throughout
the groundwater system. The upper unit beneath McClellan AFB has been divided into
the vadose (unsaturated) zone and five monitoring zones below the water table,
distinguished on the basis of general hydraulic characteristics. From shallowest to
deepest, the saturated zones are labeled A, B, C, D, and E (Radian, 1999a). The
monitoring zones at McClellan AFB were designated primarily for the purpose of
indicating the completion intervals of monitoring wells; all the monitoring zones are
hydraulically connected to a greater or lesser degree, and groundwater can move between
adjacent zones (CH2M Hill, 1994a). Generally, the zones dip to the west, and increase in
thickness from east to west. In is entirely possible for two adjacent wells screened at
different depth intervals to be completed within the same monitoring zone, or for two
wells screened at similar depths to be completed in different monitoring zones. These
local variations in the depths of monitoring zones are a consequence of the heterogeneity
2-3
S:\ES\shared\cen\MNO\EPA\McClellan\Writeup\McClellanMNODraftFinal.doc
-------
of the sediments beneath the Base, and the relative capacities of different deposits to
transmit water.
The depth to the water table beneath McClellan AFB ranges between about 90 and
110 feet below ground surface (bgs) (CH2M Hill, 1994b and 1999). At OU D, the depth
to shallow groundwater varies from approximately 99 to 102 feet bgs. As a consequence
of the relatively deep water table, surface streams are not in direct hydraulic
communication with the groundwater system beneath the Base (CH2M Hill, 1995).
Water-table elevations have declined at rates ranging from 1 to 2 feet per year during the
past 50 years. Groundwater levels are expected to continue to decline at a rate of about 2
feet per year as a consequence of large-scale groundwater production for industrial,
irrigation, and municipal uses in the Sacramento area (CH2M Hill, 1994b).
The thickness of the A monitoring zone ranges from 9 to 50 feet, and groundwater in
the A zone exists under unconfined conditions. The thickness of monitoring zone B
ranges from 40 to 75 feet, and groundwater in this zone appears to exist under partially
confined conditions. However, if the wells screened in the intermediate zone between the
A and B monitoring zones are reassigned to either the A or B zones as described in
Section 3.1, then the defined thickness of these two zones will increase. Monitoring
wells at OU D have been constructed only in the A and B monitoring zones (Radian,
1999a); therefore, no information is available regarding the deeper monitoring zones at
OUD.
Under natural conditions, prior to installation and operation of the OU D groundwater
extraction system, groundwater typically moved from northeast to south or southwest in
the A monitoring zone, and from north to south in the B monitoring zone (CH2M Hill,
1992). The local directions of groundwater movement beneath OU D currently are
influenced by the groundwater extraction system operating at the site. Groundwater flow
generally is directed radially inward toward the extraction wells (EWs).
Horizontal hydraulic gradients in the groundwater system at OU D range from 0.0008
foot per foot (ft/ft) to 0.0021 ft/ft, with the largest gradients occurring near active EWs.
2-4
S:\ES\shared\cen\MNO\EPA\McClellan\Writeup\McClellanMNODraftFinal.doc
-------
Vertical hydraulic gradients have been calculated using pairs of adjacent monitoring
wells that are completed in different monitoring zones. Vertical gradients calculated
using well pairs within that part of the groundwater system influenced by active
groundwater extraction at OU D are generally downward, similar to vertical gradients
that exist between the A and B monitoring zones in other parts of the Base (Radian,
1999a). Downward gradients range in magnitude from about 0.003 ft/ft to 0.04 ft/ft, with
the greatest vertical gradients occurring near active EWs. At distances greater than 1,000
feet from the extraction system, vertical gradients may be directed upward or downward,
depending on local potentiometric conditions. The vertical gradients between monitoring
zones A and B are usually upward in the winter and downward during the rest of the year.
Hydraulic conductivity is the property of a porous medium that describes the capacity
of the medium to transmit water. The horizontal hydraulic conductivity of the sediments
in monitoring zones A and B may range as high as 30 feet per day (ft/day); however, the
weighted average hydraulic conductivity of the sediments in the two zones is somewhat
lower (about 5.6 ft/day) (CH2M Hill, 1994a). The value of horizontal hydraulic
conductivity is about 5 to 15 times greater than the vertical hydraulic conductivity
(CH2M Hill, 1992), indicating that advective groundwater movement beneath OU D
occurs primarily in the horizontal plane. Based on the weighted average of horizontal
hydraulic conductivity for the A and B monitoring zones (5.6 ft/day), the range of
horizontal hydraulic gradients within OU D (about 0.001 to 0.002 ft/ft), and an
approximate effective porosity of 0.15 (a value consistent with the results of prior
investigations at McClellan AFB [CH2M Hill, 1992]), the calculated horizontal advective
velocity of groundwater movement in the A and B monitoring zones at OU D ranges
between about 14 and 30 feet per year (ft/year) (Parsons, 2000).
2.3 NATURE AND EXTENT OF CONTAMINATION
The contaminants of concern (COCs) in groundwater at OU D are exclusively
chlorinated aliphatic hydrocarbons (CAHs). Primary COCs include trichloroethene
(TCE), tetrachloroethene (PCE), czs-l,2-dichloroethene (DCE), and 1,2-dichloroethane
(DCA). These four compounds were identified in the GWOU IROD (CH2M Hill, 1995)
2-5
S:\ES\shared\cen\MNO\EPA\McClellan\Writeup\McClellanMNODraftFinal.doc
-------
as Base-wide groundwater COCs based on their frequency of detections, exceedance of
regulatory maximum contaminant levels (MCLs), and health risks. However, a recent
review of OU D-specific results suggests that additional volatile organic compounds
(VOCs) may be COCs in groundwater directly beneath OU D (Parsons, 2000). For
example, 1,1-DCA, 1,1-DCE, 1,1,1-trichloroethene (TCA), and vinyl chloride were
detected in groundwater beneath OU D at frequencies greater than 5 percent, and at
concentrations above their respective MCLs (Radian, 1999b).
A dissolved VOC plume in groundwater, consisting primarily of TCE at
concentrations greater than the federal MCL of 5 micrograms per liter (ug/L), has been
identified in the central and southwestern parts of OU D (CH2M Hill, 1999).
Historically, low concentrations of VOCs also have been detected occasionally in
groundwater samples collected from wells at distances up to approximately 500 feet off-
Base to the northwest. No contaminants have been detected in groundwater samples
from off-Base monitoring wells located west or northwest of OU D since 1995.
However, the dissolved CAH plume sourced at the OU D waste pits has migrated with
regional groundwater flow to the south and southwest, and historically extended off-
Base, to the west of OU D (CH2M Hill, 1994a). The off-Base extension of the plume
since has been hydraulically captured by the OU D groundwater extraction system
(Radian, 1999b).
Groundwater "hot spots" at McClellan AFB are defined as areas where contaminants
are present at concentrations greater than 100 times the applicable MCL. Such hot spots
have been identified near the former waste pits at Sites CS-2, CS-3, CS-5, CS-A, CS-S,
and CS-T, which are the sources of VOCs dissolved in groundwater. The CAHs detected
most frequently at these hot spots include PCE, TCE, czs-l,2-DCE, and 1,2-DCA. Prior
to the commencement of groundwater remediation, several VOCs were detected in
groundwater hot spots at OU D at concentrations that occasionally represented
appreciable fractions of the compounds' solubility in water. Although concentrations
greater than 1 percent of solubility can be an indication of the presence of dense
nonaqueous-phase liquid (DNAPL) (USEPA, 1994a), previous investigators had
2-6
S:\ES\shared\cen\MNO\EPA\McClellan\Writeup\McClellanMNODraftFinal.doc
-------
considered the likelihood of VOCs being present in the subsurface at OU D as a DNAPL
to be low (CH2M Hill, 1997).
Parsons (2000) reevaluated the potential presence of DNAPL at OU D as part of a
remedial process optimization (RPO) assessment. The elevated concentrations of VOCs
detected in soil-vapor and groundwater samples collected during earlier phases of
investigation and remediation activities at OU D were interpreted as evidence that CAHs
likely were present as DNAPL in the vadose zone, and possibly below the water table,
prior to implementation of interim remedial measures. Implementation of soil vapor
extraction (SVE) has removed considerable VOC mass from the vadose zone during the
past decade; however, because DNAPL is capable of migrating into "dead-end" soil-pore
spaces that are inaccessible to through-flowing fluids (e.g., air or water) (Pankow and
Cherry, 1996), it is likely that some CAH mass remains in the vadose zone as residual
DNAPL.
In addition, increases in the concentrations of TCE and 1,1-DCE in extracted
groundwater through time suggest that a free or residual DNAPL remains in the
subsurface near or below the water table in the vicinities of wells EW83 and EW87.
However, while the increased flow related to pumping may be enhancing the rate of
dissolution of CAH from a local DNAPL source, the increases in dissolved CAH
concentrations are not likely to be of sufficient magnitude to greatly affect the rate of
mass removal. Therefore, residual DNAPL near or below the water table at OU D may
persist as a continuing source of dissolved contaminants for an extended period of time.
Migration of the VOC plume has been halted as a result of the operation of the
groundwater extraction and treatment system since its installation in 1987, and the OU D
plume is currently regarded as being "fully contained" (Radian, 1999b). In general the
concentrations of dissolved CAH in groundwater have decreased appreciably during the
period of system operation. However, low concentrations of VOCs continue to be
detected sporadically at locations distal from the hot spots, west and southwest of OU D
(CH2M Hill, 1994a). According to the first quarter 2002 (1Q02) Groundwater
2-7
S:\ES\shared\cen\MNO\EPA\McClellan\Writeup\McClellanMNODraftFinal.doc
-------
Monitoring Program Quarterly Report (URS, 2002), 15 years of groundwater remediation
systems operation (1987 to 2002) have resulted in the following outcomes:
• The areal extent of the TCE plume has been reduced by 38.7 percent;
• Maximum TCE concentrations have been reduced below 100 ug/L; and
• The plume area defined by the 100-ug/L TCE isopleth has been essentially stable
during the last 11.5 years of extraction system operation.
2.4 REMEDIAL SYSTEMS
The remediation systems currently operating at OU D include an SVE and treatment
system, a groundwater extraction system, and associated monitoring networks. Pilot-
scale testing of SVE at OU D began in March 1993, and continuous system operation was
initiated in March 1994.
The primary objective of the groundwater extraction system at OU D is to prevent off-
Base migration of CAH-contaminated groundwater, thereby eliminating potential
exposure of off-Base receptors. A secondary objective of groundwater remediation, as
expressed in the Base-wide GWOU IROD (CH2M Hill, 1995), is aggressive extraction
and treatment of groundwater to remove contaminant mass.
Operation of the groundwater extraction system at OU D commenced in March 1987.
The current groundwater extraction system in OU D consists of six EWs (EW-73 and
EW-83 through EW-87), five of which are operational. Well EW-84 was removed from
service in August 1997, because its continued operation was judged to be unnecessary in
achieving the objective of plume containment (Radian, 1999a). Current plans call for
well EW-84 to remain off-line indefinitely. All EWs were installed to a depth of about
160 feet bgs, and are fully screened across both the A and B monitoring zones. The
design production rates for the individual wells in the current extraction system range
from 10 gallons per minute (gpm) to 25 gpm; the actual production rates (0 to about 11
gpm) are generally somewhat lower than the design rates. Groundwater extracted at OU
2-8
S:\ES\shared\cen\MNO\EPA\McClellan\Writeup\McClellanMNODraftFinal.doc
-------
D is conveyed via pipeline to a 1,500- gpm capacity groundwater treatment plant
(GWTP), which is centrally located at McClellan AFB and treats water extracted from
multiple OUs.
