903R93007
CBP/TRS 76/93
June 1993
PQ13-20&27?
Chesapeake Bay Coordinated
Split Sample Program
Annual Report,
1990-1991
Center (3!"
841 ChKfr
TD
225
.C54
C667
1990/91
copy 2
Chesapeake Bay Program
' Printed on recycled paper
-------
Chesapeake Bay Coordinated
Split Sample Program
Annual Report, 1990-1991
Analytical Methods and Quality Assurance Workgroup
of the Chesapeake Bay Program Monitoring Subcommittee
Produced under contract to the U.S. Environmental Protection Agency
Contract No. 68-WO-00-43
Printed by the U.S. Environmental Protection Agency for the Chesapeake Bay Program
-------
EXECUTIVE SUMMARY
The Chesapeake Bay Program is a federal-state partnership with a goal of restoring the
Chesapeake Bay. Its ambient water quality monitoring programs, started in 1984, sample over
150 monitoring stations once or twice a month. Due to the size of the Bay watershed (64,000
square miles) and the cooperative nature of the CBP, these monitoring programs involve 10
different analytical laboratories. The Chesapeake Bay Coordinated Split Sample Program
(CSSP), initiated in 1988, assesses the comparability of the water quality results from these
laboratories. This report summarizes CSSP results for 1990 and 1991, its second and third full
years of operation.
The CSSP has two main objectives: identifying parameters with low inter-organization
agreement, and estimating measurement system variability. The identification of parameters
with low agreement is used as part of the overall Quality Assurance program. Laboratory and
program personnel use this information to investigate possible causes of the differences, and
take action to increase agreement if possible. Later CSSP results will document any
improvements in inter-organization agreement. The variability estimates are most useful to
data analysts and modelers who need confidence estimates for monitoring data.
The CSSP has four components, each including three to five laboratories that analyze
samples from similar salinity regimes and concentration ranges (CBP 1991). Laboratories in
each component analyze triplicate field split samples that are usually collected quarterly. The
Mainstem Component is the only component that analyzes saline water samples. The five
laboratories in this component analyze samples from the mainstem of the Chesapeake Bay,
originally from near the mouth of the Potomac River, and later from near the mouth of the
Patuxent River. The three laboratories in the Potomac Component analyze samples from near
Key Bridge on the Potomac River. The four laboratories in the Virginia Component analyze
samples from near Hopewell, VA on the James River. The four laboratories in the Fall Line
Component analyze samples from the Susquehanna River fall line station at Conowingo, MD.
In the Mainstem Component, inter-organization agreement was high for 11 of the 17
parameters compared, but was low enough for the remaining 6 parameters to recommend
investigation. Further investigation was recommended when more than half of the inter-
organization differences were larger than within-organization precision, and there were
statistically significant inter-organization differences at the P < 0.01 level. Three of the 6
parameters that met those criteria had already been identified and studied (Bergstrom 1990,
Zimmermann et al. 1992). Two of these parameters, paniculate carbon (PC) and paniculate
nitrogen (PN) had method changes made in 1992, which should lead to higher inter-
organization agreement. The results for the third parameter, dissolved organic carbon (DOC),
appear to depend on the instrument used. The three parameters recommended for further
investigation were ammonium (NH4), total suspended solids (TSS), and silica (SI). Two
parameters previously identified as having low agreement, total dissolved phosphorus (TDP)
CSC.SA2.4/93 Coordinated Split Sample Program Annual Report, 1990-1991 • Page i
-------
and total phosphorus (TP) (Bergstrom 1990), now have high agreement since a blank
adjustment factor at one laboratory was corrected.
In the Potomac Component, inter-organization agreement was high for 9 of the 11
comparisons made. Two of the 11 parameters studied, nitrite + nitrate (NO23) and Total
Organic Carbon (TOC), had inter-organization differences that were larger than within-
laboratory precision on more than half of the sampling dates. Both parameters also had
statistically significant results from the Friedman test. Unfiltered samples at one laboratory
probably accounted for the NO23 differences, since that laboratory had higher results than the
other laboratories when they were analyzing unfiltered samples. Thus, no investigation is
recommended for NO23, since agreement improved when all three laboratories analyzed
filtered samples. The differences in TOC were probably due to different instruments at the
three laboratories.
In the Virginia Component, inter-organization agreement was high for 9 of the 13
parameters compared, but was low enough for the remaining four parameters to recommend
investigation. The four parameters identified were Orthophosphate (PO4F), Silica (SI), Total
Organic Carbon (TOC) and Paniculate Phosphorus (PHOSP). In all four cases, different
analytical methods at one of the tributary laboratories probably accounted for the inter-
organization differences.
In the Fall Line Component, inter-organization agreement was high for all 9
comparisons made. However, sample sizes were small, and the power of the statistical test
used was reduced by not being able to use replicate data.
Estimates of measurement system variability based on split sample data show that some
parameters have more variable results than others. Within-organization coefficients of variation
(CV) were generally below 20%, while inter-organization CV values were somewhat higher,
usually less than 40 to 60% depending on the component. In some cases these patterns were
consistent when results from different laboratories and sampling stations were compared.
The results from the second and third years of operation show that the CSSP is
successful at achieving its goals. The communication and cooperation among participants that
occurred was essential for getting the split sample results translated into actions that have
increased inter-organization agreement.
CSC.SA2.4/93 Coordinated Split Sample Program Annual Report, 1990-1991 • Page ii
-------
TABLE OF CONTENTS
EXECUTIVE SUMMARY i
LIST OF TABLES iv
LIST OF FIGURES v
GLOSSARY xii
I. INTRODUCTION 1
H. METHODS 2
A. SAMPLE COLLECTION AND SPLITTING 2
1. Mainstem Component 2
2. Potomac Component 3
3. Virginia Component 4
4. Fall Line Component 4
B. DATA ENTRY AND REDUCTION 5
C. ANALYTICAL CHEMISTRY METHODS 5
1. Mainstem Component 5
2. Potomac Component 5
3. Virginia Component 6
4. Fall Line Component 6
D. DATA ANALYSIS AND GRAPHING 6
1. Preliminary test of splitting randomness 6
2. Precision estimates 8
3. Assessing inter-organization agreement 8
m. RESULTS 10
A. WITfflN-ORGANIZATION PRECISION AND ACCURACY 10
1. Mainstem Component 10
2. Potomac Component 10
3. Virginia Component 24
4. Fall Line Component 24
B. INTER-ORGANIZATION PRECISION 31
1. Mainstem Component 31
2. Potomac Component 31
3. Virginia Component 31
4. Fall Line Component 31
CSC.SA2.4/93 Coordinated Split Sample Program Annual Report, 1990-1991 • Page iii
-------
C. INTER-ORGANIZATION AGREEMENT 32
1. Mainstem Component 32
2. Potomac Component 44
3. Virginia Component 53
4. Fall Line Component 53
IV. DISCUSSION 72
A. WrnnN-ORGANIZATION PRECISION AND ACCURACY 72
1. Mainstem Component 72
2. Potomac Component 72
3. Virginia Component 73
4. Fall Line Component 73
B. INTER-ORGANIZATION PRECISION 73
1. Mainstem Component 74
2. Potomac Component 74
3. Virginia Component 74
4. Fall Line Component 74
C. INTER-ORGANIZATION AGREEMENT 74
1. Mainstem Component 74
2. Potomac Component 75
3. Virginia Component 75
4. Fall Line Component 80
5. Parameters recommended for investigation in two or more
components 83
D. FUTURE DIRECTIONS FOR THE CSSP 83
1. Changes in Splitting Methods 83
2. Ensuring adequate ranges of split sample concentrations 84
V. REFERENCES 85
LIST OF TABLES
TABLE 1. Lower Detection Limits of Water Quality Parameters, Chesapeake Bay
Mainstem Monitoring Program, 1984-1992 11
TABLE 2. Lower detection limits of water quality parameters, Chesapeake Bay Fall
Line and Tributary Water Quality Monitoring Programs, 1984-1992 14
TABLE 3. Within-organization and inter-organization precision estimates, Mainstem
Component, June 1989 - December 1991 15
CSC.SA2.4/93 Coordinated Split Sample Program Annual Report, 1990-1991 • Page iv
-------
TABLE 4. Percent Recovery Data from Spiked Mainstem Component Split Samples. . . 16
TABLE 5. Standard Reference Material Percent Recovery Results from Mainstem
Component Laboratories 18
TABLE 6. Within-organization and inter-organization precision estimates, Potomac
Component, May 1989- December 1991 20
TABLE 7. Percent Recovery Data from Spiked Potomac Component Samples, 1989-
91 21
TABLE 8. Standard Reference Material Results, Potomac Component, 1990-1991 ... 22
TABLE 9. Within-organization and inter-organization precision estimates, Virginia
Component, 1990-1991 25
TABLE 10. Percent Recovery Data from Spiked Virginia Component Samples, 1990-
1991 26
TABLE 11. Standard Reference Material Percent Recovery Results from Virginia
Component laboratories, 1990-1991 27
TABLE 12. Within-organization field and laboratory precision estimates, Fall Line
Component, July 1991 29
TABLE 13. Percent Recovery Data from Spiked Fall Line Component Samples, 1990-
91 30
TABLE 14. Mainstem Component (CB5.3 & CB4.4) Spb't Sample Results, June 1989-
Dec. 1991 33
TABLE 15. Potomac Component (PMS-10) Split Sample Results (May 1989 -
December 1991). 45
TABLE 16. Virginia Component (TF5.5) split sample medians with Friedman analysis
results, 1990-1991 data 54
TABLE 17. Fall Line (CB1.0) Split Sample Results using one subsample per sample
date, October 1989-July 1991 65
TABLE 18. Summary of Friedman ANOVA results (Table 14) and plots of medians
for each sample date (Figs. 1-17), Mainstem Component, 1989-1991 76
CSC.SA2.4/93 Coordinated Split Sample Program Annual Report, 1990-1991 • Page v
-------
TABLE 19. Summary of Friedman ANOVA results (Table 15) and plots of medians
for each sample date (Figs. 18-30), Potomac Component, 1989-1991 78
TABLE 20. Summary of Friedman ANOVA results (Table 16) and plots of medians
for each sample date (Figs. 31-47), Virginia Component, 1990-1991 81
TABLE 21. Summary of Friedman ANOVA results (Table 17) and plots of medians
for each sample date (Figs. 48-59), Fall Line Component, 1989-1991 82
TABLE 22. Concentrations of selected parameters in surface and bottom layers at
Station CB4.4, 1984 - 1990 84
LIST OF FIGURES
FIGURE 1. Split sample data for Ammonium (NH4), from samples collected at Station
CBS.3 or CB4.4 (Mainstem), showing medians for each sample date with
precision bars 35
FIGURE 2. Split sample data for Nitrite (NO2), from samples collected at Station CBS.3 or
CB4.4 (Mainstem), showing medians for each sample date with precision bars.
35
FIGURE 3. Split sample data for Nitrite + Nitrate (NO23), from samples collected at
Station CBS.3 or CB4.4 (Mainstem), showing medians for each sample date
with precision bars 36
FIGURE 4. Split sample data for Total Dissolved Nitrogen (TDN), from samples collected
at Station CBS.3 or CB4.4 (Mainstem), showing medians for each sample date
with precision bars 36
FIGURE 5. Split sample data for Paniculate Nitrogen (PN), from samples collected at
Station CBS.3 or CB4.4 (Mainstem), showing medians for each sample date
with precision bars 37
FIGURE 6. Split sample data for Total Nitrogen (TN), from samples collected at Station
CBS.3 or CB4.4 (Mainstem), showing medians for each sample date with
precision bars 37
FIGURE 7. Split sample data for Orthophosphate (PO4F), from samples collected at Station
CBS.3 or CB4.4 (Mainstem), showing medians for each sample date with
precision bars 38
CSC.SA2.4/93 Coordinated Split Sample Program Annual Report, 1990-1991 • Page vi
-------
FIGURE 8. Split sample data for Total Dissolved Phosphorus (TDP), from samples collected
at Station CBS.3 or CB4.4 (Mainstem), showing medians for each sample date
with precision bars 38
FIGURE 9. Split sample data for Paniculate Phosphorus (PHOSP), from samples collected
at Station CBS.3 or CB4.4 (Mainstem), showing medians for each sample date
with precision bars 39
FIGURE 10. Split sample data for Total Phosphorus (TP), from samples collected at Station
CBS.3 or CB4.4 (Mainstem), showing medians for each sample date with
precision bars 39
FIGURE 11. Split sample data for Dissolved Organic Carbon (DOC), from samples collected
at Station CBS.3 or CB4.4 (Mainstem), showing medians for each sample date
with precision bars 40
FIGURE 12. Split sample data for Paniculate Carbon (PC), from samples collected at Station
CBS.3 or CB4.4 (Mainstem), showing medians for each sample date with
precision bars 40
FIGURE 13. Split sample data for Total Organic Carbon (TOC), from samples collected at
Station CBS.3 or CB4.4 (Mainstem), showing medians for each sample date
with precision bars 41
FIGURE 14. Split sample data for Total Suspended Solids (TSS), from samples collected at
Station CBS.3 or CB4.4 (Mainstem), showing medians for each sample date
with precision bars 41
FIGURE 15. Split sample data for Chlorophyll a (CHLA), from samples collected at Station
CBS.3 or CB4.4 (Mainstem), showing medians for each sample date with
precision bars 42
FIGURE 16. Split sample data for Phaeophytin (PHEA), from samples collected at Station
CBS.3 or CB4.4 (Mainstem), showing medians for each sample date with
precision bars 42
FIGURE 17. Split sample data for Silica (SI), from samples collected at Station CBS.3 or
CB4.4 (Mainstem), showing medians for each sample date with precision
bars 43
FIGURE 18. Split sample data for Ammonium (NH4), from Potomac samples collected at
PMS10, showing medians for each sample date with precision bars. ... 46
CSC.SA2.4/93
Coordinated Split Sample Program Annual Report, 1990-1991 • Page vii
-------
FIGURE 19. Split sample data for Nitrite (NO2), from Potomac samples collected at PMS10,
showing medians for each sample date with precision bars 46
FIGURE 20. Split sample data for Nitrite + Nitrate (NO23), from Potomac samples collected
at PMS10, showing medians for each sample date with precision bars. 47
FIGURE 21. Split sample data for Total Kjeldahl Nitrogen Whole (TKNW), from Potomac
samples collected at PMS10, showing medians for each sample date with
precision bars 47
FIGURE 22. Split sample data for Total Nitrogen (TN), from Potomac samples collected at
PMS10, showing medians for each sample date with precision bars. ... 48
FIGURE 23. Split sample data for Orthophosphate (PO4F), from Potomac samples collected
at PMS10, showing medians for each sample date with precision bars. . 48
FIGURE 24. Split sample data for Total Phosphorus (TP), from Potomac samples collected
at PMS10, showing medians for each sample date with precision bars . . 49
FIGURE 25. Split sample data for Total Dissolved Phosphorus (TDP), from Potomac samples
collected at PMS10, showing medians for each sample date with precision
bars 49
FIGURE 26. Split sample data for Total Organic Carbon (TOC), from Potomac samples
collected at PMS10, showing medians for each sample date with precision bars
50
FIGURE 27. Split sample data for Dissolved Organic Carbon (DOC), from Potomac samples
collected at PMS10, showing medians for each sample date with precision
bars 50
FIGURE 28. Split sample data for Total Suspended Solids (TSS), from Potomac samples
collected at PMS10, showing medians for each sample date with precision
bars 51
FIGURE 29. Split sample data for Silica (SI), from Potomac samples collected at PMS10,
showing medians for each sample date with precision bars 51
FIGURE 30. Split sample data for Biological Oxygen Demand 5 day (BODS), from Potomac
samples collected at PMS10, showing medians for each sample date with
precision bars 52
CSC.SA2.4/93 Coordinated Split Sample Program Annual Report, 1990-1991 • Page viii
-------
FIGURE 31. Split sample data for Ammonium (NH4), from samples collected at Station
TF5.5 (James), showing medians for each sample date with precision bars 56
FIGURE 32. Split sample data for Nitrite (NO2), from samples collected at Station TF5.5
(James), showing medians for each sample date with precision bars .... 56
FIGURE 33. Split sample data for Nitrite + Nitrate (NO23), from samples collected at
Station TF5.5 (James), showing medians for each sample date with precision
bars 57
FIGURE 34. Split sample data for Total Dissolved Nitrogen (TDN), from samples collected
at Station TF5.5 (James), showing medians for each sample date with precision
bars 57
FIGURE 35. Split sample data for Particulate Nitrogen (PN), from samples collected at
Station TF5.5 (James), showing medians for each sample date with precision
bars 58
FIGURE 36. Split sample data for Total Nitrogen (TN), from samples collected at Station
TF5.5 (James), showing medians for each sample date with precision bars 58
FIGURE 37. Split sample data for Orthophosphate (PO4F), from samples collected at Station
TF5.5 (James), showing medians for each sample date with precision bars 59
FIGURE 38. Split sample data for Total Dissolved Phosphorus (TDP), from samples collected
at Station TF5.5 (James), showing medians for each sample date with precision
bars 59
FIGURE 39. Split sample data for Particulate Phosphorus (PHOSP), from samples collected
at Station TF5.5 (James), showing medians for each sample date with precision
bars 60
FIGURE 40. Split sample data for Total Phosphorus (TP), from samples collected at Station
TF5.5 (James), showing medians for each sample date with precision bars 60
FIGURE 41. Split sample data for Dissolved Organic Carbon (DOC), from samples collected
at Station TF5.5 (James), showing medians for each sample date with precision
bars 61
FIGURE 42. Split sample data for Particulate Carbon (PC), from samples collected at Station
TF5.5 (James), showing medians for each sample date with precision bars 61
CSC.SA2.4/93 Coordinated Split Sample Program Annual Report, 1990-1991 • Page ix
-------
FIGURE 43. Split sample data for Total Organic Carbon (TOC), from samples collected at
Station TF5.5 (James), showing medians for each sample date with precision
bars 62
FIGURE 44. Split sample data for Total Suspended Solids (TSS), from samples collected at
Station TF5.5 (James), showing medians for each sample date with precision
bars 62
FIGURE 45. Split sample data for Chlorophyll a (CHLA), from samples collected at Station
TF5.5 (James), showing medians for each sample date with precision bars 63
FIGURE 46. Split sample data for Phaeophytin (PHEA), from samples collected at Station
TF5.5 (James), showing medians for each sample date with precision bars 63
FIGURE 47. Split sample data for Silica (SI), from samples collected at Station TF5.5
(James), showing medians for each sample date with precision bars .... 64
FIGURE 48. Split sample data for Ammonium (NH4), from samples collected at CB1.0 (Fall
Line), showing first subsample results with precision bars 66
FIGURE 49. Split sample data for Nitrite (NO2), from samples collected at CB1.0 (Fall
Line), showing first subsample results with precision bars 66
FIGURE 50. Split sample data for Nitrite + Nitrate (NO23), from samples collected at
CB1.0 (Fall Line), showing first subsample results with precision bars . . 67
FIGURE 51. Split sample data for Total Kjeldahl Nitrogen Whole (TKNW), from samples
collected at CB1.0 (Fall Line), showing first subsample results with precision
bars 67
FIGURE 52. Split sample data for Total Kjeldahl Nitrogen Filtered (TKNF), from samples
collected at CB1.0 (Fall Line), showing first subsample results with precision
bars 68
FIGURE 53. Split sample data for Total Nitrogen (TN), from samples collected at CB1.0
(Fall Line), showing first subsample results with precision bars 68
FIGURE 54. Split sample data for Orthophosphate (PO4F), from samples collected at CB1.0
(Fall Line), showing first subsample results with precision bars 69
FIGURE 55. Split sample data for Total Dissolved Phosphorus (TDP), from samples collected
at CB1.0 (Fall Line), showing first subsample results with precision bars. . 69
CSC.SA2.4/93 Coordinated Split Sample Program Annual Report, 1990-1991 • Page x
-------
FIGURE 56. Split sample data for Total Phosphorus (TP), from samples collected at CB1.0
(Fall Line), showing first subsample results with precision bars 70
FIGURE 57. Split sample data for Total Organic Carbon (TOC), from samples collected at
CB1.0 (Fall Line), showing first subsample results with precision bars . . 70
FIGURE 58. Split sample data for Total Suspended Solids (TSS), from samples collected at
CB1.0 (Fall Line), showing first subsample results with precision bars . . 71
FIGURE 59. Split sample data for Silica (SI), from samples collected at CB1.0 (Fall Line),
showing first subsample results with precision bars .71
CSC.SA2.4/93
Coordinated Split Sample Program Annual Report, 1990-1991 • Page xi
-------
GLOSSARY
accuracy
agreement
AMQAW
ANOVA
BODS
CBL
CBP
CBPO
CHLA
CRL
CSC
CSSP
CV
decision rule
DCLS
DCRA
DCRA/CRL
DOC
EPA
Friedman ANOVA
HRSD
MDE
MDHMH
MDL
NH4
NO2
NO23
ODU
OWML
Closeness of an analytical result to a "true" value. Usually assessed with
laboratory spike samples or by analysis of SRMs.
Tendency for different organizations (laboratories) to have similar
analytical results over time. Low inter-organization agreement results
from consistent differences with magnitudes greater than the within-
organization precision.
