903R91006
                               CBP/TRS 58/91
                                 May 1991
       Chesapeake Bay Coordinated
               Split Sample Program
          Implementation Guidelines

                         Revision 3
                          v^^^
                          Chesapeake
                                 Bay
                             Program
CB 00824

-------
        Chesapeake Bay Coordinated
             Split Sample Program
          Implementation Guidelines
                     Revision 3
                      Prepared By:

         Analytical Methods and Quality Assurance Workgroup
        of the Chesapeake Bay Program Monitoring Subcommittee
                    c/EPA
                Environmental Protection Agency
                 Chesapeake Bay Program Office
                  410 Severn Avenue-Suite 110
                  Annapolis, Maryland 21403

                      (301) 267-0061
Printed by U.S. Environmental Protection Agency for the Chesapeake Bay Program

-------
                              TABLE OF CONTENTS


LIST OF FIGURES	   ill

LIST OF TABLES	   iii

BACKGROUND	     1

PROGRAM OBJECTIVES	     1

PROGRAM DESIGN CONSIDERATIONS 	     2

SPLIT SAMPLE PROGRAM RESPONSIBILITIES 	     3
     Component Program Responsibilities 	     3
     Data Management and Reporting Responsibilities 	     3
     Coordinated Split Sample Program Oversight Responsibilities  ...     4

SPLIT SAMPLE COLLECTION AND PROCESSING PROTOCOLS	     4

LABORATORY SAMPLE HANDLING AND ANALYSIS PROTOCOLS 	     5


DATA MANAGEMENT REQUIREMENTS AND PROTOCOLS	    10
     Chain-of-custody form	    10
     Diskette submission  	    10
          Data format and parameter names 	    10
          Diskette formats	    13
          Diskette labeling . 	    13
     Hardcopy submission  .	    14
     Accompanying narrative	    14
     Data verification	    14

STATISTICAL DATA ANALYSIS AND REPORTING	    14

COORDINATED SPLIT SAMPLE PROGRAM IMPLEMENTATION 	    18

SPLIT SAMPLE COMPONENT PROGRAMS 	    19
     Chesapeake Bay Coordinated Split Sample Program  	    19
     Mainstern/Tidal Tributaries Component 	    19
     Virginia Mainstem/Tributaries Component  	    19
     Tidal Potomac River Component  	    23
     Non-tidal Tributaries/Fall-line Component  	    23

REFERENCES	    26
                                      11

-------
                               LIST OF FIGURES



Figure 1.  Sequential Field Split Sample Dispensing Order	6

Figure 2. Schematic of the Operational Flow of Analyses	8

Figure 3. Sample Chain of Custody Form	11

Figure 4. Data Submission Form	15

Figure 5. Coordinated Split Sample Program Components	20

Figure 6. Mains tern/Tidal Tributaries Component	21

Figure 7. Virginia Mains tern/Tributaries Component	22

Figure 8. Tidal Potomac River Component	24

Figure 9. Non-tidal Tributaries/Fall Line Component	25
                                LIST  OF TABLES

Table 1. Parameters and their CBP Abbreviations, Holding Times
   and Temperatures	9

Table 2. Sample data set submission format for Coordinated Split
   Sample data	12

Table 3. Analysis of variance design for testing inter-
   organization differences	16

Table 4. Analysis of variance design for testing splitting
   effectiveness	17
                                     111

-------
                                                            Split Sample Plan
                                                           Revision 3, 5/6/91
                                                                 Page 1 of 26
BACKGROUND
     In 1988, the Chesapeake Bay Program's- Monitoring Subcommittee identified
the need to assess the comparability of  the water quality data produced by the
many agencies participating  in the basinwide data  collection  programs.   The
Monitoring Subcommittee's Analytical  Methods  and  Quality  Assurance Workgroup
recommended the implementation of a basinwide  coordinated split sample program
to  address this  programmatic need.   Although  individual laboratories  can
evaluate the performance of  their  own analytical  operations against standard
reference materials,  the most  complete  mechanism  for the  evaluation of total
sampling and analysis system variability is  through the  use  of  field split
samples.  These include both  field and laboratory sources of variability.  The
Coordinated Split Sample Program  (CSSP) was started in June 1989, following the
earlier revision of  these guidelines  (CBP 1989).   This revision incorporates
changes and refinements based on the  first two years of CSSP operation.
PROGRAM OBJECTIVES

     The  major  objective  of  the  Coordinated  Split  Sample  Program is  to
establish  a   measure   of   comparability   between   sampling  and  analytical
operations for water quality monitoring basinwide.   A secondary objective is
to evaluate the in-matrix dilution of  standard EPA reference materials.  These
standard  reference  materials are  analyzed in  appropriate  matrix,  fresh  to
saline, and concentration level to match the sample.

     Continued implementation of the Coordinated Split Sample Program provides
the institutional structure  to  address three important  program coordination
needs:

     o    implemention  of a  valid  statistical approach  for  the evaluation of
          split sample results to assure  the use of  these data in data quality
          assessment;

     o    facilitation  of efforts  to  identify problems  and  achieve solutions
          in   individual  programs  as  revealed  through  the  comparability
          evaluation; and

     o    improvement  of communication of  split  sample analytical  findings
          among  organizations •through a central  computerized data  base for
          split sample  results.

The CSSP provides the  forum  and  information necessary to promote an on-going
refinement of the field  and laboratory techniques rather  than assuming that the
system is static and never  changing.   The  statistical assessment of the data
allows  the  field  and  laboratory personnel   to   improve  their  respective
techniques.   In  addition,  the description  of  the  data  quality  provides the
necessary  information   for  assessment and  application  of  the  data by  the
intended user.                                                 :

-------
                                                            Split Sample Plan
                                                           Revision 3,  5/6/91
                                                                 Page 2 of 26
PROGRAM DESIGN CONSIDERATIONS
    The Analytical Methods and Quality Assurance Workgroup determined through
a series of surveys that ,a  limited number of  split  sampling  programs were in
place.   There was considerable  diversity  in  the  objectives of  these  split
sampling operations and,  therefore,  their designs varied  accordingly.   Prior
to the  implementation  of  the Coordinated Split Sample Program,  few programs
were able to work out a well organized method of data submission,  compilation,
analysis and  timely distribution  to  participating laboratories and agencies.
Most of the other programs use only two-way split samples and therefore do not
provide the multiple comparisons that are part of the CSSP results.

     Because  of  the  difficulties associated  with  measurements  of  waters of
varying salinities and the Monitoring Subcommittee's desire to link tidal and
nontidal  monitoring  programs,  linkages were  created  between  laboratories
routinely analyzing samples  of. comparable salinities.  Through common "third
party" laboratories,  those  laboratories  analyzing only estuarine samples are
linked  with laboratories  whose  analytical  responsibilities  are  limited to
freshwater  tidal/nontidal samples.

