March 1983 i Parti
United States Environmental Monitoring Systems
Environmental Protection Laboratory
Agency Research Triangle Park NC 27711
Research and Development EPA-600/4-82-042a Jan. 1981
v>EPA Quality Assurance
Handbook for Air Pollution
Measurement Systems:
Volume V. Manual for
Precipitation Measurement
Systems
Part I. Quality Assurance
Manual
-------
Parti ii March 1983
Disclaimer
This report has been reviewed by the Environmental Monitoring Systems
Laboratory, U.S. Environmental Protection Agency, and approved for publication.
Approval does not signify that the contents necessarily reflect the views and
policies of the U.S. Environmental Protection Agency, nor does mention of trade
names or commercial products constitute endprsement or recommendation for
use.
-------
March 1983 iii Parti
Foreword
Measurement and monitoring research efforts are designed to anticipate •
potential environmental problems, to support regulatory actions by developing an
in-depth understanding of processes that impact health and the ecology, to provide
innovative means of monitoring compliance with regulations, and to evaluate the
effectiveness of health and environmental protection efforts through the monitoring
of long-term trends. The Environmental Monitoring Systems Laboratory, Research
Triangle Park, North Carolina, is responsible for development of environmental
monitoring technology and systems; EPA-wide quality assurance programs for air
pollution measurement systems; and technical support to state agencies and to
EPA's Office of Air, Noise, and Radiation; Office of Pesticides and Toxic Substances;
and Office of Solid Waste and Emergency Response.
This quality assurance manual has been developed to aid agencies, which plan to
measure precipitation, develop adequate quality assurance programs. This manual
covers both laboratory and field measurement systems currently used in precipita-
tion monitoring. Requirements of the EPA-wide quality assurance program have
been incorporated into this manual. The main goal of the EPA quality assurance
program is to provide documented environmental monitoring data of adequate pre-
cision and accuracy.
Thomas R. Mauser, Director-
Environmental Monitoring Systems Laboratory
Office of Research and Development
-------
March 1983 iv Part I
Abstract
The quality assurance manual for precipitation measurement systems was
designed to assist agencies and personnel in obtaining high quality data in
monitoring and analyzing precipitation samples. Since precipitation samples
contain trace (micromolar) quantities of constituents, they are susceptable to
appreciable error due to contamination as well as chemical and biological
degradation. Sections in this manual address requirements which should be
incorporated into every quality assurance project plan. The manual presents
guidelines for assessment of precipitation measurement data in terms of precision,
accuracy, representativeness, completeness, and comparability. The topics covered
are quality assurance objectives; organiation and planning; documentation; siting;
field operations; laboratory operations; data handling, validation and reporting; and
data quality assessment. All or most of these guidelines and protocols should be
adapted and used for any precipitation monitoring program.
-------
Parti
March 1983
Contents
Section
1.0 Introduction
2.0 Quality Assurance Objectives
3.0 Planning and Organization .
4.0 Documentation
5.0 Siting
6.0 Field Operations
7.0 Laboratory Operations
8.0 Data Handling," Validation, and Reporting
9.0 Data Quality Assessment
Appendix A Procedure Used for Hubaux and Vos
Detection Limit Calculations
Appendix B Methods Validation Study for Pb
Analysis by Flame Atomic Absorption
Pages
1
1
2
1
8
28
38
10
6
1
2
Date
1/1/81
1/1/81
1/1/81
1/1/81
1/1/81
1/1/81
1/1/81
1/1/81
1/1/81
1/1/81
1/1/81
-------
Parti vi March 1983
Acknowledgements
This quality assurance manual was prepared for the Environmental Monitoring
Systems Laboratory, Research Triangle Park, North Carolina, under the direction of
the project officer, John C. Puzak. Material in this manual has been based on
current EPA quality assurance procedures, on air and water monitoring methods,
and on procedures used in the Electric Power Research Institute (EPRI) Acid
Precipitation Study, the National Atmospheric Deposition Program (NADP), and the
Multi-State Atmospheric Power Production Pollution Study (MAP3S). The authors
extend grateful appreciation to the staff at Rockwell International's Environmental
Monitoring & Services Center, Newbury Park, California, and to the many reviewers
of the drafts of this document for their contributions.
-------
Jan. 1981
Part I-Section 1.0
1.0 Introduction
Wet deposition is one of the chief
sinks for atmospheric pollutants; these
pollutants include substances benefi-
cial as well as harmful to the environ-
ment. The harmful pollutants formed
primarily during dissolution and re-
moval of both gaseous and paniculate
nitrogen and sulfur oxides (NO. and
SO.) in rain and snow produce acid
precipitation with a pH less than 5.6—
the equilibrium value due to atmos-
pheric carbon dioxide (CO2) alone.
Increasing awareness of the poten-
tially harmful effects of acid precipita-
tion on soil, forest, and lake ecologies as
well as its corrosive effect on various
materials have made it the subject of
great scientific and social concern. One
result of these concerns has been an
increase in the number of programs for
monitoring acid precipitation. Monitor-
ing programs generally have one or
more of the following objectives:
1. To obtain a data base of variables
for evaluating acid precipitation
trends in remote areas generally
unaffected by pollutants.
. To observe and to relate acid
precipitation trends throughout a
region, including urban and non-
urban areas, to meteorological or
seasonal conditions as well as to
emissions.
3. To provide a data base for evaluat-
ing effects, for developing and
evaluating abatement strategies,
and for developing and validating
models.
Because precipitation samples collected
during the monitoring contain trace
(micromolar) quantities of constituents,
they are susceptible to appreciable error
due to small changes from contamina-
tion as well as chemical and biological
degradation. Thus stringent procedures
and protocol are essential to preserve
the chemical integrity of the sample
before analysis.
Obtaining a high quality data base
requires a quality assurance program
which includes the following phases of
monitoring:
1. Organization and planning;
2. Site selection to obtain representa-
tive samples for the region;
3. Field operations;
4. Laboratory operations;
" Data handling, validation, assess-
ment and reporting; and
„. Documentation.
Quality control and quality assurance
protocols for the above phases presented
herein can be used, or they can be
adapted for the quality assurance pro-
gram of any precipitation network. Pro-
cedures required to conduct the field
and laboratory operations are in the
companion O&M Manual (1).