Analytical Methods and Quality Assurance Workgroup (part of the CBP
Monitoring Subcommittee)
Analysis of Variance
Biological Oxygen Demand 5 day
Chesapeake Biological Laboratory (Solomons, MD)
Chesapeake Bay Program
Chesapeake Bay Program Office (Annapolis, MD)
Chlorophyll a
Central Regional Laboratory (EPA, Annapolis, MD)
Computer Sciences Corporation (contractor at CBPO)
Coordinated Split Sample Program
Coefficient of Variation (SD/mean x 100)
Used to decide which parameters have inter-organization differences that
should be investigated. Those recommended have statistically significant
differences at P < 0.01 and have more than half of the sample dates with
inter-organization differences larger than within-organization precision.
Division of Consolidated Laboratory Services (Richmond, VA)
Department of Consumer and Regulatory Affairs (Washington, DC)
DCRA staff who perform laboratory analyses at CRL
Dissolved Organic Carbon
Environmental Protection Agency
Nonparametric statistical test used as part of the assessment of inter-
organization agreement. Assumes matched (positively correlated)
samples. Results are affected by the consistency of differences over time.
Hampton Roads Sanitation District (Virginia Beach, VA)
Maryland Department of the Environment (Baltimore & Annapolis, MD)
Maryland Department of Health and Mental Hygiene (Baltimore, MD)
Method Detection Limit; usually based on within-organization precision,
calculated from 3 times the SD of 7 low-level replicates.
ammonium
nitrite
nitrite + nitrate
Old Dominion University (Norfolk, VA)
Occoquan Watershed Monitoring Laboratory (Manassas, VA)
CSC.SA2.4/93
Coordinated Split Sample Program Annual Report, 1990-1991 • Page xii
-------
PADER
PC (POC)
PHEA
PHOSP
PN(PON)
PO4F
PO4W
precision
QA
SAS
SD
SI
SRM
subsample
TDN
TDP
TKNW
TKNF
TN
TOC
TP
TSS
USGS
VIMS
VWCB
Pennsylvania Department of Environmental Resources (Hanisburg, PA)
Particulate Carbon (Paniculate Organic Carbon); the former is measured
directly (mainstem labs); the latter is calculated from TOC - DOC
(tributary labs).
Phaeophytin
Particulate Phosphorus
Particulate Nitrogen (Particulate Organic Nitrogen); the former is
measured directly (mainstem labs); the latter is calculated from TKNW -
TKNF (tributary labs).
Orthophosphate filtered
Orthophosphate unfiltered (whole water)
Repeatability of analytical measurements. Precision is high if successive
measurements are very similar, resulting in low SD and CV values.
Within-organization precision is estimated in the CSSP from the
variability among field replicates (subsamples), and inter-organization
precision is estimated from the variability among the medians over the
three subsample results from each organization. For graphing and use in
the decision rule, within-organization precision is estimated from the
larger of the MDL, or the SD of the results from the three subsamples.
Quality Assurance
Statistical Analysis System
Standard Deviation
Silica
Standard Reference Material
Replicates split in the field, soon after sample collection, either on the
sampling boat, on a dock, or in a laboratory. Each laboratory should
receive three subsamples per spb't sample. Subsamples were called
"aliquots" in some previous CSSP reports.
Total Dissolved Nitrogen; measured directly (mainstem labs) or calculated
from TKNF + NO23 (tributary labs).
Total Dissolved Phosphorus
Total Kjeldahl Nitrogen Whole
Total Kjeldahl Nitrogen Filtered
Total Nitrogen; calculated from either TKNW + NO23 (tributary labs)
or TDN + PN (mainstem labs).
Total Organic Carbon; measured directly (tributary labs) or calculated
from PC + DOC (mainstem labs).
Total Phosphorus; measured directly (tributary labs) or calculated from
TDP + PHOSP (mainstem labs)
Total Suspended Solids
United States Geological Survey (Towson, MD and other offices)
Virginia Institute of Marine Science (Gloucester Point, VA)
Virginia Water Control Board (Glen Allen, VA)
CSC.SA2.4/93
Coordinated Spb't Sample Program Annual Report, 1990-1991 • Page xiii
-------
I. INTRODUCTION
The Monitoring Subcommittee of the Chesapeake Bay Program initiated the Chesapeake
Bay Coordinated Split Sample Program (CSSP) in 1988. Its goal is to assess the comparability
of water quality results from the 10 analytical laboratories that participate in the Chesapeake
Bay Monitoring Program (Chesapeake Bay Program 1989). This goal is being achieved by
identifying any parameters that have low inter-organization agreement and estimating
measurement system variability .
Identifying parameters with low agreement enables the organizations involved to
investigate any significant differences and take actions to raise inter-organization agreement.
This might involve changing field methods, laboratory methods, or both. Because results of
field split samples are affected by both field and laboratory variability, the terms "inter-
organization" and "within-organization" are used rather than "inter-laboratory" and "within-
laboratory." The organization includes all the elements of the measurement system: field
sampling, sample handling, laboratory analysis, data handling, and the state or municipal
agency that supervise water quality monitoring program.
Estimates of measurement system variability are useful to data users such as statisticians
and modelers who need confidence bounds for monitoring data. Although split sample results
do not include sampling variability, they are the best estimate available of total system
variability for Chesapeake Bay water quality monitoring data.
The CSSP has four components, each including three to five laboratories that analyze
samples from similar salinity regimes and concentration ranges (CBP 1991). Laboratories in
each component analyze triplicate field split samples that are usually collected quarterly.
Laboratory personnel send the analytical results to the EPA Chesapeake Bay Program Office
(CBPO) in Annapolis for data transfer (or entry) and analysis by Computer Sciences
Corporation (CSC/CBPO) staff.
This report summarizes the 1990-1991 results from the four CSSP components:
1. The Mainstem and Tidal Tributaries Component is the only component that analyzes
saline water samples. This component includes three mainstem laboratories:
Chesapeake Biological Laboratory (CBL), Virginia Institute of Marine Science (VIMS),
and Old Dominion University (ODU). It also includes a Maryland tributary laboratory,
Maryland Department of Health and Mental Hygiene (MDHMH), and a Virginia
tributary laboratory, Division of Consolidated Laboratory Services (DCLS).
CSC.SA2.4/93 Coordinated Split Sample Program Annual Report, 1990-1991 • Page 1
-------
2. The Tidal Potomac River component includes three analytical laboratories: the
Maryland Department of Health and Mental Hygiene (MDHMH) in Baltimore, the
Virginia Division of Consolidated Laboratory Services (DCLS) in Richmond, and the
EPA Central Regional Laboratory (CRL) in Annapolis. DC Department of Consumer
and Regulatory Affairs (DCRA) personnel (referred to as DCRA/CRL) conduct the
analyses at CRL.
3. The Virginia Mainstem and Tidal Tributaries component of the CSSP includes four
analytical laboratories: the Virginia Institute of Marine Science (VIMS), Old Dominion
University (ODU), Division of Consolidated Laboratory Services (DCLS), and
Hampton Roads Sanitation District (HRSD).
4. The Fall Line component includes four analytical laboratories: the Maryland
Department of Health and Mental Hygiene (MDHMH) in Baltimore, the United States
Geological Survey (USGS) in Denver, the Pennsylvania Department of Environmental
Resources (PADER) in Harrisburg, and the Occoquan Watershed Monitoring
Laboratory (OWML) in Manassas, VA. The component was planned to include the
Virginia Division of Consolidated Laboratory Services (DCLS) in Richmond (CBP
1991), but logistical problems with sample splitting and distribution have prevented this.
Communication is a key element of the CSSP, enabling the staff of the participating
laboratories and program agencies to study the results and take action when appropriate.
Interim reports on each component are issued regularly by CSC/CBPO staff to the laboratory
and program personnel involved in each component. CSC/CBPO staff combine the interim
reports in a public annual report, which can include comments or discussion from the
laboratory or program personnel. The previous Annual Report (Bergstrom 1990) included
CSSP results through December 1989.
H. METHODS
A. SAMPLE COLLECTION AND SPLITTING
1. Mainstem Component
A field crew from the Maryland Department of the Environment (MDE) collected
quarterly water samples from the surface layer at Station CBS.3, near Smith Point on the
Maryland-Virginia state line. The sampling station was changed to CB4.4 in June 1990 to
facilitate getting the samples to participating laboratories. The field crew followed the splitting
procedures in the CSSP Implementation Guidelines (CBP 1991) starting in June 1989. One
large sample was stirred on the boat in a 15 gallon carboy with a paint stirrer connected to an
electric drill. Subsamples were drawn sequentially from a spigot at the bottom of the carboy.
CSC.SA2.4/93 Coordinated Split Sample Program Annual Report, 1990-1991 • Page 2
-------
The bottles for Subsample 1 were split in the sequence MDHMH-VIMS-CBL-ODU-DCLS,
followed by the bottles for subsamples 2 and 3 in the same sequence.
Before June 1990, samples were transferred from the MDE to the VIMS boat at station
CBS.3, and MDE staff filtered and distributed Maryland samples, and VIMS staff filtered and
distributed Virginia samples. As of June 1990, the MDE field crew filtered and distributed
samples to the two Maryland laboratories (CBL and MDHMH). VIMS personnel picked up
the samples in Port Royal, VA and distributed unfiltered samples to the three Virginia
laboratories (VIMS, ODU, and DCLS). Each Virginia laboratory filtered their own samples.
DCLS started analyzing Mainstem Component samples in June 1990.
Beginning in June 1989, each laboratory analyzed a minimum of four samples per
sample date: three subsamples split in the field and a laboratory replicate for one of the
subsamples. Some laboratories did more replicates. Two estimates of within-laboratory
precision were calculated from subsample (field replicate) and lab replicate data: field precision
and laboratory precision. Only field precision is reported here, because it generally includes
laboratory precision.
2. Potomac Component
A field crew from DCRA collected quarterly water samples from the surface layer at
Station PMS-10, at Key Bridge on the Potomac River. The field crew followed the splitting
procedures in the CSSP Implementation Guidelines (CBP 1991) starting in June 1989. One
large sample was stirred at the dock in a 15 gallon carboy with boat paddle. Subsamples were
drawn sequentially from a spigot at the bottom of the carboy, in the sequence MDHMH-DCLS-
DCRA/CRL. The field crew left whole water samples in ice-filled coolers at the designated
dock. Personnel from each laboratory retrieved the coolers. The DCRA field crew did not
filter any samples. Starting in December 1990, the MDE crew that picks up samples for
MDHMH began filtering samples for nutrient analysis, usually when they picked up the
samples at Blue Plains. The MDE crew used Whatman GF/F filters with a 0.7 micron pore
size.
In the laboratory, DCLS and DCRA/CRL personnel filtered samples for Ammonium
(NH4), Nitrite (NO2), Nitrite + Nitrate (NO23), Orthophosphate (PO4F), Silica (SI), and
Total Suspended Solids (TSS). DCRA/CRL and DCLS both used pre-rinsed Gelman cellulose
membrane filters with 0.45 micron pore size, smaller than the pores in the filters used by
MDE. MDHMH did laboratory filtration for TSS only. Samples were received by the
laboratories either the same day they were collected or the following day. The March 1990
samples were not picked up for DCLS, so there are no DCLS results for that split sample.
Starting in June 1989, each laboratory analyzed a minimum of four samples per
sampling date. These come from three subsamples (field replicates) split in the field and a
laboratory replicate for one of the subsamples.
CSC.SA2.4/93 Coordinated Split Sample Program Annual Report, 1990-1991 • Page 3
-------
3. Virginia Component
A field crew from the Virginia Water Control Board (VWCB) collected quarterly water
samples from the bottom layer at Station TF5.5, near Hopewell on the James River. The field
crew followed the splitting procedures in the CSSP Implementation Guidelines (CBP 1991)
starting in February 1990, filling the subsamples in the sequence HRSD-VIMS-ODU-DCLS.
The three subsamples sent to each laboratory came from three separate chum splitters, which
were filled in rotation from the sampling hose (CBP 1991). This is not the standard splitting
protocol, which calls for starting from a single well-mixed sample. The use of three separate
spb'tters appeared to affect the results from the May 1990 sample, which had very high
suspended solids (see below). VWCB staff switched to splitting from a single large churn
splitter in June 1992.
The field crew delivered iced whole water samples to each laboratory. Samples were
received by the laboratories either the same day they were collected or the following day, and
were usually filtered the morning after they were collected. DCJJS and HRSD used pre-rinsed
Gelman cellulose membrane filters with 0.45 micron pore size, while ODU and VIMS used
Whatman GF/F glass fiber filters with 0.7 micron nominal pore size, and Gelman AE filters
for PC/PN analyses.
Starting in February 1990, each laboratory analyzed a minimum of four samples per
sampling date. These come from three subsamples (field replicates) split in the field and a
laboratory replicate for one of the subsamples. In the data collected during 1988-89 the three
analytical results did not always come from three separate subsamples, but sometimes were
split in the laboratory. For this reason, the 1988-89 data were not included in the graphs or
the statistical tests.
4. Fall Line Component
A field crew from USGS-Towson sampled the Susquehanna River fall line station at
Conowingo, MD (CB1.0), and distributed samples to each laboratory. The field crew used
USGS sampling procedures, including flow-weighted cross-sectionally integrated samples
collected at 5 sections along the well-mixed turbine outflow. Splitting was done with a single
churn splitter. Field filtration was done with a 0.45 micron membrane filter and the nutrient
samples shipped to USGS were preserved with mercuric chloride and sodium chloride
according to USGS standard protocol. Samples sent to OWML, MDHMH and PADER were
not preserved. All samples were immediately placed and kept in ice-filled coolers with a 6:1
ratio of ice to sample. Samples for MDHMH and PADER were delivered to the laboratory
on the day of collection. Samples for USGS and OWML were sent via priority mail. USGS
samples usually arrived in two days. v
There were several deviations from the recommended procedures in the CSSP guidelines
(CBP 1991). Samples were not split quarterly, each laboratory received less than three
CSC.SA2.4/93 Coordinated Split Sample Program Annual Report, 1990-1991 • Page 4
-------
subsamples (field replicates) split in the field, Standard Reference Material (SRM) results were
not reported, and laboratory replicates and spikes were not performed routinely. Samples were
collected twice in 1990 and once in 1991; they could not be split quarterly due to budget
constraints at USGS. The field crew could not split enough water to provide three subsamples
due to the size of the churn splitter. Each laboratory received two subsamples in October 1989
and July 1991, and one subsample on other dates. This reduced the power of the statistical test
used, and limited the within-organization variability estimates that could be calculated. USGS
will try using three churn splitters in future samples, one for each subsample. Only MDHMH
and OWML reported any laboratory replicates for the subsamples, and percent recovery results
were incomplete: USGS did not submit any, and PADER and MDHMH only reported them
for one sample.
B. DATA ENTRY AND REDUCTION
Laboratory or program personnel submitted their data on diskette or computer tape, or
they submitted raw data on handwritten CSSP Data Submission forms. CSC/CBPO staff
entered the handwritten data, and uploaded the digital data and converted it to SAS data sets.
Data were adjusted up to the method detection limit (MDL) if they were below it because most
labs did this to their data before submission. Field and laboratory precision and medians of
the three subsamples were calculated with the SAS procedure UNrVARIATE (SAS Institute
1990). These medians of the three subsamples for each sampling date were then used to
calculate inter-organization standard deviations and coefficients of variation using the SAS
functions STD and CV respectively.
C. ANALYTICAL CHEMISTRY METHODS
1. Mainstem Component
The three mainstem laboratories in this component use different analytical methods than
the two tributary laboratories. Mainstem laboratories measure the dissolved and paniculate
fractions of most nutrients and calculate the total fractions, while tributary laboratories measure
the total and dissolved fractions and calculate the paniculate fractions (D'Elia et al. 1987). To
account for this difference in analytical methods, data analyses were done for four and five
laboratories and also for the three mainstem laboratories. Comparisons among four
laboratories were done using MDHMH data, because DCLS started analyzing samples later,
which reduced the sample size for five-way comparisons.
2. Potomac Component
Because the three laboratories included in the Potomac component use similar analytical
methods, data analyses treated them as a single group without any subgroups. All three
CSC.SA2.4/93 Coordinated Split Sample Program Annual Report, 1990-1991 • Page 5
-------
laboratories used a single acid sulfate (not persulfate) block digestion for Total Kjeldahl
Nitrogen Whole (TKNW) and Total Phosphorus (TP) analyses.
3. Virginia Component
The four laboratories included in the Virginia component used two different groups of
analytical chemistry methods. DCLS and HRSD laboratories analyze tributary samples and
perform whole water and dissolved nutrient analyses, and calculate any paniculate parameters.
Both use a single acid sulfate block digestion for Total Kjeldahl Nitrogen Whole (TKNW) and
Total Phosphorus (TP) analyses. VIMS and ODU analyze primarily mainstem Bay samples
and perform dissolved and paniculate nutrient analyses, and calculate any total parameters.
Both use separate alkaline persulfate digestions for nitrogen and phosphorus parameters (TDN
and TDP), except ODU uses acid persulfate for TDP. The Friedman test used for statistical
analysis requires a minimum of three laboratories, so data from the tributary and mainstem
laboratories could not be analyzed separately. The method differences among the four
laboratories should be taken into account when interpreting the split sample results.
There was one method change at HRSD during the time period covered by this report.
They changed from an automated (EPA method 365.1) to a manual technique (EPA method
365.3) for Orthophosphate (PO4F) in August, 1990. This reduced their method detection limit
for PO4F from 0.05 mg/1 to 0.01 mg/1 (Table 2). ODU also uses the manual procedure, while
DCLS and VIMS use the automated procedure (see Tables 1 & 2 for method detection limits).
4. Fall Line Component
In general, three of the laboratories (MDHMH, OWML, and PADER) followed EPA
standard methods, and USGS followed USGS standard methods. All four laboratories reported
the same parameters, with a few exceptions: PADER did not report NO2, SI, or TSS data, and
USGS did not report TSS data.
D. DATA ANALYSIS AND GRAPHING
1. Preliminary test of splitting randomness
Data were checked for the randomness of the splitting procedures. If splitting was done
uniformly, the results for one of the subsamples should not be consistently higher or lower than
the results for the other subsamples. Since the subsamples are split sequentially, non-random
splitting would probably result in higher results for solids and particulates in Subsample 3,
which is drawn from the lower part of the splitting vessel. Splitting randomness was checked
with the Friedman two-way non-parametric analysis of variance, comparing the results for the
three subsamples for each parameter and sampling date, using P < 0.05 to indicate
significance.
csc.sA2.4/93 Coordinated Split Sample Program Annual Report, 1990-1991 • Page 6
-------
In the Mainstem Component, sixty-one of the 63 parameter-date combinations tested
did not have statistically significant differences QP > 0.05). For the two parameters and dates
with significant differences (PHOSP, P = 0.042, and TOC, P = 0.0046, on one date each),
the magnitudes of the differences were less than the method detection limit. This shows that
splitting was done randomly.
In the Potomac Component, 97 of the 101 parameter-date combinations tested did not
have statistically significant differences (P ^ 0.05). The four significant differences (x2 =
6.0, P = 0.028) were all for NH4, and all involved the same pattern: the subsample results
were ranked from low to high, 1 < 2 < 3. The four dates affected included three in June
(1989, 1990, and 1991) and one in September 1989. Since none of the paniculate or whole-
water parameters were affected, and NH4 is filtered, these differences for NH4 could not be
due to inadequate stirring during sample splitting. They might be due to some sort of
contamination, but the regular pattern of the differences seems to make that unlikely. The
magnitudes of the differences were relatively large, and were consistent across laboratories.
For example, in June 1989, DCRA/CRL reported 0.04, 0.051, and 0.063 mg/1; MDHMH
reported 0.036, 0.060, and 0.068 mg/1; and DCLS reported 0.04, 0.06, and 0.07 mg/1. These
differences did not lead to significant results in tests for inter-organization differences in NH4
(see Results), probably because they were consistent across laboratories.
In the Virginia Component, the subsamples were not split sequentially from a single
vessel, but from three separate churn splitters. Thus, non-random splitting could result in
higher results for solids and particulates in any subsample, not just in Subsample 3.
In the Virginia Component, the results for PHOSP and TP on May 31, 1990 show a
significant effect of non-random splitting. The test for both parameters in May had P = 0.042
(X2 = 6.5), and for both parameters, all four laboratories had the highest results for Subsample
2. The difference between subsamples was as large as 0.603 to 0.831 mg/1, a 38% difference.
All laboratories except ODU had the lowest results for Subsample 1. TSS data could not be
tested because HRSD did not analyze TSS in May, but it showed the same consistent
differences in subsample results in ODU and VIMS data. The maximum TSS difference
between subsamples for TSS was 514 and 702 mg/1, a 36% difference. TSS values in this
range are an extreme test of splitting effectiveness, but since they occur at TF5.5, the splitting
procedure should not be affected by them. The splitting method was recently changed to one
using a single vessel, as recommended in the CSSP Implementation Guidelines (CBP 1991).
Although there was a probable effect of splitting order on the May 1990 PHOSP and TP data,
this had no apparent effect on the tests for inter-organization agreement (see Results). Non-
random splitting could accentuate inter-organization differences if it occurred consistently over
several sample dates.
The test for splitting randomness could not be done on Fall Line data, because three
subsamples were not analyzed.
CSC.SA2.4/93 Coordinated Split Sample Program Annual Report, 1990-1991 • Page 7
-------
2. Precision estimates
The standard deviation (SD) and coefficient of variation (CV, standard deviation/mean
x 100) of field triplicate results estimated within-organization precision. Since the field
replicates are usually split by a different organization from the one doing the laboratory
analysis, they do not measure only "within-organization" field and laboratory precision, but
they are considered to approximate it for the purposes of this report. The CSSP results also
include laboratory replicates, split in the laboratory just before analysis. These were not used
to estimate within-organization precision because they do not include field variability, and are
almost always less variable than the field replicates.