       The  Coordinated  Split  Sample  Program's  field   split   samples  and
laboratory duplicate  and spike samples  provide  an estimate of overall sampling
and analytical precision and accuracy.   Since field  split  samples are defined
as samples divided into portions following sampling,  they provide precision and
accuracy information  about all steps after sample acquisition including effects
of sample splitting procedures, sample processing, storage, shipment, analysis,
and data  processing.   CSSP field split samples are divided from  one  large
sample rather than consecutive co-located samples.  Therefore, they provide an
estimate of overall  sampling and analytical precision and accuracy assuming
analytical comparability.   Combined with routine analysis of standard reference
materials such as EPA certified materials,  results  from the Coordinated Split
Sample Program can be  used to  verify analytical comparability and provide an
independent measure of accuracy.

     The actual  structural  design of the Coordinated Split  Sample Program is
based on  a series of interconnected and interrelated split  sample component
programs  organized  around  common   geographical  areas  and  similar  sample
salinitiy   ranges.    The  four  Component  programs  are  the  Mainstern/Tidal
Tributaries Component, the Virginia Mainstem/Tributaries  Component, the Tidal
Potomac River Component, and the Non-tidal Tributaries/Fall-line Component.

     The Mains tern/Tidal Tributaries Component and Virginia Mainstem/Tributaries
Component  form  the  central core of the Coordinated  Split  Sample Program,
interrelating  laboratory  and field  operations working  the Bay  mainstem and
tidal  tributaries.    The   other  components  build  upon  this  network  of
laboratories,  linking  directly  with a  laboratory  or  group  of  laboratories
associated  with  monitoring  the mainstem of the Chesapeake Bay.

-------
                                                            Split Sample Plan
                                                           Revision 3, 5/6/91
                                                                 Page 3 of 26

SPLIT SAMPLE PROGRAM RESPONSIBILITIES

     Component Program Responsibilities

     For each component of the Coordinated Split Sampling Program, one agency
has  been  assigned  the  coordination   responsibility  for  that  component's
operation.  The responsibilities of that lead agency are as follows:

A.   Maintain contact with all  the  participating  field  and lab organizations
     to coordinate all logistics.

B.   Provide for the necessary sampling equipment, sample containers, labels,
     chain of custody paperwork,  etc. at the time of  the quarterly split sample
     collection.

C.   Collect  the  sample  and  prepare  the  splits according  to  the  protocol
     described within this document.

D.   Arrange for the  direct exchange,  transfer  or shipment of the individual
     split  samples  to  each  participating laboratory  within  the  component
     program.  This  may be accomplished by meeting  the  other organizations'
     crew at some mutually satisfactory point, personal delivery or by common
     courier.  The goal should  be  the most  rapid,   feasible  means  of sample
     delivery.  Every attempt  should be made to adhere to normal holding times,
     temperatures, preservatives and filtration arrangements followed routinely
     by  each organization.   Records  should  be  maintained  of  all  handling
     conditions and practices so that data evaluations may be facilitated.

     Data Management and Reporting Responsibilities

     The routine submission of split sample data is the responsibility of each
individual laboratory/agency  and its in-house data  management  organization.
The Chesapeake Bay Liaison Office contractor for the management and operation
of the Chesapeake  Bay Program Computer Center,  Computer  Sciences Corporation
(CSC),  is  the  designated recipient   of  all  data  generated  through  the
Coordinated  Split  Sample Program.   CSC is  responsible  for  the processing,
routine statistical analysis,  report development, and  timely  distribution of
results  back to  the participating  laboratories  and  agencies  within  their
respective component programs  when all  the  data are received from all agencies
within the individual component program.

     The Chesapeake Bay  Program's Monitoring Coordinator and Quality Assurance
Officer both review and evaluate the results of  each split sampling component
program and consult with  the  appropriate individuals in  each  organization to
determine the appropriate response to any significant findings.

-------
                                                            Split Sample Plan
                                                           Revision 3,  5/6/91
                                                                 Page 4 of 26

     Coordinated Split Sample Program Oversight Responsibilities

     The Chesapeake Bay Liaison  Office is responsible for overall coordination
of the Coordinated Split Sample Program.  The Chesapeake Bay Program's
Monitoring Coordinator  assists  organizations  in  working out any  logistical
problems which  cannot  be resolved  by  the  participating organizations.   The
Chesapeake Bay Quality Assurance Officer helps resolve any technical concerns
which arise  concerning  the  sampling and analysis of the split  samples.   The
Analytical Methods and  Quality  Assurance Workgroup reports  regularly  to the
Monitoring Subcommittee on the Program's results and any blockages to the full
implementation of the Program which are based on resource constraints.
SPLIT SAMPLE COLLECTION AND PROCESSING PROTOCOLS

     For each sample collected, the sampling crew fills a single large vessel
such as  a large  carboy  according to normal  sample  handling protocol.   For
instance, if a submersible pump is normally used this would be employed; if a
cubitainer is normally dipped manually,  this would be done.  If multiple grabs
are required to collect adequate volume for the splitting,  these multiple grabs
should be composited in a larger vessel prior to splitting.  Field triplicate
sub-samples  obtained  from a  "single  sample" potentially  reduces  additional
sources of variability caused by sampling sequentially.

     It is imperative that a  strong mixing  be applied  to the large vessel to
ensure  that  the sample  is  uniformly  mixed  and  the  sub-samples  will  be
representative.  This may be  accomplished  with a stir bar arrangement or other
form of mechanical  mixing as long as a vortex is created.   This mixing must
continue for the duration of the splitting operation.  When more particulates
are present  in  the  sample, the  splitting  operation will be more difficult to
accomplish a representative  sample therefore,  more effort must  be applied to
provide a good mix.  As  of  March 1991, the  sample  collection,  stirring and
splitting methods  used  by each  sampling  organization  (defined  in  the Split
Sample Component Program  section) were:

MDE  (Mainstem Component, at  CB4.4)—Samples are collected  by submersible pump
     into a 15 gallon carboy;  sub-samples are split on the boat from the carboy
     agitated with  paint  stirrer  turned by electric drill.

VWCB (Virginia Component, at  TF5.5)—Samples are collected  by submersible pump
     into three 9-liter churn splitters, filled in sequential sections (a few
     liters  at  a time in sequence  1-2-3-1-2-3,  etc.)  until all are filled.
     Sub-samples are split on  the boat from the churn splitters, one used for
     sub-sample  1,  one   for  2,  and  one  for  3.    This  deviates  from  the
     recommended single large  vessel.

DCRA (Potomac  Component,  at  PMS-10)—Samples  are  collected in three 5-gallon
     carboys by dipping.  In  the  lab, these are composited into one 15-gallon
     carboy, and the sub-samples are split from the large carboy'via a spigot,
     while the  carboy is  agitated manually with a paddle.