2-9
S:\ES\shared\cen\MNO\EPA\McClellan\Writeup\McClellanMNODraftFinal.doc
-------
SECTION 3
LONG-TERM MONITORING PROGRAM AT OU D
The current groundwater monitoring program at OU D was examined to identify
potential opportunities for streamlining monitoring activities while still maintaining an
effective performance and compliance monitoring program. The current monitoring
program at OU D is reviewed in the following subsections.
3.1 DESCRIPTION OF MONITORING PROGRAM
Quarterly sampling of off-Base wells begin in 1984 as part of the Installation
Restoration Program (IRP) at McClellan AFB. In 1986, the Groundwater Sampling and
Analysis Program (GSAP) was established to support ongoing remedial
investigation/feasibility study activities. In 1996, the Groundwater Monitoring Plan
(GWMP) for on- and off-Base wells was established under the Long-Term Monitoring
Program (LTMP) to update the GSAP and to support GWOU IROD activities. Under the
GWMP, groundwater samples are collected quarterly from selected wells across the
entire Base and analyzed for COCs associated with the respective plumes. However,
because groundwater sampling focus areas change from quarter to quarter, each areas of
the Base are targeted for a detailed analysis of groundwater quality once each year.
Groundwater sampling and analysis for wells in the OU D area occurs during the first
quarter of each year (URS, 2001).
Groundwater sampling data are used to monitor how plume sizes and shapes are
changing in response to EW pumping to determine if the following remedial action
objectives (RAOs), as specified in the GWOU IROD (CH2M Hill, 1995) for McClellan
AFB, are being met:
3-1
S:\ES\shared\cen\MNO\EPA\McClellan\Writeup\McClellanMNODraftFinal.doc
-------
• Prevent contaminated groundwater from migrating off-Base;
• Contain hot spots; and
• Prevent downward migration of contamination.
As described in Section 2.4, groundwater sampling data also are used to evaluate
contaminant mass removal rates.
In the OU D area, groundwater sampling is conducted to monitor areas where dissolved
VOC concentrations exceed MCLs (termed VOC target areas) in monitoring zones A and
B (there are no target areas in deeper monitoring zones in the OU D area). The field
sampling plan (FSP) identifies the wells to be sampled in the area based on the rationale
and decision logic presented in the GWMP (Radian, 1997). Based on the logic presented
in the GWMP, the monitoring frequency and sampling rationale for each well are
continually evaluated, and can change as new sampling data are obtained. Because the
OU D plume is contained by the extraction system and the plume is well defined, all of
the wells associated with the OU D plume are sampled relatively infrequently (annually
or biennially). The Monitoring and Extraction Well Baseline Sampling Frequency and
Rationale (URS, 2002) identifies 6 EWs and 45 monitoring wells to be monitored in the
vicinity of the OU D plume, based on groundwater-quality data collected through 1Q02.
Figure 3.1 shows the 51 wells currently monitored at OU D, and their monitoring
frequency. Table 3.1 lists these wells along with their assigned monitoring zone,
designated sampling frequency, and sampling rationale based on the information
provided in the FSP for 1Q02, 2Q02, and 3Q02 (URS, 2002). Wells MW-1010 and MW-
1042 are considered "cross-zone wells" (designated as "AB" in the "Original Zone"
column in Table 3.1); these wells are screened across monitoring zones A and B.
However, sampling results for these wells are assigned to a single monitoring zone based
on water-level and lithologic data, contaminant concentrations, and well construction
information for each cross-zone monitoring well compared to nearby single-zone
monitoring wells. Analytical results for samples collected from well MW-1010 are
assigned to zone B, and sampling results for MW-1042 are assigned to zone A. EWs at
3-2
S:\ES\shared\cen\MNO\EPA\McClellan\Writeup\McClellanMNODraftFinal.doc
-------
O
CN
oo
CD
CM
s^
m
£
a.
<
CD
I
m
O TT;
00
O)
c
N
GO
a>
0
o
CO
LU
O
C
a)
2
LJ_
O)
"5.
re
3
O
1 1
c 9^
< in
o
'•s
b
o
CD
•a
c
^
o
O
o
lO
CO
O
03
N
£w<"
,_ O ' UJ
5tia
Sg2|
§tg§
Q)
Qi •£ S,
c ^J
Q i
^ i
O I
CM
5
LO
CO
CD
r^
CO
CM
o
o
LO
o
LO
CN
0
0
M—
O
o
LO
JO
03
3
CT
0
^
O
-------
TABLE 3.1
CURRENT GROUNDWATER MONITORING PROGRAM
THREE-TIERED MONITORING NETWORK OPTIMIZATION
OPERABLE UNIT D
McCLELLAN AFB, CALIFORNIA
Well ID
Screened Interval
(ft msl) "
Extraction Wells
EW-73
EW-83
EW-84
EW-85
EW-86
EW-87
20.7 to -99.3
24.22 to -95.78
22.8 to -97.2
22.9 to -97.1
17.9 to -102.1
18.1 to -101.9
Zone A Wells
MW-10
MW-11
MW-12
MW-14
MW-15
MW-38D
MW-52
MW-53
MW-55
MW-70
MW-72
MW-74
MW-76
MW-88
MW-89
MW-90
MW-91
MW-92
MW-237
MW-240
MW-241
MW-242
MW-350
MW-351
MW-412
MW-458
MW-1004
MW-1026
MW-1041
MW-1042
MW-1064
MW-1073
-37.82 to -47.82
-36.31 to -46.31
-34.67 to -44.67
-35.52 to -45.52
-4 1.48 to -46.48
-57.32 to -67.32
-80.16 to -90.16
-65.47 to -75.47
-67.46 to -75.47
- 66.77 to -76.77
-58.45 to -68.45
-71.39 to -81.39
-80.23 to -90.23
-38.46 to -48.46
-38.89 to -48.89
-34.94 to -44.94
-35.85 to -45.85
-38.39 to 48.39
-41.55 to -61.55
-40.56 to -60.56
-49.24 to -69.24
-52.25 to -67.25
-32.92 to -52.92
-33.97 to -49.97
-31.76 to -51.76
-40.81 to -60.81
-30.88 to -40.88
-3 1.93 to 4 1.93
-52.87 to -62.87
-80.18 to -90.18
-30.96 to -50.96
-52.6 to -62.6
Zone B Wells
MW-19D
MW-51
MW-54
MW-57
MW-58
MW-59
MW-104
MW-1001
MW-1003
MW-1010
MW-1027
MW-1028
MW-1043
-80.16 to -90.16
-112.96 to -127.96
-81.61 to -91.61
-80.38 to -90.38
-112.63 to -122.63
-106.32 to -116.32
-113.78 to -123.78
-105.25 to -115.25
-77.72 to -87.72
-86.37 to -96.37
-70.47 to -80.47
-117 to -127
-137.09 to -147.09
Original
Zone
IAB
IAB
IAB
IAB
IAB
IAB
A
A
A
A
A
IAB
IAB
IAB
IAB
IAB
A
IAB
IAB
A
A
A
A
A
A
A
A
A
A
A
A
A
A
A
A
AB
A
A
B
B
IAB
IAB
B
B
B
B
IAB
AB
B
B
B
Zone
A/B
A/B
A/B
A/B
A/B
A/B
A
A
A
A
A
A*c/
A*
A*
A*
A*
A
A*
A*
A
A
A
A
A
A
A
A
A
A
A
A
A
A
A
A
A*
A
A
B
B
B*
B*
B
B
B
B
B*
B*
B
B
B
Sampling
Frequency
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Annual
Biennial
Annual
Annual
Biennial
Biennial
Biennial
Biennial
Annual
Biennial
Annual
Biennial
Biennial
Biennial
Biennial
Biennial
Biennial
Biennial
Annual
Annual
Biennial
Annual
Biennial
Biennial
Biennial
Biennial
Biennial
Biennial
Biennial
Biennial
Biennial
Biennial
Annual
Biennial
Biennial
Biennial
Biennial
Biennial
Biennial
Biennial
Biennial
Biennial
Biennial
ft bmsl = feet relative to mean sea level.
Sampling frequency based on 1st Quarter 2002 Ground-water
Monitoring Program Quarterly Report (URS, 2002).
* = Reassigned monitoring zone, per URS (2002).
022/742479/McCllelmTablesDraftFinal.xls/Table 3.1
3-4
-------
OU D extract water from zones A and B, and analytical data for samples from these wells
are used to characterize plumes in both zones. Ten of the monitoring wells are screened
in intermediate zones between A and B monitoring zones (designated as "IAB" in the
"Original Zone" column in Table 3.1), and concentrations from these wells have
historically not been used for plume interpretation. However, these wells are sampled
because the data may provide useful information about the effectiveness of the extraction
system. For this MNO analysis, the IAB wells were assigned to either the A or B zones
based on the recommendations listed in Table 3-3 of the 1Q02 Quarterly Monitoring
Report (URS, 2002).
The six EWs are sampled annually. Of the 32 OU D wells that monitor zone A, 22 are
sampled biennially and 10 are sampled annually. Twelve of the 13 zone B wells are
sampled biennially, and the remaining well is sampled annually. All samples from the
monitoring and extraction wells are analyzed for VOCs by USEPA Method SW8260B.
It should be noted that, in 2001, OU D wells were monitored using passive diffusion
bag samplers (PDBSs); conventional-purge sampling was performed prior to and
following the 2001 sampling event. Minor differences in analytical results may be
attributable to use of the two types of sampling techniques. The three-tiered MNO
evaluation described in this report examines the current monitoring network consisting of
51 wells (17 sampled annually and 34 biennially).
3.2 SUMMARY OF ANALYTICAL DATA
The OU D groundwater monitoring program was evaluated using results for sampling
events performed from April 1990 through August 2001. These analytical data, along
with water level and well location information was provided to Parsons by Ms. Diane
Kiota, the Air Force groundwater program manager at McClellan. The database was
processed to remove duplicate data measurements by retaining the maximum result of
the duplicate samples. The analytical database provided to Parsons contained from 5 to
18 sampling results for each of the 51 wells in the OU D monitoring program. As
discussed in Section 2.3, the primary COCs identified for the OU D plume in the GWOU
3-5
S:\ES\shared\cen\MNO\EPA\McClellan\Writeup\McClellanMNODraftFinal.doc
-------
IROD are TCE, PCE, cis-l,2-DCE, and 1,2-DCA; therefore, the MNO evaluation
focused on these constituents. Other VOCs of concern at the Base were not measured
above their MCLs in the OU D area during recent sampling events (URS, 2002), and
therefore were excluded from the MNO analysis.
Table 3.2 presents a summary of the occurrence of the four primary COCs in OU D
groundwater based on the data collected from April 1990 through August 2001. As
indicated in this table, TCE is the COC detected most frequently and at the greatest
concentrations in groundwater at McClellan AFB OU D. TCE has been detected in
approximately 63 percent of samples, and has exceeded its MCL of 5 |ug/L in
approximately 38 percent of the samples. TCE has been detected in 46 of the 51 wells in
the monitoring program, and has exceeded the MCL in 26 of these wells. One of TCE's
reductive dechlorination daughter products, cis-l,2-DCE, is the second-most prevalent
compound, and has been detected in 36 percent of the collected samples. However,
detected concentrations ofcis-l,2-DCE have exceeded the MCL of 70 ug/L in less than 1
percent of samples. PCE and 1,2-DCA have been detected at OU D in approximately 24
percent and 32 percent of samples, respectively; these compounds have been detected
within the footprint of the TCE plume. Detected concentrations of these COCs have
exceeded their MCLs in approximately 5 percent and 14 percent of the samples,
respectively.