The medians of the field triplicate results were then used to calculate the SD and CV
of the results from different organizations for each sampling date, which estimated inter-
organization precision. For the Mainstem component, precision estimates were calculated
separately for groups of four and three laboratory medians, including all the laboratories or
only the mainstem laboratories (see above). Results from DCLS were excluded for now
because they started analyzing mainstem samples at a later date.
The SD was positively correlated with the mean for several parameters, and usually the
CV was not affected by the mean. However, the CV was sometimes affected by concentration
as well. It was positively correlated with the mean in a few cases, and negatively correlated
with the mean in a few other cases. The negative correlation usually occurred when the mean
concentrations were low. Thus, neither precision estimate should be used in other analyses
without checking for concentration effects. Because the primary purpose of the CSSP is to
assess inter-organization agreement, a detailed analysis of precision estimates is beyond the
scope of this report.
The Method Detection Limit (MDL) was also used to estimate within-organization
precision, especially in the graphs of the data (see next section). At many of the laboratories,
the MDL is calculated from three times the standard deviation of seven replicates of a low-
level sample, so it estimates within-organization precision.
3. Assessing inter-organization agreement
Inter-organization agreement is the tendency for split sample analytical results from
different organizations to be consistently similar over time. Thus, any pair of laboratories with
large and consistent inter-organization differences are considered to have low agreement. A
decision rule was developed to identify which parameters had inter-organization differences that
were large and consistent enough to warrant investigation by the organizations involved. Based
on discussions by the Analytical Methods and Quality Assurance Workgroup (AMQAW) on
4/24/90 and 1/26/93, the decision was based on graphs of the data with precision bars, and the
results of statistical tests. Graphs with precision bars show the magnitude of differences, while
CSC.SA2.4/93 Coordinated Split Sample Program Annual Report, 1990-1991 • Page 8
-------
the statistical test is more sensitive to consistency of the differences over time. Investigation
was recommended if:
1) more than half of the sampling dates had pairwise inter-organization differences that
were larger than within-organization precision; and.
2) an appropriate statistical test had a probability (P) < 0.01 that the differences were
due to chance alone, equivalent to 99% confidence that the observed difference was
real.
Parameters identified by the combination of these two criteria usually have different field
and/or laboratory methods at one or more of the laboratories involved.
Graphs of the split sample results show which differences were larger than the within-
organization precision. Within-organization precision for CSSP analyses is estimated by the
larger of: 1) the Method Detection Limit (MDL, Tables 1 and 2); or, 2) the standard deviation
of the three subsamples for each sample which estimates field precision. Graphs of the
medians for each sample date for each laboratory show this estimate as "precision bars." Any
pair of laboratory medians with non-overlapping precision bars have differences that are larger
than within-organization precision. Because the overlap was sometimes difficult to assess
graphically, it was also checked with a SAS program.
A non-parametric statistical test was used to assess inter-organization agreement using
the split sample data. This test assumes matched (positively correlated) samples, since this is
inherent in the split sample design. Below detection limit data were included if they were
lower than any other results, but no comparison was made if two or more laboratories had
below detection limit data, unless the data from one laboratory could be adjusted to remove the
bias. In the Potomac data this occurred with TSS, since MDHMH and DCLS have a lower
detection limit (1 mg/1) than DCRA/CRL (4 mg/1). The bias was avoided by adjusting TSS
data that were below 4 mg/1 up to 4 mg/1 before running the Friedman test. Statistical
significance was assumed when the significance level (P) < 0.01. Standard quality control
procedures use the P = 0.01 level as the "control" or action level for precision and accuracy
charts (e.g., Montgomery 1985).
The Friedman two-way non-parametric repeated measures analysis of variance
(ANOVA) with replication within blocks (Marascuilo and McSweeney 1977) was used to test
statistically for differences among laboratory results. The Friedman program used before
(Bergstrom 1990) did not allow for replicates, so means for each sample date were used in the
previous report. This change uses more information in the data and increases the power of the
test, or its ability to detect real differences. This test was run with a SAS computer program
written by CSC/CBPO staff using the formula in Marascuilo and McSweeney (1977), including
their formula for post hoc pairwise comparisons. The program was tested with the example
in Marascuilo and McSweeney (1977). Exact P values (for N < 10) are from Siegel (1956).
csc.sA2.4/93 Coordinated Split Sample Program Annual Report, 1990-1991 • Page 9
-------
m. RESULTS
A. WITfflN-ORGANIZATION PRECISION AND ACCURACY
Two estimates of within-organization precision were used in this analysis: the Method
Detection Limits (MDLs), listed in Tables 1 and 2; and the precision of field replicates, the
three subsamples split in the field and analyzed by the same laboratory. Percent recovery data
and results from Standard Reference Materials (SRMs) estimated within-organization accuracy.
1. Mainstem Component
Table 3 lists the mean Standard Deviation (SD) and Coefficient of Variation (CV) of
field replicates. These within-organization precision estimates varied among parameters in data
from the same organization, as well as among organizations for the same parameter, with CV
values ranging from 0.8% to 114%. Over all organizations, the parameters with the highest
CVs were TSS, PHEA, PO4F, and NH4. There was a negative CV for PN from MDHMH
because the mean PN was negative, calculated from TKNW - TKNF.
Percent recovery data from spiked samples (Table 4) show that most values were near
100%. Only 3 of 260 values fell outside the range of 80-120%. Results from SRMs (Table
5) from aU laboratories except MDHMH also had percent recovery values (SRM
results/expected x 100) near 100%. All but 41 of 229 recoveries were between 90 and 110%,
and all but 18 of 229 recoveries were between 80 and 120%.
2. Potomac Component
Table 6 lists the mean Standard Deviation (SD) and Coefficient of Variation (CV) of
the three field replicates as "Within-organization precision." For parameters affected by the
change to filtered samples at MDHMH, medians were calculated before and after the method
change. The change to filtered samples had little effect on within-organization precision in
MDHMH data, although variability went down in NO23 results. Precision also changed over
time for some parameters at the other laboratories, presumably due to random variations.
These precision estimates varied among parameters in data from the same organization,
as well as among organizations for the same parameter, with CV values ranging from 0% to
25%. Over all organizations, the parameters with the highest CVs were TKNW, TSS, and
NH4.
Percent recovery data (Table 7) and results from Standard Reference Materials (SRMs,
Table 8) estimated within-laboratory accuracy. Percent recovery data are limited since DCLS
did not submit them, but most values were close to 100%. Results from SRMs from
DCRA/CRL show good agreement with the expected results, and all but three results (for
PO4F and TOC) were within the 95% confidence intervals for the SRMs.
CSC.SA2.4/93 Coordinated Split Sample Program Annual Report, 1990-1991 • Page 10
-------
TABLE 1. Lower Detection Limits of Water Quality Parameters, Chesapeake Bay
Mainstem Monitoring Program, 1984-1992.
PARAMETER
MD/OEP-MDE
(CRL then CBL)
TN .240+ (6/84-2/85)
(Calc. .2009+0/85-5/15/85)
TKNW + .031+ (5/16/85-9/86)
NO23,or .2009+(10/86-9/87)
TDN+ .0305+(10/87-)
PON)
TDN .240+ (6/84-2/85)
(Calcu- .2009+(3/85-5/15/85)
lated, .03 (5/16/85-9/86)
then .02 (10/87-)
direct)
PON .40+ (6/84-5/15/85)
(Calcu- .001 (5/16/85-9/86)
lated, .40+ (10/86-9/87)
then .0105 (10/87-)
direct)
NH4 .020 (6/84-1/85)
.040 (2/85)
.003 (3/85-4/15/88)
.005 (4/16/88-7/88)
.003 (8/88-)
NO23 .040 (6/84-2/85)
.0009 (3/85-9/87)
.00015 (10/87-8/88)
.0002 (9/88-)
N02 .01 (6/84-2/85)
.0005 (3/85-9/87)
.00015 (10/87-8/88)
.0002 (9/88-)
VA/VWCB
(ODU)
VA/VWCB
(VIMS)
(6/84-3/15/86)
(3/16/86-4/15/86)
(4/16/86-4/30/86)
(5/86-9/87)
(10/87-8/90)
(9/90-10/90)
.11+
.105+
.11+
.105+
.10+
.075+
.061+ (11/90-)
.11+ (6/84-3/15/86)
.105+ (3/16/86-4/15/86)
.11+ (4/16/86-4/30/86)
.105+ (5/86-9/87)
.05 (10/87-8/90)
.025 (9/90-)
.20+ (6/84-9/87)
.05 (10/87-10/90)
.036 (11/90-)
.01 (6/84-5/15/85)
.0056 (5/16/85-)
.01 (6/84-3/15/86)
.005 (3/16/86-4/15/86)
.01 (4/16/86-4/30/86)
.005 (5/86-6/88)
.0025 (7/88-)
.001 (6/84- )
11*+ (6/84-9/87)
124*+(10/87-4/88)
071+ (5/88-5/89)
.069+ (6/89-6/90)
.045+ (7/90-6/91)
.081+ (7/91-1/92)
.045+ (2/92-)
.12*+(6/84-10/15/86)
.ll*+(10/16/86-9/87)
.1*(.05-.462)
(10/87-4/88)
.045 (5/88-5/89)
.040 (6/89-6/90)
.026 (7/90-6/91)
.075 (7/91-1/92)
.026 (2/92-)
.20*+ (6/84-9/87)
.024M.023- .026)
(10/87-4/88)
.026 (5/88-5/89)
.029 (6/89-6/90)
.019 (7/90-6/91)
.006 (7/91-1/92)
.019 (2/92-)
.01*(.002-.051)
(6/84-4/88)
.013 (5/88-5/89)
.006 (6/89-6/90)
.004 (7/90-6/91)
.002 (7/91-1/92)
.004 (2/92-)
.01*(.001-.025)
(6/84-4/88)
.0014 (5/88-5/89)
.0021 (6/89-6/90)
.0024 (7/90-1/92)
.0008 (2/92-)
.004*(.001-.007)
(6/84-4/88)
.0008 (5/88-5/89)
.0015 (6/89-6/90)
.0006 (7/90-6/91)
.0005 (7/91-1/92)
.0002 (2/92-)
* VIMS had variable detection limits during this period, within range shown.
+ Parameter calculated during this period; MDL shown is the sum of the detection limits of the components.
All concentrations on this page are in mg/1.
CSC.SA2.4/93
Coordinated Split Sample Program Annual Report, 1990-1991 • Page 11
-------
TABLE 1 (continued). Lower Detection Limits of Water Quality Parameters, Chesapeake
Bay Mainstem Monitoring Program, 1984-1992.
PARAMETER MD/OEP-MDE
(CRL then CBL)
TP
(Di-
rect,
then
calc.)
TDP
PHOSP
(Calc.,
then
direct)
PO4F
TOC
(Di-
rect,
then
calc. )
DOC
.012
.01
.005
.0063+
.012
.0022+
.012
.01
.005
.012
.001
.024+
.02+
.010+
.0013
.024+
.0012
.012
.007
.0016
.0006
1.0
.501+
1.0
.501+
.303 +
1.0
.50
.24
(6/84-1/85)
(2/85)
(3/85-5/15/85)
(5/16/85-9/86)
(10/86-9/87)
(10/87-)
(6/84-1/85)
(2/85)
(3/85-9/86)
(10/86-9/87)
(10/87-)
(6/84-1/85)
(2/85)
(3/85-5/15/85)
(5/16/85-9/86)
(10/86-9/87)
(10/87-)
(6/84)
(7/84-2/85)
(3/85-9/87)
(10/87-)
(6/84-5/15/85)
(5/16/85-9/86)
(10/86-9/87)
(10/87-8/88)
(9/88-)
(6/84-5/15/85)
(5/16/85-8/88)
(9/88-)
VA/VWCB
(ODU)
.01
.005
.012 +
.01
.005
.02+
.015+
.01+
.007
.01
.005
1.0
1.24 +
.74 +
.63 +
1.0
.50
(6/84-12/86)
(1/87-9/87)
(10/87-)
(6/84-11/86)
(12/86-)
(6/84-11/86)
(12/86)
(1/87-9/87)
(10/87-)
(6/84-11/86)
(12/86-)
(6/84-9/87)
(10/87-8/88)
(9/88-10/90)
(11/90-)
(6/84-8/88)
(9/88-)
VA/VWCB
(VIMS)
.01*(.009-.01)
(6/84-10/87)
.02*+ (11/87-4/88)
.007+ (5/88-5/89)
.008+ (6/89-6/90)
.005+ (7/90-5/92)
.0022+ (6/92-)
.01*(.009-.012)
(6/84-4/88)
.006 (5/88-5/89)
.005 (6/89-6/90)
.002 (7/90-)
.02*+ (6/84-10/87)
.01M.009-.01)
(11/87-4/88)
.001 (5/88-5/89)
.003 (6/89-5/92)
.0002 (6/92-)
.01M-009-.013)
(6/84-7/87)
.002M.001- .004)
(8/87-4/88)
.0005 (5/88-5/89)
.003 (6/89-6/90)
.0006 (7/90-6/91)
.0008 (7/91-1/92)
.0006 (2/92-)
1.0 (ODU**, 6/84-9/87)
1.581*+(10/87-4/88)
1.099+(5/88-8/88)
.599+ (9/88-5/89)
.604+ (6/89-6/90)
.457+ (7/90-6/91)
.234+ (7/91-1/92)
.597+ (2/92-)
1.0 (ODU**, 6/84-8/88)
.50 (ODU**, 9/88-6/90)
.36 (VIMS, 7/90-6/91)
.15 (VIMS, 7/91-1/92)
.50 (VIMS, 2/92-)
* VIMS had variable detection limits during this period, within range shown.
** ODU did TOC and DOC analyses for VIMS stations until 7/90.
+ Parameter calculated during this period; MDL shown is the sum of the detection limits of the components.
All concentrations on this page are in mg/1.
CSC.SA2.4/93
Coordinated Split Sample Program Annual Report, 1990-1991 • Page 12
-------
TABLE 1 (continued). Lower Detection Limits of Water Quality Parameters, Chesapeake
Bay Mainstem Monitoring Program, 1984-1992.
PARAMETER
POC
(Calc.,
then
direct)
SI
(as SI)
TSS
CHLA
(ug/1)
PHEA
(ug/1)
2.0+
.001
1.5+
.001
.063
.1
.012
.01
4.0
1.0
1.98
1.5
1+
0.2 +
1+
0.2 +
MD/OEP-MDE
(CRL then CBL)
(6/84-5/15/85)
(5/16/85-9/86)
(10/86-9/87)
(10/87-8/88)
(9/88-)
(6/84-2/85)
(3/85-3/87)
(4/87-)
(6/84-5/15/85)
(5/16/85-9/87)
(10/87-8/88)
(9/88-)
(6/84-5/15/85)
(MDHMH, 5/16/85- )
(6/84-5/15/85)
(MDHMH, 5/16/85 -)
VA/VWCB
VA/VWCB
(ODU)
2.0+
.24
.13
.028
.023
.0281
.0234
4.0
2.0
0.2 +
1.1+
0.2 +
0.8+
(6/84-9/87)
(10/87-10/90)
(11/90-)
(6/84-5/86)
(6/86-12/90)
(1/91-4/91)
(5/91-)
(6/84-8/88)
(9/88-)
(6/84-1/91)
(2/91-)
(6/84-1/91)
(2/91-)
2.0+
.581*
.099
• .104
.097
.084
.097
.056*
.009
.007
.013
.006
.013
4.0
5.0
1.4
2.0
1.0+
3.2 +
1.32 +
1.95+
0.95+
1.0+
3.2 +
1.91+
3.43 +
1.34 +
(VIMS)
(6/84-9/87)
(.581-. 581)
(10/87-4/88)
(5/88-5/89)
(6/89-6/90)
(7/90-6/91)
(7/91-1/92)
(2/92-)
(.009-. 1)
(6/84-4/88)
(5/88-5/89)
(6/89-6/90)
(7/90-6/91)
(7/91-1/92)
(2/92-)
(6/84-4/88)
(5/88-6/91)
(7/91-1/92)
(2/92-)
(6/84-5/89)
(6/89-6/90)
(7/90-6/91)
(7/91-1/92)
(2/92-)
(6/84-5/89)
(6/89-6/90)
(7/90-6/91)
(7/91-1/92)
(2/92-)
* VIMS had variable detection limits during this period, within range shown.
+ Parameter calculated during this period; MDL shown is the sum of the detection limits of the components.
All concentrations are in mg/1 except CHLA and PHEA.
Abbreviations: TN = Total Nitrogen, TON = Total Dissolved Nitrogen, PON = Paniculate (Organic) Nitrogen,
NH4 = ammonium, NO23 = Nitrite -I- Nitrate, NO2 = Nitrite, TP = Total Phosphorus, TOP = Total
Dissolved Phosphorus, PHOSP = Particulate Phosphorus, PO4F = Orthophosphate filtered, TOC = Total
Organic Carbon, DOC = Dissolved Organic Carbon, POC = Particulate (Organic) Carbon, SI = Silica (as SI),
TSS = Total Suspended Solids, CHLA = Chlorophyll a, PHEA = Phaeophytin.
Current limits at CBL & VIMS calculated from: 3 x standard deviation of 7 replicates of the lowest concentration
sample encountered. ODU calculates this also, but uses a higher MDL equivalent to 1-2% of full scale if the
calculated value is below 1-2% of full scale. MDHMH limits shown are calculated as 2% of full scale.
Calculation methods (used when a + follows the MDL) are: TN = TKNW -I- NO23 or TDN + PON, TON =
TKNF + NO23, PON = TKNW - TKNF, PHOSP = TP - TOP, POC = TOC - DOC, TOC = DOC + POC.
CSC.SA2.4/93
Coordinated Split Sample Program Annual Report, 1990-1991 • Page 13
-------
TABLE 2. Lower detection limits of water quality parameters, Chesapeake Bay Fall Line
and Tributary Water Quality Monitoring Programs, 1984-1992.
PARA- SRBC USGS OWML HRSD
METER (PADER) (USGS) (OWML) (HRSD)
VA/VWCB MD/OEP-MDE DC/DCRA
(DCLS) (MDHMH) (CRL)
TOC 1.0
DOC
NH4 .008?
.002?
NO23 . 04
NO2 . 004
TKNW/ .20/
TKNF 1.0
TP .02
TDP .02
P04F .005?
.002?
SI
TSS 1
CHLA ?
(ug/1)
BODS ?
O.I/
1.0
~
.002/
.01
.Ol/
.10
.001/
.01
.20
.001/
.01
.001/
.01
.001/
.01
.1
1
?
?
.05
.05
.01 .05
.01 .01
.01 .01
.10 .05
.05 (9/90)
.01 .05
.01 .05
.01 .05
.01
.04
.03 9/90
1 1
1.0
1
1.0
1.0
.05
.04 (2/88-)
.05
.04 (2/88-)
.01
.10
.Ol/
.10
.Ol/
.10
.01
(8/90-)
.1
5
1 (7/88- )
3.1
1
1.0 1
0.5 (4/90-)
1.0 1
0.8 (5/89-3/90)
0.5 (4/90-)
.02
.008 (6/86 -)
.02
.002
.10
.01
.01
.01
.004 (6/86-)
.1
1
1
0.5
.0
.0
04
04
01
20
01
01
007
.2
4
?
1
All concentrations are in mg/1 except CHLA. USGS and DCLS have low and high limits for some parameters
depending on what is requested. Where the second or third limit has a date after it, the first limit applied until
that date.
CSC.SA2.4/93
Coordinated Split Sample Program Annual Report, 1990-1991 • Page 14
-------
S
i
i
u
£
•PH
1
fl
I
O
|
i
u
o
i
1
e
o
1
i
o
c
S .
c
c
V
2
0
§
a
,*
g
o
0
c.
a
4J
C
c
CO
a.
^
i
'w
u
01
L.
Q
§
re
N
*C
n
(_
o
c
.c
°5
c
01
^
c-
^J
i
CO
a.
j-
a
S
.C
^
1
3
O
U.
Z
>
_,
8
§
o
X
_J
CO
u
z
2
u
S
S
o
to
z
(_
c
z
(_
8
z
u
c
z
u
8
r- >O O O O o
«-«--» >O CM p.
«* in »» CM o o
in ro CM «-«-«-
r— in in
S«- F- O ON o
o o in CM NO
o o o o o o
0 O 0 0 0 0
in in ON in fo ro
NQ «— N* CN- OO NO
co ro in «- ro «-
ro o
o •- ro o co ro
«- o «-N- in o
o o o o o «-
o o o o o o
r>- .0 ON ON o> ON
m o •* NO ro o
in N0 rj in o 00 *O
o o o o o o
o o o o o o
fs> >O O. ON ON ON
-* o oo «- ro o
^> o o ro ro ON
^ CM ro «—
in *o NO ^ *^ r^
-j- o CO o ^t o
o o o NO o >o
000000
o o o o o o
ro
•j- CM ro z
z o O a z z
z z z t- a. •-
-t -O O ON
ro -o IN- ro
h- o o -o
in ro «- •-
«- NO o» ro
rj ro «- «*
o o o o
o o o o
0 O 0 0
^O *"~ OO 0s
3S55
ro «»
ro CM CN- CM
§<— o «-
o o o
o o o o
**? .O ON ON
OO ON 60 ON
o NO in «—
«- «- ON -t
o CM o ro
o o o o
o o o o
0000
-* NO O- O-
«- 0 ^ NO
o in NO NO
ro «-
o o o o
o o o o
o o o o
-* -O ON ON
•O O NO O
CM ro ro ^
oo in NO ro
ro N— fN- in
O «- 0 0
o o o o
o o o o
N* NO ON ON
in -» o ro
ON IN. NO NO
eo NO co NO
o o o *-
§000
000
o o o o
CL
u- CO
•-* D. O
o o z a.
a. t- o. t-
00 ON 00
ro «- ro
ro r-. ON
g«g
O 0 0
1 1 1
co ON eo
NO «- CM
N* in **
CM IN- ro
IN- -t O
«— O \—
000
OO ON CO
ro «- N-
NO IN- in
IN. IN- NO
§00 vT
in CM
»- O CM
o o o
i i i
co o co
ro co NO
in •* ro
oo eo o
o ro co
in in N*
«— o *—
000
u o
o
ro
ro
ro
«-
ON
O>
ro
o
ro
ON
ro
in
(>
r-
o>
in
r-J
o
o
00
o
CO
CO
t—
eo NO
ro eo
oo NO
Sin
CO
«- 0
1 '
00 NO
ro o
»«
SNO
ro
in in
0 0
00 NO
NO O
N* O-
«- ro
ON
«- O
CO NO
h- N*
^^
CM ro
NO oo
in in
OO *£•
o o
( f
' '
<. f
—i UJ
X Z
o a.