The purpose of this quality assurance
manual is to provide guidelines and
protocols for operators, project officers,
and program managers carrying out a
quality assurance program in all phases
of precipitation monitoring. The mate-
rial in this manual is based primarily on
the quality assurance guidelines in the
EPRI (Electric Power Research Institute)
Acid Precipitation Study in the North-
eastern United States, the NADP
(National Atmospheric Deposition Pro-
gram), and MAP3S (Multi-State Atmos-
pheric Power Production Pollution
Study). The EPA handbooks for quality
assurance of air pollution measure-
ments (2,3) were used as guides for
format and content. To have this acid
precipitation manual stand alone with-
out requiring reference to the O&M
manual (1), some duplication of material
was necessary.
References
1. Operations and. Maintenance
Manual for Precipitation Mea-
surement Systems. U.S. Environ-
mental Protection Agency, Re-
search Triangle Park, N.C., (in
press).
2. Quality Assurance Handbook for
Air Pollution Measurement Sys-
tems - Vol. I - Principles, U.S.
Environmental Protection Agency,
Research Triangle Park, N.C., EPA-
600/9-76-005 (March 1976).
3. Quality Assurance Handbook for
Air Pollution Measurement Sys-
tems • Vol. II -Ambient Air Specific
Methods, U.S. Environmental Pro-
tection Agency, Research Triangle
Park, N.C., EPA-600/4-77-027a
(May 1977).
-------
Jan. 1981
Part l-Section 2.0
2.0 Quality Assurance Objectives
The agencywide EPA quality assur-
ance policy stipulates that every moni-
toring and measurement project must
have a written and approved quality
assurance (QA) project plan (1,2). This
requirement applies to all environ-
mental monitoring and measurement
efforts mandated or supported by EPA
through regulations, grants, contracts,
or other formal means not currently
covered. To generate and report moni-
toring data of the highest quality, no
precipitation monitoring project should
be initiated without a written, approved
QA project plan.
Each agency or laboratory generating
precipitation data has the responsibility
for implementing at least minimum
procedures which assure that precision,
accuracy, completeness, and repre-
sentativeness of the data are known and
documented. In addition, each should
specify the quality levels which data
must meet to be acceptable.
All project personnel should be
familiar with the policies and objectives
'•ned in the QA project plan to assure
er interaction between the field
operations, laboratory operations, and
data management.
2.1 QA Project Plan Elements
The QA project plan should specify
the policies, organization, objectives,
functional activities, and QA and QC
activities needed to achieve the' data
quality goals of the project. The 16 items
listed below must be addressed in each
QA project plan:
1. Title page with provisions for
approval signatures;
2. Table of contents;
3. Project description;
4. Project organization and respon-
sibility;
5. QA objectives for precipitation
measurement data in terms of
precision, accuracy, complete-
ness, representativeness, and
comparability;
6. Sampling procedures;
7. Sample custody;
8. Calibration procedures and fre-
quencies;
9. Analytical procedures;
10. Data reduction, validation, and
reporting;
Internal quality control checks and
frequencies;
12. Performance and system audits
and frequencies;
13. Preventive maintenance proce-
dures and schedules;
14. Detailed routine procedures for
assessing data precision, accur-
acy, and completeness of specific
measurements;
15. Corrective actions; and
1 6. Quality assurance reports to man-
agement.
All measurement results must be
representative of the media (air, precipi-
tation, etc.) and the conditions. Data
quality objectives for accuracy and
precision established for each mea-
surement parameter should be based
on prior knowledge of the measurement
system; on method or validation studies
using replicates, spikes, standards,
calibrations, recovery studies, and blind
samples, and on requirements of the
specific project.
2.2 Specific Objectives
The QA objective is to produce data
that meet the user's requirements and
the monitoring goal. The QA activities
described in this manual are designed to
assist in generating data that are
complete, precise, accurate, repre-
sentative, and comparable.
Completeness-\n general, precipita-
tion data can be considered complete if
90% or more of the total possible
observations of events are available.
The percentage of valid usable chemical
analysis data can be slightly lower due
to sample contamination in the field
before shipment and due to container
leakage in transit.
Precision and Accuracy -Interlabora-
tory studies conducted by EPA's En-
vironmental Monitoring Systems Lab-
oratory (EMSL) on water (not precipita-
tion) systems indicate about ±10%
precision and accuracy for conductivity
and ±0.1 pH unit. Precision and
accuracy for chemical analysis are
determined separately by each labora-
tory for each analytical technique and
range.
Representativeness - Data must be
representative of the conditions being
measured. This is addressed initially in
the siting criteria (Section 5.0) and later
in discussions of continuous observa-
tions of land use and development
which can impact the data generated by
the monitoring site.
Comparability - Data reported by
different networks should be in con-
sistent units to allow for data compari-
son; the recommended units are:
Sample weight gm
Precipitation mm
pH pH units
Conductivity micromho/cm or
microS/cm
Concentration micromoles/liter
Implementation of these QA objec-
tives will produce data that are complete,
accurate, precise, representative and
comparable. Attainment of these objec-
tives requires involvement of the QA
coordinator in the planning stages of the
measurement project. This allows the
identification of areas with potentially
large negative impact on data quality
and provides a mechanism for instituting
a system of internal QC, corrective
action, data validation, and external
assessment of precision a.nd accuracy
to avoid such weak areas. Each of the
objectives is discussed extensively in
the following sections.
2.3 References
1. D.M. Costle, Administrator's Mem-
orandum. EPA Quality Assurance
Policy Statement. U.S. Environ-
mental Protection Agency, Wash-
ington, D.C., May 30, 1979.
2. D.M. Costle, Adminstrator's Policy
Statement. Quality Assurance
Requirements for all EPA Extra-
mural Projects Involving Environ-
mental Measurements, U.S. En-
vironmental Protection Agency,
Washington, D.C., June 14, 1979.
3. U.S. EPA Quality Assurance Man-
agement Staff, "Interim Guide-
lines and Specifications for Prepar-
ing Quality Assurance Project
Plans". QAMS-005/80, U.S. En-
vironmental Protection Agency,
Washington, D.C., December 29,
1980.