TCE sampling results were the primary data used to conduct the three-tiered
monitoring network optimization due to the magnitude and spatial extent of TCE
concentrations in groundwater at OU D compared to the other detected compounds.
3-6
S:\ES\shared\cen\MNO\EPA\McClellan\Writeup\McClellanMNODraftFinal.doc
-------
<<-i _e
S •!: o
^- S X
<• »*• w
53 •§ a
OJ ? ••"
"a - 1
a "S *j
s •!? «
Z £ Q
— —
o 5
^ '? —
_C5 * S
« JL ^
g S o>
z ^
Percentage of
Samples with
MCL
Exceedances
^
^J /"™V
I 3
e j« S
0> U j
BX O» t^
a "S Bio
rt ®
CS "S
•^ U
^" "S
?J Q
s«
CS i—
-S *
cc
IH
-^^
U
S
CS
CS
2
oo
, — 1
S
o
CS
CN
O
o
o^
CN
CN
OO
^
O
Q
CS
*-*
CS
CS
, — 1
o
o
f—
oo
1
r-
oo
o
o
sO
CN
VO
oo
CN
W
O
Q
CN
cfl
o
•n
m
, — i
o^
O
r-
ON
CN
O
O
sO
vq
f-r^
CN
ON
W
O
CS
^
, — 1
oo'
o
o
o
ON
O
o
sO
o^-
ON
CN
ON
W
O
H
o
o
CS
60
60
O
o
ON
ON
ft
o
OJ
60
"ft
OJ
•a
OJ
OJ
'^ .tn G
"O ^ 2
§
o
o
S) s
2 '*
a
Q
o
-------
SECTION 4
QUALITATIVE MNO EVALUATION
An effective groundwater monitoring program will provide information regarding
contaminant plume migration and changes in chemical concentrations through time at
appropriate locations, enabling decision-makers to verify that contaminants are not
endangering potential receptors, and that remediation is occurring at rates sufficient to
achieve RAOs within a reasonable time frame. The design of the monitoring program
should therefore include consideration of existing receptor exposure pathways, as well as
exposure pathways arising from potential future use of the groundwater.
Performance monitoring wells located upgradient, within, and immediately
downgradient from a plume provide a means of evaluating the effectiveness of a
groundwater remedy relative to performance criteria. Long-term monitoring (LTM) of
these wells also provides information about migration of the plume and temporal trends
in chemical concentrations. Groundwater monitoring wells located downgradient from
the leading edge of a plume (i.e., sentry wells) are used to evaluate possible changes in
the extent of the plume and, if warranted, to trigger a contingency response action if
contaminants are detected.
Primary factors to consider when developing a groundwater monitoring program
include at a minimum:
• Aquifer heterogeneity,
• Types of contaminants,
• Distance to potential receptor exposure points,
• Groundwater seepage velocity,
4-1
S:\ES\shared\cen\MNO\EPA\McClellan\Writeup\McClellanMNODraftFinal.doc
-------
• Potential surface-water impacts, and
• The effects of the remediation system.
These factors will influence the locations and spacing of monitoring points and the
sampling frequency. Typically, the greater the seepage velocity and the shorter the
distance to receptor exposure points, the more frequently groundwater sampling should
be conducted.
One of the most important purposes of LTM is to confirm that the contaminant plume
is behaving as predicted. Graphical and statistical tests can be used to evaluate plume
stability. If a groundwater remediation system or strategy is effective, then over the long
term, groundwater-monitoring data should demonstrate a clear and meaningful
decreasing trend in concentrations at appropriate monitoring points. The current
groundwater monitoring program at McClellan AFB OU D was evaluated to identify
potential opportunities to LTM optimization.
4.1 METHODOLOGY FOR QUALITATIVE EVALUATION OF
MONITORING NETWORK
The three-tiered MNO evaluation of the OU D groundwater LTM program considered
information for 51 wells in the OU D study area that currently are included in the
monitoring program. These wells, their respective monitoring zones, and their current
monitoring frequency are listed in Table 3.1, and their locations are depicted on Figure
3.1.
Multiple factors were considered in developing recommendations for continuation or
cessation of groundwater monitoring at each well. In some cases, a recommendation was
made to continue monitoring a particular well, but at a reduced frequency. A
recommendation to discontinue monitoring at a particular well based on the information
reviewed does not necessarily constitute a recommendation to physically abandon the
well. A change in site conditions might warrant resumption of monitoring at some time
in the future at wells that are not currently recommended for continued sampling.
Typical factors considered in developing recommendations to retain a well in, or remove
4-2
S:\ES\shared\cen\MNO\EPA\McClellan\Writeup\McClellanMNODraftFinal.doc
-------
a well from, the monitoring program are summarized in Table 4.1. Typical factors
considered in developing recommendations for monitoring frequency are summarized in
Table 4.2.
TABLE 4.1
MONITORING NETWORK OPTIMIZATION DECISION LOGIC
THREE-TIERED MONITORING NETWORK OPTIMIZATION
OPERABLE UNIT D
MCCLELLAN AFB, CALIFORNIA
Reasons for Retaining a Well in
Monitoring Network
Well is needed to further characterize the
site or monitor changes in contaminant
concentrations through time
Well is important for defining the lateral or
vertical extent of contaminants
Well is needed to monitor water quality at
compliance point or receptor exposure
point (e.g., domestic well)
Well is important for defining background
water quality
Reasons for Removing a Well From
Monitoring Network
Well provides spatially redundant
information with a neighboring well (e.g.,
same constituents, and/or short distance
between wells)
Well has been dry for more than 2 yearsa/
Contaminant concentrations are
consistently below laboratory detection
limits or cleanup goals
Well is completed in same water-bearing
zone as nearby well(s)
a/ Water-level measurements in dry wells should continue, and groundwater sampling should be resumed if the well
becomes re-wetted.
4.2 RESULTS OF QUALITATIVE MNO EVALUATION
The results of the qualitative evaluation of the 51 monitoring and EWs currently
included in the LTM program at OU D included are summarized in Table 4.3, shown on
Figure 4.1, and described in the following subsections. The table includes
recommendations for retaining or deleting each existing monitoring well, and for
changing the sampling frequency, and lists the rationale for the recommendations. Figure
4.1 shows the current and recommended revised sampling frequency for each well based
on the qualitative evaluation.
4-3
S:\ES\shared\cen\MNO\EPA\McClellan\Writeup\McClellanMNODraftFinal.doc
-------
TABLE 4.2
MONITORING FREQUENCY DECISION LOGIC
THREE-TIERED MONITORING NETWORK OPTIMIZATION
OPERABLE UNIT D
MCCLELLAN AFB, CALIFORNIA
Reasons for Increasing
Sampling Frequency
Groundwater velocity is high
Change in contaminant concentration
would significantly alter a decision or
course of action
Well is close to source area or operating
remedial system
Cannot predict if concentrations will
change significantly over time
Reasons for Decreasing
Sampling Frequency
Groundwater velocity is low
Change in contaminant concentration
would not significantly alter a decision or
course of action
Well is distal from source area or remedial
system
Concentrations are not expected to change
significantly over time, or contaminant
levels have been below groundwater
cleanup objectives for some prescribed
period of time
4.2.1 Monitoring Network and Sampling Frequency
The current LTM plan for OU D specifies annual sampling of the six groundwater
EWs, and annual or biennial sampling of 45 groundwater monitoring wells. These
relatively infrequent sampling frequencies are appropriate given that:
1. The current plume is well-characterized and apparently decreasing in footprint
and concentration as long as the groundwater extraction system continues to
operate;
2. Available data indicate that the advective groundwater flow velocity is
relatively slow (average of 14 and 30 feet per year in the A and B monitoring
zones, as described in Section 2.2.2); and
3. Currently, there are no sensitive groundwater receptors near the plume.
4-4
S:\ES\shared\cen\MNO\EPA\McClellan\Writeup\McClellanMNODraftFinal.doc
-------
^
O
H
^
tsS
^E4.3
ROUND WATER MONITOS
UNO NETWORK OPTIMIZ
ABLE UNIT D
I AFB, CALIFORNIA
TABI
LITATIVE EVALUATION OF G
THREE-TIERED MONITOl
OPER
McCLELLAP*
o
a
Rationale
s
o
I frl
|g|
| £ S
•3 S f* S
>,, «
•a S
a
o
I
1
"S
§ a s
a S a
U « I
o
N
O
1
85
=•
^
C
o
'«
u
§'
_
1
a
.2
1
si
o
p
ft
3
1
ment of remedial progress, and
llation of mass-removal rates, assess
1
ft
nnua
<
1
^
5
^
W
§•
ss optimization evaluat
Q
g
3
a
,ment of remedial progress, and
llation of mass-removal rates, assess
1
P-H
nnua
<
g
^
03
<
^
w
1
•°
>
0
i
3
3
Q
3
"ft
.S
1
it data are useful in ass.
.0
•1
M
1
nd B) hinders usefulness of mo:
33
1
rt
1
§
8
g
S
| Inactive well. I
3
<
3
"•
5
^J-
1
W
§•
_
1
.2
1
si
o
p
ft
3
a
ment of remedial progress, and
llation of mass-removal rates, assess
1
nnua
<
g
^
03
<
^
w
•§
if the 5-ug/L TCE isopl
•3
a
o
1
cs
1
1
M
g
i
•3
o
a
1
I
a
3
.S
[Consider discor
1
<
1
<
03
^
^
w
§•
ss optimization evaluat
(D
O
&
3
1
ment of remedial progress, and
llation of mass-removal rates, assess
[Facilitates calci
1
<
1
^
ff3
MCL in plume inte:
[Monitors TCE (
1
<
1
<
<
^
S
o
i
3
/-ely elevated TCE concentrations in
[Monitors re lath
1
<
1
<
<
^
2
/ VOC concentrations.
f
£
1
1
Relatively infrequent sampling
boundary of TCE MCL target area.
[Defines current
„
1
S
1
H
<
^
S
.2
3
3
/-ely elevated TCE concentrations in
[Monitors re lath
1
<
1
<
<
IT)
^
2
d
1
ffl
d
a
O
and effect of extraction system
ically elevated TCE concentrations,
[Monitors histor
1
<
1
<
2
<
n
i
2
a
0
1
a
§
u
O
sd by low to non-detec
*i
'M
.S
"ft
1
1
efl
northern boundary of TCE MCL tai
S
1
n
„
1
S
1
H
<
^
2
•8
S
1
"o
tl
a
8
3
JH1
1
a
.2
ta
s
*g
o
'1
a
*a
efl
a
.1
s
^
|
1
1
.2
1
o
o
o
a
a
&
•s
a
0
1
cr
o
[Consistently lo>
1
H
<
^
2
•8
S
"o
a
8
1
T3
^
3
3
a
.2
ta
-1
3
o
1
; plume magnitude or v
rt
1
ra
.S
y COCs; additional data not us<
w (< MCL) concentrations of primar
[Consistently lo>
1
S
<
£
s
w
"g
^T1
^
3
a
o
1
E.
X
Q
I
3
o
1
o
1
>
1
3
ft
*i
1
8
3
S
to
l
1
1
§
u
o
1
^
8
§
j
o
-f
ditional on having belo
§
.2
4O
1
an zone. However, this recomrr
V1W-10 monitors higher-concentratii
Clustered well I
events.
