EPA-903-R-99-024
   Chesapeake Bay Program Mainstem
Coordinated Split Sample Program Report
               1994-1998
         Chesapeake Bay Program

-------
                              Table of Contents
Introduction	      2
Methods	      4
Summary of Analytical Results	      6
Mainstem Split Results by Parameter	      9
      Total Dissolved Phosphorus	     10
      Particulate Phosphorus 	     16
      Ortho-Phosphate	     22
      Total Phosphorus	     28
      Total Dissolved Nitrogen	     32
      Particulate Nitrogen	     38
      Ammonium	     44
      Nitrate + Nitrite	     50
      Nitrite	     56
      Total Suspended Solids	     62
      Particulate Carbon	     67
      Silica	     73
      Chlorophyll-a	     79
Appendix A - Tables	     85
Appendix B - Blind Audit Report  	     95

-------
I.  INTRODUCTION

       The Monitoring Subcommittee of the Chesapeake Bay Program initiated the
Chesapeake Bay Coordinated Split Sample Program (CSSP) in 1988. It's goal is to
assess the comparability of water quality results from the 9 analytical laboratories that
participate in the Chesapeake Bay Monitoring Program. This goal is achieved by
identifying  any parameters that have low inter-laboratory agreement and by estimating
the measurement system variability.

       Identifying parameters with low agreement enables the labs and organizations
involved to investigate significant differences among their methods, and take actions to
raise their inter-laboratory agreement. This might involve changing field methods,
laboratory methods, or both. It is important to note that the split sample variability can
come from  variability in field sampling as well as lab analysis variability.  Therefore,
laboratory variability includes all elements of the measurement system: field sampling,
sample handling, lab analysis, data handling and the state or municipal agency that
supervise the water quality monitoring program.

       Estimates of the variability of the measurement system are useful to data users
such as statisticians and modelers who need confidence bounds for monitoring data.
Although split sample results do not include routine sampling variability, they are the
best measurements we have available to estimate variability of the total system of
Chesapeake Bay water quality monitoring data.

       The CSSP has three components, each including three to five labs that analyze
samples from similar salinity regimes and concentration ranges (CBP 1991).  Labs from
each component analyze triplicate field samples that are collected quarterly with the
exception of the Fall Line component, which is sampled twice a year due to budgetary
constraints. Labs  send the analytical results to the EPA Chesapeake Bay Program Office
(CBPO) in  Annapolis for data analysis.

       This report summarizes the 1994-1998 results from the mainstem component of
the Coordinated Split Sample Program. The mainstem component is the only component
that analyzes saline water samples.  This component includes two mainstem labs:
Chesapeake Biological Laboratory (CBL) and Old Dominion University (ODU). It also
includes a Maryland tributary lab, Maryland Department of Health and Mental Hygiene
(MDHMH), and a Virginia tributary lab, Division of Consolidated Laboratory Services
(DCLS). The split samples are collected by the Maryland Department of Natural
Resources.  (Figure 1)

-------
Mainstem Split
      CBL
    DHMH
    Figure 1
                  ODU

-------
II.  METHODS

A.  SAMPLE COLLECTION AND SPLITTING

       A field crew from the Maryland Department of Natural Resources (MD DNR)
collected quarterly split samples from the surface layer at station MCB4.4.  The field
crew followed the splitting procedures in the CSSP Guidelines (CBP 1991). One large
sample is stirred on the boat in a 15 gallon carboy with stirring rod connected to an
electric drill.  Subsamples were drawn sequentially from a spigot at the bottom of the
carboy into 1 liter polyethylene bottles. The bottles from subsample 1 are dispensed in
the sequence MDHMH- CBL-ODU-DCLS, followed by the bottles for subsamples 2 and
3.
B. ANALYTICAL CHEMISTRY METHODS

       Historically, the mainstem labs use different analytical methods than the two
tributary labs.  Mainstem labs measure the dissolved and particulate fractions, while the
tributary labs measure the total and the dissolved fractions and then calculate the
particulate fractions.  However, in 1995 DCLS began measuring the dissolved and
particulate fractions. All laboratories filter their samples between 8 and 10 am the day
after collection. Methods are discussed in greater detail within the sections describing the
results for each parameter.
C. DATA ANALYSIS AND GRAPHING

       Inter-laboratory agreement is the tendency for split sample results from different
labs and organizations to be consistently similar over time. Any pairs of labs that have a
large and recurring inter-laboratory difference are said to have low agreement. A
decision rule was developed to identify which parameters had inter-laboratory differences
that were large and consistent enough to warrant investigation by the organizations
involved.  Based upon 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 appropriate statistical tests.  Graphs with
precision bars will show the magnitude of the differences for any one given sample date,
while the statistical test is more sensitive to the consistency of the differences over time.
Further investigation was recommended if:

       1) more than half of the sampling dates had pairwise inter-laboratory diffferences
that were larger than within organization precision, (if the error bars don't overlap), and.

       2) an appropriate statistical test had a probability (p) < 0.01 that the differences
between labs was due to chance alone and not analytical differences.

-------
       Parameters identified by this combination of factors usually have different field
and/or laboratory methods at one or more of the laboratories involved.

       Graphs of the split sample results show which labs had results that were farther
apart than their own laboratory precision estimates. The within-laboratory precision
estimates for CSSP analysis were either 1) the standard error of the three subsamples for
each sample date; or 2) 2x the standard error of the difference between the calculated and
observed value the lab obtains when analyzing standard reference material (SRM) for the
variable in question. No labs analyzed SRMs for every parameter. See Table 1 for a
description of what value was used (standard error of the three replicates or  a
combination of the standard error of the three replicates and the standard error of the
SRM data) in determining the error bars. Graphs of the means for each sample date for
each lab were drawn showing this within-laboratory precision as  "error bars".  Any pairs
of lab means that do not have overlapping "error bars" have differences that were larger
than their within-laboratory precision.

       A multi-factor ANOVA was used to assess interlaboratory agreement using the
CSSP split sample data. Factors examined in the data were date,  replicate number, lab
and date/lab interaction. Due to the assumptions of the ANOVA, if a lab was missing all
data for a particular date, that date was dropped from the analysis. If the results of the
ANOVA suggested that there was significant interlab variability (i.e. a significant
difference among labs, p< 0.01), then the data were subjected to a Least Squares Means
analysis. The replicate factor of the ANOVA examined the data for differences
associated with replicate number (1, 2 or 3). A significant difference, as determined by
the replicate factor, could be an indication of inefficient mixing of the reservoir from
which the replicates are split.  The lab factor examined the data for variance  associated
with a lab and the date/lab factor determined if variability associated with the labs was
variable over time. The ANOVA was then rerun with the lab factor using the date/lab
mean square error term. This was done to determine if the difference among labs was
greater than the within run variability associated with each lab.

       In the Least Squares Means analysis, the mean  concentration for each lab was
compared to the  mean concentration of the other labs (for the mainstem - 4  labs for a
total of 6 comparisons). From this, it could be determined which labs were significantly
different from one another. A Least Squares Means analysis was  also conducted on the
mean of the absolute value of the residuals for each lab's split sample data. This analysis
gave some insight into the analytical precision of each lab relative to the other labs. The
Least Squares Means results are summarized graphically in the parameter sections of this
report.  An example graphic is displayed in figure 2.

                       CBL  ODU DCLSDHMH
       Figure 2. The above diagram summarizes an example Least Squares Means
       Analysis. Labs are ordered from left to right in terms of increasing
       variability. Labs underlined by the same line are not significantly different
       from one another in terms of their variability

-------
              Table 1 - Table describing what value was used for the error bars in
              the graphs of the split sample results. "SRM" indicates that the error
              bar is 2x the standard error of the difference between the calculated
              and observed value the lab obtains when analyzing standard
              reference material (SRM) for the variable in question. "Split"
              indicates that the error bar is the standard error of the three
              subsamples for each sample date.
Parameter
TDP
PP
PO4f
TP
TDN
NH4
NO23
NO2
TSS
PC
Si
PN
CBL
SRM
Split
SRM
Split
SRM
SRM
SRM
Split
Split
Split
Split
Split
ODU
SRM
Split
SRM
Split
SRM
SRM
SRM
Split
Split
Split
Split
Split
DCLS
SRM
Split
SRM
Split
SRM
SRM
Split
Split
Split
Split
Split
Split
DHMH
Split
Split
SRM
Split
Split
SRM
SRM
Split
Split
Split
Split
Split
       The ANOVA and Graphical analyses were run twice. Once on the entire 1994-
1998 data set (20 sampling dates) and once on data from September of 1997 through the
end of 1998 (6 sampling dates). The latter was conducted with the intention of detecting
recent or developing problems among the labs.

III. Summary of Analytical Results

A) Within Laboratory Precision

       Three estimates of within laboratory precision and bias were used in this analysis:
the coefficient of variation (CV) of the three field replicates from each split (precision),
the percent spike recovery and the standard reference material (SRM) percent recovery
(bias). The CV expresses the standard deviation of the three replicates as a percentage of
the mean of the three replicates. A lower CV indicates a higher degree of precision. If a
lab is consistently obtaining CVs above 25% for a given parameter, further investigation
may be required. However, CVs tend to be  related to concentration. As  concentration
decreases, variability increases and the standard deviation becomes a larger percentage of
a smaller mean.

-------
       The percent spike recovery is determined by spiking an aliquot of one of the three
replicates with a known concentration. The measured value is then expressed as a
percentage of the expected value.  SRM analyses for selected parameters are conducted
within the same run as the three split sample replicates for that parameter. SRM percent
recoveries are determined by analyzing a sample of known concentration and calculating
a percentage based on the measured value and the expected value. Percent recovery
values should be between 90 and 110%. Values less than 80% or greater than 120% are
indicative of a problem.

       Laboratory CVs for each parameter and sampling date were generally low
(summarized in the method comparison tables for each parameter;  complete data
available in Appendix A). There were a few exceptions to this however. DHMH had CVs
exceeding 25% for total dissolved phosphorus and particulate phosphorus on six of
fourteen dates and twelve of fifteen dates respectively. CBL had CVs exceeding 25% for
PO4f and NH4 on seven of sixteen dates and  six of sixteen dates respectively. DCLS had
CVs exceeding 25% for PC on seven of eight dates.

       Spike percent recoveries and SRM percent recoveries were generally good for all
labs. All of the mean and median spike percent recoveries for every parameter measured
by each lab were well within the 90 to 110%  range, also good for all labs. SRM
recoveries almost all fell within the 90 to 110% range.

B) Interlaboratory Agreement
       Of the twelve parameters analyzed, seven had significant differences between at
least two of labs according to both the Least Squares Means analysis and the graphical
analysis. These parameters were TOP, PP, TP, TON, TSS, PC and PN. When the
analyses were done on the September 1997 through December 1998 data significant
differences between at least two labs were found only for TDP, TSS and PN. These
results are summarized in Table 2.

-------
Table 2 - Table summarizes the results of the Least Squares Means analysis of means and residuals
and the graphical analysis. Significant indicates that significant differences were found
Parameter
Total
Dissolved
Phosphorus
Particulate
Phosphorus
Ortho-
Phosphate
Total
Phoshorus
Total
Dissolved
Nitrogen
Particulate
Nitrogen
Ammonium
Nitrate +
Nitrite
Nitrite
Total
Suspended
Solids
Particulate
Carbon
Silica
Chlorophyll-a
1994 - 1998
LS Means
of Means
Significant
Significant
Significant
Significant
Significant
Significant
Not
Significant
Not
Significant
Not
Significant
Significant
Significant
Not
Significant
NA
1994 - 1998
LS Means
of Residuals
Significant
Significant
Significant
Significant
Not
Significant
Not
Significant
Not
Significant
Not
Significant
Not
Significant
Not
Significant
Not
Significant
Not
Significant
NA
1994 -1998
Graphical
Analysis
Significant
Significant
Not
Significant
Significant
Significant
Significant
JNOt
Significant
Significant
Significant
Significant
Significant
Significant
NA
1997.5 - 1998
LS Means of
Means
Significant
Not
Significant
Not
Significant
Not
Significant
Not
Significant
Significant
Not
Significant
Not
Significant
Not
Significant
Significant
Not
Significant
Not
Significant
NA
1997.5 - 1998
LS Means of
Residuals
Not
Significant
Significant
Significant
Not
Significant
Not
Significant
Not
Significant
Not
Significant
Not
Significant
Not
Significant
Not
Significant
Not
Significant
Not
Significant
NA
1997.5 -1998
Graphical
Analysis
Significant
Not
Significant
Significant
Significant
Significant
Significant
Significant
Not
Significant
Significant
Significant
Significant
Significant
NA

-------
Mainstem Split Results
    By Parameter

-------
                    Parameter: Total Dissolved Phosphorus
Labs: CBL, ODU, DCLS and DHMH




Measurement (direct/indirect): All labs measure directly




         Total Dissolved Phosphorus Method Comparison - Mainstem Labs
Variable
03 fc
"a § .9
111
^ &• M
5/5 S °


z
es
%
es
3
S^




•a
o
1




.0
B
c
1
g
M
HH

OJ
=
tfS 1
h3 _£
'53

CBL
Vacuum, 0.7^m GF/F
30 ml glass test tube


Graduated cylinders.
Glass tubes cleaned w/
10% HC1, DI rinsed,
autoclaved with
potassium persulfate
before use.



Alkaline persulfate
digestion (60 min @
4psi) + EPA 365.1-
automated ascorbic acid
method. (Valderrama,
1981 & D Eliaetal,
1977)

Technicon AAII; 880
nm
37 ° Cheat bath,
50 mm flow cell




Rinsed w IN HC1 for 15
min. after analysis,
DI for 15 min.



ODU
Vacuum, 0.7^m GF/F
HOPE, 250 ml


Class A volumetric.
Dedicated glass tubes.
1/94-5/97: Dichromic
acid soak w/ ROW
rinse.
After 6/97: Liquinox ,
w/tap water rinse,
rinsed twice w/ 4N HC1
then 9X w/RGW.
1/94-5/97: EPA 365.3
Manual ascorbic acid by
std. addition)
6/97 on: Alkaline
persulfate digestion
(autoclave 30min@ 105
ec) + EPA 365.1 -auto.
ascorbic acid method.
1/94-5/97: Perkin Elmer
X-l single beam spec.
6/97o«:SkalarSANplus,
880 nm. Auto
background/ matrix
correct (lOlOnm), 75
mm flow cell, 40° C
heat bath
Rinsed w/RGW for 30
min. after analysis.
Weekly: Cartridge
cleaned w 0.5 N NaOH
for hr. and ROW for
hour. Align flowcell.
DCLS
Vacuum, 0.7^m GF/F
HOPE


Rainin auto pipet.
Digestion tubes
autoclaved with
persulfate before use.
Dedicated glassware
washed in 1:1 HC1, DI
rinsed.


Alkaline persulfate
digestion + EPA 3 65.1-
automated ascorbic acid
method.




Skalar SANplus, 880nm.
Auto background/
matrix correction w/
10 lOnm filter.
50 mm flow cell



Rinsed w/ DI water
daily.

Rinsed w/ 0.5 N NaOH
for hr. weekly

DHMH
Vacuum,0.7^m GF/F
HOPE


Disposable 30 ml
borosilicate tubes w/
polypropylene screw
caps





EPA 365. 4: Acid block
digestion (H2SO4,
K2SO4 & HgSO4 ) +
automated ascorbic acid
method.



Alpkem model 3570
with SoftPac software
660 nm heat bath.





Rinse w/DI for 15 min,
15minw/10%HCL,
20 min with DI, 30 min
with IN NaOH, 30 min
w/DI water.

                                     10

-------

OJ
M
es
(2

S
C
3
^
"g
sS
•a
c
55
Q 5B
is *f
•fS
•a
8» H^
U
-•g
11
i>
. fsl O y
1 A I?
|g a
13 si
|l
•*- >
o o
•a a
g

.-^
'3.
0.
^^
^^
U

Potassium persulfate,
boric acid for digestion.
Two reagents, DI, SDS
wetting agent for
analysis
KH2PO4 in DI H2O

3 standards & DI blanks
are digested.
Glycerophosphate
internal (check)
standard.

0.0186- 0.092 &
0.1488 -0.372 mg/L

0.001 mg/L


0.0186 mg/L

0/20



SPEX
93-109



1994-1998
Range - 93-103
Mean -97. 9
Median - 97
1997.5-1998

Range - 93 - 103
Mean -98. 3
Median -98.5
1/94-5/97: Two
reagents,
6/97 on: Combined
reagent, ROW, FFD-6
wetting agent.
1/94-5/97: KH2PO4 in
com-posite of filter.
sample water.
6/97 on: Artificial sea
water (ASW) salinity ~
sample.
5-6 standards & ASW
blanks are digested.
0.005 -0.15 mg/L


0.00 1-0.004 mg/L


0.005 mg/L

1/19



SPEX
90-104



1994-1998
Range - 90-105
Mean -98. 5
Median - 99
1997.5-1998

Range - 93 - 105
Mean- 97.5
Median - 96.5
Two reagents, FFD-6
wetting agent



KH2PO4 in DI water.
Fresh standards &
blanks are digested.
Glycerophosphate
check standard.



0.020 - 0.200 mg/L


0.01 mg/L (1994)
0.001 mg/L (95-98)

0.020

4/18



APG
91-111



1994-1998
Range - 78-104
Mean -93. 9
Median - 93
1997.5-1998

Range - 93 - 99
Mean- 95.25
Median- 94.5
Two reagents, NaCl
diluent, Dowfax 2A1
wetting agent


KH2PO4 in DI H2O

5 standards - Are DI
blanks and stds.
digested?



0.010 -0.500 mg/L


0.01 mg/L


0.010 mg/L

8/18



Not analyzed



1994-1998
Range - 81-105
Mean -98. 9
Median - 99.5
1997.5-1998

Range - 94 - 101
Mean- 98.2
Median - 98
11

-------
8.
|l
H
< 28 day sat -20 °C

< 28 day sat -20 °C

< 28 day sat -20 °C

4°Cfor48hrs., or,
< 28 days w/ HzS04

 Split Results:

1994-1998
                           ANOVA results
Effect
Rep
Lab
Date*Lab
Lab using
Date*Lab
error term
P Value
0.86
0.0001
0.0001
0.0001
LS Means Results

              Of Means

   DHMH  DCLS CBL ODU
            Of Residuals

CBL  ODU  DCLS DHMH
       The ANOVA results indicate that there is no replicate affect on variability within
the data meaning the splitting procedure was conducted properly. The remainder of the
ANOVA results indicate that there was a difference among labs, that this difference
varied through time, and that this difference was greater than the within run variability
associated with each lab.
       The LS means results indicate that ODU's TDP results were consistently different
than the other three labs through time as was the case with DHMH. DCLS and CBL were
not consistently different from one another (as indicated by the underlining below the
two), but were different from ODU and DHMH. The LS Means of the residuals indicate
that, in terms of variability around the mean, CBL and  DHMH are different from one
another but not from ODU and DCLS.
                                        12

-------
     Graphical Results
                                     Mainstem TOP
0.0600
0.0500
0.0100  -*-
o.oooo '—i
               ft   ^
            Of the seventeen dates where all four labs had data, graphical results show that
     DHMH was significantly different than CBL and ODU twelve times. CBL, ODU and
     DCLS were not significantly different from one another and DCLS was not significantly
     different from DHMH.

     Discussion of TDP Results
            The results of the LS means analysis indicate that ODU is biased low and DHMH
     is biased high. This pattern can be seen in the graph of the data where ODU frequently
     has the lowest value on each date and DHMH almost always has the highest value. CBL
     and DCLS are always in the middle and do not appear to be consistently higher or lower
     than each other. No bias is indicated by the spike recoveries of ODU and DHMH;
     ODU's SRM recoveries are good. In June of 1997, ODU switched from a manual
     method using standard additions to an automated method but the methods were
     demonstrated to be equivalent. As of this date, ODU has not found the source of this
     apparent bias.