ON
-*
^J
•~
ON
in
0
o
^
co
I
o
O
ro
"*
eo
i
o
ON
eo
ro
§
o
o
o
ON
ON
rJ
in
CM
o
0
ON
NO
-»'
in
o
0
o
t ,
to
i
S
Cfl
n
li
a
C/5
3
£
"o
«
1
5
•S
be
1
CO
•g
"e
S
t .
'g
§
•a
IS
*o
1
^B
*s
GO
1
"E,
$
^
•is
"^
1
CO
>,
^
«
i
J
3
"o
X
u
I
•S
1-4
U
Q
£
Z
II
^ *Pi
a, *-
c ^
.2 ^
N O
C co
3) §
II
*> °
*Q *w
5 "o
*-» p
3 e-
O g)
0 **
^ —
a §
g ^
^ en
ja •••
w a
IS o
? !S
| §
3 *
J.2
N
o *3
f 1
CO 1
<*- S3
2.1
•^ wT
B £
0 1*
z|
^ u
8*J
O
T^ IS
* pt
S-n
^j
O tc
ii s
II o
N™*1' trt
*4—
§ 2
11
* 0
O »3
*- (t_
C O
8
||
o 2
11 «"
^ a
- 0
c
O ye
.2 «
•S.8
«
"§
CJl
.5
"o.
S
•a
§
J
.1
C
£?
°
"3
S
&}
•S
!o
>n
a>
o\
O\
«-^
O
&
(Hi
1
c
I
I
u
"3,
VI
ex
1/3
1
rt
'S
§
u
-------
TABLE 4. Percent Recovery Data from Spiked Mainstem Component Split Samples.
Para-
meter
NH4
N02
N023
TON
TKNW
TKNF
P04F
TDP
PHOSP
Labor-
atory
CBL
DCLS
MDHMH
ODU
VIMS
CBL
DCLS
MDHMH
ODU
VIMS
CBL
DCLS
MDHMH
ODU
VIMS
CBL
ODU
VIMS
DCLS
MDHMH
MDHMH
CBL
DCLS
MDHMH
ODU
VIMS
CBL
DCLS
MDHMH
ODU
VIMS
CBL
ODU
VIMS
3/90
_
-
88
109
92
-
-
100
-
105
.
-
110
-
95
99
-
101
-
100
-
-
-
120
-
96
100
-
-
-
100
100
-
92
6/90
_
100
100 .
-
97
-
100
98
104
107
_
100
106
99
101
.
103
98
93
106
100
-
100
84
100
100
.
100
-
101
100
_
103
100
9/90
97
105
• 116
-
89
103
100
-
85
-
100
103
110
101
109
101
104
93
93
100
108
93
90
88
98
100
100
91
104
102
105
101
87
110
Percent
12/90
89
-
84
100
104
91
-
96
100
98
95
-
106
98
100
100
125
90
_
80
86
94
-
92
100
99
99
-
86
100
96
.
82
102
Recoverv
4/91
98
-
124
101
96
98
-
104
103
100
100
-
-
95
105
98
110
93
_
-
-
100
-
-
95
97
103
-
-
101
91
100
100
102
6/91
97
100
112
-
91
96
103
-
-
98
96
-
.
-
94
98
106
92
100
86
84
96
100
84
-
103
94
100
84
102
95
99
91
98
9/91
95
105
96
102
91
100
100
96
100
95
97
-
_
-
100
99
102
121
107
102
-
99
100
-
100
102
93
100
-
103
111
85
90
100
12/91
100
81
-
96
101
98
-
-
100
100
95
-
.
103
113
104
102
99
100
96
112
102
100
-
94
103
100
100
107
-
108
_
-
102
CSC.SA2.4/93
Coordinated Split Sample Program Annual Report, 1990-1991 • Page 16
-------
TABLE 4 continued. Percent Recovery Data from Spiked Mainstem Component Split
Samples.
Para-
meter
TP
DOC
TOC
SI
Labor-
atory
DCLS
MDHMH
CBL
MDHMH
ODU
VIMS
DCLS
MDHMH
CBL
DCLS
MDHMH
ODU
VIMS
3/90
.
96
.
-
-
102
-
-
83
-
-
-
98
6/90
100
-
.
112
91
94
.
113
100
100
-
98
103
9/90
91
104
108
106
115
97
_
106
98
102
140
106
90
Percent
12/90
.
80
101
-
100
99
.
-
100
-
100
96
88
Recovery
4/91
.
-
100
-
90
105
_
-
93
-
117
-
97
6/91
100
111
109
-
105
106
_
-
94
-
88
103
96
9/91
100
-
104
-
98
98
.
-
92
100
102
99
94
12/91
100
109
„
107
-
98
.
108
90
100
-
97
101
Note: Percent recovery data are only possible for directly measured parameters. There is no way to spike
parameters analyzed directly from filters (PC, PN, CHLA or TSS). Percent recovery was not calculated when
the sample concentration was censored at the detection limit. TDP = Total Dissolved Phosphorus, TP = Total
Phosphorus, SI = Silica, NO2 = nitrite, NO23 = nitrite -I- nitrate, NH4 = ammonium, PHOSP = Paniculate
Phosphorus, DOC = dissolved Organic Carbon, PO4F = Orthophosphate filtered, TON = Total Dissolved
Nitrogen.
CSC.SA2.4/93 Coordinated Split Sample Program Annual Report, 1990-1991 • Page 17
-------
TABLE 5. Standard Reference Material Percent Recovery Results from Mainstem
Component Laboratories.
Para- Date
meter
NH4
NO23
TON
TKNW
PO4F
3/90
6/90
9/90
12/90
4/91
6/91
9/91
12/91
3/90
6/90
9/90
12/90
4/91
6/91
9/91
12/91
3/90
6/90
9/90
12/90
4/91
6/91
9/91
12/91
3/90
6/90
9/90
12/90
4/91
6/91
9/91
12/91
3/90
6/90
9/90
12/90
4/91
6/91
9/91
12/91
CBL
Dist.1 (EPA)2
104
97
95
99
101
99
103
102
105
105
104
103
99
96
106
116
108
104
102
101
109
102
(0
(0
(0
-
(0
(0
(0
.
(0
(0
(0
(0
(0
(0
-
.
-
(0
(0
-
(0
(0
-
_4
-
-
-
-
-
-
-
_
(0
(0
(0
(0
(0
(0
-
.2)
.2)
.2)
.2)
.2)
.2)
.2)
.2)
.2)
.2)
.2)
.2)
.5)
.5)
.5)
.5)
-05)
.05)
.05)
-05)
.05)
.05)
DCLS
Dist
95
100
103
100
100
97
100
100
100
100
100
96
100
100
100
100
100
100
95
95
. (EPA)
(1
(2
-
-
(1
(1
(1
.
(1
(2
-
-
(1
(1
(1
_4
-
-
-
-
-
-
-
_
(1
(5
-
-
(2
(2
(2
_
(0
(0
-
-
(0
(0
(0
.9)
.0)
.0)
.0)
.0)
.42)
.0)
.0)
.0)
.0)
-6)
-0)
.5)
-5)
-5)
.35)
.39)
.25)
.2)
.2)
ODU
Dist./Est
102/ -
72/ 68
110/107
97/ 99
95/ 95
96/ 86
96/ 94
96/ 94
101/ -
101/ 94
101/ 98
101/ 96
100/ 94
100/104
100/ 97
107/ 97
100/ -
101/ 99
103/104
99/ 86
104/100
109/ 97
94/ 99
97/105
97/100
97/ 95
105/103
103/103
105/ 95
110/100
100/102
.3(EPA)
(0.04)
(0.056)
(0.04)
(0.04)
(0.04)
(0.04)
(0.04)
(0.04)
(0.04)
(0.04)
(0.04)
(0.04)
(0.04)
(0.04)
(0.04)
(0.04)
(0.5)
(0.5)
(0.5)
(0.5)
(0.25)
(0.25)
(0.25)
(0.5)
_4
-
-
-
-
-
-
-
.
(0.039)
(0.039)
(0.039)
(0.039)
(0.039)
(0.039)
(0.039)
VIMS
Dist./Est
79/ 88
75/100
73/ 85
85/107
99/107
86/ -
81/ 77
90/ 77
102/ 94
107/103
103/ 99
112/112
106/ 99
93/ -
100/105
96/105
_
100/107
71/ 66
103/ 89
113/ 89
119/122
143/ 62
114/ 62
_4
-
-
-
-
-
-
-
100/ 96
100/ 96
123/118
95/100
100/100
103/ -
105/ 87
87/ 87
. (EPA)
(2.0)
(2.0)
(0.2)
(0.2)
(0.2)
(0.2)
(0.2)
(0.2)
(2.0)
(2.0)
(0.2)
(0.2)
(0.2)
(0.2)
(0.2)
(0.2)
(5.0)
(0.25)
(0.25)
(0.25)
(0.25)
(0.25)
(0.25)
(0.5)
(0.5)
(0.039)
(0.039)
(0.039)
(0.039)
(0.039)
(0.039)
CSC.SA2.4/93
Coordinated Split Sample Program Annual Report, 1990-1991 • Page 18
-------
TABLE 5 (continued). Standard Reference Material Percent Recovery Results from
Mainstem Component Laboratories.
Para- Date CBL
meter Dist.'(EPA)2
TOP
3/90
6/90
9/90 102 (0.15)
12/90 101 (0.15)
4/91
6/91 99 (0.15)
9/91 107 (0.15)
12/91
DCLS
ODU
Dist. (EPA) Dist./Est
.
96 (0.5) 105/104
87 (1.5) 103/
105/
99/
100 (0.75) 101/
103/
100/
99
99
98
97
95
95
3 (EPA)
.
(0.
(0.
(0.
(0.
(0.
(0.
(0.
075)
15)
15)
15)
15)
15)
15)
PHOSP 3/90 -s -4
DOC
TSS
6/90
9/90
12/90
4/91
6/91
9/91
12/91
3/90 -3
6/90
9/90
12/90
4/91
6/91
9/91
12/91
3/90 -5
6/90
9/90
9/90
12/90
4/91
6/91
9/91
12/91
95/
96/
91/
96/
91/
96/
96/
-" 112/
98/
111/
103/
108/
105/
92/
97/
.5
106/
103/
90/
-
-
-
-
-
-
-
-
-
-
-
-
(0.
(0.
(0.
(0.
(0.
(0.
(0.
78)
98)
78)
78)
78)
78)
78)
VIMS
Dist./Est
107/ 99
112/ 87
85/ 72
100/ 72
104/120
95/ 91
97/ 91
_5
-
-
-
-
-
.
-
(EPA)
(1.
(0.
(0.
(0.
(0.
(0.
(0.
5)
075)
075)
075)
075)
075)
075)
- (4.10)
-
84
73
72
79
81
82
-
(6.
(9.
(9.
(3.
(3.
(3.
(3.
_
(31
12)
18)
18)
01)
06)
06)
67)
-2)
105/107
95/112 (4.
98/109
102/106
109/102
88/104
100/ 99
_5
-
(4.
1/2
(2.
(4.
(4.
(2.
(3.
10)
.05)
05)
20)
20)
05)
08)
- (278)
-
(36
-
-
-
-
-
.1)
-
-
-
-
-
-
1 SRM diluted before analysis in distilled/deionized water matrix.
2 SRM diluted before analysis in estuarine (saline) water matrix.
3 Expected (EPA) value for SRM, in mg/1.
4 Analysis not performed by this laboratory.
5 Analysis performed by this laboratory, but SRM results not reported.
CSC.SA2.4/93
Coordinated Split Sample Program Annual Report, 1990-1991 • Page 19
-------
TABLE 6. Within-organization and inter-organization precision estimates, Potomac
Component, May 1989- December 1991.
Para-
meter
NH4
NH4
N02
N02
N023
N023
TKNW
TN
P04F
P04F
TP
TOP
TOC
TSS
SI
SI
BODS
Period1
Within-organization
DCRA/CRL
U
F
U
f
U
F
A
A
U
F
F
F
A
A
U
F
A
SD
0.0068
0.0184
0.0028
0.00058
0.024
0.035
0.056
0.061
0.0014
0.0018
0.0075
0.0042
0.61
2.0
0.013
0.045
0.105
CV
9.8
23.0
17.2
4.4
1.9
2.8
10.7
3.4
4.1
4.3
12.1
8.6
12.1
13.4
0.81
1.7
7.0
N
5
5
5
5
6
5
10
10
4
5
5
4
11
9
5
5
3
precision
DCLS
SD
0.0114
0.0102
0.0
0.0012
0.014
0.019
0.023
0.038
0.0087
0.0023
0.0051
0.0039
0.14
1.9
0.055
0.0097
0.236
.CV
16.5
14.6
0.0
4.9
0.98
1.5
5.6
2.0
16.7
6.4
11.0
6.9
4.6
15.0
5.6
0.76
15.7
N
in in
in in
6
5
10
10
4
5
5
4
11
9
5
5
3
MDHMH
SD
0.0114
0.0098
0.00088
0.00058
0.038
0.013
0.115
0.127
0.0030
0.0013
0.0110
0.0077
0.43
1.3
0.029
0.0058
0.283
CV
16.5
17.1
6.1
6.4
2.3
0.75
24.9
7.8
5.7
4.7
14.3
15.4
9.1
23.5
2.6
0.24
19.8
I nter-organi zat i on
precision
Three labs
N
5
5
5
5
6
5
10
10
4
5
5
4
11
9
5
5
3
SD
0.0076
0.0111
0.0030
0.0017
0.12
0.052
0.097
0.116
0.0087
0.0040
0.0064
0.0091
1.38
3.6
0.11
0.17
0.43
CV
8.5
19.4
21.1
15.0
8.0
4.0
19.6
6.3
23.3
13.0
10.2
28.6
30.8
38.6
10.1
7.4
34.5
N
5
5
5
5
6
5
10
10
4
5
5
4
11
9
5
5
3
' U = MDHMH unfiltered (before Dec. 1990), F = MDHMH filtered (starting Dec. 1990), A = all
available data. TP used data from F period only, because DCLS lowered their MDL in December 1990.
Notes: March 1990 and DOC data were not used due to missing data.
SD= standard deviation, CV = coefficient of variation (= SD/Mean*100), N = number of sampling dates over
which mean precision was calculated.
Within-organization precision is based on the precision of three subsamples; inter-organization precision is based
on the precision of medians of the subsamples for each sampling date.
CSC.SA2.4/93
Coordinated Split Sample Program Annual Report, 1990-1991 • Page 20
-------
TABLE 7. Percent Recovery Data from Spiked Potomac Component Samples, 1989-91.
Para-
meter
NH4
NO2
NO23
TKNW
PCM
TOP
TP
DOC
TOC
SI
Para-
meter
NH4
NO2
NO23
TKNW
PO4F
TOP
TP
DOC
TOC
SI
Laboratory
DCRA/CRL
MDHMH
DCRA/CRL
MDHMH
DCRA/CRL
MDHMH
DCRA/CRL
MDHMH
DCRA/CRL
MDHMH
DCRA/CRL
DCRA/CRL
MDHMH
DCRA/CRL
DCRA/CRL
MDHMH
DCRA/CRL
MDHMH
Laboratory
DCRA/CRL
MDHMH
DCRA/CRL
MDHMH
DCRA/CRL
MDHMH
DCRA/CRL
MDHMH
DCRA/CRL
MDHMH
DCRA/CRL
MDHMH
DCRA/CRL
MDHMH
DCRA/CRL
MDHMH
DCRA/CRL
MDHMH
DCRA/CRL
MDHMH
Percent Recoverv
5/1/89
_
140
110
98
105
-
-
102
100
144
-
-
116
82
-
103
136
~
6/12/89
107
108
108
100
85
95
117
98
96
106
-
115
104
-
91
99
111
~
9/11/89
93
102
116
100
106
108
67
100
96
112
95
83
76
86
92
92
75
~
1/8/90
102
110
102
100
85
102
109
98
-
136
100
100
100
108
115
93
_
95
3/5/90
86
76
96
-
95
-
98
104
98
-
94
102
100
105
104
103
138
"
6/11/90
101
108
110
104
95
100
91
112
113
160
106
99
108
96
100
108
87
100
9/11/90
104
72
-
96
120
104
73
100
_
-
113
86
100
94
98
106
78
85
Percent Recoverv
12/10/90
115
107
110
98
90
96
107
118
97
96
93
88
93
100
85
84
100
96
112
100
4/22/91
104
-
110
-
121
-
107
98
97
-
106
-
89
97
99
-
100
-
104
~
6/3/91
101
100
106
100
96
104
124
92
101
116
108
78
108
94
.
99
-
91
-
-
9/3/91
104
-
108
98
116
-
98
82
110
100
109
-
95
76
92
-
95
99
100
105
12/2/91
110
100
106
-
98
-
-
106
103
108
-
99
-
113
_
-
-
-
116
~
Note: Percent recovery data are not possible for TSS analysis. DCLS did not report any percent recovery data.
TOP = Total Dissolved Phosphorus, TP = Total Phosphorus, SI = Silica, NO2 = Nitrite, NO23 = Nitrite +
Nitrate, TKNW = Total Kjeldahl Nitrogen Whole, NH4 = Ammonium, DOC = Dissolved Organic Carbon,
TOC = Total Organic Carbon, PO4/PO4F = Orthophosphate.
CSC.SA2.4/93
Coordinated Split Sample Program Annual Report, 1990-1991 • Page 21
-------
TABLE 8. Standard Reference Material Results, Potomac Component, 1990-1991.
Para-
meter
NH4
N023
TKNW
P04F
TOP
TP
Date
1/90
3/90
6/90
9/90
12/90
4/91
6/91
9/91
12/91
1/90
3/90
6/90
9/90
12/90
4/91
6/91
9/91
12/91
1/90
3/90
6/90
9/90
12/90
4/91
9/91
12/91
3/90
6/90
12/90
4/91
6/91
9/91
12/91
1/90
3/90
9/90
6/90
9/90
12/90
4/91
6/91
9/91
Laboratory
DCRA/CRL
DCRA/CRL
DCRA/CRL
DCRA/CRL
DCRA/CRL
DCRA/CRL
DCRA/CRL
DCRA/CRL
DCRA/CRL
DCRA/CRL
DCRA/CRL
DCRA/CRL
DCRA/CRL
DCRA/CRL
DCRA/CRL
DCRA/CRL
• DCRA/CRL
DCRA/CRL
DCRA/CRL
DCRA/CRL
DCRA/CRL
DCRA/CRL
DCRA/CRL
DCRA/CRL
DCRA/CRL
DCRA/CRL
DCRA/CRL
DCRA/CRL
DCRA/CRL
DCRA/CRL
DCRA/CRL
DCRA/CRL
DCRA/CRL
DCRA/CRL
DCRA/CRL
DCRA/CRL
DCRA/CRL
DCRA/CRL
DCRA/CRL
DCRA/CRL
DCRA/CRL
DCRA/CRL
Results
Expected
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
2
4
4
4
4
0
0
0
0
0
0
0
0
0
0
0
1
1
1
1
1
0
0
0
0
.90
.90
.90
.90
.90
.90
.90
.90
.90
.43
.43
.43
.43
.43
.43
.43
.43
.00
.78
.78
.78
.78
.40
.40
.40
.40
.35
.35
.35
.35
.35
.35
.500
.00
.03
.03
.03
.03
.120
.130
.130
.120
(ma/1)
Distilled
1
1
1
1
1
1
1
1
2
1
1
1
1
1
1
1
1
1
5
4
4
4
0
0
0
0
0
0
0
0
0
0
0
1
0
1
0
1
0
0
0
0
.84
.93
.92
.78
.85
.84
.73
.80
.02
.35
.29
.50
.51
.46
.35
.35
.40
.98
.05
.90
.72
.68
.411
.330
.460
.340
.401**
.347
.333
.342
.337
.357
.514
.00
.924
.06
.896
.06
.127
.120
.120
.130
95%
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
3
3
3
3
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
.68
.68
.68
.68
.68
.68
.68
.68
.68
.28
.28
.28
.28
.28
.28
.28
.28
.68
.70
.70
.70
.70
.089
.089
.089
.089
.33
.33
.33
.33
.33
.33
.43
.89
.89
.89
.89
.89
.087
.096
.096
.087
-
-
-
-
-
-
.
-
_
-
-
-
-
-
-
-
-
.