1
<
<
£
^
"S
fe
3
a
8
1
JH
3
a
.2
S
5
1
O
1
; plume magnitude or v
rt
1
S
.S
y COCs; additional data not use
N (< MCL) concentrations of primar
[Consistently loi
1
S
<
^
l<
1
increasing trends in VOC concentral
iPotential slight
1
<
1
"•
<
^
f^
-5
1
;loser to estimated 5-ug/L TCE
aied in same interval and is located c
|MW-89 is scree
1
H
<
i
A
§
'i
a
s
1
3.
o
a
.2
1
s
1
1
sation of 5-ug/L TCE isopleth over t
|Helps define loi
„
1
S
1
s
<
s
l<
§
'i
a
s
1
3.
o
a
1
.§
1
1
1
sation of 5-ug/L TCE isopleth over t
|Helps define loi
„
1
S
1
s
<
o
l<
-5
1
;loser to estimated 5-ug/L TCE
aied in same interval and is located c
a
1
H
<
Is
A
..d
|
;loser to estimated 5-ug/L TCE
aied in same interval and is located c
a
1
H
<
Is
A
d
ivide useful informatio]
&
o
G
g
'93. Well is too far from plume
dances since monitoring began in 19
[No MCL excee.
1
2
<
^
rs
A
1
1
1
1
o
s
o
o
'93. Well is too far from plume
dances since monitoring began in 19
[No MCL excee.
1
S
<
o
js
,<
M
^rt
0
*3
o
a
.^
•§
1
8
-g
0
S
tn
i
1
i
§
o
1
•§
8
0
,_]
U
s
ditional on having belo
0
.2
§
1
an zone. However, this recomrr
1
§
1
0
1
O
5
Clustered well I
events.
?
<
<
£
S
(D
'-£
1
0
O
"S
I
o
a
8
a
"S
o
Q
O
o
%
I
^ncem
j
U
2
g
fi
a
Q is conditional on hav:
o
g
8
e
Itration interval. However, this
veil MW-15 monitors higher-concer
its.
Paired A-zone \
monitoring evei
1
<
<
N
£
S
ul information.
1
0
O
o
I
"ft
1
[Well is too far i
1
3
<
o
IT)
i
S
ul information.
1
0
o
I
"ft
1
|Well is too far i
1
""
<
10
i
s
ul information.
1
0
o
I
"ft
1
|Well is too far i
1
H
<
r-i
js
s
1
•1
.S
&1
-1
3
^
dde definition of plumt
o
2
J
d
1 information. Wells MW-90 a
o
ft
o
1
a
"E
1
|Well is too far i
1
H
<
00
IT)
|
2
1 information.
rom plume (650 ft) to provide usefu
|Well is too far i
1
H
<
3
O
5J
S
i
M
ra
S
o
T3
§
rmation. Last MCL ex
1
3
W
1
O
o
§
a
i
<§
pgradient to cross-gradient, and too
^
3
J
•a
X
1
H
<
S
o
5J
2
S
o
S
Ol
a
o
o
o
a
•g
-H
O
a
O
u
Only 1 trace-level del
a
.2
1
|
M
I
o
i
^
&
o
0
1
1
g
o
o
ft
| Hydraulically u
1
S
<
-
o
5
S
§"*
o
2
a
o
U
8
a
&
o
a
.2
8
•§
g
"V
Only infrequent trace
a
.2
1
|
M
I
o
i
^
p- to cross-gradient and too far from
3
3
1
ffi
1
S
<
(N
O
5
S
d
.2
1
•3
|
0
§
pgradient and too far from plume (1
p
1
1
H
<
3
js
2
S
an m
«r
M
.S
0
1
8
in
U
O
O
&
a
&
^
8
1
8
u
2
O
§
1
g
*>
P
ft
o
o
o
3
1
o
0
1
•3
ft
| Hydraulically u
1
S
<
p
o
£
S
mgradient of plume area.
o
I
1
„
1
s
1
ffl
— ffl
1 n
3 ^
N 2
-------
n o
•* B!
Ed O
d Es
™ S
H 2
|
o 3
w a
Isl
g H 3
|ia
•it,
113
1°"
n
u
PS
u
pa"
&
<
Z
u
u
u
u
&
ex
ationale
PS
o
61) ij -^
'fi g "a
.-§ g, |
« £ a
•s S fe S
-^ 1
S
|
a
"S
§ a s
* 1 *
fe
o
N
n
1—1
i
g
o
infrequent monil
1
g
ons su
C detecti
8
o
o
^
rt
I
g
er quality t
ffl
1
isessment of
«t
o
o
a
PH
3
1
S
-
1
ID
S
ffl
-H
2
§
I
lelected pri:
g trends in s
s
a
1
i
1
"a
1
<
-
1
d
•^
ffl
^f
2
c3
.fi
1
1
ffl
1
o
1
o
1
*o
[W-59 suff
2
1
m
)f nearby Zo:
w.
1
1
•|
ID
S
ffl
t~~
2
•
.a
*d
1
1
.s
:ections supports
•a
O
0
o
1
level hisl
i
1
1
•1
1
S
er quality t
*
ffl
1
isessment of
«t
o
o
a
PH
3
1
S
-
•a
ID
S
ffl
oo
1
d
1
*d
0
fi
1
g
el*
I
|
to 2001, s
o
a
ry COCs from If
1
£"•
O
•el detect
"i
o
S
"d
O
1
1
o
QJ:
1
1
a
a
o
y well near c
i3
S
n
S
N
3
1
S
-
1
ID
S
ffl
Ol
2
Ol
~*
d
U
8
§
1
£
level detection o
I
-H
1
§
d
1
3
1
a
&
o
o
o
from plum
and too far
1
1
9
•a
3
•a
ffi
1
ID
S
ffl
O
2
Ol
Ol
Cs since :
O
U
;tions of primary
S
ID
T3
O
§
informal
1
1
*>
O
o
^
I
1
3
far from p]
lent and too
y cross-gradi
•a
3
•1
ffi
•a
ID
S
ffl
O
O
2
Ol
Ol
Cs since :
O
U
;tions of primary
S
ID
T3
O
§
informal
1
1
*>
O
p
o
t
I
1
3
far from p]
lent and too
y cross-gradi
•a
3
•1
ffi
•a
ID
S
ffl
S
O
S
.ry COCs.
.fi
c detections of pi
o
S
o
d
formatio
.fi
1
^
o
H
o
G
i
from plum
and too far
1
1
9
•a
3
•a
ffi
1
ID
S
ffl
O
O
2
Ol
Ol
OCs sina
U
>
ections of primal
o
s
1
1
1
1
o
-
1 a
II
P D
•^ 0
1°
'>
II
0 0
O O
° 0
i i
p p
II
go
0
idient and t
idient and t
to cross-gr;
to cross-gr;
1 1
11
§ §
13 13
3 3
•3 -3
ffi ffi
1 1
ID ID
S S
ffl ffl
co en
o o
2 2
-------
O
HI <
O
HI
HI <
on ^
u-a
Q
^
O
CM
m
CO
CD
CO
CM
O
O
LO
o
LO
CM
0
0
M—
o
o
LO
JO
03
3
CT
0
^
O
-------
The LTM objectives for monitoring and extraction wells differ. Because EWs draw
water from a larger (and sometimes unquantified) volume of the aquifer surrounding the
well, the analytical data cannot be used for plume definition to the same extent as data
from monitoring wells. However, data from EWs can be used to assess contaminant
mass-removal rates and progress toward achieving RAOs, and to facilitate system
optimization decision-making. Because the plume extent and magnitude are generally
well-characterized both laterally and vertically, and because operation of the groundwater
extraction system appears to be providing hydraulic control sufficient to cause reductions
in the areal extent of and COC concentrations within the CAH plume , continued
monitoring of a relatively small number of wells (i.e., fewer than the 51 wells currently
included in the OU D LTM plan) should be sufficient to monitor changes in the plume
extent and magnitude in the future.
4.2.1.1 Extraction Wells
The current operational status of extraction well EW-84 is not known. According to
Parsons (2000), this well was deactivated and there were no plans to reactivate it in the
foreseeable future. However, annual monitoring of this EW is continuing (URS, 2002).
The long screened interval of this well (from 22.8 feet below mean sea level [bmsl] to -
97.2 feet bmsl; see Table 3.1) hinders the usefulness of the analytical data obtained from
this well (this issue also pertains to the other extraction wells) in that it cannot be
confidently associated with either the A or B zones. However, groundwater samples
from this well exhibit COC concentrations greater than MCLs that are likely present
primarily in the uppermost part of the aquifer (i.e., the A monitoring zone) based on
groundwater quality data from both A- and B-zone monitoring wells. Continued
sampling of EW-84 will provide data useful for monitoring the magnitude of COC
concentrations in the west-central plume area and the hydraulic effects of groundwater
extraction. However, if this well remains inactive, its abandonment should be considered
given that it may be acting as a conduit for downward migration of VOCs from the A to
the B zone. As shown in Table 4.3, continued annual monitoring of the other five
(active) EWs is recommended in order to permit periodic calculation of mass removal
4-8
S:\ES\shared\cen\MNO\EPA\McClellan\Writeup\McClellanMNODraftFinal.doc
-------
rates, and to facilitate assessment of remedial progress and system optimization.
However, well EW-86 is currently outside of the 5-ug/L TCE isopleth, and consideration
should be given to discontinuing pumping of this well unless it is required for hydraulic
control of the plume.
4.2.1.2 A-Zone Monitoring Wells
Continued monitoring of only 8 of the 24 A-zone wells currently included in the LTM
program is recommended (Table 4.3 and Figure 4.1). In addition, continued monitoring
of only 2 of the 7 IAB wells reassigned to the A zone, as proposed by URS (2002), is
recommended.
Nine of the A-zone wells (MW-1004, MW-1041, MW-1064, MW-1073, MW-1026,
MW-237, MW-351, MW-412, and MW-350) and one of the lAB-zone wells (MW-1042)
recommended for removal from the LTM program are distant (approximately 650 to
2,200 feet) from the location of the 5-ug/L TCE isopleth as estimated during the 1Q02
sampling event, and also are located hydraulically upgradient or crossgradient from the
OU D plume area. These wells are too far from the plume to provide useful information
on plume magnitude and extent given the localized and hydraulically contained
distribution of elevated TCE concentrations. Well MW-240 also is 700 feet from the
plume as defined by the 5-ug/L TCE isopleth, but is located downgradient from the OU
D waste pits, based on a regional south to southwest groundwater flow direction (Section
2.2.2). Based on the decreasing trends exhibited by TCE concentrations at the
downgradient (southern) edge of the A-zone plume (e.g., at wells MW-14, MW-15, and
MW-242; see Section 5), plume expansion toward well MW-240 is not occurring, and
continued monitoring of this well is not necessary unless hydraulic conditions change.
According to URS, the distant wells listed in the previous paragraph are sampled
primarily for community relations purposes (i.e., to demonstrate that contaminant levels
in these off-Base residential areas are remaining low and are not of concern from a
human health point of view). The recommendation to discontinue monitoring of these
wells is primarily technical and does not take into account political/community relations
4-9
S:\ES\shared\cen\MNO\EPA\McClellan\Writeup\McClellanMNODraftFinal.doc
-------
concerns. The recommendations should be evaluated with these other concerns in mind
and followed as appropriate.