            The LSM of the residuals indicate that CBL and DHMHs variability (precision)
     are significantly different. This is also evident upon examination of the CV data for both
     labs. Out of twenty observations, CBL had no CVs greater than 25% while DHMH had 8

-------
of 18 greater than 25%. DHMH differences in variability may be due to using a block
digestion procedure instead of the alkaline persulfate digestion.
       The graphical analysis supports the differences detected between DHMH and
CBL and DHMH and ODU by the LSM. It does not support the other differences
detected by the LSM.
1997.5-1998
                           ANOVA results
Effect
Rep
Lab
Date*Lab
Lab using
Date*Lab
error term
P Value
0.2448
0.0001
0.0001
0.0013
LS Means Results

              Of Means

   DHMH CBL DCLS ODU
           Of Residuals

CBL  ODU  DCLS DHMH
       The ANOVA results indicate that there is no replicate affect on variability within
the data meaning the splitting procedure was conducted properly. The remainder of the
ANOVA results indicate that there was a difference among labs, that this difference
varied through time, and that this difference was greater than the within run variability
associated with each lab.
       The LS means results indicate that ODU's TOP results were consistently different
than the other three labs through time. DHMH was consistently different from DCLS and
ODU but not from CBL. DCLS and CBL were not consistently different from one
another but were from ODU. The LS Means of the residuals indicate that, in terms of
variability around the mean, there were no differences in variability among labs.
                                        14

-------
       Graphical Results
    0.05
 )
~
Q.
0
                                            Mainstem TOP
              On the four dates for which data were available for all labs, ODU was
       significantly different from DHMH on three out four dates. None of the other pairwise
       comparisons yielded significantly different results.

       Discussion:
              The negative bias detected in ODUs  1994-1998 data was also detected in the
       more recent 1997.5 - 1998 data. Also, examination of the graph of the 1997.5-1998 data
       indicates that DHMH still has a positive bias. The LSM of the residuals indicates that
       DHMHs precision improved in the 1997.5-1998 data set. The graphical analysis supports
       the difference between ODU and DHMH detected by the LSM but none of the other
       differences detected.
                                               15

-------
                      Parameter: Particulate Phosphorus

Labs: CBL, ODU, DCLS and DHMH

Measurement (direct/indirect): CBL and ODU directly, DCLS indirectly until 2/95 and
      DHMH indirectly

           Particulate Phosphorus Method Comparison - Mainstem Labs
Variable

0 =
••2 o
2
~*^ !•
^M ffrf

"2. PH
| Qg
VI





•o
o
'S
S






a
.0
B
a
1


hH


0)
O
cS
a

'53
S
"S
hH
CBL
Particulates filtered in
duplicate on 47 mm
GF/F filters are dried
at 104 °C, weighed
for TS Sand stored at
room temp until
combusted at 550°C
for 1.5 hrs.

Filters extracted in
IN HC1 for < 24 hrs
before analysis.
Combustion at 550 °C
converts all P cmpds.
to PO4, which is
extracted w/ HC1 and
measured by EPA
365.1 (automated
ascorbic acid
method).
Technicon AAII
880 nm

50mm flow cell








DI rinse for 15 min.






ODU
Particulates filtered
on 47 mm GF/F,
frozen at -20°C.
Filters dried at
104 °C, weighed for
TSS, and combusted
at 550 °C> 1.5 hrs.


Cooled filters are
extracted in IN HC1
for 24 hrs.
Combustion at 550 °C
converts all P cmpds.
to PO4, which is
extracted w/ HC1 and
measured by EPA
365.1 (automated
ascorbic acid
method).
1/94-12/95 SIC
continuous flow

analyzer.
1/96 on: SKALAR
SANplus
880nm, w/lOlOnm

background
correction. 75mm
flow cell, 40° Cheat
bath
Rinsed w/ ROW for
30 min. after
analysis. Weekly:
Cartridge cleaned w
O.SNNaOHfor
hr. and ROW for
hour. Align flowcell.
DCLS (after 2/95)
Particulates filtered
on 47 mm GF/F,
frozen until ready for
analysis.
Dried at 105 °C
overnight, muffled
for 2 hours at 550 °C.
Cooled in desiccator.

1/94 - 2/95:
Calculated
2/95 - present:
Extracted overnight
with IN HC1,
combust at 550 °C to
convert all P cmpds.
to PO4, which is
extracted w/ HC1 and
measured by EPA
365.1
SKALAR SANplus
880nm, w/lOlOnm

background
correction.
50mm flow cell






Rinsed w/ DI water
daily.

Rinsed w/ 0.5 N
NaOH for hr.
weekly

DHMH
Calculated PP

= TP-TDP

TP method is same as
TOP (EPA 365.4,
block digestion),
using an un-filtered
sample.






























                                      16

-------















umber
Z


















a
OJ
M
cS
OJ
tf

Is
•§ .5
a ,0
•2 S
2 a
i> -,
~*^ w
•f Q
U -a
-1
it "°
1 A U e 1
ON pS3 ^p ® W
^ —
'S.S j w
^ a3 S a1
2 "S sO g
55 g 0s *•

J
^
.1
S.
0.
C/5
u

M^|
•- D 1» U
1.SI-S
PH H o»
H
Two reagents, DI,
SDS




KH2PO4 in 1 N HC1


0.185- 1.48 mg/L

0.00 12 mg/L

0.185 mg/L

0/20
(all < 8% CV)

None



1994-1998
Range - 96 - 103
Mean- 99.9
Median - 100
1997.5-1998
Range - 100 - 103
Mean- 101.2
Median -101
< 28 day sat -20 °C




1/94-12/95: Two
reagents.
1/96 on: Combined
reagent, ROW, SD-
difenyl oxide
disulfonates
KH2PO4 in 1 N HC1


0.10 -3.0 mg/L

0.00 15 -0.0034 mg/L

0.10 mg/L

0/19
(all < 12% CV)

Spex aqueous
93-104


1994-1998
Range - 92 - 1 10
Mean- 100.5
Median- 100.9
1997.5-1998
Range - 97.8 - 106.5
Mean- 102.4
Median- 102.5
< 28 day sat -20 °C




Two reagents, FFD-6
wetting agent




KH2PO4 in 1 N HC1
Made fresh daily

0.010 -0.500 mg/L

0.001 mg/L

0.010 mg/L

2/19

None



1994-1998
Range- 96-113
Mean- 102.3
Median -99
1997.5-1998
Range- 96-110
Mean- 103.8
Median- 104.5
< 28 day sat -20 °C















0.02 mg/L (.01+ .01)

Not Applicable

13/19
(7/19 > 50% CV)

Not Applicable



Not Applicable







Not Applicable




17

-------
Split Analysis Results:

1994-1998
                           ANOVA results
Effect
Rep
Lab
Date*Lab
Lab using
Date*Lab
error term
P Value
0.18
0.0001
0.0001
0.0064
LS Means Results

             Of Means

   DCLS ODU CBLDHMH
            Of Residuals

ODU CBL  DCLSDHMH
       The ANOVA results indicate that there is no replicate affect on variability within
the data meaning the splitting procedure was conducted properly. The remainder of the
ANOVA results indicate that there was a difference among labs, that this difference
varied through time, and that this difference was greater than the within run variability
associated with each lab.
       The LS means results indicate that there is good agreement among DCLS, ODU
and CBL. DHMH's PP results were consistently different from DCLS and ODU but not
from CBL. The LSM of the residuals indicate that, in terms of variability around the
mean, DHMH was significantly different from the other three labs.
                                        18

-------
     Graphical Results
                                         Mainstem PP
0.045
 0.04
            Graphical results show that of the fourteen dates when data were available for all
     four labs, all labs were different from each other on more than 50% of those fourteen
     dates.

     Discussion of Particulate Phosphorus
            The difference detected by the LS means analysis between DCLS and DHMH and
     between ODU and DHMH may be due to the fact that DHMH calculates PP. This may
     also explain DHMHs high variability as detected by the LSM of the residuals and the
     number of CVs greater than 25%. No further causes were investigated.
            While the LS means indicates good agreement among CBL, ODU and DCLS, the
     graphical analysis suggests that all  of the labs are significantly different from one
     another. With the exception of DHMH, this may be due to the labs high precision with
     this parameter (as evidenced by their CVs and the results of the LSM of the residuals)
     which causes smaller error bars and reduces the likelihood of overlap.
                                            19

-------
1997.5-1998
                           ANOVA results
Effect
Rep
Lab
Date*Lab
Lab using
Date*Lab
error term
P Value
0.7829
0.0001
0.0001
0.4746
LS Means Results

              Of Means

   DCLS  ODU  DHMH  CBL
            Of Residuals

DCLS ODU CBL DHMH
       The ANOVA results indicate that there is no replicate affect on variability within
the data meaning the splitting procedure was conducted properly. These results also
indicate that there was a difference among labs, that this difference varied through time,
but, that this difference was less than the within run variability associated with each lab.
       The LS means results indicate that there were no consistent differences among
labs. The LS Means of the residuals indicate that, in terms of variability around the mean,
DHMH was significantly different from the other three labs.
                                        20

-------
       Graphical Results
                                            Mainstem PP
   0.035
    0.03
   0.025
=-  0.02
O)
£
Q.
°-  0.015
    0.01
   0.005
                                                                                  -•*- DCLS

                                                                                      1
                              >
                             CM
              Of the four dates when data were available for all labs, all labs were different
       from each other more than 50% of the time.

       Discussion
              The results indicate that from September 1997 through December 1998, there
       were differences among labs but that these differences do not appear to be consistent over
       time. This suggests and improvement for DHMH in the more recent data, however, the
       LSM of the residuals indicates that there is still a problem with precision. This may be
       due to the fact that DHMH calculates PP.
              While the LS means results indicate that there were no consistent differences
       among labs, the graphical analysis suggests that all of the labs are significantly different
       from one another. With the exception of DHMH, this may be due to the labs high
       precision with this parameter (as evidenced by their CVs and the results of the LSM of
       the residuals) which causes smaller error bars and reduces the likelihood of overlap.
                                               21

-------
                        Parameter: Ortho-Phosphate
Labs: CBL, ODU, DCLS and DHMH




Measurement (direct/indirect): All labs measure directly




                Phosphate Method Comparison - Mainstem Labs
Variable
a! -9

vn *** ®
ta 
-------

at
es
ivJ

^
•a Jg
cs g
•a ,5
§ *
55

o
S!
u
•a
W H^
U
l\
'S ITS "S.
^ A M
s .« a
a S °
2l ^ a
2 "8 g *

OJ
a
.•i 2
& ^
5/5 8
u B
^

Two reagents, DI, SDS


KH2PO4 in DI H2O






0.00372 - 0.372


0.0006 mg/L


0.00372 mg/L
9/20
(4/20 > 50% CV)



SPEX
93-109


1994-1998
Range - 92 - 104

Mean -98.4
Median - 97.5
1997.5-1998
Range - 92 - 103
Mean- 97.3
Median - 96.5
Combined reagent, DI,
SD-difenyl oxide
disulfonates
1/94-5/97: KH2PO4 in
com-posite of filter.
sample water.
6/9 7 on: AS W salinity ~
sample. 5-6 standards &
ASW blanks for std.
curve.
0.002 - 0.08 mg/L


0.0003 - 0.003 mg/L


0.002 mg/L
2/14



SPEX
75-110(1/94-5/97)
97-102 (6/97 on)

1994-1998
Range - 86 - 103

Mean- 95.3
Median -95. 5
1997.5-1998
Range - 90-98
Mean- 97.5
Median - 96.5
Two reagents, FFD-6
wetting agent

KH2PO4 in DI H2O
Made fresh daily





0.0 10 -0.100 mg/L


0.002 mg/L


0.010
1/17



APG
77-107


1994-1998
Range - 85 - 102

Mean- 93.9
Median - 94
1997.5-1998
Range- 91-95
Mean- 92
Median- 91
Two reagents, DI?or
Dowfax2A!?

KHzP04 in DI HzO
Working stds in dem.
Water.




0.004 -0.3 mg/L


0.0012 -0.0017 mg/L


0.004 mg/L
3/18



88-109


1994-1998
Range - 78 - 102

Mean- 94.9
Median -98
1997.5-1998
Range - 92 - 100
Mean- 98.3
Median- 100
23

-------
M^ IS
!•! a *
M H o»
H
< 28 day sat -20 °C

4°C~4 hrs., frozen -
20°C
< 28 days
4°C ~ 24 hrs, frozen at
-20°C<28days
4°C<48hrs.

Split Results:

1994-1998
                           ANOVA results
Effect
Rep
Lab
Date*Lab
Lab using
Date*Lab
error term
P Value
0.8741
0.0001
0.0001
0.0066
LS Means Results

              Of Means

   DHMH DCLS CBL ODU
            Of Residuals

DCLS ODU  DHMH  CBL
       The ANOVA results indicate that there is no replicate affect on variability within
the data meaning the splitting procedure was conducted properly. The results also
indicate that there was a difference among labs, that this difference varied through time,
and that this difference was greater than the within run variability associated with each
lab.
       The LS means results indicate that DHMH and ODU were consistently different
from each other. The LS Means of the residuals indicate that, in terms of variability
around the mean, CBL was significantly different from the other three labs.
                                        24

-------
      Graphical Results
                                           Mainstem PO4f
   0.035
    0.03
   0.025
O)  °'02
E
tL  0.015
    0.01
   0.005
    --DHMH
- * - DCLS
             The graphical analysis did not reveal any pairwise differences among labs
      occurring on more than 50% of the dates where there were data available for all labs.
      Visual inspection shows that DHMH frequently has a high bias relative to the other labs
      with very large error bars.

      Discussion of Orthophosphate
             The LS means analysis indicates that there is a significant bias between ODU and
      DHMH. (This was also the case for TDP.)  The bias can be seen in the graph above
      where DHMH generally has higher values for PO4f and ODU generally has lower values.
      These graphical differences, however, are not statistically significant. These biases are
      not apparent in the percent recovery data.
             Differences in DHMH results prompted an investigation that revealed they had
      not corrected for salinity (refractive index correction). Corrections were calculated and
      applied to the database in 1998.
             ODU investigated their method but did not find any reason for their differences.
                                              25

-------
       The LSM analysis of the residuals indicate that CBL has significant variability
around the mean. This is supported by the CV where nine out of CBL's twenty
observations were greater than 25%.
1997.5-1998
                           ANOVA results
Effect
Rep
Lab
Date*Lab
Lab using
Date*Lab
error term
P Value
0.0628
0.0001
0.0001
0.0582
LS Means Results

              Of Means

   DCLS CBL  DHMH  ODU
            Of Residuals

DCLS DHMH ODU CBL
       The ANOVA results indicate that there is no replicate affect on variability within
the data meaning the splitting procedure was conducted properly. The results also
indicate that there was a difference among labs, that this difference varied through time,
but that this difference was not greater than the within run variability associated with
each lab.
       The LS means results indicate that there are no consistent differences among labs.
The LS Means of the residuals indicate that, in terms of variability around the mean, CBL
was significantly different from the other three labs.
       The graphical analysis found no differences among labs.
                                       26

-------
   Graphical Results
                                        Mainstem PO4f
0.025
          Results of the graphical analysis indicate that CBL is different from ODU on
   three out of the four dates where data were available for all four labs.

   Discussion
          The LSM analysis found no significant differences among labs in the 1997.5-1998
   data. The graphical analysis did find a significant difference between CBL and ODU.
   This is due to both labs relatively small error bars.
          The LSM of the residuals indicates that CBL still exhibits significant variability
   about the mean in the more recent data set.
                                           27

-------
                          Parameter: Total Phosphorus
Labs: CBL, ODU, DCLS and DHMH

Measurement (direct/indirect): CBL and ODU measure indirectly (TDP + PP).  See
       these parameters for method descriptions. DHMH and DCLS measure total
       phosphorus directly. (Note: TDP and PP are reported to the Chesapeake
       Information Management System.) DCLS's results from the direct measurement
       are compared below.

       DHMH analyzes total phosphorus using EPA Method 365.4. The automated,
       colorimetric method is the same as that used for Total Dissolved Phosphorus, only
       an unfiltered sample is analyzed. (See TDP section for details.) Samples are
       digested at 360 C using sulfuric acid, K2SO4and HgSC>4 for several hours in  a
       block digestor.  The residue is cooled, diluted and placed on an AutoAnalyzer
       and analyzed by the ascorbic acid method.

       DCLS analyzes total phosphorus using EPA Method 365.1. In saline waters, the
       method is the same as that used for Total Dissolved Phosphorus, only an
       unfiltered sample is analyzed. (See TDP section for details.) Samples are
       digested using a manual acid persulfate digestion because high percent recoveries
       occur in saline samples with the block digestor. Digested samples are analyzed
       on a Skalar autoanalyzer.
Split Results:
1994-1998
                          ANOVA results
Effect
Rep
Lab
Date*Lab
Lab using
Date*Lab
error term
P Value
0.9468
0.0001
0.0001
0.0002
                                       28

-------
     LS Means Results

                   Of Means

         DHMH DCLS  CBL  ODD
            Of Residuals

CBL  DCLS ODU  DHMH
            The ANOVA results indicate that there is no replicate affect on variability within
     the data, which means that the splitting procedure was conducted properly. The results
     also indicate that there was a difference among labs, that this difference varied through
     time, and that this difference was greater than the within run variability associated with
     each lab.
            The LS means results indicate that DHMH was consistently different from CBL
     and ODU and ODU was consistently different from DCLS and DHMH. The LS Means of
     the residuals indicate that, in terms of variability around the mean, DHMH was
     significantly different from CBL and DCLS.

     Graphical Results
0.18


0.16


0.14


0.12
                                          Mainstem TP
                               --*--DHMH
                               -*- DCLS
0.08
0.06
0.04
0.02
      co
f   •*    in   in
O)   O)    O)   O)
in   ^    co   oo
^   ^    ^   in
00   T-    OM
                                r-   co
                                oo   ^;
                                         co
                                         O)
                                         oo
CD   CD
O)   O)
co   in
^   co
in
              O)
              Si
              CD
O)
oo
Si
r^   oo    oo
0)0)0)
00   CO    T-
                                              29

-------
Graphical analysis shows that of the fifteen dates where data are available for all labs,
only CBL and ODU were not different than each other on more than 50% of the dates.

Discussion of TP results
       The differences between the labs detected by the LS means may be due to the fact
that CBL and ODU calculate TP and DCLS and DHMH measure it directly.
       The result from the graphical analysis that CBL and ODU are not significantly
different from one another is supported by the same result from the LS means.
1997.5-1998
                           ANOVA results
Effect
Rep
Lab
Date*Lab
Lab using
Date*Lab
error term
P Value
0.5107
0.0001
0.0001
0.1299
LS Means Results

              Of Means

   DCLS DHMH CBL  ODU
            Of Residuals

ODU CBL  DHMH DCLS
       The ANOVA results indicate that there is no replicate affect on variability within
the data, which indicates that the splitting procedure was conducted properly. The results
also indicate that there was a difference among labs, that this difference varied through
time, but that this difference was not greater than the within run variability associated
with each lab.
       The LS means results indicate that there were no consistent differences among
labs. The LSM of the residuals indicate that, in terms of variability around the mean,
there were no differences among labs.
                                        30

-------
         Graphical Results
   0.18


   0.16


   0.14


   0.12


S"  0.1


K  0.08
                                            Mainstem TP
__
                                                    	•—CBL
                                                     —•—ODU
                                                     --*--DHMH
                                                     -•*- DCLS
                             aj
                                           CO
                                           0!
                                           CO
                Of the four dates where data were available for all labs, all labs were different
         from one another on more than 50% of the dates.