-------
Jan. 1981
Part l-Section 3.0
3.0 Planning and Organization
The agencywide EPA policy requires
participation by all EPA regional offices,
EPA program offices, EPA laboratories,
and States in a centrally managed QA
program, and the policy includes all
monitoring and measurement efforts.
Accordingly, a QA program plan pre-
pared by each program office, regional
office, and laboratory should include the
overall policies, organization, objec-
tives, and functional responsibilities
designed to achieve data quality goals
for monitoring and measurement activ-
ities. The 11 elements required in any
QA program plan (1) are:
Identification of organization;
Description of the organization re-
sponsibilities;
QA policy statement;
QA management;
Personnel qualifications;
Facilities, equipment and services;
Data generation;
Data processing;
P=>»a quality assessment;
Active actions; and
.ementation requirements and
schedules.
Monitoring for precipitation chemical
analysis requires strong interfaces
between field operations, laboratory
operations, and data management.
Quality assurance management should
be a separate function in the organiza-
tion.
3.1 Personnel Assignments
and Responsibilities
3.1.1 Organization
A precipitation monitoring network
involves interdependent field monitor-
ing and laboratory operations. Each
operation has its own QA aspects. The
field monitoring sites and the laboratory
can be run by independent agencies or
organizations; however, the results are
reported to and are the responsibility of
the program manager. Reporting direct-
ly to the program manager are the field
manager and the laboratory manager or
supervisor. The QA coordinator should
also report directly to the program
manager. Brief discussions of the
qualifications and duties of the program
personnel are presented below.
1 Program Management
program manager is required to
assure that high quality data are
generated within the time and funding
constraints of the program. The man-
ager must keep abreast of all develop-
ments, make necessary decisions,
review data and QA reports, issue
progress reports, and interpret the
results. The program manager should
have a degree in chemistry and some
experience in managing projects. Full
time employment is recommended. This
position can be combined with that of
the field manager.
3.1.3 QA Coordination
A QA coordinator should report
directly to the program manager with
inputs into the other program functions
and with reports on the quality of the
data generated and on corrective
actions that need to be taken.
3.1.4 Field Operations
The personnel needed to carry out the
field duties in a precipitation monitoring
network include a field manager and
station operators.
Field Manager - The field manager
may be a member of the agency or
organization operating the stations or a
member of the central laboratory staff;
should have a college degree, preferably
in chemistry; should be familiar with all
the procedures; and should have
experience in the operation of all
equipment. The duties are to solve field
problems, to notify the program man-
ager of such problems, to oversee, and
to train, or coordinate the operators.
Station Operators • The operator's
duties at the station require only his part
time presence. Station operators should
have a technical background, but need
not have a college degree. All operators
should have the training to perform at
the necessary level of knowledge and
the skill required to obtain and report
high quality data. A short-term course in
"hands-on" training is recommended;
this should be followed by on-the-job
observation immediately after the
course and by.a semi-annual inspection
thereafter. The training should cover all
pertinent aspects of the O&M manual
(2), which will be given to all personnel;
the O&M manual contains the pro-
cedures necessary to generate a high
quality data base and a schedule of
tasks. If, at any time, an operator's
performance deteriorates, additional
training must be provided by the field
manager as soon as possible, by a
refresher course or by on-site guidance
(Section 6.7).
3.1.5 Laboratory Operations
Each analytical laboratory should
have at least the following types of
employees: (1) director, (2) supervisor,
(3) analyst and (4) QC chemist. Informa-
tion on requirements, training, and
supervision of these people is presented
below.
Laboratory Director • This position is
recommended, but not required if the
laboratory has a supervisor; it may be
combined with the supervisor position
listed below. Either a laboratory director
or a laboratory supervisor is required.
The director should have a minimum of
one year's laboratory analytical experi-
ence in water/wastewater related
measurements; should have a degree in
chemistry; and should be employed full
time.
Laboratory Supervisor - This position
is not required if there is a laboratory
director, but it is recommended that a
laboratory supervisor be employed full
time, and have similar academic train-
ing and experience to that specified for
the laboratory director.
Laboratory Analyst - This person
should be employed full time, and
trained to perform with some supervi-
sion all routine chemical measure-
ments on water samples. Academic
training should be completion of at least
1 year of college chemistry and/or a
laboratory-oriented vocational course; 1
year of experience in water/waste-
water analysis is recommended, but if
this experience is lacking, a minimum of
30 days of on-the-job training in
measurements performed by the agency
is highly recommended. The analyst
must be supervised by an experienced
professional scientist—the laboratory
director, the supervisor or a similarly
trained individual.
Quality Control Chemist • This re-
sponsibility is not necessarily a full time
position; it may be part time supple-
mented by other program duties. The
percentage of the chemist's time
dedicated to QC is dictated by the size
and complexity of the program.
This individual should have a mini-
mum of a bachelor's degree in chemis-
try, engineering, or mathematics with at
least 2 years of environmental or 1 year
of QC experience. This position is under
the general supervision of the labora-
tory director, but with access to the
program manager. Duties are to imple-
ment and monitor the routine applica-
tion of QC activities in the laboratory.
-------
Part l-Section 3.0
Jan. 1981
The individual can also participate in a
formal program to train new employees
and to update skills of older employees.
Before analysts are allowed to ana-
lyze samples, they should demonstrate
their proficiency. Each new analyst
should be instructed in instrument
operation, and should be allowed to
conduct an instrument performance.
study (Section 4.3.3). Analyst perform-
ance can be evaluated by control charts
of critical QC parameters (Sections 5.5
and 5.8); if these charts indicate a
problem, the analyst should be given
further training.
3.1.6 Data Management
People involved in data acquisition,
reduction, and reporting are the field
operator, the analyst, the data entry
staff, the laboratory director, the
program manager, and the designated
QA coordinator. Presented here is an
outline of primary duties assigned to
different members of the data manage-
ment team:
Field Operator - Preparation of field
data forms, and QC of data form
preparation.
Analyst - Reading or transcription of
strip charts, and entry of data to
computer.