Cessation of monitoring at wells MW-88, MW-91, MW-92, and MW-458, which are
located closer to, but still outside of, the 1Q02 5-ug/L TCE isopleth, also is
recommended. Continued monitoring of these four wells is not recommended because
monitoring of the plume, as defined by the 5-ug/L TCE isopleth, over time can be
accomplished using sampling results from other wells located closer to the affected area
(i.e., MW-89, MW-90, and MW-14; Figure 4.1). Concentrations of the four primary
COCs (PCE, TCE, cis-l,2-DCE, and 1,2-DCA) detected in samples from well MW-89
have never exceeded MCLs, and concentrations detected in MW-90 and MW-14 have not
exceeded MCLs since 1997.
Cessation of monitoring at wells MW-72, MW-241, and MW-242 is recommended
because each of these wells is paired with another A-zone well that monitors a higher-
concentration zone. For example, clustered wells MW-10, MW-72, and MW-241 are
screened from -38 to -48 feet msl, -58 to -68 feet msl, and -49 to -69 feet msl,
respectively. TCE concentrations in samples from MW-72 have decreased steadily from
440 ug/L in April 1990 to 2.95 ug/L in February 2001. Similarly, TCE concentrations in
MW-241 have decreased from 210 ug/L in June 1993 to 2.53 ug/L in February 2001.
During the period from 1990 to 2001, TCE concentrations at MW-10 have decreased
from 1,100 ug/L to 50.9 ug/L. In this case, continued monitoring of MW-10 will allow
assessment of how maximum COC concentrations at these plume-interior locations
change over time. In the similar case of well pair MW-15/MW-242, continued
monitoring of MW-15 will enable monitoring of maximum COC concentrations at this
location.
The remaining four wells recommended for removal from the LTM program (wells
MW-53, MW-55, MW-70, and MW-74) are IAB wells that are screened in intermediate
zones between the A and B monitoring zones, and concentrations from these wells
historically have not been used for plume interpretation. As described in Section 3.1,
4-10
S:\ES\shared\cen\MNO\EPA\McClellan\Writeup\McClellanMNODraftFinal.doc
-------
URS (2002) has proposed that these wells be reassigned to the A zone for monitoring and
interpretive purposes (Table 3.1). These four wells have been included in the LTM
program because the data may provide useful information about the effectiveness of the
extraction system.
With the exception of one 6-ug/L concentration of TCE in the sample collected from
MW-53 in February 1998, detected concentrations of the primary COCs in these four
IAB wells have not exceeded MCLs. The maximum concentrations of the four primary
COCs detected in MW-52, MW-55, and MW-74 since the 1990-1991 time frame have
been 4.4 ug/L and 3.8 ug/L, respectively. The primary COCs have never been detected
in MW-70. Given the relatively uncontaminated nature of the zones monitored by these
wells, the data obtained from them does not provide very useful information about the
effectiveness of the extraction system except that it may be limiting downward migration
of dissolved contaminants from the A zone. However, water quality in the portions of the
IAB zone monitored by these five wells has been well characterized, and appears to be
stable over time. Therefore, continued monitoring is not necessary unless the hydraulic
control exerted by the extraction system decreases.
The remaining 10 A-zone wells (MW-10 through MW-15, MW-38D, MW-52, MW-
76, MW-89, MW-90) are recommended for continued sampling because they:
• Monitor relatively elevated COC concentrations within the plume interior, thereby
allowing assessment of the plume magnitude and progress toward aquifer cleanup
goals over time;
• Are useful in tracking plume stability or contraction through time (e.g., as defined
by the 5-ug/L TCE isopleth); or
• Have exhibited increasing trends for one or more of the COCs that should be
monitored over time to assess potential plume expansion either laterally or
vertically.
4-11
S:\ES\shared\cen\MNO\EPA\McClellan\Writeup\McClellanMNODraftFinal.doc
-------
4.2.1.3 B-Zone Monitoring Wells
The B-zone VOC plume is largely defined by EWs that are screened in both the A and
B monitoring zones (i.e., the analytical results for these wells are assigned to both the A
and B zones for data assessment and presentation purposes). Compared to the number of
wells screened solely in the A zone, there are relatively few monitoring wells screened
solely in the B zone beneath the A-zone plume (Table 3.1). Data from these B-zone
monitoring wells indicate that COC concentrations in B-zone groundwater beneath the A-
zone plume are relatively low (maximum detected CAH concentration of 6.45 ug/L [for
TCE] from 1990 to 2001). Due to the relatively few B-zone wells, and the importance of
documenting water quality in this zone over time, continued monitoring of four of the
five B-zone wells located in the immediate vicinity of the plume area (wells MW-51,
MW-54, MW-58, and MW-59) is recommended. Given the low magnitude of the COC
concentrations, continued biennial monitoring is appropriate for three of these wells;
annual sampling is recommended for well MW-54, where slight increasing trends for
some COCs have been observed. Continued monitoring of IAB well MW-57 is not
recommended due to its close proximity to B-zone well MW-59. Well MW-19D, located
south of and potentially downgradient from the main plume area (Figure 4.1), near the
estimated southern limit of the extraction system capture zone, serves as a sentry well,
and continued biennial monitoring of this well is recommended.
Continued monitoring of B-zone wells located both distant and hydraulically up- or
crossgradient from the plume area is not recommended. These wells, which include MW-
1003, MW-1001, MW-1043, MW-1010, MW-104, MW-1027, and MW-1028 (Table 3.1
and Figure 4.1), do not provide useful information regarding plume magnitude and
extent.
4.2.2 Laboratory Analytical Program
Groundwater samples from OU D wells are analyzed for VOCs using USEPA Method
SW8260B. Because the characterization of conditions in the OU D groundwater plume
has been largely completed, groundwater samples collected from monitoring wells could
4-12
S:\ES\shared\cen\MNO\EPA\McClellan\Writeup\McClellanMNODraftFinal.doc
-------
be analyzed for selected COCs using Method 802IB, rather than the currently-used
Method 8260B. Method 8021B can be used to analyze for the primary COCs at the site,
and could potentially result in a considerable reduction in analytical costs. Depending on
the laboratory, the cost for analysis of a groundwater sample using Method 802IB (a gas-
chromatographic [GC] method) can be substantially lower than the cost of analysis using
Method 8260B (a gas-chromatographic/mass-spectrographic [GC/MS] method),
especially if the target analyte list is reduced. USEPA Method 8260B should still be used
to analyze samples from the few wells that contain substantially elevated CAH
concentrations, in order to obtain appropriate analyte identifications.
4.2.3 LTM Program Flexibility
The LTM program recommendations made in Sections 4.2.1 are based on available
data regarding current (and expected future) site conditions. Changing site conditions
(e.g., lengthy malfunction or significant adjustment of the groundwater extraction
system) could affect plume behavior. Therefore, the LTM program should be reviewed if
hydraulic conditions change significantly, and revised as necessary to adequately track
changes in plume magnitude and extent over time.
4-13
S:\ES\shared\cen\MNO\EPA\McClellan\Writeup\McClellanMNODraftFinal.doc
-------
SECTION 5
TEMPORAL STATISTICAL EVALUATION
Temporal data (chemical concentrations measured at different points in time) can be
examined graphically, or using statistical tests, to evaluate dissolved-contaminant plume
stability. If removal of chemical mass is occurring in the subsurface as a consequence of
attenuation processes or operation of a remediation system, mass removal will be
apparent as a decrease in chemical concentrations through time at a particular sampling
location, as a decrease in chemical concentrations with increasing distance from chemical
source areas, and/or as a change in the suite of chemicals through time or with increasing
migration distance.
5.1 METHODOLOGY FOR TEMPORAL TREND ANALYSIS OF
CONTAMINANT CONCENTRATIONS
Temporal chemical-concentration data can be evaluated by plotting contaminant
concentrations through time for individual monitoring wells (Figure 5.1), or by plotting
contaminant concentrations versus downgradient distance from the contaminant source
for several wells along the groundwater fiowpath, over several monitoring events.
Plotting temporal concentration data is recommended for any analysis of plume stability
(Wiedemeier and Haas, 2000); however, visual identification of trends in plotted data
may be a subjective process, particularly if (as is likely) the concentration data do not
exhibit a uniform trend, but are variable through time (Figure 5.2).
The possibility of arriving at incorrect conclusions regarding plume stability on the
basis of visual examination of temporal concentration data can be reduced by examining
temporal trends in chemical concentrations using various statistical procedures, including
5-1
S:\ES\shared\cen\MNO\EPA\McClellan\Writeup\McClellanMNODraftFinal.doc
-------
FIGURE 5.1
TCE CONCENTRATIONS THROUGH TIME
AT WELL MW-38D
THREE-TIERED MONITORING NETWORK OPTIMIZATION
OPERABLE UNIT D
MCCLELLAN AFB, CALIFORNIA
~-, 200 -
-j
e
.9
«
TOE Concert
-, !
*
-A.
*
*• *
,*,
12(7(92 12(7(93 12(7(94 12(7(95 12(6(36 12(6(9? 12(6(38 12(6(39 12(5(00 12(5(01
Date
regression analyses and the Mann-Kendall test for trends. The Mann-Kendall
nonparametric test (Gibbons, 1994) is well-suited for evaluation of environmental data
because the sample size can be small (as few as four data points), no assumptions are
made regarding the underlying statistical distribution of the data, and the test can be
adapted to account for seasonal variations in the data. The Mann-Kendall test statistic
can be calculated at a specified level of confidence to evaluate whether a temporal trend
is exhibited by contaminant concentrations detected through time in samples from an
individual well. If a trend is identified, a nonparametric slope of the trend line (change in
concentration per unit time) also can be estimated using the test procedure. A negative
slope (indicating decreasing contaminant concentrations through time) or a positive slope
5-2
S:\ES\shared\cen\MNO\EPA\McClellan\Writeup\McClellanMNODraftFinal.doc
-------
Trend
Increasing Trend
No Trend
Confidence Factor
HIGH
Confidence Factor
LOW
Variation
LOW
Variation
HIGH
5.2
OF
IN
OU D Monitoring Network Optimization
McClellan AFB, California
draw\739732\diffusion\williamsA.cdr pg1 nap 4/3/02
-------
(increasing concentrations through time) provides statistical confirmation of temporal
trends that may have been identified visually from plotted data (Figure 5.2).
The relative value of information obtained from periodic monitoring at a particular
monitoring well can be evaluated by considering the location of the well with respect to
the dissolved contaminant plume and potential receptor exposure points, and the presence
or absence of temporal trends in contaminant concentrations in samples collected from
the well. The degree to which the amount and quality of information that can be obtained
at a particular monitoring point serve the two primary (i.e., temporal and spatial)
objectives of monitoring must be considered in this evaluation. For example, the
continued non-detection of a target contaminant in groundwater at a particular monitoring
location provides no information about temporal trends in contaminant concentrations at
that location, or about the extent to which contaminant migration is occurring, unless the
monitoring location lies along a groundwater fiowpath between a contaminant source and
a potential receptor exposure point. Therefore, a monitoring well having a history of
contaminant concentrations below detection limits may be providing little or no useful
information, depending on its location.
A trend of increasing contaminant concentrations in groundwater at a location between
a contaminant source and a potential receptor exposure point may represent information
critical in evaluating whether contaminants are migrating to the exposure point, thereby
completing an exposure pathway. Identification of a trend of decreasing contaminant
concentrations at the same location may be useful in evaluating decreases in the areal
extent of dissolved contaminants, but does not represent information that is critical to the
protection of a potential receptor. Similarly, a trend of decreasing contaminant
concentrations in groundwater near a contaminant source may represent important
information regarding the progress of remediation near, and downgradient from the
source, while identification of a trend of increasing contaminant concentrations at the
same location does not provide as much useful information regarding contaminant
conditions. By contrast, the absence of a temporal trend in contaminant concentrations at
a particular location within or downgradient from a plume indicates that virtually no
5-4
S:\ES\shared\cen\MNO\EPA\McClellan\Writeup\McClellanMNODraftFinal.doc
-------
additional information can be obtained by continued monitoring of groundwater at that
location, in that the results of continued monitoring through time are likely to fall within
the historic range of concentrations that have already been detected (Figure 5.3).