         Discussion
                Although no biases were detected by the LS means test, viewing the data
         graphically indicates two potential problem areas. First, DCLS recorded drastically
         higher values for the December 1997 and March 1998 cruises. Second, DHMH's results
         appear to be consistently higher relative to the other labs.
                The results of the graphical  analysis are not supported by the results of the LS
         means.
                                                 31

-------
                     Parameter: Total Dissolved Nitrogen
Labs: CBL, ODU, DCLS and DHMH

Measurement (direct/indirect): CBL, ODU and DCLS measure directly, DHMH
      measures indirectly (TKNf + NO23)

          Total Dissolved Nitrogen Method Comparison - Mainstem Labs
Variable
a! -9
sS JJ S
5/5 £ u
Glassware
Method
Instrumentation
Inst. Maintenance
CBL
Vacuum, 0.7^m GF/F
30 ml glass test tube
Graduated cylinders.
Glass tubes cleaned w/
10% HC1, DI rinsed.
autoclaved with
potassium persulfate
before use.
Alkaline persulfate
digestion (60 @ 4psi) +
EPA 353. 2- automated
cadmium reduction
method. (Valderrama,
1981&D Eliaetal,
1977)
Technicon AAII; 550
nm filter photometer.
50 mm flow cell
Rinsed w IN HC1 for 15
min. after analysis, DI
for 15 min.
ODU
Vacuum, 0.7^m GF/F
HOPE, 250 ml
Class A volumetric.
Dedicated glass tubes.
Liquinox , w/tap water
rinse, rinsed twice w/
4N HC1 then 9 times w/
ROW.
Alkaline persulfate
digestion (autoclave
30min @ 105 EC) +
EPA 353.2 -automated
cadmium reduction
method.
1/94-12/95: SIC
continuous flow
analyzer.
1/96 on: Skalar SANplus,
540 nm filter
photometer with 620
nm background
correction. 75 mm flow
cell
Rinsed w/ ROW for 30
min. after analysis.
Weekly: Cartridge
cleaned w 1%
hypochlorite sol n for
hr. and ROW for
hour. Align flowcell.
DCLS
Vacuum, 0.7^mGF/F
HOPE
Rainin auto pipet.
Digestion tubes
autoclaved with
persulfate before use.
Dedicated glassware
washed in 1:1 HC1, DI
rinsed.
Alkaline persulfate
digestion + EPA 353.2
automated cadmium
reduction.
Skalar SANplus, 540nm
filter photometer with
620 nm background
correction.
50 mm flow cell
Rinsed w/ DI water
daily.
Rinsed w/ IN HC1
weekly, 1% hypo-
chlorite weekly
DHMH
Vacuum, 0 . 7^mGF/F
HOPE

Calculated: TKN +
(NO2 + NO3).
Not applicable
Not applicable
                                     32

-------
0)
es
ivJ


1
es
•a
•3



c
o
'•^ o
eS gjj
l|
U
•a
W H^
U
lj
•g A § •
O 	 rt SB
^3 C O
— ^ ^. S
z 
-------
0)
es
1
•a
a
a
o
eS 6X)
*• a
•^ 3
u
"8
P
u
!j
Number of
splits with >
25% CV among
replicates.
ipi
PH£
u '£, !"
Maal
1 a s,a
III3
Helium carrier gas

1.5 mg acetanilide
(10.36%N)
None: standards run as
recovery check.

0.0123 mg/L

None
0/20
(all < 4.5% CV)
None
None

< 28 day sat -20 °C

Helium carrier gas

Chloramine-T dried for
30 min. at 50°C
0.05 mg - 1.0 mg
5 pt. calibration curve

0.007 - 0.0414 mg/L

0.05 mg
2/18
None
None

< 28 day sat -20 °C



Acetanilide
None: standards run as
recovery check.

0.01 mg/L

None
0/10
None
None







0.1 14 mg/L (0.057+
0.057)
Not applicable
Not applicable
None
None

4 °C < 48 hours

39

-------
Split Results:
1994-1998
                           ANOVA results
Effect
Rep
Lab
Date*Lab
Lab using
Date*Lab
error term
P Value
0.5814
0.0001
0.0001
0.0092
LS Means Results
              Of Means

        CBLQDUDCLS
   Of Residuals

CBL DCLS ODU
       The ANOVA results indicate that there is no replicate affect on variability within
the data which would indicate that the splitting procedure was conducted properly. The
results also indicate that there was a difference among labs, that this difference varied
through time, and that this difference was greater than the within run variability
associated with each lab.
       The LS means results indicate that CBL and DCLS were consistently different.
The LS Means of the residuals indicate that, in terms of variability around the mean,
there were no differences among labs.
                                        40

-------
       Graphical Results
                                             Mainstem PN
  0.35
   0.3
  0.25
   0.2
O)
  0.15
   0.1
  0.05
             I
             Osl
             in
O)
in
in
O)
co
I   I
in   oo
to
O)
CO
                                  in
                                  oo
I
                                                            00
                                                            O)
                 en
                 co
         oo
         Si
oo
O)
                                                                                  (D
             Graphical results show that of the eight dates when were data available for all
      three labs, all labs failed the pairwise comparisons.


      Discussion of Particulate Nitrogen
             From June 1997 on, DCLS was consistently lower than CBL and ODU indicating
      the development of a negative bias. This cause of this needs to be investigated. The
      graphical analysis indicated that DCLS and CBL were significantly different from each
      other, which is supported by the LSM.
                                               41

-------
1997.5-1998
                           ANOVA results
Effect
Rep
Lab
Date*Lab
Lab using
Date*Lab
error term
P Value
0.4268
0.0001
0.1961
0.0003
LS Means Results
              Of Means

        CBLQDUDCLS
   Of Residuals

CBL DCLS ODU
       The ANOVA results indicate that there is no replicate affect on variability within
the data which would indicate that the splitting procedure was conducted properly. The
results also indicate that there was a difference among labs, that this difference was not
variable through time, and that this difference was greater than the within run variability
associated with each lab.
       The LS means results indicate that DCLS was consistently different from CBL
and ODU. The LS Means of the residuals indicate that, in terms of variability around the
mean, there were no differences among labs.
                                        42

-------
      Graphical Results
                                           Mainstem PN
O)
      00
      Si
O)
ro
OM
                                       00
                                       O)
                                 00
                                 O)
             Of the four dates when data were available for all labs, DCLS was different from
      CBL and ODU on all dates.

      Discussion
             It appears from both the LSM and the graphical analysis that DCLS has a negative
      bias. This supports the suggestion in the 1994-1998 analysis that DCLS was developing a
      negative bias.
                                             43

-------
                           Parameter: Ammonium
Labs: CBL, ODU, DCLS and DHMH




Measurement (direct/indirect): All labs measure directly




               Ammonium Method Comparison - Mainstem Labs
Variable
a! -9
vn *** ®
ta 
-------
a
o
•3 a>
2 a
3 °
u
•a
U
_-2
-1!
* > t
•s. u |
*S 5^ "3.
^ A M
5 * °
z '? 1
— !•
« -g 1 §
d^ S?
^— i x®
CSSP spike % recovery
range
B
a1^ "8
"^ fl 1>
"3 .9 a
M H g
H
0.021 -0.168 mg/L

0.003 mg/L

0.021 mg/L

7/20
(6/20 > 50% CV)



94-111

1994-1998
Range - 88 - 106
Mean- 96.4
Median - 96.5
1997.5-1998
Range- 93-99
Mean- 96.5
Median- 96.5
< 28 day sat -20 °C



0.005 -0.15 mg/L

0.0007 - 0.0025 mg/L

0.005 mg/L

4/19 (all > 50% CV)



94-110

1994-1998
Range- 91-109
Mean- 101.9
Median- 104
1997.5-1998
Range - 99 - 109
Mean- 105.8
Median - 108
4°C ~ 4 hrs., frozen at -
20°C
< 28 days

0.0 10 -0.100 mg/L

0.004 mg/L

0.010 mg/L

2/18



92-114

1994-1998
Range - 85 - 153
Mean- 106.1
Median- 105
1997.5-1998
Range- 90-110
Mean- 100
Median -100
4 ° C ~ 24 hrs, frozen at -
20°C<28days


0.008 - 0.6 mg/L

0.0015-0. 0017 mg/L

0.008 mg/L

3/19 (all > 50% CV)



94-107

1994-1998
Range- 91-106
Mean- 98.4
Median- 98
1997.5-1998
Range - 93 - 106
Mean- 100.3
Median- 101
4°Cfor48hrs.



45

-------
Split Results:

1994-1998
                           ANOVA results
Effect
Rep
Lab
Date*Lab
Lab using
Date*Lab
error term
P Value
0.0922
O.OOOl
O.OOOl
0.5056
LS Means Results

              Of Means

   CBL ODU DHMH DCLS
            Of Residuals

CBL DHMH DCLS ODU
       The ANOVA results indicate that there is no replicate affect on variability within
the data which would indicate that the splitting procedure was conducted properly. The
results also indicate that there was a difference among labs, that this difference varied
through time, but that this difference was not greater than the within run variability
associated with each lab.
       The LS means results indicate that there were no differences among labs. The LS
Means of the residuals indicate that, in terms of variability around the mean, there were
no differences among labs.
                                        46

-------
       Graphical Results
                                            Mainstem NH4
   0.25  -
    0.2
=•  0.15
    0.1
   0.05  -
in
oo
in

I
CNi
                              in
                              S>
                              S°
                              in
                 in
                 en
CD
05
CD
CO
IB
CD
S2
!Q
co
                                                                             co
                                                                                  co
                                  CO    ?
                                  d    ^
                                  CO    CD
•*
c^
       Of the 15 dates for which data from all four labs were available, no pairwise comparisons
       resulted in labs being different from one another on more than 50% of the dates.

       Discussion of Ammonium
              Although no biases or differences were detected it should be noted that on three
       dates (11/95, 9/97 and 12/97) DHMH had results that were an order of magnitude higher
       than the other labs.  The 9/97 and 12/97 results were removed from the analyses because
       all of DHMH ammonia data during that period was deleted from the data base.
                                               47

-------
1997.5-1998
                           ANOVA results
Effect
Rep
Lab
Date*Lab
Lab using
Date*Lab
error term
P Value
0.2386
0.0023
O.OOOl
0.7359
LS Means Results

              Of Means

   CBL  ODU DHMH DCLS
            Of Residuals

DCLS CBL DHMH ODU
       The ANOVA results indicate that there is no replicate affect on variability within
the data which would indicate that the splitting procedure was conducted properly. The
results also indicate that there was a difference among labs, that this difference varied
through time, but that this difference was not greater than the within run variability
associated with each lab.
       The LS means results indicate that there were no differences among labs. The LS
Means of the residuals indicate that, in terms of variability around the mean, there were
no differences among labs.
                                        48

-------
       Graphical Results
   0.06
   0.05
   0.04
O)
E.  0.03
s
   0.02
   0.01
                                            Mainstem NH4
                                          oo
                                          S2
                                          CO
                                          d
                                          CO
              There were only two dates when data were available for all four labs therefore,
       the graphical analysis could not be performed.

       Discussion
              Although no bias was detected by the LS means analysis, after review of the
       graphical analysis it would appear that DHMH had several problems in their method,
       however they could not identify the source of the differences.
                                               49

-------
                          Parameter: Nitrate + Nitrite
Labs: CBL, ODU, DCLS and DHMH




Measurement (direct/indirect): All labs measure directly




              Nitrate + Nitrite Method Comparison - Mainstem Labs
Variable
t/} "^ O
ta 
-------
Standards
a ^

.2 x
•- a
.•2 3
U a
i> -,
~*^ M
•s 3
U -a
lj
* > «
•s. u |
*S 5^ "5.
^ A M
S ^ o
|||
^
'S.S £ w
C£ £3 o JJP
2 "=4 g g
^^ ^

§JD
a
-S 2
an &•
!/5 0
u S
^

KNO3 dried at 45 °C.
NaNO2, dried at 45 °C
& preserved w CHC13.
Stds. diluted with DI
water
NO3'std> 90%NO2-
std
0.005 - 1.40 mg/L


0.0002 mg/L

0.005 mg/L
4/20 (all > 50% CV)




95-117




1994-1998
Range- 99-115
Mean- 104.4
Median - 104
1997.5-1998
Range- 100-110
Mean- 103
Median- 101.5
KNO3 dried at 103 °C.
NaNO2, dried at 103 °C
(each preserved w
CHC13) Sds. diluted
with ASW
NO3'std> 90%NO2-
std
0.003 -0.10 mg/L


0.0002 - 0.0025 mg/L

0.003 mg/L
1/19




94-102




1994-1998
Range - 94 - 107
Mean- 99.2
Median- 99
1997.5-1998
Range- 94-99
Mean- 97
Median - 97
KNO3 dried at 105 °C.
Prepared fresh daily in
DI water
NO3/NO2 std. between
95 - 105%
0.0 10 -0.4000 mg/L


0.004 mg/L

0.010 mg/L
0/15




?




1994-1998
Range -
Mean-
Median-
1997.5-1998
Range -
Mean-
Median -
KNO dried at 110°,
preserved w/CHC!3,
Stds diluted with DI
water.
NOs-NOz, 86- 114%

0.02 -2.0 mg/L


0.002 mg/L

0.02 mg/L
1/17




96-108




1994-1998
Range - 95 - 108
Mean- 102.7
Median- 102.5
1997.5-1998
Range - 99 - 108
Mean- 104.5
Median- 105.5
51

-------
s 
-------
       Graphical Results
                                           Mainstem  NO23
   1.000
   0.900  _
   0.800  _
   0.700  _
   0.600  _
O)
.§  0.500
CM
o
   0.400  _
   0.300  _
   0.200  _
   0.100  _
   0.000

              CN
              Si
                   O)
                   in
in
O)
CO
    00
    in
in
oo
r^r^    r^r^oo   oo°0oo
0)0)0)0)0)0)0)0)
^~   O)    °0   CO   CO   ^~   ^~   ^
COCD    O)   (\j   C^   ^   O)   ^~
             T-   CO   CD       OJ
              For the 15 dates for which data were available for all labs, CBL was different
       from ODU and DCLS more than 50% and ODU and DCLS were different from one
       another on more than 50% of the dates.

       Discussion
              Although differences were detected in the graphical analysis, it does not appear
       that there is an analysis problem. However, examination of the percent recovery data
       suggests that CBL and DHMH may have slight positive bias.
                                                53

-------
1997.5-1998
                           ANOVA results
Effect
Rep
Lab
Date*Lab
Lab using
Date*Lab
error term
P Value
0.1653
0.0001
0.0001
0.0984
LS Means Results

              Of Means

   DCLS CBL  DHMHODU
            Of Residuals

DCLS DHMH CBL ODU
       The ANOVA results indicate that there is no replicate affect on variability within
the data which would indicate that the splitting procedure was conducted properly. The
results also indicate that there was a difference among labs, that this difference varied
through time, but that this difference was not greater than the within run variability
associated with each lab.
       The LS means results indicate that there were no differences among labs. The LS
Means of the residuals indicate that, in terms of variability around the mean, there were
no differences among labs.
                                        54

-------
       Graphical Results
                                           Mainstem NO23
   0.8
   0.7  -
   0.6  -
   0.5  -
O)
1  0.4
«
CM
o
   0.3  -
   0.2  -
   0.1
             O)
             S5
             Si
O)
oo
csi
             Of the four dates where data were available for all labs, there were no pairwise
       comparisons where differences occurred on more than 50% percent of the four dates.

       Discussion
             There does not appear to be any analysis problems related to NO2+3. However,
       examination of the percent recovery data suggests that CBL and DHMH may have slight
       positive bias.
                                               55

-------
                              Parameter: Nitrite
Labs: CBL, ODU, DCLS and DHMH




Measurement (direct/indirect): All labs measure directly




                   Nitrite Method Comparison - Mainstem Labs
Variable
« g |
"I1 'Is ag 'S
j/j S O
1
|
sS
3


•a
o
'S
S


=
_o
'-£
"a
£
g
1

OJ
a
"S a
"1
§
P— i
CBL
Vacuum, 0.7^m GF/F in
field, triplicate
polystyrene AA cups
Cleaned w/ 10% HC1,
DI rinsed


Automated colorimetric,
diazotization, EPA
353.2.




TrAAcs-800; 520 nm
filter photometer, 37 C
heating bath
50 mm flow cell



Rinsed w IN HC1 for 15
min. after analysis,
DI for 15 min.



ODU
Vacuum, 0.7^m GF/F in
field
HOPE
Liquinox , w/ tap water
rinse , rinsed twice w 4N
HC1 then 9 times w/
ROW.
1/94-5/97: Manual
colorimet. diazotization,
EPA 353. 3.
6/97 on: Auto.
colorimetric,
diazotization, EPA
353.2.
1/94-5/97:
spectrophotometer
6/P7o«:SkalarSANplus,
540 nm. Auto
background/ matrix
correct (620 nm filter)
75 mm flow cell
Daily: Rinse w/ ROW
for 30 min. Weekly:
Clean cartridge w 1%
hypochlorite ( hr),
ROW rinse ( hr). Align
flow cell.
DCLS
Vacuum 0.7^m GF/F in
field, HOPE.

Washed in 1:1 HC1, DI
rinsed.


Automated, colori-
metric, diazotization
EPA 353.2.




Skalar SANplus, 540nm.
Auto background/
matrix correction w/ 620
nm filter.
50 mm flow cell


Rinsed w/ DI water
daily.
Rinsed w/ IN HC1
weekly and 1%
hypochlorite weekly

DHMH
Vacuum, 0.7^m GF/F in
field, HOPE

Haemo-sol,w/tap water
rinse, DI rinse.


Automated, colori-
metric, diazotization
EPA 353.2.




Alpkem model 3570
with SoftPac software.
540nm filter, 5 mm
flow cell



Rinse w/DI for 10 min,
rinse w/ 2% HC1 30
sec, rinse w/DI 20 min,
rinse w/O.lNNaOH 60
sec, rinse w/DI for 20
min.
                                      56

-------
Reagents
Standards
=
1 a
la
u
1
u
~-g
M
-A g .
£e||
pC ^ ^ U
£
fi "=3 « M
•0 « | §
^ 3 s !"
£
g

||
a.
0.
^^
^^
U
Separate color reagents:
sulfanilamide, N-l-
naphthylethylenediamin
e dihydrochloride w
Brij-35
NaNO2, dried at 45 °C
& preserved w CHC13
Stds. diluted with DI
water.
0.0028-0.042 mg/L

0.0003 mg/L

0.0028 mg/L

3/20

None



1994-1998
Range - 96 - 105
Mean- 99.1
Median- 99
1997.5-1998
Range - 98 - 102
Mean- 99.8
Median- 100
Combined color
reagent: (sulfanilamide
&N-1-
naphthylethylene-
diamine di-
hydrochloride) w Brij-
35
NaNO2, dried at 103 °C
& preserved w CHC13.
Standardize stock
monthly. Stds. diluted
with ASW.
0.001-0.040 mg/L

0.0002-0.0010 mg/L

0.001 mg/L

1/20

None



1994-1998
Range - 95 - 108
Mean- 99.8
Median- 99
1997.5-1998
Range - 98 - 100
Mean- 99.3
Median- 99.5
Combined color
reagent: (sulfanilamide
& N-1-naphthylethyl-
enediamine dihydro-
chloride) w Brij-35, DI
water w Brij-35
NaNO2, dried at 103 °C.
Made fresh daily in DI
water.
0.0 10 -0.100 mg/L

0.002 mg/L

0.010 mg/L

0/16

None



1994-1998
Range- 70-114
Mean- 100.1
Median- 100
1997.5-1998
Range - 70 - 105
Mean- 94.3
Median- 101
Color Reagent (sulfa-
nilamide, N-1-naphth-
ylethylene-diamine di-
hydrochloride) , brij -35 ,
NHCl
KNOz dried at 110°C,
preserved w/CHCLs,
Stds diluted w/DI
water.
0.002 -0.200 mg/L

0.02-0.002 mg/L

0.002 mg/L

1/20

95-109



1994-1998
Range - 98 - 108
Mean- 102
Median- 102
1997.5-1998
Range - 98 - 102
Mean- 100.7
Median- 101
57

-------
B
"^ fl 1>
"3 .9 a
M H g
H
< 28 day sat -20 °C


4°C ~ 4 his., frozen -
20°C
< 28 days

4°C~24hrs, frozen at -
20°C<28days

4°Cfor48hrs.