Laboratory Supervisor - QC check of
strip chart reading, preparation of data
forms, QC of data form preparation, and
review of computer-generated QC
information or control chart preparation
and interpretation.
Data Processing Personnel - Input of
data from data sheets, verification of
input for keypunch errors, and update of
computer files.
QA Coordinator (Officer) - Review of
data, preparation of QA reports and
submittal of audit data, and recommen-
dations to project manager.
3.2 Designation of QA Re-
sponsibilities and Duties
3.2.1 Quality Assurance Coordina-
tor
The QA coordinator should monitor
performance as follows:
1. Review the monthly QC plots
generated for each analysis to
verify that QC data are acceptable
and to identify any consistent bias
trend.
2. Evaluate the Monthly Field Audit
Report prepared by the QC chemist
to assess the accuracy of field
analyses and to identify needs for
corrective actions.
3. Review all of QC information
presented with each set of analyti-
cal data to be reported — including
data from AC reports, ion summa-
tion values, and QC chemist's
independent internal audit report.
4. Evaluate laboratory and field
operations by conducting system
audits and reporting them to the
project manager.
5. Prepare quarterly reports to man-
agement that summarize QA activ-
ities and assess data quality in
terms of precision and accuracy
trends for both the field and the
laboratory operations.
3.2.2 The QC Chemist, the Analyst,
and the Data Clerk
Besides QA, routine QC activities are
carried out in laboratory operations by
the QC chemist, the analystandthedata
clerk.
The QC chemist introduces blind
audit samples as an independent check
on data quality, and issues a monthly
report updating control limits for all
parameters. All data given to the
laboratory director should have been
evaluated by the analyst's supervisor
and the QC chemist.
The analyst performs the analyses
and evaluates analytical performance in
real time, using readily available QC
information. As soon as possible after
analysis, the supervisor evaluates the
computer-generated QC information for
all QC parameters; reanalyzes the data
if necessary; and finally, when data are
"in control" (according to the supervisor)
releases the data for reporting to the
laboratory supervisor or director, who
submits them to the project manager.
The data clerk inputs analytical and
field data into the computer, checks and
corrects the data input, and generates
reports and graphs of QC information.
3.3 References
1. U.S. EPA Quality Assurance Man-
agement Staff, Guidelines'and
Specifications for Preparing QA
Program Plans, QAMS-004/80,
U.S. Environmental Protection
Agency, Washington, D.C., Sep-
tember 20, 1980.
2. Operations and Maintenance Man-
ual for Precipitation Measurement
Systems, U.S. Environmental Pro-
tection Agency, Research Triangle
Park, N.C., (in pres.s).
-------
Jan. 1981
Part l-Section 4.0
4.0 Documentation
4.1 Document Control
A system of document control should
be established for documentation of all
precipitation monitoring field and
laboratory operations — siting, samp-
ling, analysis, QA data handling, and
validation. The system should be
patterned after the indexing format of
the QA handbook (7) for updating
operational procedures and adding
results of special studies and other
related documents. The indexing format
at the top of each page should include:
Section number.
Revision number
Date of revision
Page number.
The elements of a precipitation mea-
surement project for document control
should include:
1. Field operations and maintenance
procedures,
2. Analysis procedures,
3. Auditing procedures,
4. Computational and data validation
irocedures,
duality assurance plan, and
_. Quality assurance manual.
4.2 Internal Documentation
A central file of all data, reports,
correspondence, and so forth should be
maintained by the project manager. In
addition, a data file should be kept by the
laboratory. Records in the files should
meet the following requirements:
1. Records should have identification
numbers and must be kept for at
least 3 years by the agency in an
orderly, accessible form; records
should include all raw data, calcu-
lations, QC data, and reports.
2. Data in laboratory records must
include the following information:
A. Sample identification number,
B. Sample type,
C. Date sample received in labora-
tory,
D. Collection data (time, date,
volume, and so forth, if labora-
tory responsibility).
E. Date of analysis,
F. Name of analyst,
G. Results of analysis (including
all raw data), and
H. Name of person receiving the
analytical data.
3. If applicable, the laboratory should
follow chain-of-custody proce-
dures from receipt of sample
through completion of analysis;
the following are guidelines:
A. Computer printouts or report
forms verified against labora-
tory records before data re-
lease,
B. Bound notebooks with num-
bered pages,
C. Sampling information records
(e.g., field data forms) with
dates, time, site location, sample
amount, and so forth.
D. An example of the data hand-
ling and reduction system ex-
amined by legal counsel to
determine soundness in possi-
ble litigation.
4.3 Reports
Each organization must periodically
assess its QA program (2). The QA
project plan provides the mechanism for
reporting to management on QA activi-
ties, on the performance of measure-
ment systems and on data quality. As a
minimum these reports should be
submitted semiannually and they should
include (3):
Periodic assessments of measurement
data accuracy, precision, and complete-
ness;
Results of performance audits;
Results of system audits; and
Significant QA problems with recom-
mended solutions.
The designated QA coordinator should
be responsible for the reports, and
should provide a separate QA section to
the final report to summarize data
quality information contained in the
periodic reports.
4.4 References
1. Quality Assurance Handbook for
Air Pollution Measurement Sys-
tems -Vol. l-Principles. U.S. En-
vironmental Protection Agency,
Research Triangle Park, N.C., EPA-
600/9-76-005 (March 1976).
2. U.S. EPA Quality Assurance Man-
agement Staff, Guidelines and
Specifications for Preparing Quality
Assurance Program Plans. QAMS-
004/80, U.S. Environmental Pro-
tection Agency, Washington, D.C.,
September 20, 1980.
3. U.S. EPA Quality Assurance Man-
agement Staff, Interim Guidelines
and Specifications for Preparing
Quality Assurance Project Plans.
QAMS-005/80, U.S. Environ-
mental Protection Agency, Wash-
ington, D.C., December 29, 1980.
-------
Jan. 1981
Part l-Section 5.0
5.0 Siting
Criteria for site selection and evalua-
tion are similar to those proposed by the
World Meteorological Organization (1).
In the design of a monitoring program,
the program objectives and network
station density must be considered.