Continued monitoring at locations where no temporal trend in contaminant
concentrations is present serves merely to confirm the results of previous monitoring
activities at that location. The relative amounts of information generated by the results of
temporal-trend evaluation at monitoring points near, upgradient from, and downgradient
from contaminant sources are presented schematically as follow:
Monitoring Point Near Contaminant Source
Relatively less information Nondetect or no trend
Relatively more information
Increasing trend in concentrations
Decreasing trend in concentrations
Monitoring Point Upgradient from Contaminant Source
Relatively less information Nondetect or no trend
Relatively more information
Decreasing trend in concentrations
Increasing trend in concentrations
5-5
S:\ES\shared\cen\MNO\EPA\McClellan\Writeup\McClellanMNODraftFinal.doc
-------
c
_o
*•>
co
o
c
o
O
Likely Future
Results
• Historic Results-
Time
5.3
OF AT
NO
IN
IS
OU D Monitoring Network Optimization
McClellan AFB, California
draw\739732\diffusion\williamsA.cdr pg2 nap 4/3/02
-------
Monitoring Point Downgradient from Contaminant Source
Relatively less information Decreasing trend in concentrations
Nondetect or no trend
Relatively more information Increasing trend in concentrations
5.2 TEMPORAL EVALUATION RESULTS
The analytical data for groundwater samples collected from the 51 wells in OU D
LTM program from April 1990 through August 2001 were examined for temporal trends
using the Mann-Kendall test. The objective of the evaluation was to identify those wells
having increasing or decreasing concentration trends for each COC, and to consider the
quality of information represented by the existence or absence of concentration trends in
terms of the location of each monitoring point.
Summary results of Mann-Kendall temporal trend analyses for COCs in groundwater
samples from wells in the TCE plume area are presented in Table 5.1. As implemented,
the algorithm used to evaluate concentration trends assigned a value of "ND" (not
detected) to those wells with sampling results that were consistently below analytical
detection limits through time, rather than assigning a surrogate value corresponding to the
detection limit - a procedure that could generate potentially misleading and anomalous
"trends" in concentrations. In addition, a value of "
-------
these samples, which is primarily an artifact of the analytical procedures, and could
generate false conclusions regarding concentration trends. The color-coding of the Table
5.1 entries denotes the presence/absence of temporal trends, and allows those monitoring
points having nondetectable concentrations, concentrations below PQLs, decreasing or
increasing concentrations, or no discernible trend in concentrations to be readily
identified. Figure 5.4 thematically displays the Mann-Kendall results for TCE by well
and hydraulic unit; the analytical results for TCE in 2000 and 2001 are also presented.
Several of the wells were only sampled once, and EW-85 and MW-1043 were not
sampled during either of the events in the 2-year period.
The basis of the decision to remove or retain a well in the monitoring program based
on the value of its temporal information is described in the "Rationale" column of Table
5.1. In general, monitoring wells at which detected chemical concentrations display no
discernible temporal trend (e.g., wells MW-53, MW-59 MW-90, MW-237, and MW-
1064 ) represent points generating the least amount of useful information, and can be
recommended for removal from the monitoring network. Monitoring wells upgradient
from the source or crossgradient from the plume (e.g., wells MW-1010, MW-1027, MW-
1041, MW-458) for which concentrations of chemicals consistently have been non-
detected or
-------
monitoring points in the LTM program, and the frequency of monitoring at particular
locations at OU D.
5-9
S:\ES\shared\cen\MNO\EPA\McClellan\Writeup\McClellanMNODraftFinal.doc
-------
n
oo
o.
Q
8
VI
H
&
VI
d
z
2
o
H
1«
Z p
o4 «<
W N
_0
"S
ft
^
TABLE
O
Z
o
d
u.
ANALYSIS O
O
Z
—
o
^
o
^
H
z
MONITORING
O
§
O
fc
OPERABLE
<
2
1
<<
u
:CLELLAN AFB
1
1
"Sb
g
•o
o
O
w
o
1
r
-t!
O
V
w
o
O
V
w
o
-t!
I
O
V
w
o
o
•o
5
o
I
H
ll
H
I
H
ll
ii
i
W
u.
O
(/i
H
of
O!
PH
V
if
H
li
8
5
(5
£
a
&
w
-------
.1 (Continued)
TABLE 5
WATER MONITORING RESULTS
Q
S OF GROUN
ANALYSI
Q
Z
H
•<
O
e-
W
H
u.
O
H
-J
PTIMIZATION
O
ING NETWORK
MONITOR
O
1
W
a
H
BLE UNIT D
OPERA
—
AFB, CALIFOR?
:CLELLAN
i
Rational
c
Z.
0
E
ft
d
r-j
i— i
QJ
*o
«
H
1
a
s
£
Ti
Low concentrations over 500 ft from TCE 5ug/L isople
~~^
i
1
g
-j
V
<
&
^
OJ
X)
d
"o
a
"T3
s
Decreasing cross-gradient. TCE non-detect or
o3
^P
^
a
No trend in source area. PCE not detected since 1 990,
?•
1
H
1
1
i
i
i
o
CQ
i>
;j
^
"S
>
QJ
bu
d
0
£H
J3
Non-detect in source area. PCE not detected since 199
?•
1
H
1
1
i
i
i
o
CQ
oo
;j
^
o
"M
.S
o
•g
S
b
'bi
n
No trend in source area. TCE non-detect or <1 ug/L thr
~?"
1
H
1
1
i|i
ft
1
«
ON
^
^
O
r^
Decreasing trend upgradient. TCE not detected since 1
~^*
1
H
1
1
i|i
1
if
^
«
;j
^
i
"3.
1
•o
o
a
0
1
?•
1
H
1
is
H
i;i
«
^_,
O
£
^
TCE non-detect since 1991 , trends no longer valuable.
?•
ii
32=
::ft
«
1
9
•
CQ
m
o
;j
^
Non-detect upgradient.
"<*"
il
|
;S
iS
&
0
0
^
^
Non-detect upgradient.
"**"
iS
1
i
ii
«
0
^
^
(1)
TCE non-detect or <1 |ig/L. No trend/non-detect upgra
?•
i
1
1
TI
1
o
«
oo
0
^
^
Non-detect/no trend upgradient.
?•
c
1-
o
!|!
1;
;-Z;
«
^
^
.5 •- .2
'» bO M te
1
O
-------
Trend Res
ll
S,,
.2
S
• r;
c
CO
S
.
§§
*88
ra
T3
c
0) J2
^•53
c^
T5 C^Coc1"^
^^ fR ^C U Q^ O O flj
0) § c ~
O UJ O
CD (jNooeoo
_l I-
i
o
N
.
=
g
f T,
0
'"S
E^
Q
5
.p
i
1
n asn
x
UJ
< < <
FIGURE 5.4
_i
2
KENDALL TEMPO
MANN-I
£
ND ANALYSIS FO
UJ
£
UJ
O
ENTRATIONS OF
O
z
O
0
c
0
N
E
"o.
O
^
o
"CD
z
O)
c
'l_
0
'c
o
^
Q
^
O
oi
1
13
q
cd
LL
<
3
o
3
§1
Qi
M!
Si
g|
i
|88
o^SSS
I o o
: o o
: CM CM
cr
ii
0^0
<
;85
: o o
: CM CM
O
o
o
lO
o
CM
0
0
O
O
ID
03
a-
0
o
-------
FIGURE 5.5
TEMPORAL TREND
DECISION RATIONALE
FLOW CHART
Monitoring Network Optimization
McClellan AFB. California
Denver, Colorado
-------
SECTION 6
SPATIAL STATISTICAL EVALUATION
Spatial statistical techniques also can be applied to the design and evaluation of
groundwater monitoring programs to assess the quality of information generated during
monitoring, and to evaluate monitoring networks. Geostatistics, or the Theory of
Regionalized Variables (Clark, 1987; Rock 1988; American Society of Civil Engineers
[ASCE] Task Committee on Geostatistical Techniques in Hydrology, 1990a and 1990b),
is concerned with variables having values dependent on location, and which are
continuous in space, but which vary in a manner too complex for simple mathematical
description. Geostatistics is based on the premise that the differences in values of a
spatial variable depend only on the distances between sampling locations, and the relative
orientations of sampling locations — that is, the values of a variable (e.g., chemical
concentrations) measured at two locations that are spatially "close together" will be more
similar than values of that variable measured at two locations that are "far apart".
6.1 GEOSTATISTICAL METHODS FOR EVALUATING MONITORING
NETWORKS
Ideally, application of geostatistical methods to the results of the groundwater
monitoring program at OU D could be used to estimate COC concentrations at every
point within the dissolved contaminant plume, and also could be used to generate
estimates of the "error," or uncertainty, associated with each estimated concentration
value. Thus, the monitoring program could be "optimized" by using available
information to identify those areas having the greatest uncertainty associated with the
estimated plume extent and configuration. Conversely, sampling points could be
successively eliminated from simulations, and the resulting uncertainty examined, to
evaluate if significant loss of information (represented by increasing error or uncertainty
6-1
S:\ES\shared\cen\MNO\EPA\McClellan\Writeup\McClellanMNODraftFinal.doc
-------
in estimated chemical concentrations) occurs as the number of sampling locations is
reduced. Repeated application of geostatistical estimating techniques, using tentatively
identified sampling locations, then could be used to generate a sampling program that
would provide an acceptable level of uncertainty regarding the distribution of COCs with
the minimum possible number of samples collected. Furthermore, application of
geostatistical methods can provide unbiased representations of the distribution of COCs
at different locations in the subsurface, enabling the extent of COCs to be evaluated more
precisely.
Fundamental to geostatistics is the concept of semivariance [y(h)], which is a measure
of the spatial dependence between samples (e.g., chemical concentrations) in a specified
direction. Semivariance is defined for a constant spacing between samples (h) by:
1 2
y(h) = — zL[g(x) - g(x + h) ] Equation 6-1
2n
Where:
y(h) = semivariance calculated for all samples at a distance h from each other;
g(x) = value of the variable in sample at location x;
g(x + h) = value of the variable in sample at a distance h from sample at location x;
and
« = number of samples in which the variable has been determined.
Semivariograms (plots of y(h) versus K) are a means of depicting graphically the range
of distances over which, and the degree to which, sample values at a given point are
related to sample values at adjacent, or nearby, points, and conversely, indicate how close
together sample points must be for a value determined at one point to be useful in
predicting unknown values at other points. For h = 0, for example, a sample is being
compared with itself, so normally y(0) = 0 (the semivariance at a spacing of zero, is
zero), except where a so-called nugget effect is present (Figure 6.1), which implies that
6-2
S:\ES\shared\cen\MNO\EPA\McClellan\Writeup\McClellanMNODraftFinal.doc
-------
FIGURE 6.1
IDEALIZED SEMVARIOGRAM MODEL
THREE-TIERED MONITORING NETWORK OPTIMIZATION
OPERABLE UNIT D
MCCLELLAN AIR FORCE BASE, CALIFORNIA
3500 -
3000
•g 2500 -
"bO
& 2000
o
1 1500 ~
0
1 1000 "
0)
g
" 500 -
0
(
^ Range ^
Sill J"^~~ • •••..-
"/T
yi
/
/* Spherical model
^ Nugget: 1,500 mg 2/L 2
Range: 1,000 ft
Sill: 1,500 mg 2/L 2
/Nugget
j 1
' i r i i i ii
) 500 1000 1500 2000 2500 3000 3500
Distance (ft)
sample values are highly variable at distances less than the sampling interval. As the
distance between samples increases, sample values become less and less closely related,
and the semivariance, therefore, increases, until a sill is eventually reached, where y(h)
equals the overall variance (i.e., the variance around the average value). The sill is
reached at a sample spacing called the range of influence, beyond which sample values
are not related. Only values between points at spacings less than the range of influence
can be predicted; but within that distance, the semivariogram provides the proper
weightings, which apply to sample values separated by different distances.