Split Results:
1994-1998
                           ANOVA results
Effect
Rep
Lab
Date*Lab
Lab using
Date*Lab
error term
P Value
0.4454
0.0001
0.0001
0.0600
LS Means Results

              Of Means

   DHMH  DCLS CBL ODU
            Of Residuals

DCLS ODU  DHMH CBL
       The ANOVA results indicate that there is no replicate affect on variability within
the data which would indicate that the splitting procedure was conducted properly. The
results also indicate that there was a difference among labs, that this difference varied
through time, but that this difference was not greater than the within run variability
associated with each lab.
       The LS means results indicate that there were no differences among labs. The LS
Means of the residuals indicate that, in terms of variability around the mean, there were
no differences among labs.
                                        58

-------
      Graphical Results
                                            Mainstem NO2
   0.035
    0.03
   0.025
_  0.02
^
O)

tM
z  0.015
    0.01
   0.005
      CBL
      ODU
    - -DHMH
- -X — DCLS
          o)O)a)a)O)O)O)a)Oo)O)a)Oooo)
                   l£>   "fr   CO   00
             Of the sixteen dates where there were data available for all four labs, all pairwise
      comparisons resulted in differences occurring in more than 50% of the dates.

      Discussion
             Although all labs failed the graphical analysis, it does not appear that an analytical
      problem with NO2 exists. Failure of the graphical analysis was apparently due to each
      labs small error bars.
                                                59

-------
1997.5-1998
                           ANOVA results
Effect
Rep
Lab
Date*Lab
Lab using
Date*Lab
error term
P Value
0.4758
0.0001
0.0001
0.4277
LS Means Results

              Of Means

   DCLS  CBLDHMH ODU
            Of Residuals

DHMH  ODU DCLS CBL
       The ANOVA results indicate that there is no replicate affect on variability within
the data which would indicate that the splitting procedure was conducted properly. The
results also indicate that there was a difference among labs, that this difference varied
through time, but that this difference was not greater than the within run variability
associated with each lab.
       The LS means results indicate that there were no differences among labs. The LS
Means of the residuals indicate that, in terms of variability around the mean, there were
no differences among labs.
                                        60

-------
         Graphical Analysis
                                            Mainstem NO2
   0.009
   0.008
   0.007
   0.006
=•  0.005
S!
Q  0.004
   0.003
   0.002
   0.001
                                           oo
                                           S2
                                           CO
                                           C!
                                           CO
oo

55
                For the four dates where data for all labs were available, all labs failed the
         pairwise comparisons.

         Discussion
                Although all labs failed the graphical analysis, this was most likely due to small
         error bars and not indicative of a problem. More importantly, no labs were consistently
         higher or lower than one another (i.e. no bias) and, therefore, there do not appear to be
         any analysis issues with NO2.
                                                  61

-------
                      Parameter: Total Suspended Solids
Labs: CBL, ODU, DCLS and DHMH




Measurement (direct/indirect): All labs measure directly




           Total Suspended Solids Method Comparison - Mainstem Labs
Variable

a
o
•a —
2 s
S es
S a

I"
cS

0) M
C .a
1 1
s§ y
3


•a
o
"S




Is!
-5 B g
W 5^ ^M
9 M 13


OJ
sS
•s g
B +*
g

CBL
0.?A«n GF/F filters dried
103-105 °C overnight,
pre-weighed to 10"4 g.
250-500 mL sample are
field vacuum filtered in
duplicate, DI rinsed.
Filters kept in Al
pouchs.

100 & 250 mL plastic
graduated cylinders,
aluminum weighing
pans.

Solids dried at 103-105
°C overnight, desiccated
& weighed to 10"4 g.
Some re-dried at 103 °C
to constant wt (±
0.5mg). Duplicates
averaged. Std. Meth.
2540D


with auto data entry



Service check





ODU
0.7 Aim GF/F filters
rinsed 3X ROW, dried at
103-105 °C for > Ihr.,
pre-weighed to constant
10"4 g. 250-1000 mL
sample are field vacuum
filtered, ROW rinsed.
Filters kept in plastic
holders.
500 mL & 1 L plastic
grad-uated cylinders.
Liquinox , w/ tap water
rinse, rinsed 2X w/ 4N
HC1 then 9X w/ ROW.
Solids dried at 103-105
°C for > 1 hr., desiccated
& weighed to 10"4 g. All
samples re-dried at 103
°C to constant wt. (±
0.5mg)


Satorius series MCI,
modelRC210S
Daily check w Class S
0.1 g Monthly check
with range of Class S
weights.
Balances are serviced at
least annually by a
qualified service
engineer. Class S
weights re-certified
annually
DCLS
0.7 Aim GF/F rinsed 3X
w 20 mL DI, dried at
105°Cfor30min,
muffled at 550 °C for 15
min. Lab filters
sufficient vol. to obtain
10 - 200 mg of residue.















AT 261 Delta Range-
Mettler
AT 250 - Mettler.



DHMH
1.5 Aim GF/F rinsed 3X
w 20 mL DI, dried at
103-105 °C for >1 hr,
pre-weighed to 10"4 g.
Lab filters 50-250 mL
sample, DI rinsed.



100 mL plastic
graduated cylinder



Solids dried at 103-105
°C for 1 hr, desiccated
& weighed to 10"4 g. All
samples re-desiccated to
constant weight. (±0.5
mg)


Mettler Toledo model
AG 204. Daily internal
calibration check, Class
S weights weekly.


Balance serviced every
two years by a qualified
service engineer.



                                     62

-------
—
"S.
s
ft
Ifl
u
0-
•a
•S*
"3
U
5B K^ 5ft
-*--» ^ ^
'•3.U t3
£ S &
o >r> a.
i. <* a
^ A M
B t« °
Z s*
- 8?
*s !S ^ u
rt 'S o **
^ "S OJ SS
li •—
'S. w a»
* 8 a*

ll2
0) aj
.§ 3
hn ~^^
s^l
2 S
O O
M H
1 rep every 10 samples



2A mg/L


0/20







None



Not Applicable


< 28 days at -20°C





1 field blank per 10
samples
1 replicate every 10
samples
Quarterly SRM
1.2 -3.3 mg/L


1/18







None



Not Applicable


4°C ~ 4 his., frozen -
20°C
< 28 days



1 rep every 10 samples



3.0 mg/L


5/20







None



Not Applicable


4°C<7days





1.0 mg/L


2/7







None



Not Applicable


4°C<7days





Split Results:
1994-1998
                           ANOVA results
Effect
Rep
Lab
Date*Lab
Lab using
Date*Lab
error term
P Value
0.2529
0.0001
0.0001
0.0001
                                        63

-------
       LS Means Results
                    Of Means

          DCLS CRT>  OPT I DHMH
                                     Of Residuals

                         CBL ODU DCLS DHMH
              The ANOVA results indicate that there is no replicate affect on variability within
       the data, which would indicate that the splitting procedure was conducted properly. The
       results also indicate that there was a difference among labs, that this difference varied
       through time, and that this difference was greater than the within run variability
       associated with each lab.
              The LS means results indicate that DCLS was consistently different from all other
       labs and that CBL and DHMH were consistently different from one another. The LS
       Means of the residuals indicate that, in terms of variability around the mean, there were
       no differences among labs.

       Graphical Results
                                            Mainstem TSS
   25
   20
   15
O)

tfl
V)
   10
                                                                                  CBL
                                                                                  ODU
                                                                           --•ic- -DHMH

                                                                           —-X-  DCLS
       [^
       s>
I
CN
2?
                O)
                2?
in
O)
Si
    00
    55
<£>   10
g>   O
C;   ro
00   r:
O)
Si
    SG
    oo
9?
iS
oo
Si
r^   oo
O)   O)
S5   S
Si   C!
T-   CO
                                                                                 00
                                                                                 O)
                                                                                      O)   i;
                                                                                          OM
                                               64

-------
       Graphical analysis show that of the dates when data were available for all labs,
only the pairwise comparisons between CBL and DCLS were not different in excess of
50% of the dates.

Discussion of Total Suspended Solids
       It appears, from the results of the LSM, that DCLS has a negative bias and,
relative to DCLS and CBL, DHMH has a positive bias. The graphical results also support
this conclusion. The positive bias was probably due to DHMH not redrying TSS samples
to a constant weight. Redrying was initiated in May 1998. The negative bias attributed to
DCLS needs to be investigated.
1997.5-1998
                           ANOVA results
Effect
Rep
Lab
Date*Lab
Lab using
Date*Lab
error term
P Value
0.2192
0.0001
0.0039
0.0030
LS Means Results

              Of Means

   DCLS CBL  ODU DHMH
            Of Residuals

CBL DCLS ODU DHMH
       The ANOVA results indicate that there is no replicate affect on variability within
the data which would indicate that the splitting procedure was conducted properly. The
results also indicate that there was a difference among labs, that this difference varied
through time, and that this difference was greater than the within run variability
associated with each lab.
       The LS means results indicate that DHMH was consistently different from all
other labs and that ODU and DCLS were consistently different from one another. The LS
Means of the residuals indicate that, in terms of variability around the mean, there were
no differences among labs.
                                        65

-------
      Graphical Results
                                          Mainstem TSS
   18
   16
   14
   12
=•  10
O)

V)
w   8
                          >
                          CM
CO
C!
CO
             Of the four dates when data was available for all labs, only the pairwise
      comparisons between CBL and ODU were not different on more than 50% of the dates.

      Discussion
             It appears, from the results of the LSM, that DCLS has a negative bias and,
      relative to DCLS and CBL, DHMH has a positive bias. The graphical results also support
      this conclusion. The positive bias was probably due to DHMH not redrying TSS samples
      to a constant weight.  Redrying was initiated in May  1998. The negative bias attributed to
      DCLS needs to be investigated.
                                             66

-------
                        Parameter: Particulate Carbon
Labs: CBL, ODU, DCLS and DHMH

Measurement (direct/indirect): CBL, ODU and DCLS measure directly, DHMH
      measures indirectly (TOC - DOC).

             Particulate Carbon Method Comparison - Mainstem Labs
Variable

=
0
1
ft
§"
a.
a
1
M
—
1"
sS
5/5



•a
o
f
S


e
"S3
"a
OJ
g
a
HH
u
a

a +*
'53

CBL
1) 25 mm GF/F muffled
at 550 °Cfor90min.

2) Particulates are field
filtered in duplicate,
placed in Al foil pouch.
3) Filters dried at 45 °C
overnight.





Filters & Al capsule
placed into nickel
sleeves & combusted at
975 °C. CxOxcmpds
are reduced to CO2(g).


Exeter CE-440
Elemental Analyzer w
Cu reduction column,
He carrier gas &
thermal conductivity
detector.

Columns renewed after

300-600 samples



ODU
1) Filters & glass vials
are muffled at 550C for
15 min& 4 hrs
respectively.
2) Particulates are field
filtered (< 50 mL
sample) on 13 mm GFF,
placed in glass
scintillation vials.
3) Filters dried at 50 °C
over-night, dessicated..
After 6/97, no
chloroform/methanol
cleaning of tin sample
cups
Filters placed into tin
sample cups are flash
combusted at 1040 °C.
A series of catalytic and
Cu reducing reactors
convert Cx Ox cmpds to
C02(g).
Carlo Erba C/N gas
chromat-ograph
equipped with
combustion & Cu
reduction columns, He
carrier gas & a thermal
conductivity detector.
Both columns renewed

after 300-600 samples


67
DCLS (after 2/95)
1) Filter prep?

2) Particulates are field
filtered (25-250 ml
sample) on 25 mm GFF,
3) Filtlers dried over-
night at 50° C
Sample cup precombust
at875°Cforlhr.




Combusted at 990 °C






Exeter Model CE-440
Elemental Analyzer, Cu
reduction column, He
carrier gas & thermal
conductivity detector.








DHMH
Calculated PC

= TOC -DOC



0.5 L polyethylene
cubitainer


Std. Methods 53 10B,

Combustion Infrared






















-------
4>
M
eS
ivJ
1
•a
c
55
c
o
'•^ o
cs e*
l|
U
•a
_4>
"3
U
l\
i>
S"sO M J
O IT) S "S
^ N §•-
_n A S "e
•D CS T
11 *
Z ^
•8.| j ^
55 g 0s *"
^ £
C/5 OJ >•
(S> -± O
0) 4)
| |
2 1"
O O
M H
Helium carrier gas
1.5 mg acetanilide
(71.09%N)

None: standards run as
recovery check.

0.0759 mg/L

None
0/20
(all < 6.3% CV)

None

None

< 28 day sat -20 °C

Helium carrier gas
Chloramine-T dried at
50°Cfor30min.

0.05 mg- 1.0 mg
5 pt. calibration curve

0.06 15 -0.196 mg/L

0.05 mg
1/18


None

None

< 28 day sat -20 °C


Acetanilide

None: standards run as
recovery check.

0.1 mg/L

None
7/11


None

None

< 28 day sat -20 °C




1.0 mg/L (0.5 + 0.5)

Not applicable
3/12


91-105 (TOC)

68-129 (DOC)

4°C< 48hrs.

68

-------
Split Results:
1994-1998
                           ANOVA results
Effect
Rep
Lab
Date*Lab
Lab using
Date*Lab
error term
P Value
0.3164
0.0001
0.0001
0.0006
LS Means Results

              Of Means

   DCLS DHMH ODD CBL
            Of Residuals

CBL ODU DCLS DHMH
       The ANOVA results indicate that there is no replicate affect on variability within
the data, which would indicate that the splitting procedure was conducted properly. The
results also indicate that there was a difference among labs, that this difference varied
through time, and that this difference was greater than the within run variability
associated with each lab.
       The LS means results indicate that DCLS and DHMH were consistently different
from ODU and CBL. The LS Means of the residuals indicate that,  in terms of variability
around the mean, there were no differences among labs.
                                        69

-------
      Graphical Results
   2.5
                                         Mainstem PC
   1.5
O)

o
Q.
   0.5
            CN
            in
                O)
                in
in
O)
co
in
O)
oo
in
                                 co
5
CO
CO
O)
00
CD
O)
CO
CD
O)
in
00
                                                          00
O)
Si
CD
O)
oo
Si
O)
ro
CN
00
O)
00
CN
CO
                                                                               00
                                                                               O)
                                                                               CD
             Graphical analysis shows that of the nine dates when there were data available for
      all labs, all labs failed the pairwise comparisons.

      Discussion of Particulate Carbon
             It appears that DCLS has a negative bias and since mid 1997 DHMH has
      developed a negative bias. DCLS also had a negative bias in the paniculate nitrogen
      method that is analyzed simultaneously. This needs to be investigated.
             DHMH reviewed their TOC and DOC methods in December 1998 and found the
      cause of high DOC results and subsequent low PC results.  Further improvements to their
      method will be implemented in 1999.
             The difference detected between CBL and ODU in the graphical analysis is
      apparently due to both labs small error bars for this parameter.
                                              70

-------
1997.5-1998
                           ANOVA results
Effect
Rep
Lab
Date*Lab
Lab using
Date*Lab
error term
P Value
0.6652
0.0001
0.0001
0.0269
LS Means Results

              Of Means

   DHMH DCLS CBL ODU
            Of Residuals

CBL ODU DCLS DHMH
       The ANOVA results indicate that there is no replicate affect on variability within
the data which would indicate that the splitting procedure was conducted properly. The
results also indicate that there was a difference among labs, that this difference varied
through time, and that this difference was greater than the within run variability
associated with each lab.
       The LS means results indicate that there were no consistent differences among
labs. The LS Means of the residuals indicate that, in terms of variability around the mean,
there were no differences among labs.
                                        71

-------
      Graphical Results
                                            Mainstem PC
   2.5
   1.5
1
O
a.
   0.5
             oo             oo
             05             CM
             Of the five dates for which data were available for all labs, all labs were different
      from one another on more than 50% of the dates except CBL and ODU.

      Discussion
             Although the LSM analysis did not detect any significant differences in the 97.5 -
      98 data, looking at the graph of the data, it does appear that both DCLS and DHMH have
      a negative bias.
                                              72

-------
                                  Parameter: Silica
Labs: CBL, ODU, DCLS and DHMH

Measurement (direct/indirect): CBL, ODU and DHMH measure directly, DCLS measures
       SiO2(Si = (SiO2/2.14))

                         Silica Method Comparison - Mainstem Labs
Variable
Sample
Filtration
&
Container
Glassware
Method
Instrumentation
Inst.
Maintenance
CBL
Vacuum, 0.7^m GF/F in
field, triplicate
polystyrene AA cups.
Low silica glassware
Cleaned w/ 10% HC1,
DI rinsed
EPA 366.0 Automated
molybdenum blue
method. Blue color is
formed by the reduction
of silicomolybdate and
ascorbic acid in acidic
conditions. Oxalic acid
elim. PO4= interference.
Technicon TrAAcs-800;
800 nm filter photometer
37 °C Heating Bath
50 mm flow cell
Rinsed w/ DI and SDS
for 10 min. after
analysis
ODU
Vacuum, 0.7,wm GF/F in
field
125 mL HOPE. (TSS
filter preparation.)
All plastic, except
pipets. Liquinox , w/
tap water rinse, rinsed
twice w/ 4N HC1 then 9
times w/ ROW.
EPA 366.0 Automated
molybdenum blue
method. Blue color is
formed by the reduction
of silicomolybdate and
ascorbic acid in acidic
conditions. Oxalic acid
eliminates PO4=
interference.
1/94-12/95: SIC
continuous flow analzer
1/96 on: Skalar SANplus,
810nm w auto
background/ matrix
correct (lOlOnm filter)
75 mm flow cell
Rinsed w/ ROW for 30
min. after analysis.
1/wk: Cartridge cleaned
w/ 0.5 N NaOH for
hr, ROW for hr, flow
cell aligned.
DCLS
Vacuum 0.7^tm GF/F in
field, HOPE.
Plastic/Nalgene used
wherever possible.
Washed in 1:1 HC1,
rinsed with DI water.
EPA 3 70.1 Automated
molybdate/ascorbic
acid. Blue color formed
by silico-molybdate +
ascorbic acid in acidic
cond. Oxalic acid elim.
PO4= interference.
Technicon AA II
660 nmwith 15 mm
flow cell
Rinsed w/ DI water
daily.
Rinsed w/ 0.5 N NaOH
for hr. weekly
DHMH
Vacuum, 0.7^tm GF/F in
field, HOPE
Hemo-Sol &
demineralized H2O.
Plastic is used wherever
possible.
EPA 3 70.1 Automated
molybdate/ascorbic
acid. Blue color formed
by silico-molybdate +
ascorbic acid in acidic
cond. Oxalic acid elim.
PO4= interference.
Technicon AA II w 15
mm x 2.0 mm flowcell
Rinsed w/DI for 30-45
min. after analysis
                                         73

-------

•8
0)
es
ivJ

£ 3
•§ I
§ •<=
55^
=
•^ O
54 6X1
II
u
1
u
-•2
11
i>
o vi S 55
!_ fS ® U
1 A ||

s •"?
Z £
*2 t?«
2 -S S *
.3
a
^ M
ll
a.
0.
^^
^^
U
Oxalic, H2SO4, ascorbic
acids,
ammon. molybdate .
Isopropanol baseline
sol n, NaOH wash
water. SDSinH2SO4
Na2SiF6 dried® 45 °C
in DI H2O


0.28 1-2. 10 mg/L



0.01 mg/L

0.281 mg/L

1/20





None

1994-1998
Range- 91-97
Mean- 93.9
Median- 93.5
1997.5-1998

Range- 91-96
Mean- 93.5
Median - 94
Oxalic, H2SO4 &
ascorbic acids, ammon.
molybdate.
ASW wash water.
FFD6inH2SO4&
ascorbic acid
Sodium metasilicate
nona-hydrate (Na2SiO3 -
9H2O) in ASW matrix
water.
0.023 -1.169 mg/L



0.0000 - 0.0013 mg/L

0.002 mg/L

1/19





None

1994-1998
Range- 81-109
Mean- 98.4
Median- 99
1997.5-1998

Range - 88 - 103
Mean- 95
Median -95. 5
Oxalic, H2SO4 &
ascorbic acids, ammon.
molybdate.
DI water wash
steol wetting agent
Na2SiO3-9H2OinDI
H2O


0.1 -10.0 mg/L



0.1 mg/L

0.1 mg/L

11/19





84-112

1994-1998
Range - 93 - 102
Mean- 99.5
Median - 100
1997.5-1998

Range- 93-101
Mean- 98.3
Median - 99.5
Oxalic acid, H2SO4,
ascorbic acid,
ammonium molybdate.


