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

-------