-------
                                                            Split Sample Plan
                                                           Revision 3, 5/6/91
                                                                 Page 5 of 26

USGS Towson  (Fall  Line Component,  at  CBl.O)—They  have  used a  large churn
     splitter, but  did  not have enough water  to  split  three sub-samples for
     each laboratory.  They may get a cone splitter.

     From the single large vessel of continuou5~sly well  mixed sample, all four
organizations  draw  three  sub-samples   for  each laboratory  in  rotational
succession (see Figure 1).  These sub-samples  were  previously  called "aliquots"
(CBP 1989),  and  the name  was changed to indicate  that  they  are  split in the
field, not in the laboratory.

     For example, if there are three laboratories  receiving the split samples,
the sample in the large vessel would be divided into nine sub-samples in nine
bottles,  three  bottles  for  each   laboratory.    The first  bottle for  each
laboratory would be filled first, then the second and the third,  resulting in
the first laboratory receiving bottles #1, #4 and #7.  Data from this type of
sample collection would indicate if there was any type of bias caused by non-
uniform mixing resulting  in samples that are  not  representative.   Splitting
effectiveness is checked statistically in CSSP reports (see Data Analysis).

    Each  laboratory evaluates these  three  sub-samples  as discrete  samples.
This approach provides an  estimate  of  the variability of  the material in the
composite container which must be  established to  provide assurance  that any
inter-organization  variability  observed  is  not a  function of  a  poorly mixed
initial  sample.    Three  sub-samples are  statistically  the minimum  number
required to permit an  effective analysis of variance.   For similar reasons, the
minimum number of laboratories involved in any  split  sampling operation should
be three wherever possible.

     Once the splitting  is  completed, the sampling  organization processes their
three sub-samples through  their  normal handling  and  preservation procedures.
The remaining sub-samples  for the other organizations are iced down until they
are delivered to the  appropriate organization.   Samples  are  delivered to the
participating laboratories as rapidly as possible.  The  chain of custody form
is sent with the samples  (see Data Submission).    Normal  sample  handling and
preservation  methods  and  holding  times  for  CBP  samples are adhered  to  as
closely as possible.  Any  deviations from normal procedures should be noted on
the chain of custody form and sent  to CSC.
LABORATORY SAMPLE HANDLING AND ANALYSIS PROTOCOLS

     Samples  should be  analyzed  as soon  as possible  after  arrival  at  the
laboratory  to minimize  holding  time  effects.    In  some components  (e.g.,
Virginia Component), participants  have agreed to filter the samples the morning
after collection,  so that all are  analyzed at  approximately the  same time.
This would not be possible when some samples are  delivered by mail (in the Fall
Line Component).  Participants in each  component  should discuss and agree on
when samples will be delivered and analyzed.

-------
£

'8

2L
is
Q
_a>
a.


i
a>

iZ

15

1
a>
J>

                                                                               "
                                                                                w
                                                                              a. a>


                                                                              I 2
                                                                              W Q.
                                                                              a> o

•E8

IS

If
o- E


I8
« a>

£-"i
«= 1
.0 Q.
"O a>

§m

g'-s
                                                                                2
                                                                                cB
                                        00

-------
                                                            Split Sample Plan
                                                           Revision 3,  5/6/91
                                                                 Page 7 of 26


     All samples are analyzed for the parameters listed in Table 1 within the
recommended Chesapeake  Bay Monitoring Program  holding'times.    If  there are
parameters in Table  1  which are not  routinely  measured in CBP  samples  by a
participating laboratory,  these parameters  do  not need  to be  analyzed and
reported solely for  the purposes of  the  CSSP, unless  specific  parameters are
requested for comparison to results from other laboratories in the component.
CSC computes any calculated parameters used in the reports, using the methods
outlined in D'Elia et al.  (1987).  Currently the calculated parameters include
Total  Nitrogen  (TN)  for  all  laboratories,  and  Total  Phosphorus  (TP),
Particulate Phosphorus  (PHOSP), Total Dissolved Nitrogen  (TON), and Particulate
Nitrogen (PN) for those laboratories that do not calculate them directly.

     A complete schematic  of the operational  flow  of  analyses  is outlined in
Figure 2.   Within the  laboratory,  at  least  one of the three  sub-samples is
subjected to the normal  quality control (QC) routine.  If a duplicate and spike
sample are  analyzed  every tenth analysis,  the  sub-sample identified  for QC
should be analyzed as a routine  quality control  sample.  This lab QC sample is
divided into a duplicate and  spiked with  the appropriate standard.  Additional
QC samples from any source can be analyzed  to  fit the lab QC sample frequency.
If the laboratory  elects to run more than  one sub-sample for additional quality
control, these data are to be submitted for evaluation as well.

     To  perform adequate   diagnostics  in  the event that  significant  inter-
organization differences are found,  it is essential that quality control data
are  available  for each laboratory  on their system's  performance with the
specific matrix under  consideration in  this  program.   It will  be  useful to
compare  the  precision  and accuracy  during analysis of the sub-samples  (and
their  matrix)   against  the  routine  precision  and  accuracy   limits  of  the
laboratory over all matrices.

     To supplement the  analyses  of the three sub-samples and the respective QC
sample, EPA standard reference material for each parameter are analyzed where
available.   The analysis   of standard reference materials provides  a  strong
measure of comparability between all laboratories and within one laboratory's
analytical  system over  time.   Quarterly  analysis  of  standard  reference
materials  is  the most  independent   evaluation  of  laboratory  performance
available at this  time.   It  is a critical element  of any  diagnostic efforts
associated with the Coordinated Sample Split Program.

     The   standard   reference  material  (SRM)   should   be    diluted   with
deionized/distilled water  in all components.   In  the Mainstem Component, an
additional standard  diluted  in the  appropriate concentration  saline  matrix
should be analyzed.  The concentration of the estuarine dilution water should
be subtracted as a blank value.

-------
  LARGE
  VESSEL
(see Figure 1
for dispensing
order)
                    Figure 2. Schematic of the Operational Flow of Analyses,
                            Coordinated Split Sample Program
                                                 Split Sample Plan
                                                 Revision 3, 5/6/91
                                                 Page 8 of 26
                                       Triplicate Subsamples
                                      (Field precision estimate)
           Normal Laboratory
            Quality Control
              Procedures
                   Analyze for
                Routine Parameters
          Analyze for
       Routine Parameters
                                              Spike Sample
                                                                  Laboratory  QC
    Analyze for
Routine Parameters
   Analyze for
Routine Parameters
      Analyze for
   Percent Recovery
(Lab accuracy estimate)
          (Lab precision estimate)
          Deionized/distilled
            water dilution
                                     EPA Standard
                                   Reference Material
                                   (Accuracy estimate)
                                            Matrix water
                                              dilution
             Analyze for
           SRM Parameter
                                             Analyze for
                                           SRM Parameter