Sodium Silicate (Fisher)



1-5 mg/L



0.000- 0.10 mg/L

Img/L

0/18





100-105

1994-1998
Range- 90-101
Mean- 95.7
Median - 95
1997.5-1998

Range- 91-100
Mean- 96.8
Median- 97.5
74

-------
s
e* -S
is a b «
^ .9 o,
n H s
H
< 28daysat4°C




< 28daysat4°C




<28daysat4°C




<28daysat4°C




Split Results:
1994-1998
                           ANOVA results
Effect
Rep
Lab
Date*Lab
Lab using
Date*Lab
error term
P Value
0.1553
0.0001
0.0001
0.1256
LS Means Results

              Of Means

   CBL ODU DCLS DHMH
            Of Residuals

DHMH ODU CBL DCLS
       The ANOVA results indicate that there is no replicate affect on variability within the data
which would indicate that the splitting procedure was conducted properly. The results also
indicate that there was a difference among labs, that this difference varied through time, but that
this difference was not greater than the within run variability associated with each lab.
       The LS means results indicate that there were no differences among labs. The LS Means
of the residuals indicate that, in terms of variability around the mean, there were no differences
among labs.
                                           75

-------
   Graphical Results
                                              Mainstem Si
   1.8
   1.6  -
   1.4  -
   1.2
O)
W  0.8  -
   0.6
   0.4  -
   0.2
O)
Si
                          in
                          O)
                          ro
                  C;   ro
                  oo   r:
O)
ro
                           r-
                           O)
                                                              00
                                                              O)
00
O)
                                                                  9?
                                                                  (3
                      00
                      Si
          Of the seventeen dates for which data were available for all labs, all labs failed the
   pairwise comparisons.

   Discussion
          The does not appear to be an analysis issue with Si. The failure of all labs in the graphical
   analysis is due to the small error bars. The spike recovery data indicates that CBL may have a
   negative bias with this parameter.
                                                76

-------
1997.5-1998
                           ANOVA results
Effect
Rep
Lab
Date*Lab
Lab using
Date*Lab
error term
P Value
0.3675
0.0001
0.0001
0.2181
LS Means Results

              Of Means

   DHMH DCLS CBL ODU
            Of Residuals

DHMH ODU DCLS CBL
       The ANOVA results indicate that there is no replicate affect on variability within the data
which would indicate that the splitting procedure was conducted properly. The results also
indicate that there was a difference among labs, that this difference varied through time, but that
this difference was not greater than the within run variability associated with each lab.
       The LS means results indicate that there were no differences among labs. The LS Means
of the residuals indicate that, in terms of variability around the mean, there were no differences
among labs.
                                           77

-------
Graphical Results
                                           Mainstem Si
1.6
1.4  -
1.2  -
0.8
0.6
0.4  -
0.2  -
                                              CBL

                                              ODU
                                            - -DHMH

                                        —-X- DCLS
                         r-
                         O)
          00
          Si
00
O)
(55

-------
                              Parameter: Chlorophyll
Labs: CBL, ODU, DCLS and DHMH

Measurement (Spectrophotometric/fluorometric): ODU, DCLS and DHMH measure
      spectrophotometrically; CBL measures fluorometrically.

                   Chlorophyll Method Comparison- Mainstem Lab

Variable









5«
£
M
s
•o
%
2
a.
"O
"3
u-







W Bfl
U 8X1
S 5<5 S3
a | -c |
o «(• e fe
Rll°
— ffi
CBL
(Fluorometric)
NA






















Not applicable




ODU
Samples collected in 1L
brown HOPE bottles.
Each bottle is sample
rinsed (3x). MgCO3 is
immediately added (1ml
per 1L sample). Filter
pad moistened w/ DI.
Grad. cyl. Rinsed 3x
with DI and 2x with
sample after inverting
sample 20x. Sample
inverted again. Vac 12
psi. Filtration time
limited to 5 min. and
generally only 300-500
mis of water filtered,
depending on water
turbidity. Filter folded in
half and placed in foil
and frozen immediately
or placed on ice and
frozen ASAP.
Field Filtered.
If 750 nm absorbance
>0.007, re-centrifuged
for 5 min at 2300 rpm


DCLS
Samples received day
after collection at 4° C in
opaque bottles w/
MgCO3 added. Sample
filtered immediately in
semi darkness @ <2.9
psi. Amount filtered
determined by color of
filter and turbidity of
water. Filter folded,
stored in glass tubes and
frozen until extraction
(next day at the latest)

Lab Filtered.








If 750 nm absorbance
>0.005 AU, sample
filtered through glass
fiber syringe filter

DHMH
Samples collected in
sample rinsed plastic
containers. Container is
vigorously shaken prior
to filtration and graduate
cylinder is sample
rinsed. Sufficient vol.
(100-1500 ml) is filtered
to solidly color the filter
pad. Vac. pressure <4.9
psi To the last 25 ml
filtered, ~1 ml of
concentrated MgCO3 is
added. Filter pad is
folded in half and placed
in foil pouch and stored
on ice until they reach
the field office where
they are frozen.

Field Filtered.


If the 750 nm
absorbance is >0.005, re-
centrifuged.

                                       79

-------





K
M «
••sf
•si
o £






C
.2
sS
O
g
JO
'3



o
o
"S
o


s
s
H
°8 ^
•a -g •—
lie
£ "S *
O w ^
!. a,
1^
"3 N
U X
-^i OJ 1>
S **
*« S 2 9J
"K "3 S S

,
o
.= •-
§"-5
I*
Filter pad is briefly
thawed, placed in a 15
mL glass centrifuge tube.
10 mL 90% acetone is
added, pad is ground
against the side of the
tube using a pestle.









3-4 drops 5%HC1. Final
Normality = 0.018-0.022



Baker Analyzed ACS
Reagent Grade (98.8%
Acetone) diluted to 90%
with water.




15 mL glass centrifuge
tubes

refrigerated overnight




Whatman OFF, 47mm,
0.7 |jm
NA

Known volume of
acetone added.


Sequoia Turner
Fluorometer Model 112;
Turner Designs Model
TD700
Pad placed in grinding
tube, 3-4 ml 90%
acetone added, pad
ground at -500 rpm with
a tissue homogenizer. A
TFE-fluorocarbon to
glass pestle is used to
fully macerate pad and
cells







2 drops IN HC1; wait 1
min but no longer than
two to take final reading.
Final Normality = 0.02

Baker Analyzed HPLC
solvent (99.7%) diluted
to 90% with ultrapure
water. 5 drops IN
sodium bicarb added per
liter.


glass


overnight at 4 ° C; or
frozen @ -20 to
-70 ° C until analysis


4.25 cm Whatman GF/F
0.7 |jm
1.0cm

Subtract filter/sediment
plug volume from total
volume in centrifuge
tube.
Perkin-Elmer Model
559A dual beam spec
l.Onm band pass

Pad placed into Pyrex
tube. 2.0 ml aqueous
acetone added to tubes
and filter is ground for 1
min w/ Teflon pestle
@500 rpm. 8.0 ml
aqueous acetone used to
rinse pestle into tube.
Tubes capped and
shaken and placed in
chilled ultrasonic bath
for 5 min. Tubes are
mixed for 10 sec in
vortex mixer, placed in
light proof box and
frozen until analysis
150 nlO.lNHCl, mixed
w/ thin tube disp.
pipette, 90 s wait. Final
Normality = 0.03

Fisher OPTIMA grade
(HPLC/Spec and Gas
Chrom grade); 100 ml
DI and enough acetone
to make 1000 ml
solution


16X150 Pyrex, washed
then rinsed w/ acetone

frozen up to 1 week




47 mm Whatman GF/F
0.7 |jm
2.0cm

Record exact amount of
acetone added for
grinding and extraction.

VarianMS-200@2.0
nm bandwidth


Filter pads removed
from freezer and allowed
to warm for 10 min. Pad
placed in tissue grinder,
2-3 ml 90% acetone
added. Sample ground
for 2-3 min until
homogenous, quant.
transfer to cent, tube w/
acetone, transfer rinses
to cent, tube, add
acetone until vol is 15
ml. Capped tubes store
in freezer.


3 drops IN HCL,
inverted to mix, 90 s.
wait. Final Normality =
0.011

Spectranalyzed acetone
and certified ACS
Sodium bicarbonate.
90% solution prepared
by adding 40 ml DI to
3600 ml acetone.
Solution buffered w/ 2ml
IN sodium bicarb.
polypropylene, 15 ml,
screw cap, acetone
resistant
frozen




Whatman OFF, 47mm,
0.7 |jm
5 cm

14mL - Bring acetone
volume up to 15 mL,
subtract 1 mL to account
for volume of filter.
DU-65 Beckman;
Bandwidth resolution is
2 nm from 200-600nm

80

-------

I W
Ji u a «
ri« M •• HM
aw g 3
iPl
111 3
<*> g




C
O
1

s
u
13
U
&.
££
•c 5
"S w
So
u
%
+- 2
pS -5
.sr-a
J §
u







Fluorometric









5 min at 1760 rpms,
rinse down tubes, then
centrifuge again


no light in hood where
analysis is conducted but
regular lighting in lab.
Keep extracted samples
in a box.
monthly verification of
wavelength accuracy
using NIST SRM
(holium oxide) filters.
Periodic evaluation of
slopes of calibration
curves.
Ca = Chlorophyll corrected for
PheoDhvtin ( ue/l=r26.7(abs664nm-
abs665nni)]xtvol(mr)
samp vol L

abs664 = optical density before
acidification
abs665 = optical density after aciiication
xtvol = extract volume
samp vol = sample volume

20 min at 2300rpm at 4°
C


Lights are not dimmed,
samples are kept in a
cooler on ice


Major parameters of
instrument performance
checked monthly.
Absorbance verified
weekly


Chlaftng/m3)=26.7fOD664b -
V2xL

where:
OD665a = optical density after acid
OD664b= optical density before acid
VI = volume of extract, L
V2 = volume of sample, m3


-500G for 20 min @
room temp


Subdued light; light
proof box in freezer.



Light bulb replaced as
needed. Manufacturer
called for major
problems



Chla(mg/m3)=
26.7(OD664b -OD665a)xVl
V2xL
where:

OD665a = optical density after acid
OD664b= optical density before acid
VI = volume of extract, L
V2 = volume of sample, m3
L = light path length??

1st time - after extraction
- 30min3000rpm
2nd time - before
analyzinglS min 3000
rpm
Subdued light in work
area, no direct light
exposure, samples
covered w/ aluminum
foil
A Note on Chlorophyll
      In the 1994-1998 period, chlorophyll was not measured by enough of the labs
participating in the mainstem split to conduct a split sample analysis. In 1997, a disparity was
observed between the chlorophyll values measured by the Maryland Department of Health and
Mental Hygiene MDMH) and the Academy of Natural Sciences (ACNATSCI). The Academy
conducts in vivo fluorescence monitoring in the mainstem of the Chesapeake Bay. As part of the
calibration procedure, they collect water samples to measure chlorophyll spectrophotometrically.
These calibration samples are drawn at the  same time as are the samples which are sent to
DHMH for analysis in the water quality monitoring program. Because the two labs were getting
different chlorophyll results for water samples taken at the same time,  the matter was brought to
the attention of the Analytical Methods  and Quality Assurance Workgroup.
      AMQAW requested that three splits be conducted. Seven labs (CBL, ODU, DHMH,
DCLS, ACNATSCI, VCU [Virginia Commonwealth University], and  CRL [EPA Central
Regional Lab]) participated in October and December of 1997 and in May of 1998. Due to a  lab
accident, DHMH's results for the October split were not used and no results were obtained from
CRL in the May Split). The October split consisted of 5 replicates each from two stations in the
Patuxent (XDE4892 and PXT0402). The December split consisted of 5 replicates from station
LE2.3 at the mouth of the Potomac. The May Split was prepared by the Chesapeake Biological
Laboratory and consisted of 10 reps each of a low level natural and a high level cultured sample.
The results of each of these splits are displayed graphically below. No consistent differences
were detected between labs for any of the splits. All labs participating  agreed that chlorophyll is
                                          81

-------
a highly variable parameter by nature and that the most benefit would be gained by focusing their
efforts on using consistent methods.
      The graphs below depict the results of these splits.
  50
 40
  30
 20
  10
                       October 1997 - Station XDE4892
                                           VCN
                                                              vcu
                            JCBL
JDCLS
                      ACNATSCI
                                                    ^> CRL
            £ODU
  90
  75
  60
  45
  30
  15
                        October 1997 - Station PXT0402
                                                             ..vcu
                              CBL
                    IDCLS
             $ODU
                                           ACNATSCI
                                 «• CRL
                                       82

-------
15.0
12.0
 9.0
 6.0
 3.0
 0.0
                    December 1997 - Station LE2.3
                           CBL
                                                      «• VCU
                   DCLS
      i
                                       *
                                      ACNATSCI
                     «• CRL
           .. ODU
                                  DHMH
             May 1998 - Low Concentration Natural Sample
 14
 12
 10
                  JDCLS
JCBL
                                        «• ACNATSCI
                                  DHMH
           .. ODU
                                                      ..VCU
                                  83

-------
            May 1998 - High Concentration Culture Sample
4000





3500





3000





2500





2000





1500





1000





 500





   0
£DCLS
       i
                     j • ACNATSCI
              IDHMH
                                    «• VCU
                                  84

-------
Appendix A
     85

-------
Table A  - Percent recovery data from spiked sample for Old Dominion University. Values are percentages of concentrations measured reletive to
concentrations expected. Percentages should be >90% or <110%. Percentages <80% or >120% are indicative of a problem.
                        PP
           TON
            Si
           NH4
           TOP
           PO4f
           NO23
           NO2
           DOC
CO
o
o
•4
o
o

CO
o
o
N.
o
o
^
Mean
Median
Min
Max
Stand. Dev.
Mean
Median
Min
Max
Stand. Dev.
100.5
100.9
91.5
109.9
4.6
102.4
102.5
97.8
106.5
2.8
101.0
100.0
89.0
118.0
6.3
100.0
100.0
98.0
101.0
1.1
98.4
99.0
81.0
109.0
7.9
95.0
95.5
88.0
103.0
5.4
101.9
104.0
91.0
109.0
5.9
105.8
108.0
99.0
109.0
4.1
98.5
99.0
90.0
105.0
3.7
97.5
96.5
93.0
105.0
4.0
95.3
95.5
86.0
103.0
4.3
94.0
94.0
90.0
98.0
2.8
99.2
99.0
94.0
107.0
3.0
97.0
97.0
94.0
99.0
2.1
99.8
99.0
95.0
108.0
3.6
99.3
99.5
98.0
100.0
0.8
100.3
101.0
93.0
104.0
3.2





              3/7/94
              5/2/94
             8/15/94
            11/14/94
             2/13/95
              5/8/95
              8/7/95
            11/13/95
              4/8/96
             5/13/96
              8/5/96
            11/20/96
              3/4/97
              6/9/97
              9/8/97
             12/8/97
             3/23/98
             6/11/98
              9/1/98
            12/15/98
95
92
96
101
110
99
102
101
94
100
104

105
96
103
102
103
107
102
98
108
 95
 89
111
118

101
102
105
97.2
 99
 99
101
 94
100
 98
100
101
100
101
109

100
102
106
109
106
103

106
81
90
94
95
88
90
98
95
103
96
99

99
96
93
104
97
109
109
104
97
105
91
108

99
105
108
109
99

97
97
103
100
100
102
100
103
100
95
101
90
97
98
96
105
96
93
100
98
98
98
98
98
95
95
103
88
95
86
94
90
92
96
98
94
98
99

107
103
100
97
99
100
101
96
103
99
99
97

99
99
94
96
95
108

101
102
104
99
107
98
98
98
95
100
95
99
100
100
98
100
99
104
99.5
101
101
101
102
101
 93

-------
Table B  - Percent recovery data from spiked sample for Chesapeake Biological Laboratory. Values are percentages of concentrations measured
reletive to concentrations expected. Percentages should be >90% or <110%. Percentages <80% or >120% are indicative of a problem.
                        PP
           TON
            Si
           NH4
           TOP
           P04f
           N023
           N02
           DOC

CO
o
o

4
o
o
CO
o
o
*7
N.
o
o
Mean

Median
Min
Max
Stand. Dev.
Mean
Median
Min
Max
Stand. Dev.
99.9

100
96
103
2.0
101.1
101
100
103
1.2
99.7

99
95
108
3.3
101.1
100.5
95
108
4.4
93.9

93.5
91
97
1.7
93.5
94
91
96
1.8
96.4

96.5
88
106
4.1
96.5
96.5
93
99
2.1
97.9

97
93
103
2.9
98.3
98.5
93
103
4.2
98.4

97.5
92
104
3.9
97.3
96.5
92
103
4.1
704.4

704
99
775
4.4
103
101.5
100
110
3.9
99.7

99
96
705
2.3
99.8
100
98
102
1.7
702.2

702
700
704
7.7





              3/7/94
              5/2/94
             8/15/94
            11/14/94
             2/13/95
              5/8/95
              8/7/95
            11/13/95
              4/8/96
             5/13/96
              8/5/96
            11/20/96
              3/4/97
              6/9/97
              9/8/97
             12/8/97
             3/23/98
             6/11/98
              9/1/98
            12/15/98
100
102
99
96
99
96
100
98
101
102
99
102
101
100
101
100
103
103
96
97
100
96
96

103
99
98
100
102
98
99
100
104
99
95
101
108
93
97
93
93
93
93
93
91
93
96
95
95
95
96
96
91
94
94
94
92
88
100
97
92
92
101
99
96
106
96
90
97
99
96
96
93
99
98
97
96
99
98
97
101
95
94