When a semivariogram is calculated for a variable over an area (e.g., concentrations of
TCE in the groundwater plume at OU D), an irregular spread of points across the
semivariogram plot is the usual result (Rock, 1988). One of the most subjective tasks of
6-3
S:\ES\shared\cen\MNO\EPA\McClellan\Writeup\McClellanMNODraftFinal.doc
-------
geostatistical analysis is to identify a continuous, theoretical semivariogram model that
most closely follows the real data. Fitting a theoretical model to calculated semivariance
points is accomplished by trial-and-error, rather than by a formal statistical procedure
(Davis, 1986; Clark, 1987; Rock, 1988). If a "good" model fit results, then rfh) (the
semivariance) can be confidently estimated for any value of h, and not only at the
sampled points.
6.2 SPATIAL EVALUATION OF MONITORING NETWORK AT OU D
TCE was used as the indicator chemical for the spatial evaluation of the groundwater
monitoring network at OU D because this COC has the largest detection percentage and
spatial distribution of measurements that exceeded groundwater MCLs. Although the A
and B zones are hydraulically connected, the A-zone wells were considered separately
from the B-zone wells for the spatial analysis because the A and B zone wells are
generally screened in shallower and deeper portions of the aquifer, respectively, and have
historically have been used to create different plume distribution maps.
A kriging analysis was not conducted for the wells because this zone contains too few
wells for a valuable spatial analysis. The monitoring network includes a total of 32
designated A-zone monitoring wells (Table 3.1). Additionally, although the EWs have
historically been used to define the plume extent in both the A and B zones, data from
active EWs are not appropriate for use in a kriging analysis because they represent COC
concentrations averaged over the area within the well's capture zone, and thus are not
point specific, nor temporally discrete; the EWs are also screened across both Zones A
and B. Therefore, the active EWs were excluded from the spatial analysis. The most
recent validated analytical data available at the start of this MNO evaluation (February
2000 or March 2001) were used in the kriging evaluation because a spatial "snapshot" is
required in order to conduct the geospatial statistical analysis. Thus, 2000 and 2001 TCE
measurements from the 32 A-Zone monitoring wells were used to develop the
semivariogram model. The commercially available geostatistical software package
Geostatistical Analyst™ (an extension to the ArcView® geographic information system
6-4
S:\ES\shared\cen\MNO\EPA\McClellan\Writeup\McClellanMNODraftFinal.doc
-------
[GIS] software package) (Environmental Systems Research Institute, Inc. [ESRI], 2001)
was used to develop a semivariogram model depicting the spatial variation in TCE
concentrations in groundwater for the 32 wells completed in the A zone in the OU D
area.
As semivariogram models were calculated for TCE (Equation 6-1), considerable
scatter of the data was apparent during fitting of the models. Several data
transformations (including a log transformation) were attempted to obtain a
representative semivariogram model. Ultimately, the concentration data were transformed
to "rank statistics," in which the 32 wells were ranked according to their 2000 or 2001
TCE concentration from 1 to 32 (tie values were assigned the median rank of the set).
Transformations of this type can be less sensitive to outliers, skewed distributions, or
clustered data than semivariograms based on raw concentration values, and thus may
enable recognition and description of the underlying spatial structure of the data in cases
where ordinary data are too "noisy".
The TCE rank statistics were used to develop a semivariogram that most accurately
modeled the spatial distribution of the data. Figure 6.2 shows the semivariogram model
in comparison to the site data. The best-fit semivariogram had the following parameters:
« Spherical Model
. Range: 600 feet
. Sill: 70
. Nugget: 10
After this semivariogram model had been developed, it was used in the kriging system
implemented by the Geostatistical Analyst™ software package (ESRI, 2001) to develop
kriging realizations (estimates of the spatial distribution of TCE in groundwater at OU
D), and to calculate the associated kriging prediction standard errors. The median kriging
standard deviation was obtained from the standard errors calculated using the entire 32-
well A-zone monitoring network for OU D. Next, each of the 32 wells was sequentially
6-5
S:\ES\shared\cen\MNO\EPA\McClellan\Writeup\McClellanMNODraftFinal.doc
-------
FIGURE 6.2
TCE A-ZONE SEMVARIOGRAM MODEL
THREE-TIERED MONITORING NETWORK OPTIMIZATION
OPERABLE UNIT D
MCCLELLAN AIR FORCE BASE, CALIFORNIA
1.48-
1.11
0.74 -
0.37-
0:
-0.37
0
i_
• • *
I
§siP
-•• e.« '* e
B »: *
_fl
TW
S <•> -B
e » 9
100 200 300 400 500 600 700 800
Distance, h
removed from the network, and for each resulting well network configuration, a kriging
realization was completed using the TCE concentration rankings from the remaining 31
wells. The "missing-well" monitoring network realizations were used to calculate
prediction standard errors, and the median kriging standard deviations were obtained for
each "missing-well" realization and compared with the median kriging standard deviation
for the "base-case" realization (obtained using the complete 32-well monitoring
network), as a means of evaluating the amount of information loss (as indicated by
increases in kriging error) resulting from the use of fewer monitoring points.
Figure 6.3 illustrates the spatial-evaluation procedure by showing kriging prediction
standard-error maps for three kriging realizations. Each map shows the predicted
standard error associated with a given group of wells based on the semivariogram
parameters discussed above. Lighter colors represent areas with lower spatial
uncertainty, and darker colors represent areas with higher uncertainty; regions in the
vicinity of wells (i.e., data points) have the lowest associated uncertainty. Map A on
Figure 6.3 shows the predicted standard error map for the "base-case" realization in
6-6
S:\ES\shared\cen\MNO\EPA\McClellan\Writeup\McClellanMNODraftFinal.doc
-------
O
O
D)
C
'(/)
CO
T3
C
0)
ra
0)
Q.
.1
-t->
CD
N
a)LU
c
E
o
?CO
.g
a a
O L1J (£ E .2
«^§2S|
°? (0 Q Ql -£ ro
(O = LJJ LJJ o O
UJ S ? Q ^ af
SSzS5*
2 S o 2 f i
U- y (0 Z o 0
|J2!g
^ liJ (0 ^ 2
^ Q
ID
O
if!
Z
0
HI
a
n
(0
T3
-------
which all A-zone wells are included. Map B shows the realization in which well MW-
1004 was removed from the monitoring network, and Map C shows the realization in
which well MW-55 was removed. Figure 6.3 shows that when a well is removed from
the network, the predicted standard error in the vicinity of the missing well increases (as
indicated by a darkening of the shading in the vicinity of that well). If a "removed"
(missing) well is in an area with several other wells (e.g., well MW-55; Map C on Figure
6.3), the predicted standard error may not increase as much as if a well (e.g., MW-1004;
Map B) is missing from an area with fewer surrounding wells.
If removal of a particular well from the monitoring network caused very little change
in the resulting median kriging standard deviation (less than about 1 percent), that well
was regarded as contributing only a limited amount of information to the LTM program.
Likewise, if removal of a well from the monitoring network produced larger increases in
the kriging standard deviation, this was regarded as an indication that the well contributes
a relatively greater amount of information, and is relatively more important to the
monitoring network. At the conclusion of the kriging realizations, each well was ranked
from 1 (providing the least information) to 32 (providing the most information), based on
the amount of information (as measured by changes in median kriging standard
deviation) the well contributed toward describing the spatial distribution of TCE, as
shown in Table 6.1. Wells providing the least amount of information represent possible
candidates for removal from the monitoring network at the OU D.
6.3 SPATIAL STATISTICAL EVALUATION RESULTS
6.3.1 Kriging Ranking Results
Figure 6.4 and Table 6.1 present the ranking of monitoring locations based on the
relative value of recent TCE information provided by each well, as calculated based on
the kriging realizations. Examination of these results indicate that monitoring wells in
close proximity to several other monitoring wells (e.g., red and yellow color coding on
Figure 6.4) generally provide relatively lesser amounts of information than do wells at
greater distances from other wells, or wells located in areas having limited numbers of
6-8
S:\ES\shared\cen\MNO\EPA\McClellan\Writeup\McClellanMNODraftFinal.doc
-------
•o
c
E .2
O HI O £
a:
X < O o
5 BE N «s
o o o ~ z
U_ I U_ Z „,
O WzQ g»
co J2 u. t b
b ^ o OQ -I
^ i HI 5 O
D j_ ^
t £o
o
(O
HI
in
B
2
in
CO
©
c\i
O)
CO
r^
o
8
CD
C\l
6 o
O0"
CO
OM
o
o
o
LO
CM
0
0
M—
O
o
LO
JO
03
3
CT
0
^
O
-------
TABLE 6.1
RESULTS OF GEOSTATISTICAL EVALUATION RANKING OF ZONE A
WELLS BY RELATIVE VALUE OF TCE INFORMATION
OPERABLE UNIT D
THREE-TIERED MONITORING NETWORK OPTIMIZATION
McCLELLAN AFB, CALIFORNIA
All Zone A Wells
Well ID
MW-10
MW-88
MW-91
MW-90
MW-89
MW-1042
MW-1041
MW-76
MW-74
MW-72
MW-70
MW-55
MW-53
MW-38d
MW-242
MW-241
MW-15
MW-14
MW-12
MW-92
MW-52
MW-351
MW-11
MW-458
MW-412
MW-1004
MW-1064
MW-350
MW-237
MW-1073
MW-1026
MW-240
Kriging Ranking ^
1
2
5C/
5
5
5
5
13.5
13.5
13.5
13.5
13.5
13.5
13.5
13.5
13.5
13.5
13.5
13.5
21.5
21.5
21.5
21.5
24
25
26
27
29.5
29.5
29.5
29.5
32
OU D TCE Plume Area Zone A Wells
Well Id
MW-92
MW-74
MW-72
MW-241
MW-10
MW-76
MW-70
MW-53
MW-38d
MW-242
MW-52
MW-11
MW-458
MW-14
MW-12
MW-88
MW-91
MW-90
MW-55
MW-89
MW-15
Kriging Ranking b/
1
3.5
3.5
3.5
3.5
8
8
8
8
8
11.5
11.5
13
14.5
14.5
16
17
18.5
18.5
20
21
Remove
V
V
V
V
V
d/
-
-
-
-
-
-
-
-
-
-
-
Retain
—
-
-
-
-
-
-
-
-
-
-
-
V
V
V
V
1= least relative amount of information; 32= most relative amount of information.
1= least relative amount of information; 21= most relative amount of information.
c Tie values receive the median ranking of the set.
Wells in the "intermediate" range and receive no recommendation for removal or retention.