-
-
-
-
-
-
-
_
-
-
-
-
-
-
_
-
-
-
-
-
-
-
"
CI*
2
2
2
2
2
2
2
2
2
1
1
1
1
1
1
1
1
2
5
5
5
5
0
0
0
0
0
0
0
0
0
0
0
1
1
1
1
1
0
0
0
0
.12
.12
.12
.12
.12
.12
.12
.12
.12
.56
.56
.56
.56
.56
.56
.56
.56
.32
.77
.77
.77
.77
.836
.836
.836
.836
.37
.37
.37
.37
.37
.37
.57
.21
.21
.21
.21
.21
.169
.180
.180
.169
CSC.SA2.4/93 Coordinated Split Sample Program Annual Report, 1990-1991 • Page 22
-------
TABLE 8 (continued). Standard Reference Material Results, Potomac Component, 1990-
1991.
Para-
meter
TOC
Date
1/90
3/90
6/90
9/90
12/90
4/91
6/91
9/91
12/91
Laboratory
DCRA/CRL
DCRA/CRL
DCRA/CRL
DCRA/CRL
DCRA/CRL
DCRA/CRL
DCRA/CRL
DCRA/CRL
DCRA/CRL
Results
Exoected
6.1
6.1
17.2
10.0
10.0
6.12
6.1
6.1
12.2
(ma /I)
Distilled
7.5
6.9
17.7
10.2
11.9**
5.91
6.0
7.5
16.5**
95%
3.50 -
3.50 -
15.4 -
9.0 -
9.0 -
5.31 -
3.50 -
3.50 -
10.83 -
CI*
9.32
9.32
19.8
11.8
11.8
7.60
9.32
9.32
14.13
* Provided by EPA Environmental Monitoring and Support Laboratory (EMSL) Cincinnati.
** Measured value was outside the 95% confidence interval.
Note: DCLS and MDHMH have not reported any SRM results. NH4 = ammonium, NO23 = Nitrite + Nitrate,
TKNW = Total Kjeldahl Nitrogen Whole, PO4F = Orthophosphate filtered, TOP = Total Dissolved
Phosphorus, TP = Total Phosphorus, TOC = Total Organic Carbon.
CSC.SA2.4/93
Coordinated Split Sample Program Annual Report, 1990-1991 • Page 23
-------
3. Virginia Component
Means of within-organization precision, based on field replicates, are listed in Table 9.
The mean CV values were generally low (less than 16%) except for PN at HRSD, which had
CV= 26%.
Percent recovery data from laboratory spikes (Table 10) and results from Standard
Reference Materials (SRMs, Table 11) estimated within-laboratory accuracy. VIMS and ODU
also diluted SRMs in the sample matrix, although the salinity at TF5.5 is usually zero. Percent
recovery values (Table 10) were almost all (96%, 219/227 values) within the range 80-120%
recovery. Results from SRMs (Table 11) generally show good agreement with the expected
results; percent recoveries were within the 80-120% range for 90% of the reported values
(139/155).
4. Fall Line Component
Within-organization precision was estimated from the standard deviation and coefficient
of variation of field replicates (Table 12). Within-organization coefficients of variation ranged
from 0% for several parameters to 52% for total dissolved phosphorus (TDP) results from
USGS.
Percent recovery data and results from Standard Reference Materials (SRMs) are used
to estimate within-organization accuracy in the CSSP. Percent recovery data (Table 13) were
only reported by PADER, MDHMH, and OWML; most values were close to 100%. None
of the laboratories reported SRM results.
csc.sA2.4/93 Coordinated Split Sample Program Annual Report, 1990-1991 • Page 24
-------
O
CO
2*
S
g -
cfl
0) (A >
«. JQ CJ
Q a
01
"v t- o
a .c co
M »—
a
L.
0
fc.2"
v a
c —
t.
C 3
SO O
u. CO
c.
c
c z
o
M
U C
0> Z>
aS
•— Crt
a
fsl
c
1C Z
D
0
coo
— CO
•M X
5
C
O
ru
o
o
ro
o
^_
o
r~i
O
O
ro
oo
,_
o
o
0
o
ro
o
0
0
o
ro
in
O
o
o
o
ru
o
z
oo
.
.
in
ro
o
o
eo
o
ro
g
0
o
o
00
CM
,_
ru
ru
o
o
0
00
«*
ru
00
o
0
o
00
in
X}
o
o
o
§
Z
**
£
ro
0
-o
in
CO
in
o
o
•*
-T
ro
•O
CM
o
o
^*
•0
eo
in
o
o
1
'
0
H-
ro
ro
ro
o
ru
•o
in
o
o
i
1
in
•
,-
CM
g
0
^
ro
ru
o
o
o
in
-O
ro
ro
0
0
ru
^
c>
ru
0
o
in
ro
ru
ro
o
CM
o
o
z
-
•
1
oo
o
ro
in
o
0
CM
N.
O
O
ru
0
o
o
r^
ro
^.
R
o
o
0
0-
CD
«—
00
R
o
o
o
r-
o
o
o
o
3
-
'
CM
O
O
o
0
in
ro
t~\
o
o
r^
CM
*o
in
CM
o
o
o
»
o-
§
0
0
a
in
•o
ru
ru
o
o
r-
ro
^
§
o
o
a.
CO
O
X
r- in in oo
oo ro
ru in
ro
r»- o
oo oo
O «-
i • i •
«- o
ru v»
jo ' r-' '
N.
*o *o
o ro
o in «-
oo ru -to
ro
K eo
in ro oo
CM Is- CO O
O «- CM «-
o o «- o
N. in in eo
1^ ^- ro ~t
ro ro •* »-
o o in
o ru oo o
o o «- o
in
NJ- •<-
^- •&
«—
o
ru o
o o
o «-
r^ in in eo
^ in ^3 o^
ro o' in ^
ro ro
O- o »-
o >o ro ro
o o «- o
U CO
a. o co «
*~ *~~
M
S CO
S "o
P P
If
^ s
n
o ep
S 9
5 S
O.S
|s
* 1*
<*• s
0 g
|i
& o
g g
§ II
•*• en ^
§— - -S
en Q*
^? ctt
QU CO
»u (—
^ ^ o
u ° S
^^ C i-
o .2 J2
_O _oi ^™
'•S "o SJ
'C ^"S,
^H «^
** " Zt
g ™
"o "S -2
s a .§
4> & i—
o .52 *0
^ *^
II g u
> is ^
^ v O
a o. 2
•2 c J
.^ c "§
^ KS S
-------
TABLE 10. Percent Recovery Data from Spiked Virginia Component Samples, 1990-1991.
Para-
meter
NH4
N02
NO23
TKNW
TKNF
TON
PO4F
TOP
PHOSP
TP
DOC
SI
Laborato
DCLS
HRSD
ODU
VIMS
DCLS
HRSD
ODU
VIMS
DCLS
HRSD
ODU
VIMS
DCLS
HRSD
HRSD
ODU
VIMS
DCLS
HRSD
ODU
VIMS
DCLS
HRSD
ODU
VIMS
ODU
VIMS
DCLS
HRSD
ODU
VIMS
DCLS
ODU
VIMS
rv
2/90
97
99
98
96
100
98
96
105
92
98
95
-
92
100
101
102
101
100
100
95
98
100
98
102
97
101
96
91
94
98
-
98
100
96
Percent Recoverv from Laboratory Soikes
5/90
104
-
101
-
100
-
-
-
107
-
99
-
100
-
-
103
81
100
-
99
-
93
-
101
105
97
-
100
-
96
103
102
103
-
9/90
100
100
92
103
100
100
99
98
92
98
102
104
93
60
60
100
97
100
100
98
104
100
100
105
105
97
95
100
120
104
97
100
101
94
1/91
80
105
94
-
105
99
102
-
90
101
101
-
104
100
102
90
-
110
95
102
-
58
101
100
95
100
-
93
99
105
-
101
90
5/91
107
93
99
95
100
95
104
98
122
103
100
98
93
99
100
102
94
100
100
97
97
100
92
102
111
96
113
92
99
94
94
103
101
92
6/91
94
90
99
88
100
90
102
95
117
90
98
95
100
52
121
100
100
100
100
-
106
100
97
103
-
97
94
100
92
96
94
100
101
90
9/91
96
89
-
91
100
105
-
91
105
100
-
103
100
99
101
-
112
100
93
-
95
100
102
-
74
-
84
100
86
-
100
100
-
95
12/91
85
85
-
77
100
105
-
102
108
99
-
109
100
107
111
-
115
92
98
-
105
92
95
-
118
-
97
92
86
-
104
102
-
94
Note: Percent recovery is not possible for TSS, PC or PN. TOP = Total Dissolved Phosphorus, TP = Total
Phosphorus, SI = Silica, NO2 = Nitrite, NO23 = Nitrite + Nitrate, TKNW = Total Kjeldahl Nitrogen Whole,
TON = Total Dissolved Nitrogen, NH4 = Ammonium, DOC = Dissolved Organic Carbon, PO4F =
Orthophosphate. When percent recovery was done on more than one subsample, the first value for that date is
listed.
csc.sA2.4/93 Coordinated Split Sample Program Annual Report, 1990-1991 • Page 26
-------
TABLE 11. Standard Reference Material Percent Recovery Results from Virginia
Component laboratories, 1990-1991.
Para- Date
meter
NH4
N023
TON
TKNW
PO4F
2/90
5/90
9/90
1/91
5/91
6/91
9/91
12/91
2/90
5/90
9/90
1/91
5/91
6/91
9/91
12/91
2/90
5/90
9/90
1/91
5/91
6/91
9/91
12/91
2/90
5/90
9/90
1/91
5/91
6/91
9/91
12/91
2/90
5/90
9/90
1/91
5/91
6/91
9/91
12/91
DCLS
Dist.'(EPA)2
96
95
100
93
98
98
98
100
101
90
96
92
102
94
94
96
104
100
104
96
100
100
100
88
92
100
100
(0.28)
(1.9)
(2.0)
-
(11.0)
(1.0)
(1.0)
(1.0)
(0.14)
(1.43)
(2.0)
-
(1.3)
(1.0)
(1.0)
-
_4
-
-
-
-
-
-
-
(0.32)
(0.32)
(5.0)
-
(2.5)
(1.0)
(2.5)
(2.5)
(0.05)
(0.035)
(0.39)
-
(0.48)
(0.25)
(0.2)
(0.2)
HRSD
Dist
92
85
100
96
102
103
108
103
100
97
92
102
98
91
100
105
112
56
111
98
91
112
110
103
113
103
105
102
. (EPA)
(2.
(0.
(0.
(0.
(0.
(0.
(0.
-
(5.
(0.
(0.
(0.
(0.
(0.
(0.
-
_4
-
-
-
-
-
-
-
(5.
(1.
(0.
(0.
(1-
(1-
(0.
-
(1.
(0.
(0.
(0.
(0.
(0.
(0.
35)
48)
2)
5)
5)
89)
201)
34)
38)
33)
5)
5)
55)
268)
0)
02)
76)
5)
0)
0)
99)
1)
46)
039)
039)
039)
058)
055)
ODU
Dist./Est
98/ 95
95/ 90
109/105
96/ 98
101/ -
101/103
96/ 94
104/ 96
102/102
101/103
111/107
100/101
103/ 94
104/ 99
99/ -
98/104
98/103
84/106
103/102
98/100
112/102
109/ 99
94/ 99
97/105
103/100
97/100
97/ 96
103/ 95
92/ 98
100/ 98
97/ 97
.3(EPA)
(0.
(0.
04)
04)
(0.04)
(0.
(0.
(0.
(0.
(0.
(0.
(0.
(0.
(0.
(0.
(0.
(0.
(0.
(0.
(0.
(0.
(0.
(0.
(0.
(0.
(0.
_4
-
-
-
-
-
-
-
(0.
(0.
(0.
(0.
(0.
-
(0.
(0.
04)
04)
04)
04)
04)
04)
04)
04)
04)
04)
04)
04)
04)
5)
25)
5)
5)
25)
25)
25)
5)
04)
04)
039)
039)
039)
039)
039)
Dist./Est
79/ 88
79/ 91
83/ 76
81/ 83
85/ 90
86/ 83
81/ 77
81/ 90
102/ 94
100/100
103/ 99
97/ 75
92/105
93/ -
100/105
109/ 96
_
-
133/104
123/108
79/100
119/122
143/ 62
84/114
_4
-
-
-
-
-
-
-
100/ 96
100/ 92
123/ 97
100/ 87
121/ 90
103/ -
105/ 87
105/ 87
VIMS
. (EPA)
(2.0)
(2.0)
(0.2)
(0.2)
(0.2)
(0.2)
(0.2)
(0.2)
(2.0)
(2.0)
(0.2)
(0.2)
(0.2)
(0.2)
(0.2)
(0.2)
(0.25)
(0.25)
(0.25)
(0.25)
(0.25)
(0.25)
(0.5)
(0.5)
(0.039)
(0.039)
(0.039)
(0.039)
(0.039)
(0.039)
CSC.SA2.4/93
Coordinated Split Sample Program Annual Report, 1990-1991 • Page 27
-------
TABLE 11 (continued). Standard Reference Material Percent Recovery Results from
Virginia Component Laboratories, 1990-1991.
Para- Date
meter
TOP
PHOSP
TP
DOC
TSS
2/90
5/90
9/90
1/91
5/91
6/91
9/91
12/91
2/90
9/90
1/91
5/91
6/91
9/91
12/91
2/90
5/90
9/90
5/91
6/91
9/91
2/90
5/90
9/90
1/91
5/91
6/91
9/91
12/91
9/90
5/91
12/91
DCLS
Dist.'(EPA)2
110 (0.1)
100 (0.1)
100 (1.5)
-
96 (0.75)
100 (0.3)
-
HRSD
Dist. (EPA)
_5
ODU
Dist./Est.
105/
98
- 105/104
-
-
-
-
-
102/
105/
100/
100/
97/
99
99
96
96
98
105/102
_4
-
-
-
-
-
-
110 (0.1)
100 (0.1)
100 (1.5)
101 (0.75)
100 (0.3)
-
_4
-
-
-
-
-
-
-
.5
-
-
_4
-
-
-
-
-
-
87 (1.8)
96 (0.98)
96 (0.5)
-
93 (0.3)
88 (0.765)
_4
-
-
-
-
-
-
-
_5
-
-
87/
96/
95/
99/
91/
96/
96/
103/
96/
104/
91/
92/
98/
99/
97/
94/
103/
.
-
-
-
-
-
-
86
81
88
89
74
91
78
87
97
.3(EPA)
(0
(0.
(0.
(0.
(0.
(0.
(0.
(0.
(0.
(0.
(0.
(0.
(0.
(0.
(0.
_4
-
-
-
-
-
(4.
(6.
(9.
(9.
(3.
(3.
(3.
(0.
(31
.15)
075)
15)
15)
15)
15)
15)
15)
39)
98)
98)
78)
78)
78)
78)
5)
12)
79)
18)
06)
06)
06)
19)
.5)
VIMS
Dist./Est. (EPA)
104/ 95
96/ 96
101/101
104/120
95/ 91
105/ 97
105/107
100/111
99/109
101/ 96
101/ 96
109/102
100/ 99
.
-
(0
(0
(0
(0
(0
(0
_5
-
-
-
-
-
-
_4
.
-
-
-
-
_
(4
(8
(2
(4
(4
(4
(3
_5
.075)
.075)
.075)
.075)
.075)
.075)
.1)
.2)
.05)
.1)
.1)
.2)
.08)
- (113.0)
-
-
1 SRM diluted before analysis in distilled/deionized water matrix.
2 SRM diluted before analysis in estuarine (saline) water matrix.
3 Expected (EPA) value for SRM, in mg/1. When the EPA values for the distilled and estuarine dilutions were
different, the value for the distilled dilution is shown.
4 Analysis not performed by this laboratory.
5 Analysis performed by this laboratory, but SRM results not reported.
CSC.SA2.4/93
Coordinated Split Sample Program Annual Report, 1990-1991 • Page 28
-------
¥«^
Os
Os
rM
„>»
3
S
8
8,
g
§
u
0>
.S
s
oj
fe
J8
B
«
tt
.2
!ss
w
a
^>
o
1
e
^
2
•o
«
TB
"3
^5
0
^rj
0$
-S
i
o
1
K.^
J^
.
rH
S
H
C to
O X)
•H id £
AJ i-H
nt
N ^
11
0 >
S 6
1 •
JH c
(U O
4J -H
C 01
•H -H Q
U CO
0)1 ft
S
6
|
0
Q
c/
m
"u a
^
2s ^
*W B
"
C Q
0 co
-H
4J
id
N
"s *
id
CT
o >
'CO U
c o
"2 w
5°
-H
> Q
CO
c
(d
S
z
* E;
[d
§
S
Q
co
0)
4J
f
a
CN
^
in
CO
CN
0
O
CN
O
o
0
•
0
CN
T
in
tN
^
O
0
o
CN
1<
in
ro
o
0
•
0
CN
rH
ID
H
r~
o
o
•
o
i
CN
r-
3
_
"t*
CO
o
0
o
CN
O
o'
o
•
0
CN
O
0
O
o
0,
•*'
ao
rH
O
O
•
0
o
(d
4J
id
•0
O
d
0
CN
°
^
[^
VD
0
O
CN
f)
o
• in
0
o
•
0
M
H
CN
in
m
o
o
o
CN
0'
O
0
CN
0
O
0
•
O
S
CN CN
f- W
i> in
ro o
ID rH
rH rH
O O
CN CN
r~ co
H in
rH f-
O rH
o o
• •
o o
CN CN
a> o
H CN
f
oo en
oo in
O H
0 0
oo' o'
rH
rH
^4
rH O
• •
O O
CN CN
r~ oo
CN in
H
H U)
0 0
• *
0 0
1 1
fN O
°.
1™^ **— g
g « «2
^S « *5
*c ^
> u §
*° 2 S
1 a "
S | M
8 '| -J3
|j* o, f
^. ^ «
U g J3
J^
« 2
•a "K ^
1 '| S
"" «"(!
Sll 1
"•'«*' ^
ill i
,2 9 w ""*
o fj ^ c
S *S S .2
S > fi ^
> o *o
111 1
^1 1 •*
•1*2 w>
^ s •= .s
TD > .-i co
Is i
•a g g
E S OT 5
•~ .TJ u
E3 J3 3
.S .a .2 -o
•t; -^ e to
y ^ g U
*- "- 'C
41 h .2 o
s 1 § S
n r i
£ "2>! £
* -S S *-
cn rs S —
« & M S
t5 E g o
60 ^ jj rt
1*1 'if S
rt c u o
M 1 "3. -p
g C | §
z z S
ON
rvi
ID
ON
ON
ON
t
e
(U)
s
(U
"S,
rt
00
ex
00
rt
u
S
0)
U
S
-------
TABLE 13. Percent Recovery Data from Spiked Fall Line Component Samples, 1990-91.
Parameter Laboratory Percent Recovery
NH4
N02
NO23
TKNW
TKNF
PO4F
TOP
TP
DOC
TOC
SI
PADER
MDHMH
OWML
PADER
MDHMH
OWML
PADER
MDHMH
OWML
PADER
MDHMH
OWML
PADER
MDHMH
OWML
PADER
OWML
PADER
MDHMH
OWML
PADER
MDHMH
OWML
MDHMH
OWML
PADER
MDHMH
OWML
OWML
3/28/90
100
100
100
96
100
101
106
100
128
100
102
97.5
100
100
100
94
100
112
-
8/1/90
101
99
86
112
104
79
104
100
103
117
102
101
101
7/24/91
107.9
100.5
104.9
97.0
102.0
104.0
103.6
90.4
90.8
120.5
104.4
Note: Spiked samples, and thus percent recovery data, are not possible for TSS analysis. DCLS and USGS did not
report any percent recovery data. TOP = Total Dissolved Phosphorus, TP = Total Phosphorus, SI = Silica, NO2
= Nitrite, NO23 = Nitrite + Nitrate, TKNW = Total Kjeldahl Nitrogen Whole, TKNF = Total Kjeldahl Nitrogen
Filtered, NH4 = Ammonium, DOC = Dissolved Organic Carbon, TOC = Total Organic Carbon, PO4F =
Orthophosphate.
csc.sA2.4/93 Coordinated Split Sample Program Annual Report, 1990-1991 • Page 30
-------
B. INTER-ORGANIZATION PRECISION
1. Mainstem Component
Table 3 lists the standard deviation (SD) and coefficient of variation (CV) of the mean
results from split samples for each organization on each sampling date, under "Mean inter-
organization precision." For example, in three-way comparisons, these estimates represent the
mean variability among results from CBL, VIMS, and ODU. The data user should decide which
precision estimate is best for a particular application (see Methods for a definition and discussion
of each estimate).
Although parameters with low inter-organization precision might be expected to have low
inter-organization agreement, this correspondence was not found. Three parameters had high
inter-organization CV means (over 50% for four-way and three-way comparisons): NH4, PO4F,
and PHEA (although there were no four-way comparisons for PHEA). Only one of these, NH4,
was identified as having low inter-organization agreement (see next section). These three
parameters (NH4, PO4F, and PHEA) also had among the highest within-organization CV means
(see previous section).
2. Potomac Component
The coefficient of variation (C V) of laboratory medians for each sampling date estimated
inter-organization precision (Table 6). The mean CV over 3-10 sample dates ranged from 4%
for NO23 to 39% for TSS, but more data are needed to determine if there are consistent
differences among parameters in CV. TP data were excluded when DCLS data were below
detection limits (before December 1990). The parameter that had the lowest inter-organization
agreement based on graphing and ANOVA, NO23, did not have the highest inter-organization
CV values.