Programs can be grouped as baseline
monitoring, long-term trend monitoring
and research. Stations may be charac-
terized as remote, rural, and urban;
however, they should yield representa-
tive results for the region although this
characterization is somewhat over-
simplified. Precipitation composition,
long-range transport of acid precipita-
tion and precipitation comtaminants,
and the impact of a source shown by its
distance or time scale are not well
known. Many large point sources are in
rural areas. Thus, choosing optimum
site locations to monitor either the
impact of a specific source or regional
effects not related to any source is
difficult. Instead of remote, regional,
urban, and so forth, a complete descrip-
tion of the site and its surroundings
'd be given so the user of the data
d decide what the station is measur-
ing or what it represents. However, to
aid the decision and for convenience,
site area categories have been used. In
all categories, the criteria are of prime
importance. If no site in an area meets
the criteria, a site in an alternate area
must be selected.
Network station density, which helps
define the spatial and temporal resolu-
tions of the data obtained, isdetermined
by the program objectives, the area
meteorology and topography, and the
budgetary constraints.
5.1 Network Design Con-
siderations
In designing a precipitation monitor-
ing network, stations are located either
singly or in combination, according to
the objectives of the program.
1. Measurement of baseline (remote
area precipitation),
2. Measurement of representative
regional (rural) precipitation, or
3. Measurement of urban area (local
impact pollutant emission sources)
on precipitation.
These three are generally differentiated
by concentration levels. The background
•emote station should show con-
mation primarily due to natural
processes; the regional station would be
affected primarily by long-range trans-
port; and the urban, site would show
high concentrations due to a polluted
local environment. To select station
locations, it is necessary to have
detailed information on the locations of
emission sources, the regional variabil-
ities of ambient pollutant concentra-
tions, the precipitation amounts, the
prevailing winds, and other meteorolog-
ical data. Thus, the design of a network
needs to address details such as the
number, location, and type of sampling
stations and equipment to be used. Due
to the variation in acid precipitation
caused by terrain, meteorological
conditions and -demographic features,
each network should be designed
individually after considerimg physical
evidence, economic factors, and pro-
gram objectives.
5.2 Site Selection Criteria
Selecting a precipitation monitoring
site is as important as selecting a
measurement technique to obtain
representative data. The variabilities
and the long-range transports of pol-
lutants make it difficult to determine
whether a site iscollectingprecipitation
data representative of an area. The
station should collect samples repre-
sentative of both the amount and the
composition of the precipitation in the
area.
The transport and diffusion of air
pollutants and their resultant concentra-
tions in precipitation are complicated by
topography, minor topographical fea-
tures may exert small effects, but major
features such as valleys or mountain
ranges can affect large areas. In
mountainous regions, precipitation
tends to be unevenly distributed due to
topographical lifting of clouds and
deflecting of airflows; such unevenness
should be recognized for site selection.
Near a large body of water, sea or land
breeze motions may change direction;
in the daytime, the winds may come
from the water, but at night they may
come from land. Because of these
changes, area sampling representation
would be affected. In such areas, more
than one station should be installed but
the number of sites should depend on
site availability and on funding level.
To optimize site locations for different
station categories, the following selec-
tion criteria are listed in order of
importance. In the baseline and regional
categories, the site must meet the first
three as well as all those in Section 5.3.
5.2.1 Baseline Scale
The station should be in a location
where the effects of human activities
are negligible. It is difficult to find
locations which meet all of the following
criteria (2,3,4) but each should meet the
following as closely as possible:
1. The station should be in an area
where no significant changes in
land-use practices within a 100 to
1000 km (depending on prevailing
wind direction) from the station are
anticipated during the study period.
2. The station should be far away
from major population centers,
major highways, industries, air
routes, and large natural sources
(e.g. geysers); it should be in
remote, uninhabited, or sparsely
inhabited areas. If an isolated
island is used, data corrections for
sea salt aerosols should be made.
3. The site should not have a history
of frequent local natural phenom-
ena such as forest fires, dust and
sand storms, or volcanic activities.
4. The site should have provisions for
setting up a meteorological and
aerometric monitoring station
(2,3,4,5).
5. The site should be readily accessi-
ble on a flat or gently sloping
terrain (less than 20°), and shel-
tered from strong winds.
5.2.2 Regional Scale
Site selection criteria for a regional
monitoring network are (1,6,7,8,9):
1. The general area should be free
from influences of large anthro-
pogenic sources such as cities or
towns; industrial, sewage or power
plants; refineries, commercial areas,
and airports; and large local natural
sources. Such pollutant sources
should be distant enough for
pollutants not to unduly effect the
precipitation chemistry. If, the site
location must be near a large
source (e.g., within 50 km), the
station should be in the prevailing
upwind direction from the source.
2. The immediate area of the site
should be separated from local
sources such as houses, active
farmlands or orchards, marshes
and swamps, landfills, roads; such
sources should be several kilo-
meters distant, and preferably
downwind.
3. If stations are near pollutant
sources, the site location should
-------
Part l-Section 5.0
Jan. 1981
avoid undue influence by a single
pollutant source.
4. If an area is characterized by a
common type of land use, the
collector can be near the common
pollution source.
5. The site should have provisions for
setting up meteorological and
aerometric monitoring equipment
(2,3,4,5).
6. The site should be readily accessi-
ble on a flat or gently sloping
terrain (less than 20°), and shel-
tered from strong winds.
The selected site should be evaluated
for representativeness and for local
contaminations by installing a temporary
grid of neighboring satellite samplers
around it; sampling procedures for this
temporary network should be compara-
ble to those in the original site; data
collection by the network is recom-
mended for a few months, not to exceed
1 year. The permanent site should be
selected after evaluating the site's
representativeness by analyzing data
from stations in the temporary network.
For evaluating the effects of long-
range transport on acid precipitation,
the average interstation distance can be
several to several hundred kilometers.
5.2.3 Urban Scale (or Local Scale)
To study urban areas or the effects of
point sources on precipitation, an array
of monitoring stations should be near
the source in the area of interest. The
stations must not be near other sources
or at sites that do not follow the criteria
(Section 5.3). The network station
density depends on the objective of the
monitoring program and on budgetary
considerations. For studying the local
effect due to a single point or area
source, the average interstation dist-
ance should be of the order of kilo-
meters. However, actual station density
and interstation distance should be
decided by the desired spatial resolu-
tions. The following criteria apply:
1. If an area is characterized by a type
of land use, the site can be near
such pollution sources.