022/742479/McCllelanTablesDraftFinal.xls/Table 6.1
6-10
-------
monitoring points (e.g., blue color coding on Figure 6.4). This is intuitively obvious, but
the analysis allows the most valuable and least valuable wells to be identified
quantitatively. In this analysis, the A-zone wells distant from the TCE plume, as defined
by the 5-ug/L concentration isopleth (MW-1004, MW-1064, MW-350, MW-237, MW-
1073, and MW-240) were identified as providing the greatest relative amount of spatial
information. However, as discussed in Section 4.2.1.2, these wells are located
approximately 650 to 2,200 feet from the 5-ug/L TCE isopleth (Figure 6.4), and most
also are located hydraulically up- or crossgradient from the plume sources (i.e., the OU D
waste pits); they are too far from the plume to provide useful information on plume
magnitude and extent given the very localized and contained nature of dissolved COC
present at concentrations above the respective MCLs. Thus, although these wells provide
spatial information, they not in the region of interest for the OU D plume.
Therefore, a revised kriging analysis was conducted only for those monitoring wells
within 500 feet of the OU D plume, as defined by the 5-ug/L TCE isopleth. Although the
EWs are in this region, they were excluded from the analysis because their screened
intervals span both the A and B zones, and monitor water drawn from the respective EW
capture zones, while the monitoring well provide "point" data representative of water
flowing past the well. Thus, the revised analysis examined the relative spatial
information provided by the 21 monitoring wells in the vicinity of the OU D plume using
the same procedure described in Section 6.2. The best-fit semivariogram for the revised
21-well network had the following parameters:
« Circular Model
. Range: 600 feet
. Sill: 20
« Nugget: 19
Figure 6.5 and Table 6.1 present the ranking of monitoring locations based on the
relative value of TCE information provided by each well, as calculated based on the
kriging realizations for the select group of zone A wells. For example, Table 6.1
6-11
S:\ES\shared\cen\MNO\EPA\McClellan\Writeup\McClellanMNODraftFinal.doc
-------
o
-
CD
E
W oo
O) CD
£' 0)
5
C
0 _c
§* •? °
CN
00
'00
'-4-J O)
CD c
oo •-
O -T3
E CD
"§
T- o
CN .E
.0"
(0
<^rf
c: <
UJ
F L-. s > c «J
S3S <••§.£
p uj O uj O £
^ 0- LL 7 v =
Ul
_CD =
Q. ^
o 5
UJ <
O '
in
co
o
O) c
^ o
CN CD
t~* CD
s
00
°?
CM CO
*T 3-
Q
r^ u^ ©
o
LU
O
LJJ
O
n
- A
a o
O
CO
o
o
8
i I
;2
; CM
.O
CM
6 o
oer
g
CN
O
O
JO
CD
0
-------
identifies the five wells ranked at or below 3.5 (wells MW-10, MW-241, MW-92, MW-
74, and MW-72) that provide the relative least amount of information, and the four wells
ranked at or above 18.5 (wells MW-15, MW-55, MW-89, MW-90) that provide the
greatest amount of information regarding the occurrence and distribution of TCE in
groundwater in the zone A wells in the region of interest surrounding the OU D TCE
plume. The five lowest-ranked wells are potential candidates for removal from the OU D
groundwater monitoring program, and the four highest-ranked wells are candidates for
retention in the monitoring program, intermediate ranked wells receive no
recommendation for removal or retention in the monitoring program based on the spatial
analysis.
6-13
S:\ES\shared\cen\MNO\EPA\McClellan\Writeup\McClellanMNODraftFinal.doc
-------
SECTION 7
SUMMARY OF THREE-TIERED MONITORING NETWORK
EVALUATION
The 51 wells included in the groundwater monitoring program at OU D were
evaluated using qualitative hydrogeologic and extraction-system information, temporal
statistical techniques, and spatial statistics. At each tier of the evaluation, monitoring
points that provide relatively greater amounts of information regarding the occurrence
and distribution of COCs in groundwater were identified, and were distinguished from
those monitoring points that provide relatively lesser amounts of information. In this
section, the results of the evaluations are combined to generate a refined monitoring
program that potentially could provide information sufficient to address the primary
objectives of monitoring, at reduced cost. Monitoring wells not retained in the refined
monitoring network could be removed from the monitoring program with relatively little
loss of information. The results of the evaluations were combined and summarized in
accordance with the following decision logic:
1. Each well retained in the monitoring network on the basis of the qualitative
hydrogeologic evaluation is recommended to be retained in the refined
monitoring program.
2. Those wells recommended for removal from the monitoring program on the
basis of all three evaluations, or on the basis of the qualitative and temporal
evaluations (with no recommendation resulting from the spatial evaluation)
should be removed from the monitoring program.
7-1
S:\ES\shared\cen\MNO\EPA\McClellan\Writeup\McClellanMNODraftFinal.doc
-------
3. If a well is recommended for removal based on the qualitative evaluation and
recommended for retention based on the temporal or spatial evaluation, the final
recommendation is based on a case-by-case review of well information.
The results of the qualitative, temporal, and spatial evaluations are summarized in Table
7.1. These results indicate that 30 of the 51 monitoring wells could be removed from the
groundwater LTM program with little loss of information. The justifications for the
recommendations for the four wells that fall into case 3 of the decision logic are as
follow:
• Well MW-55 should be retained in the monitoring program based on its
contribution to spatial plume-definition information.
• Wells MW-72 and MW-241 are recommended for removal from the monitoring
network based on the qualitative and spatial evaluations, and for retention based on
the temporal evaluation. They should be removed from the monitoring network
because they are monitoring lower concentration portions of the A zone than
clustered well MW-10. However, this recommendation is conditional on having
below MCL concentrations of COCs during three most recent consecutive
monitoring events.
• Well MW-242 is recommended for removal from the monitoring network based on
the qualitative and spatial evaluations, and for retention based on the temporal
evaluation. It should be removed from the monitoring network because it is
monitoring lower concentration portions of the A zone than clustered well MW-
15. However, this recommendation is conditional on having below MCL
concentrations of COCs during three most recent consecutive monitoring events.
A refined monitoring program, consisting of 21 wells (13 to be sampled annually, and
8 to be sampled biennially) would be adequate to address the two primary objectives of
monitoring. This refined monitoring network would result in an average of 17 sampling
events per year, compared to 34 events per year in the current monitoring program.
7-2
S:\ES\shared\cen\MNO\EPA\McClellan\Writeup\McClellanMNODraftFinal.doc
-------
O
o
2i
0.
O
O
H
o
I
§
£
w
H
sa
2^
%%
Is
hJ H
t/!
S
a -a o
0 0
„ 01
s .3
fa "H-
= S
ffl
ffl
ffl
ffl
ffl
ffl
ffl
ffl
ffl
ffl
ffl
ffl
ffl
ffl
ffl
ffl
ffl
ffl
ffl
ffl
ffl
£9
£9
gq
gq
£9
£9
u;
5
-------
H
sa
Jj
ca
a
i/j
a
n.
_o
fa
o
n.
S
H
_o
a
"S
ecommen
VIonitorii
Frequenc
«
a
S
^
i
ta
a
i
*
1
o
ia
1
Remove
+1 OC >,
fa* & 3
= 1 '
rj « «
fe
a
e
^2
£?
1?
!
^
Z
^
-
c3
^
o
i
S
!
^
Z
^
-
c3
3
-------
Implementing these recommendations for optimizing the LTM monitoring program
at OU D at McClellan AFB could reduce current LTM annual monitoring costs events
by 50 percent. Based on analytical costs alone, implementing these recommendations
could save $2550 per year based on an estimate of $150 per sample analysis.
Additional cost savings could be realized if groundwater samples collected from select
wells (e.g., wells along the lateral and upgradient plume margins) were analyzed for a
short list of halogenated VOCs using USEPA Method SW8021B instead of USEPA
Method SW8260B.
7-5
S:\ES\shared\cen\MNO\EPA\McClellan\Writeup\McClellanMNODraftFinal.doc
-------
SECTION 8
REFERENCES
American Society of Civil Engineers (ASCE) Task Committee on Geostatistical
Techniques in Hydrology. 1990a. Review of geostatistics in geohydrology - I.
Basic concepts. Journal of Hydraulic Engineering 116(5):612-632.
ASCE Task Committee on Geostatistical Techniques in Hydrology. 1990b. Review of
geostatistics in geohydrology - II. Applications. Journal of Hydraulic
Engineering 116(6): 63 3-65 8.
CH2M Hill. 1992. Field Investigation Work Plan, OU D. Draft Final. Prepared for
McClellan AFB, CA. August.
CH2M Hill. 1994a. Groundwater Operable Unit Remedial Investigation/ Feasibility
Study. Prepared for McClellan AFB. June.
CH2M Hill. 1994b. Operable Unit D Remedial Investigation Report. Final Copy. June.
CH2M Hill. 1995. Groundwater Operable Unite Interim Record of Decision. Prepared
for McClellan AFB, CA.
CH2M Hill. 1997. Groundwater Operable Unit, Phase 2 Work Plan. Prepared for
McClellan AFB, CA. August.
CH2M Hill. 1999. Basewide VOC Feasibility Study Report, Volume 1. Final (Version
3). Prepared for McClellan AFB, CA. July.
Clark, I. 1987. Practical Geostatistics. Elsevier Applied Science, Inc., London.
Environmental Systems Research Institute, Inc. (ESRI). 2001. ArcGIS Geostatistical
Analyst Extension to ArcGIS 8 Software, Redlands, CA.
Gibbons, R.D. 1994. Statistical Methods for Groundwater Monitoring. John Wiley &
Sons, Inc., New York.
Norris, R.M. and R.W. Webb. 1990. Geology of California. John Wiley & Sons, Inc.
New York, NY. 2nd ed.
8-1
S:\ES\shared\cen\MNO\EPA\McClellan\Writeup\McClellanMNODraftFinal.doc
-------
Pankow, J.F., and J.A. Cherry. 1996. Dense Chlorinated Solvents and Other DNAPLs in
Groundwater—History, Behavior, and Remediation. Waterloo Press, University of
Waterloo. Guelph, Ontario (Canada).
Parsons. 2000. Remedial Process Optimization Report for Operable Unit D, McClellan
Air Force Base, California. Draft. Prepared for US Air Force Center for
Environmental Excellence Technology Transfer Division and Department of the
Air Force. October.
Radian. 1997'. Groundwater Monitoring Plan. Final. Prepared for McClellan AFB, CA.
September.
Radian. 1999a. Final Groundwater Monitoring Program Quarterly Report, First
Quarter 1999. June.
Radian. 1999b. Final Groundwater Monitoring Program Quarterly Report, Third
Quarter 1999. June.
Rock, N.M.S. 1988. Numerical Geology. Springer-Verlag. New York, New York
URS. 2001. 1st Quarter 2001 (1Q01) Groundwater Monitoring Program Quarterly
Report. July.
URS. 2002. 1st Quarter 2002 (1Q02) Groundwater Monitoring Program Quarterly
Report. July.
US Environmental Protection Agency (USEPA). 1994a. DNAPL Site Characterization.
R.S. Kerr Environmental Research Laboratory and the Office of Solid Waste and
Emergency Response. Publication 9355.4-16S, EPA/540/F-94/049. September.
USEPA. 1994b. Methods for Monitoring Pump-and-Treat Performance. Office of
Research and Development. EPA/600/R-94/123.
Wiedemeier, T.H., and P.E. Haas. 2000. Designing Monitoring Programs to Effectively
Evaluate the Performance of Natural Attenuation. Air Force Center for
Environmental Excellence (AFCEE). August.
8-2
S:\ES\shared\cen\MNO\EPA\McClellan\Writeup\McClellanMNODraftFinal.doc
-------
a)
0) (/)
-
O 0)
0) P
O 0)
E E
O LU
-------