-------
                                                            Split Sample Plan
                                                           Revision 3, 5/6/91
                                                                 Page 9 of 26
Table 1.  Parameters and their CBP code, holding times and temperatures
Parameter (me/1 except CHLA)
Total phosphorus as P
Total dissolved phosphorus as P
Particulate phosphorus as P
Dissolved orthophosphate as P
Total nitrogen as N
Total dissolved nitrogen as N
Particulate nitrogen as N
Total Kjeldahl nitrogen as N
Dissolved Kjeldahl nitrogen as N
Ammonium as N (filtered)
N02 + N03 as N (filtered)
Nitrite as N (filtered)
Nitrate as N (filtered)
Total organic carbon
Dissolved organic carbon
Particulate organic carbon
Particulate carbon
Silica as Si (filtered)
Total suspended solids
Chlorophyll a (ug/1)
Pheaophytin
Biological Oxygen Demand 5 day
HOLDING
CODE TIME (davs)
TP
TOP
PHOSP
P04F
TN
TON
PN
TKNV
TKNF
NH4
N023
N02
N03
TOC
DOC
POC
PC
SI
TSS
CHLA
PHEA
BODS
28
28
28
28
N/A
28
28
28
28
28
28
28
28
28
28
28
28
28
7
30
30
N/A
TEMPERATURE (°C)
-20
-20
-20
-20
N/A
-20
-20
-20
-20
-20
-20
-20
-20
-20
-20
-20
-20
4
4
-20
-20
N/A
Notes:

1. Report  all  parameters that are measured directly.   Calculated parameters
will be calculated by CSC using the outline in D'Elia et al. 1987.

2. If there are parameters  noted  above  which  are  not  routinely measured by a
participating  laboratory,  these  parameters do  not need  to be analyzed  and
reported solely for the purposes of the Coordinated Split Sample Program.   In
some cases, laboratories may  be requested  to  analyze  parameters performed by
other laboratories  in the  component.   Please  report  all of  the parameters
listed that you routinely analyze.

3. Please report any deviations from these maximum holding times  in a narrative
accompanying the submitted data.

4.  Parameters without   holding   times  and temperature were  determined  by
calculating the concentration from parameters  that were measured directly (per
D'Elia et al. 1987).

-------
                                                            Split Sample Plan
                                                           Revision 3, 5/6/91
                                                                Page 10 of 26

     EPA standard  reference  materials are available  through  the Cooperative
Research and Development  Agreement   (CRADA)  of the  U.S.  EPA  Environmental
Monitoring  and  Support Laboratory,  Cincinnati, Ohio.  • These  EPA certified
materials are obtained from the company awarded the specific contract by EPA.
Standards are or will be available for organic compounds, toxic and hazardous
compounds,  pesticides and  inorganic compounds.  Any questions in  regards to the
EPA standard  reference  materials  should be directed  to  Claudia Walters,  CBP
Quality Assurance Officer, CBLO, at (301) 267-0061 or (800) 523-2281.
DATA MANAGEMENT REQUIREMENTS AND PROTOCOLS

     Chain-of-custody form

    Upon sampling, the chain of custody  form  is  filled out completely by the
sampling agency to convey the necessary information to the other participants
regarding  the  sampling site,  time of collection,  handling and  any special
observations which might affect the  results.   A  sample chain of custody form
for the Mainstem Component (Fig. 3) should be adapted for other components if
not already  in  use.   This  form should be sent with  the  split  samples to the
laboratories, and sent to CSC with the data submission.  Each laboratory should
perform  the requisite  analyses  and  report  concentration  values  for  each
determination on each sub-sample  of  the  split sample,  with the associated QC
and SRM data.

     Diskette submission

     Data  sets are submitted to CSC as an ASCII text file on an IBM diskette,
unless other arrangements  have been made  with CSC.   CSC  has  no keypunching
staff, since virtually all of the  data they receive are on diskettes or tapes.
The provided Data Submission form  (Fig. 4)  is  optional  (see below), and may be
used to prepare the data for keypunching.   If questions arise concerning data
submission,  please  contact  Peter  Bergstrom,  CSC, at  (301)  267-0061  or (800)
523-2281.

           A. Data format and parameter names

     The preferred  data  formal: is standard columnar  text  (ASCII), with each
variable in  a separate column,  and columns separated by spaces.  This can be
created and  edited  with  a word processor if  it  is saved  as  an ASCII or text
file.  If  you use LOTUS or dBASE please  send  an ASCII file (.PRN  from Lotus).
Include all  the variables listed above, in the same order;  values for  PARAM are
in  Table   1.   Please report values  that are below  detection  limits  as you
normally report them, and be sure  they  get the '<' flag.  Report  missing values
for numeric variables as a period.  The data received to date will fit in a 132
column width.   If you cannot  fit  your data  into a  132  column width, please
submit  them as  two  files,  each  less  than  132  columns  wide,  on  the same
diskette.   A sample format with hypothetical data is in Table  2/.   This format
is available on diskette from CSC  to  facilitate data entry.

-------
                                Figure 3.
                                                             Split Sample Plan
                                                             Revision 3, 5/7/91
                                                                   page 11 of 26
                                                     COLLECTED FOR:  Q  P 1C

                                                     BOTTLE NUMBER:  Q i . ^

                   MAIN  BAY SPLIT SAMPLE CUSTODY LOG
COLLECTION  DETAILS:  DATE:
                                         TIME:  / V / fa   DEPTH:  Q, £

                     LOCATION:  ftl £ vHtH          SALINITY:    I \ , ^ T
COMMENTS:  (unusual conditions,  problems, floating algae, high  solids,  etc.)
SPLITTING DETAILS:
COMPOSITE CONTAINER
FILLED BY:
multiple grabs
pump ^-"
other
COMPOSITE SUBSPLIT BY:^-
sequential bottles *-^"^
cone splitter
other
TRANSFER SEQUENCE:
SPLITTING SEQUENCE BOTTLE LABELLED
bottle
bottle
bottle
bottle
bottle
bottle
bottle
bottle
bottle
bottle
bottle
bottle
bottle
bottle
bottle
DATE TIME
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15

MDE
VIMS
CBL
VWCB
MDE
VIMS
CBL
CODU
VWCB
MDE
VIMS
CBL
^OP.U
VWCB
BY WHOM?
- Al
- Bl
- --C1
- El
- A2
- B2
- C2
'-~E2
- A3
- B3
- C3
""- E3
TEMP. OF SAMP
composite  collected & split  ,'^L-lQ-^/J

Subsamples picked up        / -2-~l&~7&
                                                                    (circle one)
                                                                 0°C  4° C   ambie
Subsamples  delivered to lab  17-fQ-QQ  2Q,
          0°C
                                                                            ambie