3. Virginia Component
The coefficient of variation (CV) of laboratory medians for each sampling date estimated
inter-organization precision (Table 9). The mean CV over 3-8 sample dates ranged from 4%
for NO23 to 40% for TDP and PO4F.
4. Fall Line Component
The standard deviation (SD) and coefficient of variation (CV) of results from three or
four organizations for each sampling date estimated inter-organization precision (Table 12). The
inter-organization CV ranged from 88 % for TDP to 4.7% for SI, based on August 1990 and July
1991 results. In almost all cases, inter-organization SD and CV means were higher than any of
the within-organization precision estimates for the same parameters. More data are needed to
determine if consistent differences exist among parameters in CV.
csc.sA2.4793 Coordinated Split Sample Program Annual Report, 1990-1991 • Page 31
-------
C. INTER-ORGANIZATION AGREEMENT
1. Mainstem Component
Friedman ANOVA was used to assess inter-organization agreement and determine which
parameters had statistically significant inter-organization differences. The ANOVA results
(Table 14) show that 16 out of 17 parameters had statistically significant inter-organization
differences (P < 0.01) for at least one comparison. The medians over all sample dates are
shown for comparison purposes only, and are not used in the statistical test. Comparing the
three mainstem laboratories only, there were significant differences in 11 of the 15 parameters
compared. In the last CSSP report (Bergstrom 1990), only 4 out of 14 parameters had
statistically significant differences in the Mainstem Component, 2 of which were among the three
mainstem laboratories (PN and PC). This increase in the number of parameters with significant
differences probably reflects the increased power of the Friedman test used in this report. The
current Friedman test uses data from three subsamples per sample date rather than the means
used before, effectively tripling the sample size compared to the test used before. The number
of sample dates included was similar in both reports, since the earlier report covered 1987-1989
data, and this report includes 1989-1991 data.
Because the Friedman test does not consider the magnitudes of inter-organization
differences, time plots of the data with precision bars were used to see if the differences were
larger than within-organization precision. Figures 1-17 show the medians for each sample date
with precision bars for each parameter. These graphs show that 6 of the 17 parameters graphed
had more than half of the sample dates with non-overlapping error bars, showing that the
differences were larger than within-organization precision for those parameters. These 6
parameters were ammonia (NH4), paniculate nitrogen (PN), dissolved organic carbon (DOC),
paniculate carbon (PC), total suspended solids (TSS), and silica (SI).
csc.sA2.4/93 Coordinated Split Sample Program Annual Report, 1990-1991 • Page 32
-------
TABLE 14. Mainstem Component (CB5.3 & CB4.4) Split Sample Results, June 1989-Dec.
1991. Medians over all dates with complete data are shown.
Para- N1 Medians over all dates (mo/1 except CHLA & PHEA) Friedman results2
_CBL
NH4
NH4
NH4
N02
N02
NO2
N023
NO23
N023
TON
TDK
PN
PN
TN
TN
TN
PO4F
PO4F
PO4F
TOP
TOP
TOP
7
7
4
8
6
4
11
9
5
10
7
10
9
10
8
5
4
4
4
10
6
6
0.0070
0.0070
B
0.0150
0.0064
A
0.0064
0.0050
0.0970
A
0.0970
AB
0.0970
0.440
A
0.430
0.178
A
0.185
A
0.610
A
0.607
A
0.599
A
0.0037
0.0037
0.0037
0.0086
0.0127
0.0127
0.0098
0.0098
0.0080
0.0065
A
0.0065
0.0040
0.0881
B
0.0881
c
0.0881
0.372
B
0.364
0.143
B
0.126
B
0.506
B
0.526
B
0.461
B
0.0075
A
0.0075
0.0075
0.0085
0.0070
B
0.0070
0.0090
0.0090 0.0400
A
0.0057 0.0400
0.0055
B
0.0055 0.0075
B A
0.0031 0.0060
0.0950
0.0950 0.1000
BC A
0.0950 0.1000
0.412
A
0.409 0.370
0.138
B
0.140 0.150
B
0.556
A
0.556 0.550
0.508 0.500
0.0017
B
0.0017 0.0040
0.0017 0.0040
0.0100
0.0130 0.0190
A
0.0130 0.0170
4.0 <0.20
35.2 <0.001
0.0450 - MDL3
17.7 <0.001
48.4 <0.001
0.0100 - MDL
25.9 <0.001
48.9 <0.001
0.0400 - MDL
22.4 <0.001
19.4 <0.001
47.3 <0.001
31.5 <0.001
35.7 <0.001
20.5 <0.001
0.540 25.4 <0.001
A
26.6 <0.001
MDL
0.0100 - MDL
3.1 <0.30
39.5 <0.001
0.0100 - MDL
CSC.SA2.4/93
Coordinated Split Sample Program Annual Report, 1990-1991 • Page 33
-------
TABLE 14 (continued). Mainstem Component Split Sample Results.
Para-
meter
PHOSP
PHOSP
PHOSP
TP
TP
TP
DOC
PC
TOC
TOC
TSS
TSS
TSS
CHLA"
PHEA4
SI
SI
SI
N1
9
9
6
9
9
6
8
10
8
5
9
9
6
8
6
9
9
5
Medians
CBL
O.Q150
0.0150
0.0146
B
0.0244
0.0244
0.0262
2.81
B
1.19
A
3.92
3.76
4.80
B
4.80
B
4.70
B
0.260
B
0.260
B
0.730
B
over all dates (mcr/1 exceot CHLA
ODU
0.0140
0.0140
0.0140
B
0.0220
0.0220
0.0215
B
2.84
B
0.920
B
3.75
B
3.43
B
9.30
A
9.30
A
8.35
8.37
1.30
A
0.293
A
0.293
A
0.906
A
VIMS MDHMH
0.0160
•0.0160 0.0155
0.0146 0.0175
0.0275
0.0275 0.0440
0.0241 0.0450
A
3.52
A
0.964
B
4.52
A
4.43
A
12.7
A
12.7 10.0
A A
8.80 10.5
A A
8.44 8.61
0.985 0.150
B
0.323
A
0.323 0.350
A A
0.821 0.950
A
& PHEA) Friedman results'
DCLS y2
6.4
6.5
0.0200 23.4
A
2.1
6.5
0.0350 32.7
A
36.9
45.4
28.1
3.06 20.2
B
31.5
42.1
4.50 28.8
6.8
14.0
30.0
40.6
0.794 40.5
P
<0.048
<0.10
<0.001
<0.40
<0.10
<0.001
<0.001
<0.001
<0.001
<0.001
<0.001
<0.001
<0.001
<0.038
<0.001
<0.001
<0.001
<0.001
Number of sample dates with complete data. For parameters with more than one line of results, the first
line is usually three-way comparisons among CBL, VIMS, and ODU (except for CHLA and PHEA), the
second line is four-way comparisons adding MDHMH (except for TOC), and the third line is five-way
comparisons adding DCLS. The sample sizes may vary depending on which laboratories were included.
Underlined values were statistically significant (P < 0.01), Friedman two-way ANOVA. Pairs of medians
that have different letters below them also had statistically significant pairwise differences (A > B > C,
P < 0.01). Medians that have no letter below them, or have a letter in common, did not differ
significantly.
Too many values were below the method detection limit to analyze.
Units are ug/1, not mg/1, for CHLA and PHEA.
CSC.SA2.4/93
Coordinated Split Sample Program Annual Report, 1990-1991 • Page 34
-------
FIGURE 1. Split sample data for ammonium (NH4), from Mainstem samples collected at
CB5.3 or CB4.4, showing medians for each sample date with precision bars.
FIGURE 2. Split sample data for nitrite (NO2), from Mainstem samples collected at CB5.3
or CB4.4, showing medians for each sample date with precision bars.
MDHMH ~B~ VIMS
ODU
35
-------
FIGURE 3. Split sample data for nitrite + nitrate (NO23), from Mainstem samples collected
at CB5.3 or CB4.4, showing medians for each sample date with precision bars.
0.6
0.55-j
g> 0.45-j
TO 0.35-i
z °-3i
+ 0.25 -j
f °'2~:
Z 0.15-j
0.1 4
0.05-j
0
-*- CBL -t
-O— DCLS -<
fc- MDHMH -E
h- ODU
3- VIMS
£
O
UJ LJ
CO Q
FIGURE 4. Split sample data for total dissolved nitrogen (TDN), from Mainstem samples
collected at CBS.3 or CB4.4, showing medians for each sample date with precision bars.
1
O) 0.9-3
~ 0.8-3
d>
§ 0.6^
TJ ni
S "
O 0.4-3
0)
Q °"'
« 0.2-
° 0.1-
o>
o>
00
0.
UJ
CO
O
00
O
UJ
Q
o
£
UJ
CO
o
8
UJ
O
CD O) O) O>
a. i ui uj
< -a co Q
36
-------
FIGURE 5. Split sample data for particulate nitrogen (PN), from Mainstem samples col-
lected at CB5.3 or CB4.4, showing medians for each sample date with precision bars. Preci-
sion bars for MDHMH are not shown since they were large (0.2 mg/1).
0.35
FIGURE 6. Split sample data for total nitrogen (TN), from Mainstem samples collected at
CBS.3 or CB4.4, showing medians for each sample date with precision bars.
37
-------
FIGURE 7. Split sample data for orthophosphate (PO4F), from Mainstem samples collected
at CBS.3 or CB4.4, showing medians for each sample date with precision bars.
FIGURE 8. Split sample data for total dissolved phosphorus (TDP), from Mainstem samples
collected at CBS.3 or CB4.4, showing medians for each sample date with precision bars.
38
-------
FIGURE 9. Split sample data for paniculate phosphorus (PHOSP), from Mainstem samples
collected at CB5.3 or CB4.4, showing medians for each sample date with precision bars.
Precision bars for MDHMH are not shown since they were large (0.02 mg/1).
^ ^
"§>
-0.02
FIGURE 10. Split sample data for total phosphorus (TP), from Mainstem samples collected
at CBS.3 or CB4.4, showing medians for each sample date with precision bars.
0.08
S2* 0.07-
E, 0.06-
S 0.05-
o
g- 0.04-
CO
o
£ 0.03^1
75
O 0.02:
0.01 -
MDHMH -Q- VIMS
ODU
I
O> O>
00 00
Z Q.
3 UJ
-D (0
I
O)
8
UJ
Q
O
O
0.
UJ
O
en
O
UJ
O
o>
DC
Q.
en
o>
Q.
UJ UJ
V) Q
39
-------
FIGURE 11. Split sample data for dissolved organic carbon (DOC), from Mainstem samples
collected at CB5.3 or CB4.4, showing medians for each sample date with precision bars.
There was only one date with data from MDHMH (December 1991).
FIGURE 12. Split sample data for paniculate carbon (PC), from Mainstem samples col-
lected at CB5.3 or CB4.4, showing medians for each sample date with precision bars.
2.2
2-
« 1.24
Q) 1 -
3 0.8-
o
•-E 0.6 ^
CO
0. 0.4-
0.2-
0
o>
00
s s
o. o
UJ UJ
(/> Q
I
0.
UJ
en
o
UJ
o
O)
o>
DC Z
o- r>
O)
a.
UJ
0)
O)
o
UJ
Q
40
-------
FIGURE 13. Split sample data for total organic carbon (TOC), from Mainstem samples
collected at CB5.3 or CB4.4, showing medians for each sample date with precision bars.
D)
Vw*
c
O
.fi
8
CO
O)
FIGURE 14. Split sample data for total suspended solids (TSS), from Mainstem samples
collected at CB5.3 or CB4.4, showing medians for each sample date with precision bars.
41
-------
FIGURE 15. Split sample data for chlorophyll a (CHLA), from Mainstem samples collected
at CBS.3 or CB4.4, showing medians for each sample date with precision bars.
22-
20-
a 14-
f 12-
£10-
O 8-
6 6
4-
o>
00
o>
CO
Q.
HI
at
oo
O
HI
O
o
0>
CC
Q.
UJ
0)
O
o>
o
UJ
o
o>
cc
o>
=>
UJ
O)
8
UJ
Q
FIGURE 16. Split sample data for phaeophytin (PHEA), from Mainstem samples collected
at CBS.3 or CB4.4, showing medians for each sample date with precision bars. Only posi-
tive precision bars are shown for VIMS data since they were large.
42
-------
FIGURE 17. Split sample data for silica (SI), from Mainstem samples collected at CB5.3 or CB4.4,
showing medians for each sample date with precision bars.
C0
0)
CO
O)
CO
o
35
43
-------
2. Potomac Component
Friedman ANOVA performed on results from the three subsamples assessed inter-
organization agreement and determined which parameters had statistically significant inter-
organization differences. The results (Table 15) show that 7 of the 11 parameters analyzed
(NO23, TKNW, PO4, TDP, TOC, TSS, and SI) had statistically significant inter-organization
differences at the P < 0.01 level. Three of these, NO23, PO4 and SI, could have been affected
by the change to filtered samples in MDHMH data in December 1990. NO23 results were
significant only in early data (before December 1990), SI results were significant only in late
data, and PCM results were significant in both time periods. Three other parameters, TKNW,
TOC, and TSS, had large enough sample sizes to test 1991 data separately, to see if the
differences persisted. Only TOC still had statistically significant differences in 1991 data (Table
15). Missing data ruled out analysis of DOC results.
Medians for each sample date with precision bars for each parameter plotted against time
show inter-organization agreement (Figures 18-30). The length of the precision bars are the
MDL (Tables 1 & 2) or the standard deviation of three field subsamples for that sample,
whichever was greater. On sample dates with non-overlapping precision bars, the inter-
organization differences were larger than the witnin-organization precision.
Four parameters, nitrate + nitrite (NO23), orthophosphate (PO4F/PO4W), Total Organic
Carbon (TOC), and Total Phosphorus (TP), had more than half of the sampling dates with non-
overlapping precision bars. NO23 samples (Fig. 20) and PO4 samples (Fig. 23) were from
unfiltered samples at MDHMH (until December 1990) and filtered samples at the other two
laboratories. For NO23, there were eight sampling dates with non-overlapping precision bars,
and all but two, December 1990 and June 1991, were during the period when MDHMH received
unfiltered samples. PO4 had three sample dates with non-overlapping precision bars, all before
December 1990. TOC results had non-overlapping precision bars between DCLS and
DCRA/CRL results on six dates (Fig. 26). Total Phosphorus (TP, Fig. 24) had several non-
overlapping precision bars due to high detection limits in early DCLS data. The other 9
parameters had inter-organization differences that tended to be smaller than within-laboratory
precision.
csc.sA2.4/93 Coordinated Split Sample Program Annual Report, 1990-1991 • Page 44
-------
TABLE 15. Potomac Component (PMS-10) Split Sample Results (May 1989 - December
1991).
Parameter
N
Friedman Results*
NH4 (W) **
NH4(F)
NO2 (W) **
N(. 2 (F)
NO23 (W) **
NO23 (F)
TKNW
TKNW++
TN
P04 (W) **
P04 (F)
TP***
TDP
TOC
TOC++
TSS
TSS++
SI(W)**
SI(F)
BODS
5
4
5
5
6
5
10
4
10
4
5
5
4
11
4
9
3
5
5
3
DCRA/CRL
0.081
0.075
0.013
0.011
1.26
B
1.38
0.536
A
0.490
1.915
0.032
B
0.032
0.059
0.048
4.30
A
5.25
A
11.0
A
8.0
2.89
2.54
1.4
DCLS
0.090
0.070
0.010
0.010
1.46
B
1.39
0.500
B
0.450
1.925
0.040
A
0.030
0.060
0.035
3.30
B
3.30
B
12.0
4.0
2.71
2.38
1.0
MDHMH
0.072
0.074
0.015
0.010
1.60
A
1.40
0.425
B
0.400
2.015
0.042
A
0.024
0.062
0.048
3.44
3.95
7.0
B
6.0
2.70
2.40
1.2
Y2
4.4
1.0
0.84
0.52
29.5
3.5
12.3
0.69
5.08
15.0
8.9
2.1
8.87
17.0
13.4
12.5
3.6
2.5
8.4
1.8
P
>0.10
>0.70
>0.70
>0.70
<0.001
>0.18
0.70
>0.05
<0.001
<0.01
>0.3
<0.01
<0.001
<0.01
<0.001
>0.20
>0.20
<0.01
>0.50
***
Medians over the number of sample dates shown (N). Pairs of medians that have different letters below
them also differed significantly (P < 0.01, A > B).
Probability associated with Chi-square value with 2 df, testing for consistent differences among the three
organizations with Friedman ANOVA. Statistically significant values (P < 0.01) are underlined.
(W) used data through September 1990 only, when MDHMH analyzed unfiltered samples, and (F) includes
data from December 1990 on, when MDHMH analyzed filtered samples. CRL and DCLS always analyzed
filtered samples.
Using 1991 data only, to see if significant differences persisted.
Using TP data from December 1990 onward, after the MDL was lowered for DCLS analyses.
CSC.SA2.4/93
Coordinated Split Sample Program Annual Report, 1990-1991 • Page 45
-------
FIGURE 18. Split sample data for ammonium (NH4), from Potomac samples collected at
PMS10, showing medians for each sample date with precision bars. MDHMH samples were
filtered starting in December 1990.
0.2
^ 0.16-
^ 0.14-
g^ 0.12 -j
i 0.1^
0.08-
0.06-
0.04-
0.02-
0
o
<
o>
§2
O)
s
to
g
i
uj
0)
o
UJ
Q
o>
DC
O.
0)
UJ
0)
O)
&
Q
FIGURE 19. Split sample data for nitrite (NO2), from Potomac samples collected at
PMS10, showing medians for each sample date with precision bars. MDHMH samples were
filtered starting in December 1990. Negative error bars were omitted for clarity.
0.04
0.035-
0.03^
O) 0.025 ••
J 0.02-;
2 0.015-:
0.01-i
0.005-
i i
< 3
O)
UJ
CO
O O
o> 2
Z C
< <
O
°*
UJ
(/)
O)
c
Q.
o>
Z
=>
o>
Q.
UJ
0)
o>
a
o
46
-------
FIGURE 20. Split sample data for nitrite + nitrate (NO23), from Potomac samples collected
at PMS10, showing medians for each sample date with precision bars. MDHMH samples
were filtered starting in December 1990.
2.2
2-
^ 1.8-
O) 4 c
E 1.6-
g 1.2-
Z .
+
£ 0.8-
i«-
0.4-I
0.2-
0
DCRA/CRL
DCLS
MDHMH
o>
UJ
(0
I
<
UJ
(0
o>
DC
Q.
O)
s:
HI
(0
o>
O
LU
O
FIGURE 21. Split sample data for total Kjeldahl nitrogen whole (TKNW), from Potomac
samples collected at PMS10, showing medians for each sample date with precision bars.
47
-------
FIGURE 22. Split sample data for total nitrogen (TN), from Potomac samples collected at
PMS10, showing medians for each sample date with precision bars. MDHMH samples for
N023, one of the components of TN, were filtered starting in December 1990.
c
O)
g
z
75
o
2.75 i
2.5-
2.25-
1.75 4
1.5-j
1.25-j
0.75 \
0.5-
0.25 '-.
0
o>
§2
at
o>
UJ
(0
|
<
HI
en
$
O
Ul
Q
o>
DC
0.
O)
UJ
o>
O
UJ
Q
FIGURE 23. Split sample data for orthophosphate (PO4F), from Potomac samples collected
at PMS10, showing medians for each sample date with precision bars. MDHMH samples
were filtered starting in December 1990.
0.08
0.06^
0-05^
§•0.04-3
O
§• 0.03-
g 0.02 H
0.01-
O)
O)
o>
§
z
UJ
0)
O
UJ
Q
o>
OC
Q.
o>
UJ
o>
a
o
48
-------
FIGURE 24. Split sample data for total phosphorus (TP), from Potomac samples collected
at PMS10, showing medians for each sample date with precision bars.
£ 0.04
o
FIGURE 25. Split sample data for total dissolved phosphorus (TDP), from Potomac samples
collected at PMS10, showing medians for each sample date with precision bars.
^ 0.12
O) 0.11-
<=- 0.1-
co
0.09-
o
CO
O
£
i
o
CO
CO
s
o
0.08-
0.07-
0.06-
0.05-
0.04-
0.03-
0.02-
0.01-
0
DCRA/CRL
DCLS
MDHMH
r
o>
s
I 1
O)
Ul
CO
o
O>
O
a>
01
CO
a
o
o>
c
o.
O)
O)
Q.
UJ
CO
o>
O
UJ
o
49
-------
FIGURE 26. Split sample data for total organic carbon (TOC), from Potomac samples
collected at PMS10, showing medians for each sample date with precision bars.
FIGURE 27. Split sample data for dissolved organic carbon (DOC), from Potomac samples
collected at PMS10, showing medians for each sample date with precision bars.
9
O) oj
E 8^
O
O 5-
Sl
•S 2^
2
2 1-
Q
r i
O) O) O>
00 00 00
> z o.
< => w
2 -9 (0
o
O)
Z
c
I
LLI
(0
S^ T-
O> CD
U K Z
UJ D. D
Q < -3
UJ
0)
en
O
UJ
o
50
-------
FIGURE 28. Split sample data for total suspended solids (TSS), from Potomac samples
collected at PMS10, showing medians for each sample date with precision bars.
50
O) 45-
•5- 40-
co
2 35-
£ 30-=
I 2'
I 20-
co
5 V
« 10^
DCRA/CRL
DCLS
MDHMH
I
<
o>
o>
JU
(0
<
8T- T- ^- 1-
O> O> O) O>
UJ
(0
o
UJ
o
0.