2. The site should have provisions for
setting up meteorological and
aerometric monitoring equipment
(2,3,4,5).
3. The site should be readily accessi-
ble on a flat or gently sloping terrain
(less than approximately 20°), and
sheltered from the strong winds.
5.3 Sampler and Rain Gauge
Siting Criteria
Placements of precipitation samplers
and rain gauges should assure the
adequacy of the site to collect unbiased
samples. Samplers and rain gauges
should stand far enough from trees,
hills, overhead power lines and other
obstructions to minimize interference
with sampling. Natural and manmade
obstructions may cause turbulence
and/or contamination, and may cause
nonrepresentative samples; thus no
object (even if smaller than the collec-
tor) should be within a few meters of the
collector, and no object should cast a
rain shadow on the collector. An open,
flat, grassy area, surrounded by trees at
least 100 m away, but near no sources
(unless a local source is to be studied)
would be an ideal site.
Criteria for placement of collectors
(both samplers and rain gauges) are:
1. The horizontal distance between a
large obstruction and the collector
should be at least twice the
obstruction height, or the viewing
angle of an obstruction from the
collector should be less than 30°
above the horizon.
2. The horizontal distance between
an obstruction and a collector of
comparable height should be at
least one unit height of the taller of
the two.
3. The collector should be far from
mobile pollution sources; routine
air, ground, or water traffic should
not come within 100 m of the
collector site.
4. The distance between any over-
head wires and the site must be far
enough for the samples not be
affected; thus, criterion 1 (above)
should assure that there are no
overhead wires.
5. The collector should be at least
100m from open storage of agri-
cultural products, fuels, or other
foreign materials.
6. The ground surface around the
collector should not be loose soil; it
should have a grass cover or gravel
to minimize splash and airborne
ground surface particulates that
contaminate samples.
7. Wet/dry collectors should be
oriented parallel to the prevailing
wind direction with the wet bucket
upwind of the dry bucket (so that
the dry bucket does not obstruct
the wet bucket).
8. The rain gauge should be posi-
tioned parallel to the collector and
the prevailing wind direction. If the
gauge has an access door (to a
recorder, weighing or drive mech-
anism), the door should be kept
closed and the gauge should be
mounted with the door facing
away from the wind to prevent
precipitation entering the com-
partment.
The distance between certain ob-
struction (e.g., growing trees and newly
erected structures) and the colle'
should be checked periodically. Cri\
2 should minimize the sampling area
required when more than one collector
or when a sampler and rain gauge are
on the site, and should minimize the
effect of turbulence caused by obstruc-
tions.
5.4
tion
Site Evaluation Descrip-
5.4.1 Station Identification
All stations must be identified by
documentation of site characteristics to
facilitate evaluation of data generated at
that site and interpretation of the
monitoring data. Typically, the site.
identification record should contain the
following:
1. Data acquisition objective (base-
line, trend, or research monitoring).
2. Station location (address, map
coordinates, elevation, etc.).
3. Scale of representativeness (re-
mote, regional, or area type; i.e.,
industrial, agricultural, forest,
urban, etc.).
4. Instrumentation checklist (manu-
facturer, model number, measure-
ment technique, etc.).
5. Important pollutant sources (p'
and area sources; their emis
concentrations, proximities, poi-
lutants, etc.).
6. Topography description (trees,
hills, valleys, bodies of water; type,
size, proximity, orientation, etc.);
photographs of the monitoring site
covering a 360° view from the
precipitation collector are recom-
mended).
7. Site diagram properly scaled (equip-
ment configuration, vegetation,
manmade structures, access road,
electrical powerlines, etc.).
The site identification should be docu-
mented by filling out a Site Description
Report (Section 5.5).
5.4.2 Station Classification and Eval-
uation
Each station should be initially
classified by using the information in
the Site Description Report. Section 5.5
gives a modified version of one devel-
oped by NADP(8). The five site classes
defined depend on the sophistication of
instrumentation in the station as well as
on the satisfaction of siting criteria.
Class I
1. Station satisfies all siting criteria
(Section 5.3),
2. On-site instrumentation indue1
automatic precipitation collectOi,
a recording rain gauge,
pH and conductivity meters,
-------
Jan. 1981
Part l-Section 5.0
meteorological sensors (wind-
speed and direction), and aero-
metric analyzers (SOz and NO/
NO,).
Class II
1. Station satisfies all siting criteria
(Section 5.3).
2. On-site instrumentation includes
automatic precipitation collector, a
recording rain gauge, pH and con-
ductivity meters, meteorological
sensors (wind-speed and direction),
and aerometricanalyzers (SOa and
NO/NO,).
Class II
1. Station satisfies all siting criteria
(Section 5.3).
2. On-site instrumentation includes
automatic precipitation collector,
a recording rain gauge, and pH
and conductivity meters.
Class III
1. Station satisfies all siting criteria
(Section 5.3).
2. On-site instrumentation includes
automatic precipitation collector,
nonrecording rain gauge, and pH
and conductivity meters.
Class IV
1. Station does not satisfy all siting
criteria (Section 5.3).
On-site instrumentation identical
to Class I stations.
Class V
1. Station does not satisfy all siting
criteria (Section 5.3).
2. On-site instrumentation identical
to Class II stations.
5.5 Site Description Report
All monitoring stations should be
properly identified and classified using
the report forms on the next pages.
After initial classification, an on-site
visit should be made by the QA coordi-
nator to evaluate and certify each
monitoring station as soon as possible
after the start of operation to assure the
quality of monitoring data. Classes IV
and V must be only temporary, since
they are in noncompliance with siting
criteria. The project manager (or a
designee) must be sure that siting
deficiencies are corrected within a
reasonable time. Reasonableness de-
pends on the deficiencies found; most
should be corrected within 30 days but
for deficiencies requiring longer time, a
schedule must be established for
compliance attainment. When correc-
tions are made, documentation should
be provided to the QA coordinator and
program manager and the station
classification should be changed by
them.