                                                                        CV.  ambie
FIELD  PROCESSING INFORMATION
 BOTTLE
               FIELD PROCESSING DONE  ON SAMPLE
DATE/TIME
BY
  D  I
NOTE: PLEASE SEND A COPY OF  THIS  COMPLETED FORM TO:  PETER BKRGSTROM,  CSC.
SEVERN AVENUE,  SUITE 110, ANNAPpLIS,  MD,  21403, (800) 523-2281  -

-------
                                                             Split  Sample Plan
                                                           Revision 3,  5/6/91
                                                                 Page 12 of 26
Table 2. Sample data set submission  format, for Coordinated Split Sample data
STATION DATE LAB  PARAM APC MDL SUEISAMP REP MUM RESULTS PCRECOV SPIKE SRM EPA SRM DE
CB5.3  890515  DCLS PHOSP
CBS.3  890515  DCLS PHOSP
CBS.3  890515  DCLS PHOSP QQ
CBS.3  890515  DCLS PHOSP
CBS.3  890515  DCLS TKNW
CBS.3  890515  DCLS TKNW
CBS.3  890515  DCLS TKNW
CBS.3  890515  DCLS TKNW  AA
CBS.3  890515  DCLS SI
CBS.3  890515  DCLS SI
CBS.3  890515  DCLS SI
CBS.3  890515  DCLS SI
 1
 2
 3
1
 1
 2
 3
 1
 1
 2
 3
1
2
1
1
1
2
1
1
1
2
1
1
0 .010
0 .011
0.009
0 .013
0 .754
0.712
0 .657
0 .774
2.42
2.44
2.17
2.39
106

.
t
92
105
.

.
Ill


0 .023
.

t
1.11
1.56

.
.
3.45
,

                               0.016
0.645
 1.65
                                      0.024
                                           SRM_MA
                                            0.019
       0.623  0.554
        1 .54
              1.62
The parameter names  to use and  their meanings  are:
     STATION    is  the  sampling station number (CBS.3,  PMS-10,  etc.)

     DATE       is  the  sampling date,  as YTMMDD

     LAB        is  the  abbreviation for  the  analysis laboratory  (DCLS,  OWML,
                etc.),          not the collecting agency

     PARAM      is  the  CBP abbreviation  for  the parameter (see revised  Table
                1).  Please do not  use  other abbreviations.

     APC        (Analysis  problem  code): One  or two letters, from Table  20 of
                the CBP   Data Management Plan for Water  Quality  Data (latest
                          version  is Revision 2, July 1990).

     MDL        (Method Detection  Limit): For values below the Method Detection
                limit record  '<'.   For samples requiring a dilution  record '>'
                (list the dilution factor in the comments section or narrative).
                Otherwise  leave this blank.

     SUBSAMP    is  the sub-sample number (1, 2, or 3).  Previously this variable
                was called  'ALQ' and included  the lab replicate designation, but
                the codes used by different  submitters have been so diverse that
                identification of  the  replicates was uncertain.  In the future,
                indicate ONLY the field replicate or sub-sample number here, and
                show the lab  replicate number with  the  following new variable;

     REP_NUM    is  being requested for the  first time in CSSP submissions.  It
                indicates  the LAB replicate number (usually  1 or  2,  sometimes
                3,  A,  or 5)  just  as it does  in  other  CBP submissions.   This
                needs to be a separate number for data analysis.-

-------
                                                            Split Sample Plan
                                                           Revision 3, 5/6/91
                                                                Page 13 of 26
     RESULTS   is  the  concentration  for  that  sub-sample.   Report the results
               the  way you normally  repor-t them  to  thfe CBP  (corrected for
               dilution if necessary).

     PCRECOV   is the percent recovery from the spiked  sample.  List it in the
               row  of  the  sub-sample that was spiked.   This  can be done for
               more  than one  sub-sample  if desired;  please report all spikes
               run.

     SPIKE     is the amount of spike added.  This value will be added to the
               RESULTS in that row to get the theoretical total concentration
               used  for-calculation of percent recovery.

     SRM_EPA   is the EPA value for the Standard Reference Material.  The SRM
               results can be put  in any row for that parameter.  Please attach
               a copy of the  "Answer  Sheet" that  came  from the EPA with each
               standard.

     SRM_DE    is  the  lab value  for that  SRM,  diluted in  deionized water.
               Please  list  the  dilution  used in  the Comments  section,  or in
               your  narrative.

     SRM_MA    is  the  lab  value  (if  done) for  the  SRM  diluted  in lowest
               concentration saline matrix.
          B. Diskette formats

     CSC/CBLO staff are currently able to  read the following diskette formats:

     IBM: 5.25", 360KB (PC/XT) and 1.2MB  (AT); OR   3.5", 720KB and 1.44MB

     Macintosh; 3.5", 800KB
          C. Diskette labeling

     All diskettes submitted must be labeled  in the following manner (based on
page 4-4 of  the Chesapeake Bay Program's Data  Management Plan for Water Quality
Data):

     a. Data  format  used and the  software and  version number used  for data
          storage
     b. Creation date
     c. Submitting individual, organization,  and telephone number
     d. Names and a brief description of all files on the diskette

Please contact Peter Bergstrom  of CSC at (301) 267-0061  (or  toll  free (800)
523-2281 or FTS 691-6873) if you have any questions about data submission.

-------
                                                            Split  Sample Plan
                                                           Revision 3,  5/6/91
                                                                Page 14 of 26
     Hardcopy submission
     A hardcopy submission is optional.  Use the data submission form provided
(Fig. 4).  Data entry errors will be minimized  if  each  laboratory  puts their
own data on the standard form.  Please double check all decimal points and all
numbers for clarity, and put decimal points in a separate box.


     Accompanying narrative

     Please attach  to  the data sheets  a narrative detailing  the methods and
procedures followed in  each  step of the analysis process.  Describe how the
sample was handled before and after you received  it and any sample preparations
including any digestions,  preservatives added or filters  used.   Give details
of each analytical method  used (include method  code  if  known),  any dilutions
used, and any unusual conditions or problems.    This information will be very
useful if differences are found in the results.   This narrative can be included
on the diskette (if used) as an ASCII text file if desired.

     Once the narrative has been  sent, future submissions only need to include
any changes in procedures.  Please indicate whether the changes are temporary
or permanent.

     Please attach copies of  the original lab sheets if there are any ambiguous
parameters or other features of  the data  that may  need  clarification.  These
are currently received  from MDHMH, and have been requested from DCLS.

     Data verification

     CSC/CBLO staff will upload the submitted  data  to the VAX  computer and
combine the data for each component into a single data set.  Since errors can
be introduced in  this process, each laboratory  should  verify  that  their data
are  correct  in  the combined data set.  Normally a printout with  a  Data Set
Checklist will be sent with each Interim Report,  requesting verification of the
new  data  in the  report.  The graphs  and analyses  in  the report will help
identify  any  outliers  in the data.  The  data should be  checked  against the
original  lab sheets, and any changes sent  to  CSC before the Annual Report is
produced.  Any data submitted between  the Interim Report and the Annual Report
will have to be .verified before the  Annual  Report is finalized.  Although data
verification is tedious work, the use of the data requires  that the numbers be
as correct as possible.
STATISTICAL DATA ANALYSIS AND REPORTING

     The  data  analysis  and  reporting scheme was developed by Peter Bergstrom
in consultation with  members  of the Analytical Methods and Quality Assurance
Workgroup  (AMQAW)  during 1990.   The  current  approach uses  two  main tools:
graphs of  the  data with  "precision  bars," and  two-way  analysis of variance.