98
99
96
102
97
97
97
96
101
103
93
95
102
97
96
94
101
104
101
103
92
96
97
100
97
104
102
101
103
95
95
98
92
103
99
115
107
110
110
104
101
106
100
107
99
104
104
110
101
102
105
100
100
98
105
96
97
99
98
100
99
97
96
100
96
101
101
101
102
101
99
98
98
103

100
101
104
101
104

-------
Table C - Percent recovery data from spiked sample for Division of Consolidated Laboratory Services. Values are percentages of concentrations
measured reletive to concentrations expected. Percentages should be >90% or <110%. Percentages <80% or >120% are indicative of a problem.
                        PP
           TON
           SiO2
           NH4
           TOP
           PO4f
            TP
           NO2
           TKNw
CO
o
o
*r
o>
o

CO
o
o
N.
o
o
Mean
Median
Min
Max
Stand. Dev.
Mean
Median
Min
Max
Stand. Dev.
702.3
99.0
96.0
773.0
6.7
103.8
104.5
96.0
110.0
5.9
91.0
93.0
62.0
103.0
10.6
95.5
95.0
89.0
103.0
6.0
99.5
100.0
93.0
102.0
2.5
98.3
99.5
93.0
101.0
3.6
106.1
105.0
85.0
153.0
16.1
100.0
100.0
90.0
110.0
10.0
93.9
93.0
78.0
104.0
6.5
95.3
94.5
93.0
99.0
2.9
93.9
94.0
85.0
102.0
4.5
92.0
91.0
91.0
95.0
2.0
99.9
99.0
82.0
129.0
11.0
104.0
99.0
95.0
118.0
12.3
100.1
100.0
70.0
114.0
8.7
94.3
101.0
70.0
105.0
16.3
97.8
100.0
94.0
100.0
3.0





             3/7/94
             5/2/94
            8/15/94
           11/14/94
            2/13/95
             5/8/95
             8/7/95
           11/13/95
             4/8/96
            5/13/96
             8/5/96
           11/20/96
             3/4/97
             6/9/97
             9/8/97
            12/8/97
            3/23/98
            6/11/98
             9/1/98
           12/15/98
97
96
99
108
97
113
98
109
98
110
103
96
106
95
99
94
88
91
62
90
97
93
89
103
102
100
102
100
94
100
102
98
100

100
100
100
100
93
100
99
101
110

108
98
104
109
108
99
153
105

85
100
90
110
100
100
91
92
78
98
97
84
91
88
93
100
104

96
93
99
93
96
100
90
88
93
95
95
97
93
102
90
98
100
85
91
95
91
91
100
100
91
92
100
82
92
92
100
108
108
98
95
129

118
99
95
100
98
100
100
95
102
103
98
105
99
100
100
110
114
70
100
102
105
                                                                                              100
                                                                                              100
                                                                                               94
                                                                                              100
                                                                                               95

-------
Table D - Percent recovery data from spiked sample for Maryland Department of Health and Mental Hygene. Values are percentages of concentrations
measured reletive to concentrations expected. Percentages should be >90% or <110%. Percentages <80% or >120% are indicative of a problem.
                       DOC
           N023
            Si
           NH4
           TOP
           P04f
            TP
           N02
          TKNw
CO
o
o
T
o
o
CO
o
o
T
o
o
Mean
Median
Min
Max
Stand. Dev.
Mean
Median
Min
Max
Stand. Dev.
101.4
104.0
68.0
129.0
14.0
98.5
98.0
93.0
106.0
5.2
102.7
102.5
95.0
108.0
3.5
104.5
105.5
99.0
108.0
3.4
95.7
95.0
90.0
101.0
3.6
96.8
97.5
91.0
100.0
3.4
98.4
98.0
91.0
106.0
5.0
100.3
101.0
93.0
106.0
5.9
98.9
99.5
81.0
105.0
5.3
98.2
98.0
94.0
101.0
2.9
94.9
98.0
78.0
102.0
6.3
98.3
100.0
92.0
100.0
3.2
98.7
100.0
76.0
113.0
8.9
97.0
96.0
92.0
105.0
5.0
102.0
102.0
98.0
108.0
2.4
100.7
101.0
98.0
102.0
1.6
105.2
106.5
90.0
116.0
8.0
95.8
97.0
90.0
101.0
4.1
             3/7/94
             5/2/94
            8/15/94
           11/14/94
            2/13/95
             5/8/95
             8/7/95
           11/13/95
             4/8/96
            5/13/96
             8/5/96
           11/20/96
             3/4/97
             6/9/97
             9/8/97
            12/8/97
            3/23/98
            6/11/98
             9/1/98
           12/15/98
71
113
129
68
112
107
110
105
104
104
98
117
106
93
101
102
106
93
95
94
100
104
107
95
102
99
103
101
100
108
99
101
104
104
107
106
99
105
108
102
90
95
99
95
93

90
95

99
91
98
95
101
98
100
91
97
95
100
97
97
100
106
100
92
99
94
91
106
94
98
98
106
93
104
81
105

100
104
103
103
98
98
98
101
98
99
102
101
94

98
97
101
98
102
99
92
100
90
96
86

90
78
90
92
100
92
100
100
98
100
100
76
110

104
108
102
101
100
100
113
98
85
92
103
105
92

94
98
96
104
104
104
101
99
100
100
102
106
100
102
102
108
104
102
102
100
100
102
98
109
103

112
108
99
99
109
115
115
105
110
116
114
97
97

94
101
90

-------

-------
Table E - SRM data for Old Dominion University. The SRM_EPA values are the known concentrations of the SRMs, SRM_DE values are the
concentrations that were measured and the % Recov values are the percentages of the SRM_Des relative to the SRM_EPAs. Percentages should be
>90% or <110%. Percentages <80% or >120% are indicative of a problem.


3/7/94
5/2/94
8/15/94
11/14/94
2/13/95
5/8/95
8/7/95
11/13/95
4/8/96
5/13/96
8/5/96
11/20/96
3/4/97
6/9/97
9/8/97
12/8/97
3/23/98
6/11/98
9/1/98
12/15/98


3/7/94
5/2/94
8/15/94
11/14/94
2/13/95
5/8/95
8/7/95
11/13/95
4/8/96
5/13/96
8/5/96
11/20/96
3/4/97
6/9/97
9/8/97
12/8/97
3/23/98
6/11/98
9/1/98
PP
Samp
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00

0.00

0.00

0.00
0.00
0.00
0.00
NH4
Samp
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
PP
SRM EPA
0.50
0.50
0.50
0.50
0.50
0.50
0.50
0.50
0.50
0.50
0.50

0.50

0.75

0.75
0.75
0.75
1.00
NH4
SRM EPA
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
0.03
0.03
0.03
0.03
0.03
PP
SRM DE
0.52
0.48
0.48
0.51
0.48
0.47
0.48
0.50
0.49
0.46
0.49

0.51

0.77

0.77
0.78
0.72
0.99
NH4
SRM DE
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
0.03
0.03
0.03
0.03
0.03
PP
% Recov
103.40
96.40
95.60
101.20
95.20
93.60
95.00
100.20
97.00
92.60
98.20

101.60

103.29

103.05
103.85
96.05
99.30
NH4
% Recov
110.00
94.00
107.75
96.50
96.75
94.25
102.25
96.50
102.75
102.75
98.00
97.00
99.50
94.50
98.39
100.97
100.00
95.16
105.48
TON
Samp
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
TOP
Samp
0.00
0.00
0.01
0.006
0.013
0.003
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
TON
SRM EPA
0.25
0.4
0.4
0.4
0.4
0.4
0.4
0.4
0.4
0.4
0.4
0.4
0.4
0.4
0.4
0.4
0.4
0.21
0.21
0.21
TOP
SRM EPA
0.15
0.15
0.15
0.15
0.15
0.15
0.15
0.15
0.15
0.15
0.15
0.15
0.12
0.12
0.12
0.12
0.06
0.105
0.105
TON
SRM DE
0.235
0.412
0.408
0.374
0.4
0.401
0.381
0.404
0.417
0.395
0.397
0.397
0.399
0.4
0.407
0.398
0.403
0.1967
0.198
0.197
TOP
SRM DE
0.154
0.148
0.159
0.158
0.16
0.149
0.145
0.155
0.154
0.156
0.149
0.135
0.123
0.1218
0.1203
0.1216
0.0591
0.1005
0.1007
TON
% Recov
94.00
103.00
102.00
93.50
100.00
100.25
95.25
101.00
104.25
98.75
99.25
99.25
99.75
100.00
101.75
99.50
100.75
93.67
94.29
93.81
TOP
% Recov
102.67
98.67
99.38
101.28
98.16
97.39
96.67
103.33
102.67
104.00
99.33
90.00
102.50
101.50
100.25
101.33
98.50
95.71
95.90
PO4f
Samp
0.00
0.00

0.00
0.005
0.001
0.00
0.001
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
N023
Samp
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
PO4f
SRM EPA
0.02
0.02

0.02
0.02
0.02
0.02
0.02
0.02
0.02
0.02
0.02
0.04
0.04
0.04
0.0375
0.0375
0.0375
0.0375
0.04
N023
SRM EPA
0.04
0.04
0.04
0.04
0.04
0.04
0.04
0.04
0.04
0.04
0.04
0.04
0.20
0.20
0.04
0.04
0.035
0.035
0.035
PO4f
SRM DE
0.02
0.0185

0.022
0.027
0.022
0.021
0.019
0.018
0.02
0.021
0.015
0.039
0.04
0.0391
0.0378
0.0382
0.038
0.0376
0.0387
N023
SRM DE
0.04
0.04
0.04
0.04
0.04
0.04
0.04
0.04
0.04
0.04
0.04
0.04
0.20
0.20
0.04
0.04
0.0358
0.0345
0.0348
PO4f
% Recov
100.00
92.50

110.00
108.00
104.76
105.00
90.48
90.00
100.00
105.00
75.00
97.50
100.00
97.75
100.80
101.87
101.33
100.27
96.75
N023
% Recov
96.50
99.75
97.75
94.00
99.25
100.00
100.75
101.75
96.00
98.50
99.00
98.75
98.80
99.50
100.86
101.14
102.29
98.57
99.43

-------
12/15/98
            0.00
                       0.08
                                 0.08
                                            97.50
                                                             0.00
                                                                          0.105
                                                                                      0.0979
                                                                                                  93.24
                                                                                                                    0.00
                                                                                                                                 0.08
                                                                                                                                             0.0778
                                                                                                                                                         97.25

3/7/94
5/2/94
8/15/94
11/14/94
2/13/95
5/8/95
8/7/95
11/13/95
DOC
Samp
0.22
0.49
0.49
0.15
0.015
0.028
0.03
0.17
DOC
SRM EPA
6.15
6.15
6.15
6.15
6.15
6.15
6.15
6.15
DOC
SRM DE
6.39
6.6
6.6
6.52
6.52
6.18
6.27
6.4
DOC
% Recov
100.31
99.40
99.40
103.49
105.76
100.03
101.46
101.27

-------
Table F - SRM data for Chesapeake Biological Laboratory. The SRM_EPA values are the known concetrations of the SRMs, SRM_DE values are the
concentrations that were measured and the % Recov values are the percentages of the SRM_Des relative to the SRM_EPAs. Percentages should be
>90% or <110%. Percentages <80% or >120% are indicative of a problem.



3/7/94
5/2/94
8/15/94
11/14/94
2/13/95
5/8/95
8/7/95
11/13/95
4/8/96
5/13/96
8/5/96
11/20/96
3/4/97
6/9/97
9/8/97
12/8/97
3/23/98
6/11/98
9/1/98
12/15/98


3/7/94
5/2/94
8/15/94
11/14/94
2/13/95
5/8/95
8/7/95
11/13/95
4/8/96
5/13/96
8/5/96
11/20/96
3/4/97
6/9/97
9/8/97
12/8/97
3/23/98
6/11/98
9/1/98
12/15/98
NH4
Samp

0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
NO23
Samp
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0

0
0
0
NH4
SRM_EPA

0.2
0.2
0.2
0.2
0.2
0.2
0.2
0.2
0.2
0.2
0.2
0.2
0.2
0.2
0.2

0.155
0.155
0.155
0.2
NO23
SRM EPA
0.2
0.2
0.2
0.2
0.2
0.2
0.2
0.2
0.2
0.2

0.2
0.2
0.2
0.2
0.2

0.175
0.175
0.2
NH4
SRM_DE

0.192
0.208
0.195
0.198
0.188
0.203
0.199
0.199
0.203
0.196
0.203
0.192
0.205
0.204
0.198

0.172
0.16
0.157
0.194
NO23
SRM DE
0.198
0.199
0.201
0.2
0.199
0.21
0.191
0.201
0.195
0.197

0.206
0.205
0.204
0.191
0.196

0.192
0.205
0.225
NH4
% Recov

96.00
104.00
97.50
99.00
94.00
101.50
99.50
99.50
101.50
98.00
101.50
96.00
102.50
102.00
99.00

110.97
103.23
101.29
97.00
NO23
% Recov
99.00
99.50
100.50
100.00
99.50
105.00
95.50
100.50
97.50
98.50

103.00
102.50
102.00
95.50
98.00

109.71
117.14
112.50
TON
Samp

0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
TOP
Samp
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0

0
0
TON
SRM_EPA

0.5
0.5
0.5
0.5
0.5
0.5
0.5
0.5
0.5
0.5
0.5
0.5
0.5
0.5
0.5
0.5
0.42

0.42
0.375
TOP
SRM EPA
0.15
0.15
0.15
0.15
0.15
0.15
0.15
0.15
0.15
0.15
0.15
0.15
0.15
0.15
0.15
0.15
0.205

0.205
0.15
TON
SRM_DE

0.502
0.52
0.54
0.54
0.55
0.55
0.55
0.51
0.53
0.5
0.55
0.5
0.55
0.51
0.52
0.56
0.43

0.43
0.41
TOP
SRM DE
0.153
0.152
0.148
0.145
0.1509
0.1596
0.1482
0.1517
0.139
0.141
0.164
0.148
0.142
0.141
0.164
0.155
0.198

0.198
0.1498
TON
%
Recov
100.40
104.00
108.00
108.00
110.00
110.00
110.00
102.00
106.00
100.00
110.00
100.00
110.00
102.00
104.00
112.00
102.38

102.38
109.33
TOP
PO
San

0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0

0
0
0

% Recov
102.00
101.33
98.67
96.67
100.60
106.40
98.80
101.13
92.67
94.00
109.33
98.67
94.67
94.00
109.33
103.33
96.59

96.59
99.87




















                                                                                       PO4f    PO4f
                                                                              Samp SRM_EPA SRM_DE
                                                                                      0.039
                                                                                      0.039
                                                                                      0.039
                                                                                      0.039

                                                                                      0.039
                                                                                       0.05
                                                                                       0.05
                                                                                       0.05
                                                                                       0.05
                                                                                       0.05
                                                                                       0.05
                                                                                       0.05
                                                                                       0.05
                                                                                       0.05
                                                                                       0.05
0.0406
0.0359
0.0384
0.0367

0.0365
0.0501
0.0497
0.0474
0.0498
0.0511
0.0493
0.0508
 0.05
 0.051
0.0475
 PO4f

Recov
104.10
 92.05
 98.46
 94.10

 93.59
100.20
 99.40
 94.80
 99.60
102.20
 98.60
101.60
100.00
102.00
 95.00
                                                                                      0.075    0.0811  108.13
                                                                                      0.075    0.0815  108.67
                                                                                       0.05     0.0527  105.40

-------
Table G - SRM data for the Division of Consolidated Laboratory Services. The SRM  EPA values are the known concentrations of the SRMs, SRM DE
values are the concentrations that were measured and the % Recov values are the percentages of the SRM_Des relative to the SRM_EPAs. Percentages
should be >90% or <110%. Percentages <80% or >120% are indicative of a problem.


3/7/94
5/2/94
8/15/94
11/14/94
2/13/95
5/8/95
8/7/95
11/13/95
4/8/96
5/13/96
8/5/96
11/20/96
3/4/97
6/9/97
9/8/97
12/8/97
3/23/98
6/11/98
9/1/98
12/15/98


3/7/94
5/2/94
8/15/94
11/14/94
2/13/95
5/8/95
8/7/95
11/13/95
4/8/96
5/13/96
8/5/96
11/20/96
3/4/97
6/9/97
9/8/97
12/8/97
3/23/98
6/11/98
9/1/98
12/15/98


3/7/94
5/2/94
8/15/94
11/14/94
2/13/95
NH4
Samp
0.00
0.00

0.00
0.00

0.00


0.00
0.00



0.00
0.00


0.00
0.00
TP
Samp
0.00

0.00
0.00
0.00



0.00
0.00
0.00
0.00
0.00
0.00

0.00


0.00
0.00
TKNw
Samp
0.00
0.00
0.00

0.00
NH4
SRM EPA
0.07
0.07

0.30
0.30

0.30


0.06
0.30



0.30
0.30


0.046
0.046
TP
SRM EPA
0.75

0.75
0.75
1.50



4.76
4.76
4.76
4.76
2.10
0.31

0.23


0.046
0.046
TKNw
SRM EPA
2.50
2.50
2.50

6.50
NH4
SRM DE
0.08
0.078

0.317
0.324

0.319


0.066
0.276



0.325
0.331


0.049
0.051
TP
SRM DE
0.72

0.68
0.75
1.60



4.66
4.55
4.31
4.77
2.24
0.30

0.23


0.046
0.047
TKNw
SRM DE
2.40
2.40
2.00

6.40
NH4
% Recov
114.29
1 1 1 .43

105.67
108.00

106.33


110.00
92.00



108.33
110.33


106.52
110.87
TP
% Recov
96.00

90.67
100.00
106.67



97.90
95.59
90.55
100.21
106.67
96.77

100.00


100.00
102.17
TKNw
% Recov
96.00
96.00
80.00

98.46
TON
Samp












0.00

0.00
0.00


0.00
0.00
TOP
Samp
0.00

0.00
0.00








0.00
0.00
0.00
0.00


0.00
0.00







TON
SRM_EPA












0.48

0.48
0.48


0.48
0.48
TOP
SRM EPA
0.75

0.75
0.75








0.046
0.046
0.046
0.046


0.046
0.046







TON
SRM_DE












0.497

0.491
0.493


0.456
0.478
TOP
SRM DE
0.72

0.68
0.75








0.051
0.049
0.049
0.046


0.046
0.049







TON
% Recov












103.54

102.29
102.71


95.00
99.58
TOP
% Recov
96.00

90.67
100.00








110.87
106.52
106.52
100.00


100.00
106.52







Si
Samp








0.00
0.00
0.00

0.00
0.00
0.00
0.00


0.00
0.00
P04f
Samp
0.00

0.00
0.00
0.00

0.00
0.00

0.00
0.00



0.00
0.00


0.00
0.00







Si
SRM_EPA








0.50
0.50
0.50

0.50
0.50
0.50
0.50


0.50
0.54
P04f
SRM EPA
0.10

0.195
0.075
0.075

0.075
0.075

0.015
0.075



0.075
0.075


0.06
0.06







Si
SRM_DE








0.55
0.51
0.42

0.53
0.55
0.49
0.56


0.54
0.50
P04f
SRM DE
0.077

0.192
0.073
0.077

0.068
0.074

0.016
0.073



0.073
0.075


0.062
0.062







Si
% Recov








110.00
102.00
84.00

106.00
110.00
98.00
112.00


108.00
92.59
P04f
% Recov
77.00

98.46
97.33
102.67

90.67
98.67

106.67
97.33



97.33
100.00


103.33
103.33








-------
Table H - SRM data for the Maryland Department of Health and Mental Hygene. The SRM_EPA values are the known concentrations of the SRMs,
SRM_DE values are the concentrations that were measured and the % Recov values are the percentages of the SRM_Des relative to the SRM_EPAs.
Percentages should be >90% or <110%. Percentages <80% or >120% are indicative of a problem. SRM data for 1998 were not available at the time of
this report.