<
UJ
(0
o>
o
UJ
o
FIGURE 29. Split sample data for silica (SI), from Potomac samples collected at PMS10,
showing medians for each sample date with precision bars. MDHMH samples were filtered
starting in December 1990.
co
(0
f
4.5
4-
7Z 3.5-
3-
2.5-
1-1
0.5 r
0
DCRA/CRL
DCLS
MDHMH
i i s
< 3 UJ
I
<
I
UJ
(0
o
UJ
o
O)
oc
0.
o>
S!
UJ
(0
O)
O
UJ
o
51
-------
FIGURE 30. Split sample data for biological oxygen demand 5 day (BODS), from Potomac
samples collected at PMS10, showing medians for each sample date with precision bars.
52
-------
3. Virginia Component
Friedman ANOVA performed on results from the three subsamples assessed inter-
organization agreement and determined which parameters had statistically significant inter-
organization differences. The results (Table 16) show that 9 of the 13 parameters analyzed had
statistically significant inter-organization differences at the P < 0.01 level over the 1990-1991
period. Analyses were also done using 1991 data only, when 8 of the 13 parameters had
statistically significant differences. Three parameters, NO23, PN, and TSS, had no significant
differences over both time periods.
Medians for each sample date with precision bars plotted against time show inter-
organization agreement (Figures 31-47). Four parameters were graphed that were not analyzed
statistically because data were reported by only two laboratories (PC, DOC, CHLA, and
PHEA).
Four of the 17 parameters graphed, PO4F, PHOSP, TOC, and SI, had inter-organization
differences that were larger than within-organization precision on more than half of the sampling
dates. The other parameters had inter-organization differences that tended to be smaller than
within-organization precision.
4. Fall Line Component
The results from the Friedman ANOVA are shown in Table 17, with medians for each
laboratory over all sampling dates analyzed. The results show that none of the 9 parameters had
statistically significant differences.
Graphs of the results with precision bars (Figures 48-59) also show that most of the
parameters had high inter-organization agreement. One parameter, nitrite + nitrate (NO23), had
three samples with non-overlapping precision bars, due to higher results from MDHMH.
CSC.SA2.4/93 Coordinated Split Sample Program Annual Report, 1990-1991 • Page 53
-------
TABLE 16. Virginia Component (TF5.5) split sample medians with Friedman analysis
results, 1990-1991 data. Data from 1991 were analyzed separately where applicable.
Parameter1 N2
Laboratory Medians (mq/1)
Friedman results3
NH4
NH4
N02
N02
N023
NO23
TON
PN
TN
TN
PO4F
PO4F
TOP
TOP
PHOSP
PHOSP
TP
TP
TOC
8
5
3
2
8
5
4
3
5
4
7
4
7
4
7
4
7
4
5
DCLS
0.0750
0.0800
0.050
0.035
0.365
0.400
.
-
0.800
1.150
0.020
c
0.020
B
0.030
B
0.030
0.090
A
0.110
A
0.160
0.165
3.87
B
HRSD
0.0950
A
0.1000
0.080
A
0.055
0.385
0.420
0.623
B
0.450
0.710
B
1.070
B
0.030
AB
0.028
0.060
A
0.056
0.090
0.101
0.169
0.169
0
0
0
0
0
0
0
0
0
1
0
0
0
0
0
0
0
0
7
ODU
.0728
B
.0739
.046
B
.027
.374
.421
.737
.258
.799
B
.072
B
.037
A
.036
A
.036
.032
.078
.097
.127
.138
.18
A
VIMS
0
0
0
0
0
0
0
0
0
1
0
0
0
0
0
0
0
0
7
.0760
.0800
.052
.032
.385
.410
.878
A
.315
.933
A
.158
A
.025
BC
.025
B
.033
.033
.064
B
.083
B
.136
.138
.43
A
18
6
15
12
1
7
11
1
26
21
39
30
24
5
15
15
8
13
21
Y2
.2
.1
.6
.7
.1
.8
.2
.4
.8
.1
.4
.0
.5
.3
.1
.9
.7
.1
.3
P
<0
<0
.001
.20
<0.01
<0
>0
<0
•0!
.70
.10
<0.01
>0
<0
<0
<0
<0
<0
<0
<0
.30
.001
.001
.001
.001
.001
.20
.01
<0.01
<0
.05
<0.01
<0
.001
CSC.SA2.4/93
Coordinated Split Sample Program Annual Report, 1990-1991 • Page 54
-------
TABLE 16 (continued): Virginia Component (TF5.5) split sample medians with Friedman
analysis results, 1990-1991 data. Data from 1991 were analyzed separately where
applicable.
Parameter1 N2 Laboratory Medians (mq/1)
TSS
TSS
SI
SI
5
4
8
5
DCLS
43.0
45.0
3.28
B
2.90
B
HRSD ODU
49.0 48.7
53.5 49.1
3.56
A
3.28
A
VIMS
49.3
50.7
3.48
A
3.16
A
Y2
4.4
6.3
35.7
17.4
P
>0.20
<0.10
<0.001
<0.001
1 Dissolved Organic Carbon (DOC), Paniculate Carbon (PC), Chlorophyll (CHLA), and Phaeophytin
(PHEA) could not be analyzed due to missing data. Where two lines of results are shown for a parameter,
the first includes all complete data, and the second includes only 1991 data. Where there is only one line,
the complete data were all from 1991.
2 Number of sampling dates with complete data. In one case (NO2), dates were excluded due to below
detection limit data.
3 Underlined P values were statistically significant (P < 0.01, Friedman 2-way ANOVA on three replicates
per sample date). Laboratory medians with different letters below them had statistically significant pairwise
differences (P < 0.01, A>B>C), otherwise they did not differ significantly.
CSC.SA2.4/93 Coordinated Split Sample Program Annual Report, 1990-1991 • Page 55
-------
FIGURE 31. Split sample data for ammonium (NH4), from Virginia samples collected at
TF5.5, showing medians for each sample date with precision bars.
O)
o
<
FEB90 MAY90 SEP90 JAN91 MAY91 JUN91 SEP91 DEC91
FIGURE 32. Split sample data for nitrite (NO2), from Virginia samples collected at TF5.5,
showing medians for each sample date with precision bars.
FEB90 MAY90 SEP90 JAN91 MAY91 JUN91 SEP91 DEC91
56
-------
FIGURE 33. Split sample data for nitrite + nitrate (NO23), from Virginia samples collected
at TF5.5, showing medians for each sample date with precision bars.
0.6-
^ 0.5-
D)
S
0.3-3
0)
0.2-
0.1-
DCLS —*— ODU
HRSD ~B~ VIMS
FEB90 MAY90 SEP90 JAN91 MAY91 JUN91 SEP91 DEC91
FIGURE 34. Split sample data for total dissolved nitrogen (TDN), from Virginia samples
collected at TF5.5, showing medians for each sample date with precision bars.
0)
.
i
o
CO
0)
FEB90 MAY90 SEP90 JAN91 MAY91 JUN91 SEP91 DEC91
57
-------
FIGURE 35. Split sample data for paniculate nitrogen (PN), from Virginia samples col-
lected at TF5.5, showing medians for each sample date with precision bars.
FE
MAY90 SEP90 JAN91 MAY91 JUN91 SEP91 DEC91
FIGURE 36. Split sample data for total nitrogen (TN), from Virginia samples collected at
TF5.5, showing medians for each sample date with precision bars.
3
o>
0)
D>
2
2
o
2.75-
2.5-
2.25-
2-j
1.75-
1.5-
1.25-j
1^
0.75-
0.5-
0.25-
DCLS —«— ODU
HRSD C VIMS
FEB90 MAY90 SEP90 JAN91 MAY91 JUN91 SEP91 DEC91
58
-------
FIGURE 37. Split sample data for orthophosphate (PO4F), from Virginia samples collected
at TF5.5, showing medians for each sample date with precision bars.
FEB90 MAY90 SEP90 JAN91 MAY91 JUN91 SEP91 DEC91
FIGURE 38. Split sample data for total dissolved phosphorus (TDP), from Virginia samples
collected at TF5.5, showing medians for each sample date with precision bars.
FEB90 MAY90 SEP90 JAN91 MAY91 JUN91 SEP91 DEC91
59
-------
FIGURE 39. Split sample data for paniculate phosphorus (PHOSP), from Virginia samples
collected at TF5.5, showing medians for each sample date with precision bars.
O>
£L
0)
o
CL
0)
O
•E
£
FE890 MAY90 SEP90 JAN91 MAY91 JUN91 SEP91 DEC91
FIGURE 40. Split sample data for total phosphorus (TP), from Virginia samples collected at
TF5.5, showing medians for each sample date with precision bars.
1.1
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
O)
o
JC
Q.
0)
O
£
T5
FEB90 MAY90 SEP90 JAN91 MAY91 JUN91 SEP91 DEC91
60
-------
FIGURE 41. Split sample data for dissolved organic carbon (DOC), from Virginia samples
collected at TF5.5 showing medians for each sample date with precision bars.
D>
C
O
£
co
O
.o
CO
s>
O
•o
I
8
0)
0
8-
7-
6-
5-
4-
3-
2-i
FEB90 MAY90 SEP90 JAN91 MAY91 JUN91 SEP91 DEC91
FIGURE 42. Split sample data for paniculate carbon (PC), from Virginia samples collected
at TF5.5, showing medians for each sample date with precision bars.
20
18
16-
14-
12-
10-
8
4-j
2\
D)
C
O
£
CO
O
£
(D
£
ODU
VIMS
FEB90 MAY90 SEP90 JAN91 MAY91 JUN91 SEP91 DEC91
61
-------
FIGURE 43. Split sample data for total organic carbon (TOC), from Virginia samples
collected at TF5.5, showing medians for each sample date with precision bars.
25
FEB90 MAY90 SEP90 JAN91 MAY91 JUN91 SEP91 DEC91
FIGURE 44. Split sample data for total suspended solids (TSS), from Virginia samples
collected at TF5.5, showing medians for each sample date with precision bars.
FEB90 MAY90 SEP90 JAN91 MAY91 JUN91 SEP91 DEC91
62
-------
FIGURE 45. Split sample data for chlorophyll a (CHLA), from Virginia samples collected
at TF5.5, showing medians for each sample date with precision bars.
O>
CO
o.
2
.o
6
FEB90 MAY90 SEP90 JAN91 MAY91 JUN91 SEP91 DEC91
FIGURE 46. Split sample data for phaeophytin (PHEA), from Virginia samples collected at
TF5.5, showing medians for each sample date with precision bars.
ODU -B- VIMS
FEB90 MAY90 SEP90 JAN91 MAY91 JUN91 SEP91 DEC91
63
-------
FIGURE 47. Split sample data for silica (SI), from Virginia samples collected at TF5.5,
showing medians for each sample date with precision bars.
5.5
CO
5-j
4.5-:
4-i
^' ^ •
*§> 3-i
= 2-
1.5-E
0.5-=
FEB90 MAY90 SEP90 JAN91 MAY91 JUN91 SEP91 DEC91
64
-------
TABLE 17. Fall Line (CB1.0) Split Sample Results using one subsample per sample date,
October 1989-July 1991.
Para-
meter1
NH4
NH4
N02
N023
NO23
TKNW
TKNW
TKNF
TKNF
TN
TN
PO4F
PO4F
TDP
TDP
TP
TP
TOC
TOC
TSS
SI
N2
4
2
0
4
2
4
2
4
2
4
2
4
2
4
2
4
2
4
2
0
0
Laboratorv Medians, Subsamole 1
PADER
0.0900
0.1250
•
1.225
0.995
0.470
0.470
0.380
0.360
1.73
1.47
0.0020
0.0020
0.025
0.025
0.055
0.050
3.05
3.10
-
•
uses
0.0650
0.0850
-
1.150
0.950
0.600
0.650
0.350
0.350
1.80
1.60
0.0050
0.0025
0.0060
0.0060
0.035
0.035
2.70
2.70
-
•
MDHMH
0.096
0.134
•
1.300
1.050
0.450
0.625
0.325
0.250
1.90
1.68
0.0040
0.0040
0.038
0.044
0.038
0.028
2.68
3.14
•
•
(ma/1) Friedman results3
OWML *2
1.5
0.105 4.2
.
5.4
0.970 1.1
0.4
0.430 0.6
1.5
0.355 1.1
1.5
1.40 3.6
0.5
0.0150 4.7
6.0
0.015 4.2
4.9
0.025 4.1
1.6
3.15 0.6
.
.
P
0.65
0.38
-
>0.069
>0.83
>0.93
0.96
0.65
0.83
0.65
0.46
0.93
>0.21
0.069
0.38
>0.069
>0.38
0.65
0.96
-
-
NH4 = ammonium, NO2 = Nitrite, NO23 = Nitrite + Nitrate, TKNW = Total Kjeldahl Nitrogen
Whole, TKNF = Total Kjeldahl Nitrogen Filtered, TN = Total Nitrogen (= TKNW + NO23), PO4F =
Orthophosphate filtered, TDP = Total Dissolved Phosphorus, TP = Total Phosphorus, TOC = Total
Organic Carbon, TSS = Total Suspended Solids, SI = Silica (as SI). Missing data prevented running the
Friedman test for NO2, TSS, and SI. The first line for each parameter shows three-way comparisons
(without OWML) and the second line shows four-way comparisons.
Number of sample dates with complete data.
Probability that the observed differences were due to chance (accepting the null hypothesis), based on
Friedman 2-way ANOVA (Siegel 1956).
CSC.SA2.4/93
Coordinated Split Sample Program Annual Report, 1990-1991 • Page 65
-------
FIGURE 48. Split sample data for Ammonium (NH4), from samples collected at CB1.0 (Fall Line),
showing first subsample results with precision bars.
0.18
0.16-
0.14-
| 0.12^
I 0.08-
| 0.06-
0.04,
0.02-
NH4_PAD
NH4 USG
NH4_MDH
NH4 OWM
OCT89
MAR90
AUG90
JUL91
FIGURE 49. Split sample data for Nitrite (NO2), from samples collected at CB1.0 (Fall Line),
showing first subsample results with precision bars.
0.18
0.16-
0.14-
c. 0.124
o>
£ 0.14
g 0.084
Z }
0.06-j
0.04-
0.024
NO2_PAD
NO2 USG
NO2_MDH
NO2 OWM
OCT89
MAR90
AUG90
JUL91
66
-------
FIGURE 50. Split sample data for Nitrite + Nitrate (NO23), from samples collected at CB1.0 (Fall
Line), showing first subsample results with precision bars.
1.8
1.6
~ 1.4
O)
£1.2
+ 0.8-
S
* 0.64
0.4,
0.2 J
0
NO23_PAD
NO23 USG
N023_MDH
NO23 OWM
OCT89
MAR90
AUG90
JUL91
FIGURE 51. Split sample data for Total Kjeldahl Nitrogen Whole (TKNW), from samples collected
at CB1.0 (Fall Line), showing first subsample results with precision bars.
o>
0)
"o
0)
§>
(0
•o
0)
I
1
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
TKNW_PAD
TKNW USG
TKNW_MDH
TKNW OWM
OCT89
MAR90
AUG90
JUL91
67
-------
FIGURE 52. Split sample data for Total Kjeldahl Nitrogen Filtered (TKNF), from samples collected
at CB1.0 (Fall Line), showing first subsample results with precision bars.
TKNF_PAD
TKNF USG
TKNF_MDH
TKNF OWM
OCT89
MAR90
AUG90
JUL91
FIGURE 53. Split sample data for Total Nitrogen (TN), from samples collected at CB1.0 (Fall
Line), showing first subsample results with precision bars.
OCT89
MAR90
AUG90
JUL91
68
-------
FIGURE 54. Split sample data for Orthophosphate (PO4F), from samples collected at CB1.0 (Fall
Line), showing first subsample results with precision bars.
PO4F_PAD
PO4F USG
PO4F_MDH
PO4F OWM
OCT89
MAR90
AUG90
JUL91
FIGURE 55. Split sample data for Total Dissolved Phosphorus (TDP), from samples collected at
CB1.0 (Fall Line), showing first subsample results with precision bars.
0.1
1,0.09-:
^ O-08-!
o 0.07-
g 0.06 4
^ 0.05-
! 0.04-
| 0.03 4
O
a 0.02 ^
^
•" 0.01-
TDP_PAD
TDP_USG
TDP_MDH
TDP OWM
OCT89
MAR90
AUG90
JUL91
69
-------
FIGURE 56. Split sample data for Total Phosphorus (TP), from samples collected at CB1.0 (Fall
Line), showing first subsample results with precision bars.
0.1
0.09-
=• 0.08^
£ 0.071
w
S 0.06-
o
§• 0.05-:
£ 0.04 J
| 0.03-i
0.02-:
0.01-^
0
TP_MDH
TP OWM
OCT89
MAR90
AUG90
JUL91
FIGURE 57. Split sample data for Total Organic Carbon (TOC), from samples collected at CB1.0
(Fall Line), showing first subsample results with precision bars.
0.5-
0
TOC_PAD
TOC USG
TOC_MDH
TOC OWM
OCT89
MAR90
AUG90
JUL91
70
-------
FIGURE 58. Split sample data for Total Suspended Solids (TSS), from samples collected at CB1.0
(Fall Line), showing first subsample results with precision bars.
14
o
to
10-
8-
v
i H
CO
^ 4^
*«
TSS_PAD
TSS USG
TSS_MDH
TSS OWM
OCT89
MAR90
AUG90
JUL91
FIGURE 59. Split sample data for Silica (SI), from samples collected at CB1.0 (Fall Line), show-
ing first subsample results with precision bars.
2.2
2-
1.8-
1.2-
£
8
= 0.8^
CO
0.6^
0.4-
0.2 J
0
SI_PAD
SI USG
SLMDH
si OWM
OCT89
MAR90
AUG90
JUL91
71
-------
IV. DISCUSSION
Readers should keep in mind that all the components include some laboratories that
analyze a large number of samples per day comprising a wide range of sample concentrations-
from estuarine samples to wastewater treatment plant samples. The number of samples they
analyze reduces the time available for researching advanced techniques and equipment.
In addition, in some components the concentrations encountered are much higher or lower
than those usually encountered by some of the laboratories in their Chesapeake Bay sampling.
This occurred in the Virginia Component with ODU and VIMS, so their samples from TF5.5
often required considerable dilution. Laboratories encountering samples below their usual range
of concentrations may find they are near or below their detection limit. This introduces
additional sources of variability not encountered in their routine sampling.
A. WrnnN-ORGANIZATTON PRECISION AND ACCURACY
1. Mainstem Component
The estimates of within-organization precision generally show similar results from the
different organizations involved. One exception is the two parameters calculated by subtraction
in the Mainstem component, PN and PHOSP results from MDHMH (Table 3). These
parameters were more variable than the PN and PHOSP results calculated directly by the other
three organizations. The same pattern was noted by D'Elia et al. (1987) and used as an
argument for using the direct methods. On the other hand, one parameter, TDN, was calculated
by addition in the MDHMH data and directly by the other organizations, and it had similar
precision among the four organizations that reported it.
The estimates of within-organization accuracy generally showed similar results from the
different organizations involved for spike recovery (Table 4), but there were a few differences
among organizations for Standard Reference Material (SRM) recovery (Table 5). VTMS had
more variable SRM recoveries for TDN than those reported by CBL or ODU, but all the VIMS
values were within the confidence limits (+/- 1 SD) supplied by EPA, which correspond to
TDN SRM recoveries between 53 % and 184 %. ODU showed slightly lower recoveries of DOC
SRMs diluted in saline matrix than VIMS, but this is consistent with the known differences
between the DOC instruments used.
2. Potomac Component
The estimates of within-organization precision (Table 6) generally show similar results
from the different organizations involved.
CSC.SA2.4/93 Coordinated Split Sample Program Annual Report, 1990-1991 • Page 72
-------
The estimates of within-organization accuracy generally showed similar results from the
different organizations involved for spike recovery (Table 7). Standard Reference Material
(SRM) results (Table 8) could not be compared because only DCRA/CRL reported them.
3. Virginia Component
Paniculate Nitrogen (PN) data at HRSD was probably more variable because it is
calculated by subtraction (TKNW-TKNF). Parameters calculated by subtraction tend to be more
variable than directly measured parameters (D'Elia et al. 1987).
As in the Mainstem Component, VIMS had more variable SRM results for TDN than
ODU, but all the VIMS results were within the +/- 1 SD limits provided by EPA (53-184%
recovery). Another pattern seen in Mainstem SRM results, lower DOC results from ODU in
estuarine matrix, was also found in the Virginia Component data (mean SRM recoveries were
86% from ODU and 103% from VIMS). This is consistent with the known tendency for the
Shimadzu analyzer used by VIMS to achieve higher DOC recoveries.
4. Fall Line Component
The within-organization precision estimates are quite tentative, since they are based on
duplicates of only two samples. The one high CV value, from TDP results from USGS, resulted
from divergent values for two low-concentration samples, which tends to inflate the CV. These
estimates will be more reliable when sample sizes are larger.
B. INTER-ORGANIZATION PRECISION
The estimates of inter-organization precision are designed to assess the measurement
system variability in each component after sampling has occurred. They should be used with
caution for that purpose, since both SD and CV values can be affected by concentration.
The estimates of inter-organization precision should not be used to assess inter-
organization agreement. Unless there are consistent inter-organization differences over several
sampling dates, larger CV values may not indicate any problem with inter-organization
agreement. Some tests give more variable results than others, and this variability may not be
under the laboratory's control. In the discussion that follows, the parameters with the highest
inter-organization CV values are identified, and this list is compared to the list of parameters
with high within-organization CV values, and the list of parameters with low inter-organization
agreement.