-------
page 15 of 2b








^_
o —
•o

CM
• 0>
O) CNJ
o TT
0. 5
°- o
LLJ
>




_
^*
^
Q£
_ 0
^™ i*^
a 2l tt

J*f LLJ C^
O — ' LU
Z£ Q_
QL ~ f
Q. 3> *"
" •*. ^^ y -J
-d" x~ ^ i^














^^

•
CO

o.
JT
^J
P
~^~
O
O














o

?i
CO



..J
Ou
^j
^H
CO










^
2
£*
a

Ul
LU
D.

CO
LU
h™













• -
•™
CO

^
M)
z
in ^.^
Z.

X













..

g
E
o


*-
>•
D




fid




"c" ^
<5 o
5 « p
s'x-S
Ge £ C
?i
v*'*-'


-
Q-
°|T3

.U
B



f

< 5
5 -^
» ^
x
&






UJ
^
&5





>

5
a

P

O) w
£•0


S *
X' «
8
t£

csu,^r35
<-0
So--
0EI
oT^O



^ TfT
<^-S
^ ^
oi *^^



























































^>
^





































































0
r-






































































CM





































































CM
CM






































































"O





































































CM
0


















-





-.



j

j
J







I

1


































. .
t?
2?
LU
^.
^
^£
O
U








































1

1

1












1-
1"





































































CN
1—





































































^—
CM





































































CM
CM
























-



















































































































CM








































































1




t?
^_
LU
^>
^
«^
O
u -

























































•~
•—





































































CN
'-






































































CM





































































CM







































































r>





































































CM











































































.
c/
^
LI


^
c


-------
                                                            Split Sample Plan
                                                           Revision 3,  5/6/91
                                                                Page 16 of 26

     Graphs of the split sample results show which differences were larger than
the within-laboratory precision.  Based on a discussion with AMQAW members on
A/24/90, within-laboratory precision for th.e graphs is estimated by the larger
of: 1) the Method Detection Limit (MDL); or, 2) the standard deviation of the
three  sub-samples for  each  sample,  which  estimates  analytical  and  field
precision.  In most cases,  the MDL  is  larger  than  the field precision so the
MDL usually sets  the  precision limit.   Graphs of  the cruise means  for each
laboratory show this estimate as "precision bars."   Any laboratory means with
non-overlapping precision bars have differences that  are  larger than within-
laboratory precision.

     The Friedman two-way non-parametric repeated measures  analysis of variance
(ANOVA) with replication within blocks  (Marascuilo and McSweeney 1977) is used
to  test   statistically  for  differences   among   results  from   different
organizations.  In this  design,  the organizations are treatments, the sampling
dates are blocks,  and  the  three sub-samples  are replicates within blocks.  The
null  hypothesis  is   that   there  are  no  consistent  differences  among  the
concentrations measured by  the different  organizations.   The ANOVA design is
shown in Table 3.
Table  3.    Analysis  of  variance  design  for  testing  inter-organization
differences, Coordinated Split Sample Program.
                                   TREATMENTS  (Organizations)
                                   (Potomac Component shown)

                                   CRL/DCRA        DCLS        MDHMH

BLOCKS
(Sampling
dates)

March |
1
June |
1
September |
1
December |
1
1
1
1
1
2
2
2
2
3* |
3 1
1
3 1
3 1
1
1
1
1
1
2
2
2
2
3 1
3 1
1
3 1
1
3 1
|
1
1
1
1
2
2
2
2
3
3
3
3
                                   *Replicates within
                                   blocks  (sub-samples)


     The  Friedman program  used before  (Bergstrom 1990)  did  not  allow for
replicates, so  cruise means  had to be  used in previous reports.   This change
to using  replicate  data uses more information in  the  data and increases the
power of the test, or its ability to detect real differences.  The  test is done
with a computer program written by Peter Bergstrom  in SAS using .the formula in
Marascuilo and McSweeney (1977), including their  formula for post hoc pairwise

-------
                                                            Split Sample Plan
                                                           Revision 3,  5/6/91
                                                                Page 17 of 26

comparisons.   The  program was  tested with  the example  in Marascuilo  and
McSweeney (1977).

     The  Friedman  test  is also  used for  a preliminary  test   of  splitting
effectiveness.  This is done to check whether the results from one sub-sample
were consistently  higher or lower  than  the results from  other  sub-samples.
This would most likely occur if the  later sub-samples, drawn from lower in the
carboy,  had more  sediment in  them or  if some  of the  sample  bottles  are
contaminated.  This test is done separately for each parameter  and sampling
date,  using the  three  sub-samples  as  treatments  and  the organizations  as
blocks.  The null hypothesis is that  there are no consistent differences among
the concentrations  of  the three sub-samples.   The  ANOVA design is  shown in
Table 4.  If the results do not show any differences caused by splitting,  the
analysis for inter-organization differences is done.  If there are significant
effects of  splitting, the  data  must  be examined to  see  if  these  effects will
bias the tests of inter-organization differences.
Table 4.   Analysis of  variance  design for testing  splitting effectiveness,
Coordinated Split Sample Program.

                                TREATMENTS (Sub-samples)

                               1             23
Blocks
(Organi-
zations)


DCLS
HRSD
ODU
VIMS
1
1
1
1
1 2
2
2
2
3 1
3 1
3 1
3 1
                          (Note: No replicates within blocks.)
     Statistical  significance is  assumed when  P <  0.01 (Bergstrom  1990).
Standard quality control procedures use the P = 0.01 level as the "control" or
action level for precision and accuracy  charts (e.g.,  Montgomery  1985).   The
P < 0.01 standard  is now attainable for most of the parameters most components,
due to larger sample sizes and the use of replicates in the Friedman test.