2/8/94
3/7/94
5/2/94
11/14/94
2/13/95
5/8/95
8/7/95
11/13/95
4/8/96
5/13/96
8/5/96
11/20/96
3/4/97
6/9/97
9/8/97
12/8/97


2/8/94
3/7/94
5/2/94
11/14/94
2/13/95
5/8/95
8/7/95
11/13/95
4/8/96
5/13/96
8/5/96
11/20/96
3/4/97
6/9/97
9/8/97
12/8/97


2/8/94
3/7/94
5/2/94
11/14/94
2/13/95
5/8/95
8/7/95
11/13/95
4/8/96
5/13/96
8/5/96
11/20/96
3/4/97
6/9/97
9/8/97
19/Q/Q7
TOC
Samp
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
NH4
Samp
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
NO23
Samp
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
n
TOC
SRM_EPA







40.9
40.9
40.9
40.9
40.9
40.9
40.9
40.9
40.9
NH4
SRM EPA
2.0
2.0
2.0
2.0
2.0
2.0

2.0
2.0
2.0
2.0
2.0

2.0
2.0
2.0
NO23
SRM EPA
0.93
2.0
2.0
2.0
2.0
2.0

2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.0
9 n
TOC
SRM_DE







40.9
42.8
41.3
42.8
40.2
38.1
37.6
40.7
43.0
NH4
SRM DE
2.04
1.97
1.88
1.925
2.035
1.9

1.935
1.955
1.955
1.965
2.07

1.985
1.99
2.14
NO23
SRM DE
1.005
2.115
2.153
2.14
2.02
1.94

2.01
2.02
1.96
1.93
2.06
2.02
2.06
2.15
9 n9
TOC
% Recov







100.00
104.65
100.98
104.65
98.29
93.15
91.93
99.51
105.13
NH4
% Recov
102.00
98.50
94.00
96.25
101.75
95.00

96.75
97.75
97.75
98.25
103.50

99.25
99.50
107.00
NO23
% Recov
108.06
105.75
107.65
107.00
101.00
97.00

100.50
101.00
98.00
96.50
103.00
101.00
103.00
107.50
•mi nn
TP
Samp



0
0
0
0
0
0
0
0
0
0
0
0
0
NO2
Samp




0
0
0
0
0
0
0
0
0
0
0
0
TKNw
Samp
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
n
TP
SRM_EPA



1.5

1.5
1.5
1.5
1.5
1.5
1.5
1.5
5
5
1.5
1.5
NO2
SRM_EPA




0.113
0.102

0.102
0.102
0.102
0.102
0.102
0.102
0.102
0.102
0.113
TKNw
SRM_EPA



5.0
5.0
1.5

5.0
5.0
5.0
5.0
5.0
5.0
5.0
5.0
R n
TP
SRM_DE



1.56

1.61

1.53
1.45
1.35
1.58
1.35
4.86
4.54
1.46
1.47
NO2
SRM_DE




0.113
0.097

0.101
0.1
0.103
0.104
0.105
0.106
0.101
0.111
0.114
TKNw
SRM_DE



4.865
4.635
1.54

5.001
5.001
4.323
5.102
4.351
4.546
4.767
5.05
AQ
TP
% Recov



104.00

107.33
0.00
102.00
96.67
90.00
105.33
90.00
97.20
90.80
97.33
98.00
NO2
% Recov




100.00
95.10

99.02
98.04
100.98
101.96
102.94
103.92
99.02
108.82
100.88
TKNw
Si
Samp













0
0
0
PO4f
Samp
0
0
0
0
0
0

0
0
0
0
0
0
0
0
0

Si
SRM_EPA













1
4
1
PO4f
SRM EPA
0.39
0.39
0.39
0.39
0.39
0.39

0.39
0.39
0.39
0.39
0.39
0.5
0.5
0.5
0.5

Si
SRM_DE













1.05
3.99
1.05
PO4f
SRM DE
0.4
0.425

0.385
0.4
0.385

0.405
0.39
0.38
0.36
0.345
0.465
0.465
0.47
0.46

Si
% Recov













105.00
99.75
105.00
PO4f
% Recov
102.56
108.97
0.00
98.72
102.56
98.72

103.85
100.00
97.44
92.31
88.46
93.00
93.00
94.00
92.00

% Recov



97.30
92.70
102.67

100.02
100.02
86.46
102.04
87.02
90.92
95.34
101.00
QR nn

































































-------
Table I - Coefficients of Variation (CVs) by parameter, lab and date. CVs are the standard deviation of the three replicates analyzed
by a lab on a particular date expressed as a percentage of the mean of those three replicates.CVs should be consitently less than 25%
across dates.

CBL
3/7/94 14.5
5/2/94 7.1
8/15/94 9.4
11/14/94 8.7
2/13/95 13.4
5/8/95 5.4
8/7/95 4.2
11/13/95 4.6
4/8/96 5.4
5/13/96 0.7
8/5/96 2.2
11/20/96 8.6
3/4/97 10.7
6/9/97 1 .4
9/8/97 4.6
12/8/97 3.5
3/23/98 7.7
6/11/98 6.6
9/1/98 2.5
12/14/98 17.1

CBL
3/7/94 8.6
5/2/94 2.1
8/15/94 3.6
11/14/94 4.2
2/13/95 3.8
5/8/95 7.3
8/7/95 1 .3
11/13/95 3.1
4/8/96 3.3
5/13/96 5.4
8/5/96 1.1
11/20/96 2.1
3/4/97 3.5
6/9/97 1 .8
9/8/97 2.2
12/8/97 4.2
3/23/98 3.0
6/11/98 3.3
9/1/98 0.8
12/14/98 7.7
TOP
ODU
13.6
ND
2.3
10.0
6.0
22.9
7.9
0.0
0.0
12.4
4.0
62.4
13.3
0.0
5.5
7.0
19.9
1.4
3.5
9.2
TP
ODU
5.2
24.0
1.9
4.2
2.0
9.2
2.9
4.9
0.0
2.7
3.7
ND
5.3
2.0
1.0
3.3
3.2
2.0
1.3
5.0

DHMH
ND
34.6
17.0
8.7
ND
15.1
27.2
9.1
20.4
58.3
7.4
29.7
34.6
23.3
41.6
18.7
25.0
29.4
1.4
11.3

DHMH
ND
47.0
9.3
14.6
32.8
18.0
4.0
3.7
10.2
44.2
11.8
13.6
14.1
4.6
15.5
7.8
11.0
11.2
1.7
3.9

DCLS
0.0
ND
21.7
34.6
43.3
10.8
2.6
3.2
74.8
16.7
3.3
27.0
22.3
18.4
2.7
0.0
8.7
ND
1.9
0.0

DCLS
0.0
0.0
0.0
0.0
0.0
0.0
9.1
0.0
0.0
17.3
0.0
10.8
0.0
12.4
ND
10.8
13.3
ND
0.0
15.6
NO23
CBL
3/7/94 0.5
5/2/94 1 .4
8/15/94 61.7
11/14/94 8.1
2/13/95 2.1
5/8/95 0.6
8/7/95 6.6
11/13/95 4.4
4/8/96 2.0
5/13/96 0.8
8/5/96 64.3
11/20/96 1.0
3/4/97 12.3
6/9/97 0.2
9/8/97 64.0
12/8/97 1 .0
3/23/98 0.3
6/11/98 0.6
9/1/98 57.5
12/14/98 17.1
ODU
2.8
0.7
ND
0.5
1.8
0.4
8.9
1.0
17.1
6.1
17.3
9.1
0.6
2.5
16.7
13.2
0.4
0.6
56.8
15.6
DHMH
2.6
19.9
ND
9.3
0.4
1.0
11.1
2.7
0.7
0.4
ND
0.6
3.5
1.2
ND
1.2
1.2
0.2
34.6
0.0
DCLS
0.3
0.6
ND
0.0
0.5
0.5
ND
24.7
0.6
0.5
ND
0.6
1.1
0.8
ND
1.0
2.2
ND
0.0
0.0
CBL
5.8
3.7
5.3
3.5
0.8
9.8
4.8
0.6
2.1
7.8
4.2
0.8
4.5
3.9
0.8
5.7
5.1
2.4
3.9
7.7

CBL
5.5
4.0
22.8
4.2
3.2
3.9
9.8
6.2
1.7
3.1
4.7
0.7
3.4
6.2
4.9
4.2
2.8
3.1
6.4
17.7

CBL
3.5
2.3
17.3
20.7
2.4
16.0
7.5
2.5
32.6
2.2
68.5
0.8
0.8
2.1
68.7
30.3
5.1
20.8
23.1
4.7
PP
ODU
0.0
2.8
2.2
7.1
5.3
0.0
3.3
11.5
0.0
3.5
5.9
ND
4.5
2.9
5.3
0.0
0.0
2.4
2.0
5.2
TON
ODU
1.8
2.6
1.8
2.9
3.4
0.0
4.6
3.0
1.7
0.5
8.3
2.3
0.7
1.1
3.5
0.8
0.2
1.4
2.7
0.2
NO2
ODU
5.8
0.2
0.0
1.8
0.0
2.2
2.5
6.9
8.3
5.8
10.8
0.7
2.5
3.7
50.0
6.4
0.5
1.3
0.0
11.1
P04F
DHMH
ND
86.7
10.8
141.4
32.8
59.3
86.9
16.4
40.1
55.5
46.7
53.7
66.1
45.1
20.0
33.1
8.3
8.9
5.8
24.7

DHMH
5.5
24.7
ND
2.9
3.3
4.0
4.5
ND
1.2
3.3
ND
7.1
1.3
4.9
ND
22.5
1.4
1.1
5.8
1.7

DHMH
7.1
31.0
0.0
0.0
10.2
0.0
0.0
0.0
8.3
2.3
0.0
2.4
0.0
0.0
21.7
0.0
0.0
10.8
0.0
0.0
DCLS
0.0
21.7
17.3
43.3
15.7
29.3
4.0
2.4
2.0
0.8
4.3
0.8
6.7
6.9
4.7
0.0
0.0
ND
2.7
2.1

DCLS
ND
ND
ND
ND
ND
6.3
3.8
0.3
17.6
2.0
3.7
3.4
2.6
0.1
2.1
3.5
1.5
ND
6.3
2.0

DCLS
0.0
0.0
ND
0.0
0.0
0.0
0.0
0.0
4.3
1.9
ND
0.0
0.0
0.0
ND
21.7
0.0
ND
0.0
0.0
CBL
45.3
1.9
18.2
22.9
29.6
15.5
6.2
71.3
87.6
85.4
3.0
41.4
7.4
4.2
42.8
0.0
50.8
9.4
22.1
46.8

CBL
25.0
1.4
104.4
22.2
95.2
51.6
8.5
11.3
12.9
6.3
20.1
32.7
1.3
5.1
55.3
13.9
4.7
50.7
48.5
0.0

CBL
20.4
14.3
8.7
8.8
7.8
13.6
8.9
18.6
9.4
6.0
8.2
12.8
3.4
13.5
6.1
8.6
13.0
7.5
7.0
14.5
ODU
0.0
ND
14.3
11.9
9.1
24.7
4.3
12.5
ND
34.6
6.9
ND
ND
ND
4.6
ND
14.7
1.0
1.6
41.7
NH4
ODU
17.0
9.8
ND
4.3
82.2
40.1
9.0
2.8
14.8
1.1
15.6
50.9
1.5
9.9
58.5
15.7
2.9
14.8
3.9
7.8
TSS
ODU
12.0
1.7
11.4
3.9
5.1
ND
32.4
1.2
11.0
1.1
6.5
ND
19.6
4.0
5.1
10.1
12.4
4.7
13.2
20.7
DHMH
52.7
16.7
0.0
0.0
ND
0.0
20.0
9.9
0.0
11.9
ND
52.9
45.4
7.4
0.0
ND
0.0
6.7
5.6
15.7

DHMH
ND
74.5
ND
8.6
24.7
46.9
5.8
5.5
0.0
0.0
5.9
7.5
9.2
17.6
2.8
7.5
1.4
86.6
0.0
3.3

DHMH
10.2
11.5
69.6
22.3
2.8
34.3
24.1
24.1
43.3
10.6
22.9
18.3
20.8
11.9
43.3
6.2
19.9
15.7
32.8
0.0
DCLS
ND
0.0
13.3
0.0
0.0
0.0
0.0
0.0
0.0
13.3
6.7
10.8
ND
33.3
10.8
15.7
0.0
ND
0.0
0.0

DCLS
4.8
4.6
35.0
8.7
11.4
8.7
1.0
9.8
4.2
9.0
24.7
16.7
7.0
49.9
9.4
23.6
5.5
ND
ND
13.9

DCLS
ND
ND
ND
ND
ND
ND
ND
19.9
ND
0.0
ND
0.0
ND
60.3
ND
0.0
0.0
ND
36.5
ND

-------

3/7/94
5/2/94
8/15/94
11/14/94
2/13/95
5/8/95
8/7/95
11/13/95
4/8/96
5/13/96
8/5/96
11/20/96
3/4/97
6/9/97
9/8/97
12/8/97
3/23/98
6/11/98
9/1/98
12/14/98
CBL
1.7
2.0
1.7
5.1
0.8
0.9
3.4
4.2
3.0
0.8
1.4
6.3
3.8
0.0
0.9
0.9
2.3
0.9
2.0
1.8
PC
ODU
6.8
7.8
8.2
2.0
2.3
5.2
5.1
10.5
ND
4.3
0.5
ND
12.9
2.3
5.4
6.2
37.1
13.4
3.2
18.0
DHMH
ND
ND
ND
ND
ND
ND
ND
ND
0.1
6.9
4.2
1.9
0.3
1.9
3.6
4.8
46.9
30.6
24.0
60.1
DCLS
ND
ND
ND
ND
ND
ND
ND
ND
28.0
47.0
53.7
8.1
30.6
36.8
25.5
64.4
4.4
ND
1.0
5.8
CBL
1.2
1.2
0.4
0.0
1.3
4.7
0.5
2.0
1.0
0.0
0.0
1.3
0.0
2.5
1.3
34.6
0.6
0.0
0.8
9.1
Si
ODU
0.6
ND
0.7
0.9
0.7
4.8
0.6
1.8
0.8
0.5
3.3
0.4
0.5
2.9
0.4
14.8
0.6
0.3
1.1
1.6
DHMH
3.5
22.2
0.0
0.0
0.0
0.0
2.4
3.2
2.7
ND
0.0
0.0
6.7
4.0
1.2
ND
0.6
0.0
0.4
3.0
DCLS
3.1
0.0
0.0
0.0
0.0
24.7
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
2.7
ND
0.0
43.3
CBL
1.5
2.2
1.4
4.3
0.6
1.8
3.7
3.1
1.3
4.5
1.2
5.3
4.2
1.3
0.3
0.9
1.7
0.7
1.6
3.2
PN
ODU
3.4
10.0
1.4
12.4
3.0
0.6
38.4
12.1
ND
37.3
2.5
ND
18.8
3.3
3.9
6.0
9.7
2.8
3.2
20.5
DCLS
ND
ND
ND
ND
ND
ND
ND
ND
1.2
7.5
4.4
1.5
0.7
1.9
4.0
4.6
ND
ND
2.6
5.6

-------
   Appendix B



Blind Audit Report
       95

-------
     Chesapeake Bay Program

     Blind Audit Nutrient Results

       January and June 1998
          November 1998
         Carl Zimmermann
          Carolyn Keefe
Nutrient Analytical Services Laboratory
  Chesapeake Biological Laboratory
       Solomons, MD 20688

        November 20, 1998
               96

-------
INTRODUCTION:

The purpose of this Blind Audit Program is to provide samples of specific nutrient analytes at
concentrations commonly found in estuarine systems for analysis by laboratories who analyze water
samples collected from the Chesapeake Bay and its tributaries. The concentrations of these samples,
which are unknown to the recipient analysts, are compared to their true concentrations.

In the early years of the Chesapeake Bay Program, the U.S. EPA  provided blind audit samples on an
irregular basis to laboratories analyzing Chesapeake Bay water samples. However, these  audit samples
were designed for waste water/drinking water applications rather than estuarine water applications.
Consequently, the concentrations were much higher than normally occur in the Bay and did not provide
a reasonable  estimate of accuracy for low level nutrient analyses. For example, a blind audit
concentration of 1.0 mg NH4-N/L would be comparable for NPDES water samples but would be an order
of magnitude  greater than concentrations normally occurring in most parts of Chesapeake Bay.

The only continuous program  providing an estimate of laboratory performance has been the
Chesapeake Bay Coordinated Split Sample Program (CSSP). Data generated from this program provide
the only long term QA/QC data base that compare nutrient measurements provided by laboratories
analyzing  water samples collected from Chesapeake Bay and its tributaries. Samples for the CSSP are
natural water  samples collected from Chesapeake Bay or a tributary. Briefly, a common unfiltered water
sample is distributed to the various field/laboratory personnel who in turn subsample into dissolved and
particulate fractions. These are analyzed and the results compared to those of other participating
laboratories. Resulting data analysis can show how field filtration techniques and/or laboratory practices
affect data variability. The CSSP samples are each subject to cumulative errors of analytical
determinations from variation in both field and laboratory  procedures. Also, these data sets cannot
definitively determine the accuracy of laboratory analyses.

The current Blind Audit Program was designed to complement the CSSP. Blind Audit particulate samples
distributed to  participants have few cumulative errors associated with field filtering and subsampling
procedures. Prepared concentrates of dissolved substances, whose concentrations are unknown to the
analysts, are provided  so that laboratory accuracy can be assessed.

There have been no blind audit assessments within the Chesapeake Bay Program for the  past nine
years. It is the intent of this Blind Audit Program to continually provide unknown, low level dissolved and
particulate nutrient samples to laboratories analyzing Chesapeake Bay Program nutrients, as well as to
other laboratories interested in participating in the Blind Audit Program.

MATERIALS AND METHODS

Blind Audit samples were sent to participating laboratories in January (27 January 1998) and June (15
June 1998) 1998. Those participating laboratories and contact personnel are found in Table  1.

Parameters measured during the January audit were: total dissolved nitrogen, total dissolved
phosphorus, nitrate+nitrite, ammonium and phosphate. A high and a  low concentration sample were
provided for each of these analytes. Particulate carbon, nitrogen and phosphorus samples were also
provided for those laboratories that routinely analyze these parameters.

Dissolved Blind Audit concentrates were prepared by careful dilution  of high quality standards using 18.3
megohm deionized water. The concentrates were sealed  in  10 ml ampules for shipment to the
participants. One ampule contained a concentrate of an organic nitrogen compound and an organic
phosphorus compound to be diluted for the analysis of low level total dissolved nitrogen and total
dissolved phosphorus. A second ampule contained a concentrate of organic nitrogen and organic
phosphorus to be diluted for the analysis of higher level total dissolved nitrogen and total  dissolved
phosphorus. A third ampule contained a concentrate to be diluted for the analysis of low level inorganic
nutrients (ammonium,  nitrate and phosphate). A fourth ampule contained a concentrate to be diluted for
                                              97

-------
the analysis of higher level inorganic nutrients. At each participating laboratory, an aliquot from each
ampule was diluted and analyzed according to accompanying instructions for preparation and dilution.
Blind Audit samples were then inserted randomly in a typical estuarine sample set. Final concentrations
were reported for each diluted concentrate according to the dilution instructions provided.