All sites should be evaluated again
a year to establish that they
itain compliance. All aerometric
and meteorological instrumentation
should conform to standard ambient
monitoring guidelines (4,10).
-------
Part l-Section 5.0
Jan. 1981
A. Data Acquisition Objective (Description)
B. Site Category
1. Station Identification
4. Latitude"
5. Longitude*
2. County .
3. State
6. Elevation
7. Station environment: Remote.
Suburban
Industrial
Rural.
(m)
Urban .
Commercial.
8. Name of official
9. Mailing address -
position .
(number and street)
10. Phone ( )
(city)
(state)
(2ip)
C. Instrumentation
1. Precipitation Co/lector Type: Automatic
Non-automatic .
Manufacturer
Model
Diameter (I.D.) of
Sample Bucket _
(cm) Serial No.
2. Raingauge:
3. NO, Monitor:
4. S02 Monitor:
5. Other Aerometric:
1. Recording
2. Type: Weighing
3. Manufacturer _
4. Funnel Size:
Tipping Bucket
Non-recording
Other .
Model
(cm) Serial No.
1. Recording
2. Type: Chemiluminescent
3. Manufacturer
-Non-recording
-Other
1. Recording
2. Type: Fluorescent
3. Manufacturer
Model/Ser. No.
Non-recording
Other
/. Sensor:
Non-recording
2. Type
Model/Ser. No.
Recording
3. Manufacturer
Serial No.
Model
6. Other Meteorological Instrument:
1. Wind Speed Sensor:
2. Wind Direction Sensor:
3. Temperature Sensor:
4. Solar Radiation Sensor:
Type
Manufacturer
Type
Serial No.
Model.
Model.
Manufacturer
Serial No.
Type
Model.
Manufacturer
Serial No.
Type
Serial No. /Model
7. pH Meter:
8. Conductivity Meter:
1. Type
2. Manufacturer
Manufacturer
Temp.
Compensated:
Model/Ser. No.
1. Type
2. Manufacturer
Temp.
Compensated:
Model/Ser. No.
*To be reported in xx.yy.zz format corresponding to deg., min., sec.
-------
Jan. 1981 6 Part l-Section 5.0
D. Site Documentation:
V Local topographic map (Scale
1:250.000)
Identify the site location and
major sources on the map.
-------
Jan. 1981
Part l-Section 5.0
2. Sketch a map to document the
environment within a Vi mile
radius of the site. Include the
following information on the draw-
ing where applicable.
Site Diagram and Equipment Con-
figuration at Center of Drawing
Roadways with Names (paved and
unpaved)
Parking Areas (paved and unpaved)
Stationary Sources (NEDS#)
Buildings (number of stories)
Undeveloped Land (ground cover)
Tree Lines or Clusters
High Power Lines
Topographical Features
(Valleys, Hills etc.)
Bodies of Water
North Direction
-------
Part l-Section 6.0
Jan. 1981
3. Site photographs, labelled to in-
dicate the four compass directions.
-------
Jan. 1981
Part l-Section 6.0
5.6 References
1. WMO Operational Manual for
Sampling and Analysis Tech-
niques for Chemical Constituents
in Air and Precipitation, World
Meteorological Organization Pub.
No. 299(1974).
2. Quality Assurance Handbook for
Air Pollution Measurement Sys-
tems. Vol. II - Ambient Air Specific
Methods. EPA-600/4-77-027a.
Research Triangle Park, NC (1977).
3. Guide to Meteorological Instru-
ment and Observing Practices.
World Meteorological Organiza-
tion Pub. No. 8. TP8J1971).
4. Ambient Monitoring Guidelines
for Prevention of Significant Deteri-
oration (PSD). EPA-450/1 -78-019
(1978).
5. Quality Assurance Handbook for
Air Pollution Measurement Sys-
tems. Vol. I - Principles, EPA-
600/9-76-005, Research Triangle
Park, NC (1976).
6. Galloway, J.N., C. Hall, and G.E.
Likens, The Collection of Precipi-
tation for Chemical Analysis,
Tellus30, 71 (1978).
7. Granat, L, Principles in Network
Design for Precipitation Chemistry
Measurements, Proc. Symp. on
Atmospheric Contribution to the
Chemistry of Lake Waters, J. Great
Lakes Res. 2(1), 42(1976).
8. Site Selection and Certification.
National Atmospheric Deposition
Program (1970).
9. Bowersox, V.C., Acid Precipitation
at a Rural Central Pennsylvania
Site, MS Thesis, Dept. of Mete-
orology, Pennsylvania State Uni-
versity (1980).
10. EPA Ambient Air Quality Surveil-
lance Regulations, Code of Federal
Regulations, Title 40 Part 58 (May
10, 1979).
-------
Jan. 1981
Part I-Sect ion 6.0
6.0 Field Operations
The primary aims of quality assurance
requirements for field operations are to
maximize sample collection and to
collect and handle samples in a way that
preserves their integrity and identity.
Precipitation samples are very dilute,
and thus very susceptible to contami-
nation. The goal of this section is to
present requisites for obtaining high
quality data; discussed are facilities,
method selection, acceptance testing,
equipment installation and operation,
sampling methodology, field measure-
ments, network evaluation, and docu-
mentation.
6.1 Facilities
All field facilities should be accessible
and should have a clean shelter, a level
table, deionized or distilled water
(conductivity less than 1.5 micromho/
cm), and a sink or drain. If deionized or
distilled water cannot be produced at
the site, it can be purchased at local
supermarkets or drug stores. A refriger-
ator and a 110v AC outlet are highly
•able; the former is necessary for
( or sequential sampling to pre-
serve the samples until they are
shipped. For weekly sampling, the
samples should be shipped within 24 h
of collection so refrigeration is not
important. Since samples can be at
ambient temperature in the collector for
as long as 1 week, another few hours
after collection should not matter.
The precipitation collector and the
recording rain gauge can be run on
either 12v DC by storage batteries or
110v AC. Both means have advantages
and disadvantages. With batteries there
can be no loss of operation due to a
power failure, and sites can be located
in areas with no power. However,
batteries must be periodically re-
charged, and power output drops in cold
weather can result in collector down-
time. Furthermore, if large current
usage is required (e.g., for heating),
batteries are not recommended.