     Accuracy data,  from percent recoveries  and Standard  Reference  Material
(SRM) analysis,  are included  in tables for each report,  but  are not the subject
of  statistical  tests.   These data are used to  supplement the split  sample
results, and for diagnostic  purposes  if  the split  sample  results  show inter-
organization differences  that should  be  investigated.   Confidence limits for
SRMs  (provided  by EPA)  are  included only when  the  standards were  analyzed

-------
                                                            Split Sample Plan
                                                           Revision 3,  5/6/91
                                                                Page 18 of 26

without dilution, since the confidence  limits do not apply to diluted samples.
SRMs  analyzed  after dilution  are reported  as the diluted  value, with  the
expected value based on the dilution used.-

     Variability  data  are  reported   for  within-organization  and   inter-
organization  variability,  as  the standard  deviations  and  coefficients  of
variation (CV).  Within-organisiation  variability comes from the variability of
the triplicate sub-samples  (field replicates)  for each  sampling date,  using
REP_NUM=1 if there were laboratory replicates done, using the MEANS procedure
in SAS.  Variability of laboratory replicates is used  for diagnostic purposes
only,  since it includes fewer sources of variability than the field replicates
(sub-samples).   Inter-organization variability is calculated among  the mean
concentrations of the  three sub-samples reported  by each laboratory for each
sampling date, using  the  MEAN  and CV functions is SAS.   All the variability
estimates are reported in one table  to facilitate comparisons.

     Each quarterly  meeting of AMQAW  includes  an update on  the latest CSSP
results by CSC/CBLO staff.  In addition, CSC/CBLO staff contact data submitters
as soon as possible after receipt of the  CSSP  data if  there  appear to  be any
errors or outliers in the data or other problems with the data.

     Written results of the statistical analysis of the split sample data are
routinely  distributed  to  agencies  and  laboratories  within the  individual
program components.   These are  called Interim Reports.  The original quarterly
reporting schedule has proved  impractical  due  to  slow  submission of data and
the time needed  for report  preparation.   The current schedule is one Interim
Report per year per  component,  including data from two or three sampling dates
from that year.  It may be possible  to prepare  two Interim Reports a year per
component if data submission and reporting are streamlined.  An Annual  Report
is prepared summarizing the results of  all  components and any responses  to the
Interim Reports  (Bergstrom 1990).   The Annual Report  is approved  by  AMQAW,
distributed  to  all  participating agencies  and  laboratories,   and  formally
presented to the Monitoring Subcommittee.
COORDINATED SPLIT SAMPLE PROGRAM IMPLEMENTATION

     Implementation  of the  Coordinated  Split Sample  Program is  an ongoing
process, and  the  program requirements change slightly over  time.   Since the
program's inception in June 1989, most organizations have  made changes in their
protocols  to  achieve  full  implementation.   The  few remaining gaps  in full
implementation are the result  of  problems  with  obtaining splitting equipment
or Standard Reference Materials.  Hopefully these  will  be remedied  in the near
future.
     Samples  are  split  quarterly.    Discussions  of CSSP  implementation and
statistical analysis of the data  occur  during the  quarterly AMQAW meetings
attended by the coordinating agency,  field  members and  lab representatives for
each component and  the CBP Quality Assurance Officer  and  the OBP Monitoring
Coordinator.

-------
                                                            Split Sample Plan
                                                           Revision 3, 5/6/91
                                                                Page 19 of 26
SPLIT SAMPLE COMPONENT PROGRAMS

          Chesapeake Bay Coordinated Split Sample Program (See Figure 5)

Coordinating Agency: U.S. EPA Chesapeake Bay Liaison Office
Coordinated Split Sample Program Coordinator: Joe Macknis (800) 523-2281
Technical Program Coordinator: Claudia Walters (800) 523-2281
Data Management Coordinator: Peter Bergstrom, CSC (800) 523-2281
Statistical Analysis Coordinator: Peter Bergstrom, CSC (800) 523-2281

The overall Coordinated  Split Sample Program will be coordinated through the
Chesapeake Bay  Liaison Office (CBLO).   Computer Sciences  Corporation  (CSC)
staff  at  CBLO will  have  the  major   responsibility  for  centralized  data
management,  statistical  analysis  and  reporting   on   behalf  of  all  the
participating agencies and laboratories.

          Mainstem/Tidal Tributaries Component (See Figure 6)

Coordinating and Field Sampling Agency:  Maryland Department of the Environment
Component Program Coordinator: Bruce Michael

Contact Persons:
Robert Magnien, MDE (301) 631-3680 (MD MONITORING PROGRAMS)
Bruce Michael, MDE (301) 631-3680 (MD MAINSTEM FIELD)
Carolyn Keefe, CBL (301) 326-4281 (MD MAINSTEM LAB)
Sally Bowen, MDE (301) 974-3238 (MD TRIBUTARY FIELD)
Alvin Sober, MD DHMH (301) 225-6200 (MD TRIBUTARY LAB)
Frederick Hoffman, WCB (804) 367-6683 (VA MONITORING PROGRAMS/VA TRIE FIELD)
Betty Salley, VIMS (804) 642-7213 (VA UPPER MAINSTEM FIELD AND LAB)
Steve Sokolowski, ODU (804) 683-4524 (VA LOWER MAINSTEM FIELD AND LAB)
Norma Roadcap, DCLS (804) 786-4853 (VA TRIBUTARY NUTRIENT LAB)
Robert Potts, DCLS (804) 786-4826 (VA TRIBUTARY CARBON/TSS/BOD LAB)

The  Mainstem/Tidal Tributaries  Component  forms  the central core  of  the
Coordinated Split Sample Program, interrelating laboratory and field operations
working the Bay tidal  mainstem and  tributaries.  It is the only component that
analyzes saline  samples.   Sampling  is  performed by  a  field crew  from MDE,
currently at Station CB4.4, and before June 1990 at Station CBS.3.

          Virginia Hainstem/Tributaries Component (Figure 7)

Coordinating and Field Sampling Agency:  Virginia Water Control Board
Component program Coordinator: Frederick Hoffman

Frederick Hoffman, WCB (804) 367-6683  (VA MONITORING PROGRAMS/VA TRIE FIELD)
Betty Salley, VIMS (804) 642-7213 (VA UPPER MAINSTEM FIELD AND LAB)
Steve Sokolowski, ODU (804) 683-4524 (VA LOWER MAINSTEM FIELD AND LAB)

-------
JB
ID in
CO
CM
to §>
W DC 0.
 41

 I

 O
U


 2

 2
a,
 V
"a

 ea
 a

 •a

 "eS

 •3
 s

                                                                      - - "S
                                                                     |||
                                                                     o is 03
                                                                     CL .9 o
                                                                     ssi-
                                                                     Q) co £
                                                                     :— c: O
                                                                     m
                                                                     o
                                                                   ID

-------
S2 9?