CSC.SA2.4/93 Coordinated Split Sample Program Annual Report, 1990-1991 • Page 73
-------
1. Mainstem Component
The highest inter-organization CV means were over 50% (Table 3): NH4, PO4F, and
PHEA. These three parameters also had among the highest within-organization CV means (see
previous section). Only one of these three, NH4, was identified as having low inter-organization
agreement (see next section).
2. Potomac Component
The highest inter-organization CV means were over 30% (Table 6): TOC, TSS, and
BODS. Contrary to the results in the Mainstem Component, these three parameters did not have
the highest within-organization CV means (see previous section). Only one of these, TOC, was
identified as having low inter-organization agreement (see next section).
3. Virginia Component
The highest inter-organization CV means were over 30% (Table 9): PN, PO4F, TOP,
and TOC. Contrary to the results in the Mainstem Component, these three parameters did not
have the highest within-organization CV means (see previous section), except for PN (PON) at
HRSD. Only two of these, PO4F and TOC, were identified as having low inter-organization
agreement (see next section).
4. Fall Line Component
The inter-organization CV results (Table 12) will be more useful when they can be
compared to past data from the same program. Current results are similar to inter-organization
variation estimates from the Mainstem Component, but sample sizes are too limited to draw any
conclusions.
C. INTER-ORGANIZATION AGREEMENT
1. Mainstem Component
The results in Table 14 and Figures 1-17 are summarized in Table 18, showing what
actions are recommended based on these results. Eleven parameters (NO2, NO23, TDN, TN,
PO4F, TDP, PHOSP, TP, TOC, CHLA, and PHEA) had inter-organization agreement that was
high enough that no action is recommended. Two of these, TP and TDP, were investigated
following the last CSSP report (Bergstrom 1990), and the differences were reduced by using a
more accurate blank correction.
Two parameters (PN and PC) currently show inter-organization differences, but should
show higher inter-organization agreement in 1992 data, after field method changes were
CSC.SA2.4/93 Coordinated Split Sample Program Annual Report, 1990-1991 • Page 74
-------
implemented in January 1992. One parameter (DOC) is the subject of an extensive comparison
study by VIMS, documenting the magnitudes of differences caused by the use of different
instruments.
Three parameters (NH4, TSS, and SI) are recommended for initial study to determine
possible cause(s) of the inter-organization differences found. The higher NH4 results from
MDHMH were probably related to the autoanalyzer used there; agreement should improve in
1993 after a new autoanalyzer was obtained. For SI, the inter-organization differences mainly
occurred when concentrations were high. The five dates with non-overlapping error bars were
September 1989, June through December 1990, and September 1991. These dates all had SI
concentrations over 0.8 mg/1, except September 1991.
2. Potomac Component
The results from graphing (Figures 18-30) and the Friedman test (Table 15) are combined
in Table 19. Two of the 11 parameters graphed, NO23 and TOC, had inter-organization
differences that were larger than within-laboratory precision on more than half of the sampling
dates. Both parameters also had statistically significant results from the Friedman test.
Unfiltered samples at MDHMH probably accounted for the NO23 differences, since
MDHMH had higher results than the other laboratories during the early period. Thus, no
investigation is recommended for NO23, since agreement improved after December 1990.
The differences in TOC involved higher results from DCRA/CRL, lower results from
DCLS, and intermediate results from MDHMH. Different instruments at the three laboratories
may account for these differences. DOC results showed a similar pattern, but missing data
prevented any statistical analysis. Which results are more accurate cannot be estimated, because
only DCRA/CRL reported SRM results for TOC. The differences are continuing, since some
of the larger differences were in 1991 data. Investigation of possible causes of the differences
in TOC results is recommended.
The other five parameters that had statistically significant results from the Friedman test
(TKNW, PO4F, TDP, TSS, and SI) did not meet the criteria for recommending action, because
the magnitude of the differences was smaller than within-organization precision (Table 19).
Agreement of TKNW, PO4F, and TSS results appears to be improving over time, based on
graphing and Friedman results. More data are needed to determine whether the small
differences in TDP and SI results will persist.
3. Virginia Component
Four of the 13 parameters that were analyzed statistically, PO4F, PHOSP, TOC, and SI,
had inter-organization differences that were larger than within-organization precision on
CSC.SA2.4/93 Coordinated Split Sample Program Annual Report, 1990-1991 • Page 75
-------
TABLE 18. Summary of Friedman ANOVA results (Table 14) and plots of medians for
each sample date (Figs. 1-17), Mainstem Component, 1989-1991.
More than half of
Friedman
Parameter
Ammonium
(NH4)
Nitrite
(NO2)
Nitrite +
Nitrate (NO23)
Total Dissolved
Nitrogen (TON)
Particulate
Nitrogen (PN)
Total
Nitrogen (TN)
. Orthophosphate
(PO4F)
Total Dissolved
Phosphorus (TOP)
Particulate Phos-
phorus (PHOSP)
Total Phosphorus
(TP)
Dissolved Orga-
nic Carbon (DOC)
Particulate
Carbon (PC)
Total Organic
Carbon (TOO
3-wav
no
yes
yes
yes
yes
yes
yes
no
no
no
yes
yes
yes
ANOVA sionificant1
4-wav 5-wav
yes MDL2
yes MDL
yes MDL
yes
yes
yes yes
MDL MDL
yes MDL
no yes
no yes
-
-
yes
Differences > Recommended
than MDL Action
Yes (MDHMH > CBL) ,
7/7 dates
No
No
No
Yes (CBL > others) ,
7/10 dates
No
No
No
No
No
Yes (VIMS >
others) , 6/8 dates
Yes (CBL > others) ,
8/10 dates
No
study3
none
none
none
method
changes4
none
none
none5
none
none3
comparisoi
study*
method
changes4
none
CSC.SA2.4/93
Coordinated Split Sample Program Annual Report, 1990-1991 • Page 76
-------
TABLE 18 (continued). Summary of Friedman ANOVA results (Table 14) and plots of
medians for each sample date (Figs. 1-17), Mainstem Component, 1989-1991.
More than half of
Friedman ANOVA significant' Differences > Recommended
Parameter 3 -way 4 -way 5 -way than MDL Action
Total Suspended yes yes yes Yes (Others > CBL) , study3
Solids (TSS) 8/9 dates
Chlorophyll a no No none
(CHLA)
Phaeophytin
(PHEA)
Silica (SI)
yes
yes yes yes
No
Yes (Others > CBL) ,
5/9 dates
none
study3
1 Three-way comparisons include CBL, ODU, and VIMS; four-way add MDHMH; and five-way add DCLS.
Sample sizes were usually smaller for five-way comparisons, because DCLS data started in June 1990. '
2 Too many values were below the Method Detection Limit (MDL) for a comparison.
3 Study begins with a comparison of MDLs and field and laboratory methods to look for possible cause(s) of
differences, and may later include a method comparison study if needed.
4 Method changes were implemented in January 1992 to make field methods more consistent and to reduce or
eliminate these differences, which were caused by different filtration methods.
5 After the 1989 CSSP Mainstem report showed that MDHMH results were high for these two parameters, a more
accurate blank correction was found, and MDHMH data were corrected for this report.
6 Study done by VIMS documents the magnitude of these differences, which are apparently due to the use of
different instruments.
CSC.SA2.4/93 Coordinated Split Sample Program Annual Report, 1990-1991 • Page 77
-------
TABLE 19. Summary of Friedman ANOVA results (Table 15) and plots of medians for
each sample date (Figs. 18-30), Potomac Component, 1989-1991.
A. Parameters affected by the change to filtered samples at MDHMH in December 1990
Parameter
More than half of
Friedman ANOVA significant' Differences > Recommended
unfiltered filtered than MDL Action
Ammonium no
(NH4)
Nitrite no
(NO2)
Nitrite + yes
Nitrate (N023)
Orthophosphate yes
(PO4F, PO4W)
Silica (SI) no
no
no
no
yes
yes
no
no
yes2,
8/11 dates
no
no
none
none
none
none
none
B. Parameters not affected by the change to filtered samples at MDHMH
Parameter
More than half of
Friedman ANOVA significant' Differences > Recommended
all data 1991 data than MDL Action
Total Kjeldahl yes
Nitrogen Whole
(TKNW)
Total Nitrogen no
(TN)
Total Phosphorus3 no
(TP)
Total Dissolved yes
Phosphorus (TOP)
Total Organic
Carbon (TOO
Dissolved Organic
Carbon (DOC)
yes
no
no
no
no
no
yes
yes,
6/11 dates
none
none
none
none
study4
none
csc.sA2.4/93 Coordinated Split Sample Program Annual Report, 1990-1991 • Page 78
-------
TABLE 19 (continued). Summary of Friedman ANOVA results (Table 15) and plots of
medians for each sample date (Figs. 18-30), Potomac Component, 1989-1991.
B. (continued). Parameters not affected by the change to filtered samples at MDHMH.
More than half of
Friedman ANOVA significant1 Differences > Recommended
Parameter all data 1991 data than MDL Action
Total Suspended yes no no none
Solids (TSS)
Biological Oxygen -5 - none
Demand 5 day
(BODS)
1 P < 0.01.
2 All, or almost all, of the differences larger than the MDLs were before December 1990, when MDHMH
analyzed unfiltered samples. Thus, no action is recommended.
3 Using data from December 1990 onward, after the MDL was lowered for DCLS analyses.
4 Study begins with a comparison of MDLs and field and laboratory methods to look for possible cause(s) of
differences, and may later include a method comparison study if needed.
5 Missing data prevented statistical analysis for DOC and BODS.
CSC.SA2.4/93 Coordinated Split Sample Program Annual Report, 1990-1991 • Page 79
-------
more than half of the sampling dates (Table 20). All of these parameters also had
statistically significant results (P < 0.01) from the Friedman test (Table 16), so they meet
the criteria for recommending investigation.
The PO4F differences appear to have been caused by higher results from ODU,
where PO4F calibrations are by standard additions. ODU also had higher PO4F results in
Mainstem Component samples (Table 14), although the magnitudes of the differences were
small (Table 18). The different matrix and higher PO4F concentrations at Hopewell (TF5.5)
may contribute to the higher results from ODU in the Virginia Component. HRSD changed
methods and lowered their method detection limit (MDL) in August, 1990.
The PHOSP differences appear to be due to higher results from DCLS, and to a
lesser extent from HRSD. Since both DCLS and HRSD calculate PHOSP by subtraction,
from TP - TDP, and VIMS and ODU measure it directly, no further investigation appears to
be necessary.
The TOC results from DCLS averaged about 3.5 mg/1 lower than results from VIMS
or ODU. The lower results were probably due to the use of a different digestion method at
DCLS. Until 1992, DCLS used a Dorman DC-180 TOC analyzer with UV persulfate
digestion, which tends to provide less complete digestion when particulates are present
compared to other TOC methods (R. Potts pers. comm.). They started using a Dorman DC-
190 with high-temperature digestion (680 °C) in March 1992, which should increase
agreement in 1992 data. The differences in SI results might be due to different methods at
DCLS; DCLS results averaged about 0.3 mg/1 lower than results from ODU or VIMS.
However, SI agreement appears to have increased slightly during 1991 (Table 16, Fig. 47).
The other parameters that had statistically significant results from the Friedman test
did not meet the criteria for recommending investigation, because the magnitude of the
differences was smaller than within-organization precision. Additional data are needed to
determine whether the agreement of these parameters warrants further study.
4. Fall Line Component
Inter-organization agreement was high for all 9 parameters analyzed (Table 21). The
one parameter with more than half of the sample dates with non-overlapping precision bars
(NO23) did not have statistically significant differences. However, the small sample sizes,
and resulting low power of the test used, mean that there may be some significant differences
when sample sizes are larger. Also, the higher MDLs in this component, compared to the
Mainstem Component, make it more likely that the precision bars will be overlapping in this
component. The differences among laboratories in sample preservation, detection limits, and
analytical methods did not cause any detectable differences in results. These differences
include preserved samples at OWML and USGS, but not at MDHMH or PADER; and USGS
analytical methods at USGS, and EPA methods at the other three laboratories.
CSC.SA2.4/93 Coordinated Split Sample Program Annual Report, 1990-1991 • Page 80
-------
TABLE 20. Summary of Friedman ANOVA results (Table 16) and plots of medians for
each sample date (Figs. 31-47), Virginia Component, 1990-1991.
Parameter
More than half of
Friedman ANOVA significant' Differences > Recommended
all data 1991 data than MDL Action
Ammonium (NH4) Yes
Nitrite (N02) Yes
Nitrite + No
Nitrate (N023)
Total Dissolved Yes
Nitrogen (TON)
Particulate No
Nitrogen (PN)
Total Yes
Nitrogen (TN)
Orthophosphate
(P04F)
Yes
Total Dissolved Yes
Phosphorus (TDP)
Particulate Phos- Yes
phorus {PHOSP)
Total Phosphorus No
(TP)
Total Organic Yes
Carbon (TOC)
Total Suspended No
Solids (TSS)
Silica (SI)
Yes
No
Yes
No
No
No
No
No
No
none
none
none
Yes No none
No No none
Yes No none
Yes Yes (ODU>VIMS, 6/7 study2
dates, + ODU>DCLS,
4/7 dates)
No No none
Yes Yes (DCLS>VIMS, none3
4/7 dates)
Yes No none
Yes Yes (Others > study2
DCLS, 5/5 dates)
none
Yes Yes (Others > study2
DCLS, 6/8 dates)
1 Data for 1991 were analyzed separately to see if differences persisted.
2 Study begins with a comparison of MDLs and field and laboratory methods to look for possible cause(s) of
differences, and may later include a method comparison study if needed.
3 Higher PHOSP results from DCLS are probably due to determination by subtraction by TP - TDP; HRSD also
used this method and had high results.
CSC.SA2.4/93
Coordinated Split Sample Program Annual Report, 1990-1991 • Page 81
-------
TABLE 21. Summary of Friedman ANOVA results (Table 17) and plots of medians for
each sample date (Figs. 48-59), Fall Line Component, 1989-1991.
Parameter
More than half of
Friedman ANOVA significant' Differences >
than MDL
Recommended
Action
Ammonium
(NH4)
Nitrite
(N02)
Nitrite +
Nitrate (NO23)
Total Kjeldahl
Nitrogen Whole
(TKNW)
Total Kjeldahl
Nitrogen Filtered
(TKNF)
Total Nitrogen
(TN)
Orthophosphate
(P04F)
Total Dissolved
Phosphorus (TOP)
Total Phosphorus
(TP)
Total Organic
Carbon (TOO
Total Suspended
Solids (TSS)
Silica (SI)
No
No
No
No
No
No
No
No
No
No
No
No
No
No
Yes,
3/4 dates
No
No
No
No
No
No
No
No
No
none
none
none
none
none
none
none
none
none
none
none
none
P < 0.01.
CSC.SA2.4/93
Coordinated Split Sample Program Annual Report, 1990-1991 • Page 82
-------
5. Parameters recommended for investigation in two or more components
Two parameters, Total Organic Carbon (TOC) and Silica (SI), were recommended for
investigation in two components each. The patterns of their differences were compared to
see if the same laboratories were involved.
Investigation was recommended for TOC in the Potomac and Virginia components.
TOC also showed significant differences in the Mainstem Component, although most of the
differences were smaller than the precision estimates. In all three components, TOC results
from the Division of Consolidated Laboratory Services (DCLS) in Richmond tended to be
lower than results from other laboratories. This tendency was probably due to the use of the
Dorman DC-180 TOC instrument used at DCLS, which used UV persulfate digestion. TOC
agreement should increase in 1992 data after DCLS started using a Dorman DC-190 with
high-temperature digestion.
Investigation was recommended for SI in the Mainstem and Virginia components. In
the Mainstem Component, Chesapeake Biological Laboratory (CBL) tended to have lower
results, and in the Virginia Component, DCLS tended to have lower results. CBL did not
participate in the Virginia Component, but DCLS results were close to CBL results in the
Mainstem Component (see Table 14). The SI methods and instruments used should be
studied to see if they could be a cause of these patterns.
D. FUTURE DIRECTIONS FOR THE CSSP
1. Changes in Splitting Methods
Two changes in splitting methods have been made or will be made soon, although
they did not affect 1990-1991 data. The Virginia Component changed from using three
separate churn splitters to a single large churn splitter (30 liters, about 8 gallons) in June
1992. The Fall Line Component will change from using a single churn splitter, which was
not large enough to provide the subsamples needed, to using three separate churn splitters.
Although these changes appear inconsistent, each is suited to the logistics and the
concentrations encountered in each component. The Virginia Component encounters much
higher TSS concentrations, and data analyzed in this report showed that the three separate
splitters did not split evenly under those conditions. The Fall Line Component encounters
lower TSS concentrations, so the three splitters should be adequate there. The change to
using three churn splitters, and thus providing three subsamples to each Fall Line laboratory,
will increase the power of the Friedman test, making it easier to detect any differences that
are present. Especially if sampling cannot be done quarterly, providing three subsamples is
very important. Also, the Fall Line Component uses USGS field procedures, which include
using churn splitters, but do not include splitting from a large carboy. USGS staff have been
CSC.SA2.4/93 Coordinated Split Sample Program Annual Report, 1990-1991 • Page 83
-------
unable to obtain a single churn splitter that was large enough to provide the required number
of subsamples.
2. Ensuring adequate ranges of split sample concentrations
Low ambient concentrations may be a problem in any split sample program. They
can lead to two data problems: the occurrence of below detection limit results, and a lack of
data on inter-organization agreement at higher concentrations. Both are a problem in the
Mainstem Component, and to a lesser extent for certain parameters in other components.
Below detection limit values tend to limit the usefulness of the split sample results:
they may rule out assessing inter-organization agreement, and they usually prevent
calculation of percent recovery from spiked samples. Also, the low concentrations of some
parameters such as TSS in Mainstem samples (30 mg/1 or less, Figure 14) limits our
knowledge of inter-organization agreement at higher concentrations.
The simplest way to expand the concentration range of Mainstem Component split
samples would be to start using bottom samples from CB4.4. Mainstem splits have always
used surface samples, but the VIMS-ODU two-way split samples used bottom samples.
Higher TSS values in bottom samples might make it more difficult to split samples
accurately, but extremely high concentrations do not occur at CB4.4. The maximum TSS
concentrations during 1984-1990 were less than 93 mg/1 at CB4.4 (Table 22).
TABLE 22. Concentrations of selected parameters in surface and bottom layers at
Station CB4.4, 1984 - 1990.
SURFACE
BOTTOM
Parameter
Mean
Min.
Max.
Mean
Min.
Max.
SALIN 13 .6
(Salinity)
6.6
18.9
20.2
15.9
24.2
TSS
CHLA
P04F
NH4
N023
5
10
0
0
0
.2
.7
.0043
.024
.21
2
1
0
0
0
.5
.8
.0015
.003
.0009
18
33
0
0
0
.2
.0
.011
.14
.96
14
8
0
0
0
.9
.1
.019
.15
.064
2
0
0
0
0
.0
.4
.0018
.003
.001
92.6
59.8
0.084
0.48
0.32
CSC.SA2.4/93
Coordinated Split Sample Program Annual Report, 1990-1991 • Page 84
-------
Using bottom samples would expand the concentration range of most of the
parameters that have a limited range in CB4.4 surface samples (Table 22): TSS, CHLA,
PO4F, and NH4. The range might be slightly reduced for NO23, but there are no agreement
problems for NO23. The salinity would be increased, but would still be within the range
encountered by the two tributary laboratories at the mouths of rivers. Higher concentrations
might increase the number of samples that required dilution before analysis by some of the
mainstem laboratories, however. Bottom samples could be used for every split sample, or
surface and bottom samples could be used in alternation or chosen at random (e.g., with a
coin toss). Bottom sample data from CB4.4 would increase the usefulness of mainstem
CSSP results with little or no increase in effort or cost.
V. REFERENCES
Bergstrom, P. 1990. Chesapeake Bay Coordinated Split Sample Program Annual Report,
1989. CBP/TRS 51/90, Chesapeake Bay Program, Annapolis, MD.
Chesapeake Bay Program. 1991. Chesapeake Bay Coordinated Split Sample Program
Implementation Guidelines, Revision 3. Chesapeake Bay Program Technical Report
Series, CBP/TRS 59/91, Annapolis, MD.
D'Elia, C., R. Magnien, C. Zimmermann, P. Vaas, N. Kaumeyer, C. Keefe, D. Shaw, and
K. Wood. 1987. Nitrogen and phosphorus determinations in estuarine waters: A
comparison of methods used in Chesapeake Bay Monitoring. CBP/TRS 7/87, EPA
Chesapeake Bay Program, Annapolis, MD.
Marascuilo, L. A., and M. McSweeney. 1977. Nonparametric and distribution-free
methods for the social sciences. Brooks/Cole Publishing Co., Monterey, CA.
Montgomery, D. C. 1985. Introduction to statistical quality control. Wiley & Sons, NY.
SAS Institute, Inc. 1990. SAS Procedures Guide, Version 6, Third Edition. SAS Institute,
Gary, NC.
Siegel, S. 1956. Nonparametric statistics for the behavioral sciences. McGraw-Hill Book
Co., NY.
Zimmermann, C., C. Keefe, K. Wood, and N. Kaumeyer. 1992. Comparison of
instrumentation and filters used for the analysis or paniculate carbon and nitrogen in
estuarine waters. Chesapeake Biological Laboratory, Solomons, MD, CBP/TRS
83/92.
csc.sA2.4793 Coordinated Split Sample Program Annual Report, 1990-1991 • Page 85
------- |