Particulate analytes are measured by analyzing suspended material concentrated on filter pads. There
are no commercially available suspensions of pure carbon, nitrogen or phosphorus compounds, so a
natural sample was subsampled onto filter pads for analysis by participating laboratories. A batch water
sample was collected off the CBL pier in January and June, and subsampled for particulate samples of
carbon, nitrogen and phosphorus. Particulate C/N samples were filtered from the batch sample with care
being taken to shake the sample before each filtration to ensure homogeneity. Four 25 mm GF/F pads
were sent to each laboratory for analysis. One  laboratory s instrument requires that only 13 mm filters be
utilized. For that laboratory, four 13 mm GF/F pads were provided. Samples were dried completely
(overnight at 47°C) before shipment. Vacuum filtration was used  to process the 25 mm filters, but
positive pressure was used to filter the 13 mm  filters. Our laboratory did not have the facilities necessary
to vacuum filter these small filters.

The same general procedure was followed for particulate phosphorus samples which were concentrated
by vacuum filtration on 47 mm GF/F pads.

Particulate concentrations for the January Blind Audit were estimated as closely as possible by analyzing
at least eight replicates of each analyte by Chesapeake Biological Laboratory. These  calibration
replicates  also provided  an estimate of variability due to the cumulative  effect of filtering and other
processing errors. Filter pads were sent to each laboratory  for the analysis of particulate C, N and P. The
volume of sample filtered  was noted in the instructions so that each laboratory could report values in
mg/L.

For the June Blind Audit, two samples concentrated on filters were supplied to each laboratory for each
particulate analysis. One laboratory analyzed a second pair of filters because the first pair was rejected
when the analyst noticed a marked visible difference between the replicates. The standard deviations
determined for the January particulate fractions also were used to assess the variability of the June data.

Analysis of chlorophyll a samples was added to the suite of nutrients in June 1998. Samples were filtered
onto 47 mm GF/F glass fiber filters and two were then sent to each  laboratory.

For both audits, samples were sent in coolers via next day  carrier to the participating laboratories. In
June, when chlorophyll samples were sent, a cold temperature was required, so frozen cold packs were
packed in those coolers.

RESULTS

JANUARY 1998 DISSOLVED FRACTION

Figures summarizing all results are found at the end of the  report.

Total Dissolved Nitrogen:  The true low level concentration was 0.35 mg N/L and reported concentrations
ranged from 0.27-0.40 mg N/L. The true high level concentration was 1.05 mg N/L and reported
concentrations ranged from 0.97-1.15 mg N/L.  All laboratories reported concentrations that were within
0.10 mg N/L of the respective total dissolved nitrogen concentrations.
Total Dissolved Phosphorus: The true low level concentration was 0.024 mg P/L and reported
concentrations ranged from 0.020-0.040 mg P/L. The true high level concentration was 0.096 mg P/L
and reported concentrations ranged from 0.050-0.110 mg P/L.. All laboratories except one reported
concentrations within 0.005 mg P/L of the true  concentration for the low level total dissolved phosphorus
sample. All laboratories except one reported concentrations within 0.015  mg P/L of the true concentration

                                              98

-------
for the higher level total dissolved phosphorus concentration.

Ammonium: The true low level concentration was 0.063 mg N/L and reported concentrations ranged
from 0.060-0.081 mg N/L. The true high level concentration was 0.330 mg N/L and reported
concentrations ranged from 0.320-0.364 mg N/L. All laboratories except one reported concentrations
within 0.006 mg N/L of the true low level ammonium concentration. All laboratories reported
concentrations within 0.034 mg N/L of the true higher level ammonium concentration.

Nitrate+nitrite: The true low level concentration was 0.112 mg N/L and reported concentrations ranged
from 0.110-0.126 mg N/L. The true high level concentration was 1.15 mg N/L and reported
concentrations ranged from 1.12-1.23 mg N/L.. All laboratories reported concentrations within 0.014 mg
N/L of the true low level nitrate concentration, and within 0.08 of the true higher level nitrate
concentration.

Phosphate: The true low level concentration was 0.031 mg P/L and reported concentrations ranged from
0.020-0.040 mg P/L. The true high level concentration was 0.310 mg P/L and reported concentrations
ranged from 0.298-0.335 mg  P/L. All  laboratories except two reported concentrations within 0.003 mg
P/L of the true low level phosphate concentration. All  laboratories reported concentrations within 0.025
mg  P/L of the true higher level phosphate concentration.

JANUARY 1998  PARTICULATE FRACTION

Again, it should be noted that these samples were filtered from a common water sample and,
consequently, are not true blind audit samples made from pure constituents; rather, a concentration
range around a mean was established by the analysis of 12 replicate particulate C/N samples and 8
replicate  particulate phosphorus samples. This still provides a verification of measurement processes in
routine analytical conditions at participating laboratories, without the potential variability associated with
differing field filtration techniques.

Particulate Nitrogen: The mean concentration of the 12 replicate samples was 0.078 mg N/L ± 0.004
(S.D.) and all but one of the responding laboratories reported the mean concentration of their four
replicates within 0.078 mg N/L ±  0.012, i.e., 3 X S.D.  .

Particulate Carbon: The mean concentration  of the 12 replicate samples was 0.411 mg C/L ± 0.050
(S.D.) and all responding laboratories reported the mean concentration of their four replicates within
0.411 mg C/L ± 0.150, i.e., 3  X S.D..

Particulate Phosphorus: The  mean concentration of the 8 replicate samples was 0.0318 mg P/L ± 0.0010
(S.D.) and all responding laboratories reported the mean concentration of their four replicates within
0.0318 mg P/L ± 0.0030, i.e., 3 X S.D..


JUNE 1998 DISSOLVED FRACTION

The concentrations of some Blind Audit samples were reduced for the June audit.  Low level total
dissolved N and P concentrations remained unchanged  from the January concentrations, but the  higher
level concentrations were halved from those of January. Low level ammonium concentrations were also
halved, as were the low level  phosphate concentrations. The higher level concentration phosphate
samples were reduced by a factor of five from the June samples. Basically, for the June Blind Audit, the
true concentrations remained unchanged or were substantially reduced from January levels.

Total Dissolved Nitrogen: The true low level concentration was the same as in January, 0.35 mg N/L and
reported  concentrations ranged from  0.205-0.42 mg N/L. The true high level concentration  was 0.53 mg
N/L and reported concentrations  ranged from 0.39-0.62  mg N/L. All laboratories reported concentrations

                                              99

-------
within 0.15 mg N/L of the true concentration of the respective total dissolved nitrogen concentrations.

Total Dissolved Phosphorus: The true low level concentration was 0.024 mg P/L (the same as January)
and reported concentrations ranged from 0.020-0.030 mg P/L. The true high level concentration was
0.048 mg P/L and reported concentrations ranged from 0.030-0.0513 mg P/L. All laboratories reported
concentrations within 0.006 mg P/L of the true low level total dissolved phosphorus concentration. All
laboratories except one reported concentrations within 0.006 mg P/L of the true higher level total
dissolved phosphorus concentration.

Ammonium: The true low level concentration was 0.035 mg N/L and reported concentrations ranged
from 0.025-0.040 mg N/L. The true high level  concentration was 0.280 mg N/L and  reported
concentrations ranged from 0.2645-0.281 mg  N/L. All laboratories reported concentrations within 0.010
mg N/L of the true low level ammonium concentration, and within 0.020 of the true  higher level
ammonium concentration.

Nitrate+nitrite: The true low level concentration was 0.175 mg N/L and reported concentrations ranged
from 0.160-0.210 mg N/L.  The true high level concentration was 0.600 mg N/L and reported
concentrations ranged from 0.550-0.594 mg N/L. All laboratories except one reported concentrations
within 0.015 mg N/L of the true low level nitrate concentration. All laboratories reported concentrations
within 0.050 mg N/L of the true higher level nitrate concentration.

Phosphate: The true low level concentration was 0.0186 mg P/L and reported concentrations ranged
from 0.0190-0.0203 mg P/L. The true high level concentration was 0.0620 mg P/L and reported
concentrations ranged from 0.0600-0.0672 mg P/L. All laboratories reported concentrations within 0.0020
mg P/L of the true low level phosphate concentration, and within 0.0060 of the true  higher level
phosphate concentration.

JUNE 1998 PARTICULATE FRACTION

Particulate  Nitrogen: The mean concentration of the samples analyzed by the five participating
laboratories was 0.307 mg  N/L. Each reported mean from any  participating laboratory was within 0.307
mg N/L ± 0.012, i.e., 3 X S.D. of the 12 January calibration replicates.


Particulate Carbon: The mean concentration of the samples analyzed by the five participating
laboratories was 1.60 mg C/L. Each reported  mean from any participating laboratory was within 1.60 mg
C/L ± 0.15, i.e., 3 X S.D. of the 12 January calibration replicates.

Particulate Phosphorus: The  mean concentration of the samples analyzed by the five participating
laboratories was 0.0454 mg P/L..  Each reported mean from any participating laboratory was within
0.0454 mg P/L ± 0.0030, i.e., 3 X S.D.  of the 8 January calibration replicates.

Chlorophyll: There was quite large variation between laboratories in the chlorophyll  a concentrations
reported. CBL and DCLS reported nearly identical concentrations, while the Academy of Natural
Sciences was more than 7 //g/L greater, and VIMS and ODD reported concentrations substantially lower.

DISCUSSION

Three important issues should be considered when assessing whether individual Blind Audit results  are
within acceptable limits.

Variation Associated With An Analytical Method: A certain amount of analytical variability is associated
with any quantitative determination. The method detection limit (three times the standard deviation of
seven low level replicate natural samples) is often used to express that level of variation. Total dissolved

                                              100

-------
nitrogen data provide a good example. The detection limit at CBL has been determined to be 0.02 mg
N/L. Any total dissolved nitrogen measurement has a potential 0.02 mg N/L variability associated with it.
This variability, when expressed as a percent of the true  concentration, can be extremely large for low
level concentrations and fairly low for higher concentrations. For example, a 0.20 mg N/L concentration
has an analytical variability of 10% associated with it; whereas, a 1.20 mg N/L concentration has an
analytical variability of 2%.

Reporting Significant Figures: The number of significant figures  used by a laboratory to report analytical
results can significantly affect data interpretation in a blind audit study. If a laboratory reports only two
significant figures (for whatever reasons) and an audit sample has a true concentration expressed in
three significant figures, then substantial under or over estimates of the true concentration can  be
reported. For example, if a  true  value of 0.035 mg P/L has been prepared and a  laboratory only reports
two significant figures, i.e., 0.03 mg P/L, then the results expressed are 86% of the expected true value.

Preparation of True Standards: Companies that prepare large quantities of unknowns assign acceptable
confidence limits around the  true  value. In one case (SPEX, CertiPrep), the mean recovery and
standard deviation are later reported along with the true concentration and the 95% confidence interval
(Cl). The 95% Cl represents the mean  recovery ±  2 standard deviations and was developed from
regression  equations from Water Pollution Performance Evaluation Studies. A recently purchased set of
these standards gave a true total P value of 3.00 mg P/L with a  95% Cl of 2.47-3.42 mg  P/L.. The lower
end of the 95% Cl recovery allows 82% recovery of the true concentration. This type of statistical
analysis was not performed on the Blind Audit Program samples prepared for this study.

With the above issues in mind and even though only two rounds of the Blind Audit  Program have been
completed, some consistent patterns have been observed that warrant discussion or further
investigation:

1. Reported concentrations  of all analytes except total dissolved phosphorus and chlorophyll a are similar
between laboratories participating in the Blind Audit Program. Except for total dissolved phosphorus, no
laboratory reported concentrations for an individual analyte that  were consistently different from the
range of the other reported concentrations. This probably indicates that all participating laboratories
execute these measurements with accuracy and precision.

2. If possible, all participants should report data from future Blind Audits to three significant figures to
facilitate concentration  comparisons.

3. A 95% Confidence Interval for each  concentration level of every analyte should  be established,
possibly with the assistance of EPA statisticians.

4. One laboratory reported consistently lower concentrations for  total dissolved  phosphorus in both the
low and higher level samples. Although other laboratories reported concentrations for the low level
sample that were similar, none reported similar concentrations for the higher level samples.

5. Reported chlorophyll a concentrations were quite variable. In  connection with these data and other
CBP chlorophyll a data anomalies, the  CBP Quality Assurance Officer is contacting all participants with
respect to methodology-spectrophotometric-one wave length/trichromatic/fluorometric; type of grinding;
use of buffers; etc.
                                              101

-------
Table 2 lists concentrations of analytes where the difference between the reported concentration and the
true concentration was more than two times a typical MDL in both the January and June Blind Audits.
These differences may not be cause for concern since 95% confidence intervals have not been
assigned.
Table 2. Consistent differences noted in 1998 Blind Audit results
Total Dissolved Nitrogen; Low Concentration (mg N/L)
January
Lab.
CBL
HPL
PADER
True
0.35
0.35
0.35
Reported
0.27
0.281
0.40
% of True
77%
80%
114%
June




True
0.35
0.35
0.35
Reported
0.30
0.205
0.42
% of True
86%
59%
120%
Total Dissolved Nitrogen; High Concentration (mg N/L)
January
Lab.
PADER
True
1.05
Reported
1.15
% of True
109%
June


True
.53
Reported
.62
% of True
117%
Total Dissolved Phosphorus; Low Concentration (mg P/L)
January
Lab.
CBL
HPL
PADER
True
.024
.024
.024
Reported
.0285
.020
.02
% of True
119%
83%
83%
June




True
.024
.024
.024
Reported
.0205
.021
.02
% of True
85%
87%
83%
Total Dissolved Phosphorus, High Concentration (mg P/L)
January
Lab.
PADER
True
.096
Reported
.05
% of True
52%
June


True
.048
Reported
.03
% of True
62%
                                             102

-------
        Total Dissolved N
      January 1998 Blind Audit
0.5
0.4
0.3
0.2
0.1
    Illllllll
                   I Data A
   DCLS  CBL  ODU  TECH  TRUE
            X-Axis
                                Total Dissolved N
                              January 1998 Blind Audit
       1.2
       1.1
        1
       0.9
       0.8
    Illllllll
I Data A I
                            DCLS  CBL  ODU  TECH TRUE
                                    X-Axis
        Total Dissolved P
         January 1998
                                Total Dissolved P
                              January 1998 Blind Audit
 0.05
_, 0.04
5: 0.03
^ 0.02
 0.01
   0
Illllllll
I Data A
                               0.15
      D)
      E 0.05
                        I Data A I
    DCLS  CBL  ODU  TECH  TRUE
            X-Axis
                            DCLS  CBL  ODU TECH TRUE
                                    X-Axis
          Ammonium
      January 1998 Blind Audit
                                  Ammonium
                               January 1998 Blind Audit
  0.1
  0.08
  0.06
  0.04
  0.02
Illlllllll
I Data A
 0.38
 0.36
 0.34
E 0.32
  0.3
 0.28
     D)
                               Illlllllll
'Seriesl
    DCLS  CBL ODU TECH PADER
            X-Axis
                            DCLS  CBL
                ODU  TECH PADER
                 X-Axis
                            103

-------
         Nitrite+nitrate-N
      January 1998 Blind Audit
  0.13
g 0.12
u>
E 0.11
  0.1
fflffiffi
1
   'Data A
    DCLS CBL  ODU  TECH PADER
             X-Axis
                                    Nitrite+nitrate-N
                                  January 1998 Blind Audit
1.25
 1.2
1.15
 1.1
1.05
t
                                       Mini

                     • Data A
                               DCLS  CBL  ODU TECH PADER
                                        X-Axis
           Phosphate
      January 1998 Blind Audit
0.05
0.04
0.03
0.02
0.01
  0
     Illlllllll
    DCLS CBL  ODU  TECH PADER
             X-Axis
                     I Data A
                                        Phosphate
                                   January 1998 Blind Audit
           0.34
         _i 0.32
         o  0.3
         E 0.28
           0.26
   Illlllllll
                     I Data A
                               DCLS  CBL  ODU TECH PADER
                                        X-Axis
                             104

-------
  0.1000
  0.0800
  0.0600
  0.0400
  0.0200
  0.0000
           Particulate Nitrogen
         January 1998 Blind Audit
                    • Data A
         DCLS      CBL      HPL
                    X-Axis
         P articulate Carbon
       January 1998 Blind Audit
  0.6000

=3 0.4000

E 0.2000

  0.0000
         DCLS
  0.0325
  0.0320
  0.0315
  0.0310
  0.0305
  0.0300
          Particulate Phosphorus
          January 1998 Blind Audit
                    I Data A I
         DCLS
CBL
  X-Axis
ODU

-------
     Total Dissolved Nitrogen
       June 1998 Blind Audit
U)
0.5
0.4
0.3
0.2
0.1
 0
Illllllll
                        I Data A
   DCLS VIMS  HPL  TECH  TRUE
       Participating Laboratories
                                Total Dissolved Nitrogen
                                 June 1998 Blind Audit
                              0.8
                              0.6
                              0.4
                              0.2
                              0
Illllllll
• Data A
                               DCLS  VIMS  HPL  TECH TRUE
                                    Participating Laboratories
   Total Dissolved Phosphorus
     June 1998 Blind Audit
                              Total Dissolved Phosphorus
                                 June 1998 Blind Audit
0.04
0.03
0.02
0.01
 0
    Illllllll
                       I Data A
                           0.06

                          £ 0.04
                          u>
                          E 0.02
 Illllllll
• Data A
    DCLS  VIMS  HPL  TECH  TRUE
        Participating Laboratories
                                DCLS  VIMS   HPL  TECH TRUE
                                    Participating Laboratories
          Ammonium
       June 1998 Blind Audit
                                      Ammonium
                                   June 1998 Blind Audit
0.05
0.04
0.03
0.02
0.01
 0
    Illllllll
                       I Data A
                           0.05
                           0.04
                           0.03
                           0.02
                           0.01
 Illllllll
I Data A
    DCLS  VIMS  HPL  TECH  TRUE
        Participating Laboratories
                                DCLS  VIMS   HPL  TECH TRUE
                                    Participating Laboratories
                            106

-------
      Nitrite+Nitrate
    June 1998 Blind Audit
0.25
 0.2
0.15
 0.1
0.05
  0
 Illllllll
I Data A
DCLS  VIMS   HPL  TECH   TRUE
     Participating Laboratories
                  Nitrite+Nitrate
               June 1998 Blind Audit
                                    I Data A I
                                       DCLS  VIMS  HPL   TECH
                                            Participating Laboratories
                               TRUE
        Phosphate
    June 1998 Blind Audit
                         I Data A
DCLS  VIMS  HPL  TECH  TRUE
     Participating Laboratories
                   Phosphate
               June 1998 Blind Audit
         0.021
         0.02
         0.019
         0.018
         0.017
                                      I
                                           Illllll
t
    I Data A
            DCLS  VIMS  HPL   TECH TRUE
                 Participating Laboratory
                            107

-------
            Particulate Carbon
           June 1998 Blind Audit
   1.8000
_,  1.7000
O  1.6000
^  1.5000
   1.4000
   1.3000
I Data A
          DCLS  CBL  VIMS  ODU  HPL
              Participating Laboratories
                            Chlorophyll a
                        June 1998 Blind Audit
         U)
25
20
15
10
 5
 0
                                                 I Data A
                                    ODU
                DCLS   CBL   VIMS         ANSP
                      Participating Laboratories
   0.048
   0.046
   0.044
   0.042
    0.04
          P articulate Phosphorus
           June 1998 Blind Audit
1 Data A
         DCLS  CBL   VIMS  ODU  ANSP
              Participating Laboratories
                     P articulate Nitrogen
                     June 1998 Blind Audit
0.3400
0.3200
0.3000
0.2800
0.2600
                                      • Data A
                   DCLS  CBL   VIMS  ODU  HPL
                       Participating Laboratories
                                              108

-------