°-!x«>
& IO CM

9-co o



W-2 w
w W m
S > O)
Q. O) TO

w tr Q.
 NC



 a
 6JD
                                                                                   1
UJ

O
o^n
o I—I

-------
J5
a> m CM
  -co'B
•=  >  0>

CO CC D!
                                                                                     C/)
                                                                                                                                    w c
                                                                                                                                       8.
                                                                                                                                 O 0
                                                                                                                                 W CO
                                                                                                                                 CC CC
                                                                                                                                 I X
  I

D

-------
                                                            Split Sample Plan
                                                           Revision 3, 5/6/91
                                                                Page 23 of 26

Norma Roadcap, DCLS (804) 786-4853 (VA TRIBUTARY NUTRIENT LAB)
Robert Potts, DCLS (804) 786-4826 (VA TRIBUTARY CARBON/TSS/BOD LAB)
Drew Francis, HRSD (804) 460-2261 (HRSD FIELD AND LAB) '

The  Virginia Mainstem/Tributaries  Component  links  the  field agencies  and
laboratories  involved  in sampling the  Virginia Chesapeake Bay  mainstem and
tributaries, building on the existing, routine split sampling program between
the  Virginia Institute of  Marine Science and  the Old  Dominion  University.
Sampling is done by a field  crew from VWCB at  Station TF5.5,  in  the James River
at Hopewell, VA.


          Tidal Potomac River Component (Figure 8)

Coordinating Agency:  Metropolitan Washington Council of Governments
Component Program Coordinator: Tom An
Field Sampling Agency: DC Department of Consumer and Regulatory Affairs

Contact Persons:
Tom An, MVCOG (202) 962-3366 (POTOMAC COORDINATED MONITORING PROGRAM)
Morris Hennesy, MDE (301) 974-3677 (MD TRIBUTARY PROGRAM)
Sally Bowen, MDE (301) 974-3677 (MD TRIBUTARY FIELD)
Alvin Bober, MD DHMH (301)  225-6200 (MD TRIBUTARY LAB)
Hamid Karimi, DC DCRA (202) 404-1120 (DC MONITORING PROGRAMS)
Sheila Besse, DC DCRA (202) 404-1120 (DC FIELD)
Al Robertson, DC DCRA/CRL (301) 266-9180 (DC LAB)
Frederick Hoffman, VWCB (804) 367-6683 (VA MONITORING PROGRAMS)
Jeff Talbott, VWCB/NRO (703) 490-7352 (VA TRIBUTARY FIELD)
Norma Roadcap, DCLS (804) 786-4853 (VA TRIBUTARY NUTRIENT LAB)
Robert Potts, DCLS (804) 786-4826 (VA TRIBUTARY CARBON/TSS/BOD LAB)

Building upon  the  existing Potomac Regional Monitoring  Program's co-located
split  sampling program,  the  laboratories  and  field agencies   involved  in
sampling  the Potomac River are  incorporated  into  the Baywide split  sample
program through the Tidal Potomac River Component.  The samples are collected
by a field crew from DCRA at Station PMS-10,  at Key Bridge.


          Non-tidal Tributaries/Fall-line Component (Figure 9)

Coordinating and Field Sampling Agency: USGS Mid-Atlantic Regional Office
Component Program Coordinator: Joel Blomquist

Contact Persons:
Dwayne Womer, PA DER (717)  787-9637 (PA PROGRAM)
Ken Walizer, PA DER (717) 787-8184 (PA FIELD)
Lynn Schaffer, PA DER (717) 783-1998 (PA LAB)
Bruce Michael, MDE (301) 631-3680 (MD FALL-LINE PROGRAM)
Linda Zynjuk, USGS Towson (301) 828-1535 (MD FALL-LINE FIELD)  -
Joel Blomquist, USGS Towson (301) 828-1535 (MD FALL-LINE LAB)

-------

  in
   to
g-to o
E c Tt
~ > 0)
Q. 03 CD

CO DC Q.
   2

   S
   a.

   a
   a.
  OJ

  •o
   4»
  •S *•*
   2 B
   C «   03

                                                                 £*&
                                                                 W° TJ
                                                                 £2 S

                                                                 e|i
                                                                 £53 »
                                                                 o>r g
                                                                 p re O

                                                                 £i*
                                                                 its
                                                                 Lu < •
                                                                   w
00

Ss
> Q

-------
_0>
       CM
    .
    .52  o>
=  >  o>
 Q. 0>  TO
W  DC D.
       §
       a
       OJD

-------
                                                            Split Sample Plan
                                                           Revision 3,  5/6/91
                                                                Page 26 of 26

Alvin Bober, MD DHMH (301) 225-6200 (MD TRIBUTARY LAB)
Tom Grizzard, OWML (703) 361-5606 (MD POTOMAC FALL-LINE PROGRAM)
Harold Post, OWML (703) 361-5606 (MD POTOMAC FALL-LINE FIELD)
David Sirois, OWML (703) 361-5606 (MD POTOMAC FALL-LINE LAB)
Frederick Hoffman, VWCB (804) 367-6683 (VA MONITORING PROGRAMS/TRIB FIELD)
Norma Roadcap, DCLS (804) 786-4853 (VA TRIBUTARY NUTRIENT LAB)
Robert Potts, DCLS (804) 786-4826 (VA TRIBUTARY CARBON/TSS/BOD LAB)
Donna Belval, USGS Richmond (804) 771-2427 (VA FALL-LINE FIELD/LAB)

The Non-tidal Tributaries/Fall-line Component  links  those  field agencies and
laboratories involved in sampling the  fall line stations of all the Chesapeake
Bay tributaries.  The Susquehanna River Basin Commission's monitoring programs
are  also  linked  to  this   component  via  the  Pennsylvania  Department  of
Environmental Resources laboratory.  All  samples are  collected by a field crew
from  USGS Towson  at  the Susquehanna  River   fall  line  station  (CB1.0)  at
Conowingo, MD.
REFERENCES
Bergstrom, P.   1990.   Chesapeake  Bay  Coordinated  Split  Sample Program Annual
     Report, 1989.   CBP/TRS 51/90, EPA Chesapeake Bay Program, Annapolis, MD.

Chesapeake Bay Program.  1989.  Chesapeake Bay Coordinated Split Sample Program
     Implementation  Guidelines,   Revision  2.   EPA  Chesapeake Bay  Program,
     Annapolis, MD.

Chesapeake Bay  Program.   1990.  Data  Management  Plan for  Water quality Data.
     EPA Chesapeake Bay  Program,  Annapolis, MD.

D'Elia, C., R.  Magnien, C.  Zimmermann, P. Vaas, N.  Kaumeyer,  C. Keefe, D. Shaw,
     and K. Wood.  1987.  Nitrogen and phosphorus determinations in estuarine
     waters:  A  comparison of  methods   used  in  Chesapeake  Bay  Monitoring.
     CBP/TRS 7/87, EPA Chesapeake Bay Program, Annapolis, MD.

Marascuilo, L.  A., and M.  McSweeney.   1977.   Nonparametric and distribution-
     free  methods for  the  social sciences.    Brooks/Cole  Publishing  Co.,
     Monterey,  CA.

Montgomery, D.  C.  1985.  Introduction to statistical quality control.  Wiley
     & Sons, NY.

-------
         OS
     (U  ••=;
     (ft  c
 >.~  i>
 ro  S  F
CD  Q.  £
ro  *>
v  -S
Q.  c
CO -=:
O
               (N
               03
               O
               O

               m
               o

               Q
               Q

-------