6.2 Method Selection
Similar equipment and supplies are
required at a station for event, daily, or
.weekly sampling; weekly sampling is
most common at present. For sequential
sampling, a different type of collector is
npeded. For event, daily, or sequential
iling, a refrigerator, polyethylene
,es with caps, polyfoam insulated
shipping containers, and freeze-gel cold
packs are recommended for sample
storage and shipment. A list of equip-
ment for typical weekly sampling is in
Table 6-1; for other types of sampling,
the list should be changed to include a
different type of collector, a more
sensitive balance (2.6 kg) capacity,
polyethylene bottles and caps, and
means for shipping the sample in a cold
state.
6.2.1 Precipitation Collectors and
Rain Gauges
The rain gauge and the precipitation
collector serve different functions. The
rain gauge measures the amount of
precipitation. The precipitation collector
collects a sample for chemical analysis.
The two devices are not interchange-
able.
6.2.1.1 Precipitation Collectors—The
precipitation collectors must meet the
following criteria:
1. Reliable automatic operation -
collector container opens at start
of precipitation and closes after
event ends.
2. Prevention of contamination of
wet sample by dry deposition.
3. Minimization of evaporation.
4. Container inertness to sample
constituents of interest.
Collectors, which adequately meet
these criteria, are available. For event,
daily, or longer-time rain sampling, the
most satisfactory collector is one based
on the design of the Department of
Energy's Health and Safety Laboratory
(HASL) (1); however, the sampler is less
efficient for snow collection.
The first three criteria are met by
means of a precipitation sensor and a
motor-driven tightfitting lid for the
collector container. When the grid and
plate of the sensor are shorted by a drop
of water, the motor is activated, lifting
the lid from the container. The sensor
has two heating circuits: one goes on to
melt snow or ice (on the sensor) when
the temperature is below 2°C, and the
other is activated when the lid lifts off
the sample bucket to heat the sensor to
about 55°C and to increase the rate of
evaporation of water from the sensor.
Heat hastens the sealing of the sample
by the lid after precipitation ceases, thus
minimizing the exposure times to dry
fallout and to snow blowout from the
collector. A seal between the container
and the lid is achieved by a plastic foam
gasket under the lid and by a spring load;
however in strong winds, the lid will
wobble and allow some contamination
to enter the sample bucket.
To ensure inertness to major con-
stituents in acid precipitation, use
polyethylene sample buckets (1,2). Use
only plastic (or Teflon) containers for
inorganic constituents. For cost, dura-
bility, and availability, high-density
linear polyethylene containers are
generally employed. Glass or metal can
affect inorganic sample integrity, but
either should be used if organics are
studied.
For subevent or sequential samplers,
the same requirements as above hold,
but mixing between sequential samples
should be minimized because these
samplers separate samples on either a
volume or time-of-collection basis. In
volume-based sampling, precautions
must be taken to minimize mixing or
carry over of samples. For time-based
sampling, there should be provision for
overflow during heavy rainfalls. For
sequential sampling, the time cor-
responding to each subevent specimen
must be known for correlations with
other data. Sequential samplers vary
greatly in sophistication from a series of
connected bottles to completely auto-
mated and electrically operated designs
(3-8). One of the latter is available
commercially (8).
Table 6-1. Field Equipment List for Weekly Sampling
Automatic precipitation sampler (Aerochem Metrics #301)
Collection containers 13.5 gal) for collector
Recording rain gauge with event marker
Rain gauge mount
pH meter, electrode
Buffers. pH 3.0. 4.0, 6.0. 7.0 and 8.0 (1 liter each)
Conductivity meter and cell
Standard KCI solution, 74 micromhos/cm {500 mil
Temperature probe
Balance (20 kg capacity) or graduated cylinder (2 liter)
Set of attachment weights for balance (1,2.2.5.10 kg)
No. /site
1
3
1
1
1
1
1
1
1
1
1
-------
Part l-Section 6.0
Jan. 1981
6.2.1.2 Rain Gauges—In case of
collector malfunction and to reference
all the precipitation amounts against a
standard, a rain gauge is used to record
the quantity of precipitation. Recording
rain gauges are of two basic operational
designs—the weighing type and the
tipping bucket type. Both types of
gauges should be capable of measuring
precipitation to approximately 0.25 mm
(0.01 in.) and they should be accurate to
a few percent. Weighing gauges mea-
sure within about ±0.76 mm (0.03 in.),
but their accuracy of about 1% of full
scale is independent of precipitation
rate. The generally accepted accuracy
for tipping bucket gauges is 1% for
precipitation rates of 25 mm/h (1 in/h)
or less, 4% for 75 mm/h (3 in/h), and
6% up to 150 mm/h (6 in/h); rates are
either measured directly or derived from
the cumulative data. Data from the
tipping bucket are amenable to tele-
metry.
The recording rain gauge should have
an event marker pen to indicate when
the collector is open and closed. Such a
pen typically .rises from its baseline
when the collector lid opens, and
remains actuated until the lid closes and
the pen falls to its baseline position. To
prevent an event marker pen from
interfering with a sample trace pen, the
two must be offset on the time axis; thus
only one pen can be set at the correct
time, and care must.be taken not to use
the incorrect beginning or ending times.
Since the operator is seldom present to
observe collector behavior during an
event, the event pen marker is invalua-
ble in indicating malfunction.
6.2.2 pH and Conductivity Apparatus.
Temperature Probe
The pH and conductivity of a 20 ml
aliquot of the sample should be mea-
sured at the station as soon as possible
after the sampler is collected and after it
has reached the same temperature as
the calibration solutions. The sample
and the aliquot should be protected from
contamination during this time.
6.2.2.1 pH Apparatus—The pH meter
and electrode must be capable of
measuring within ±0.03 unit precision
with an accuracy of 0.05 unit. Meters
should have an impedance of at least
10" ohms. A combination glass and
reference electrode of the nongel type
with an unprotected membrane bulb is
preferred. The combination electrode
requires less sample and fewer wash-
ings than two separate electrodes;
electrodes with shielded bulbs are more
difficult to clean, and are thus more
prone to yield errors. When a new
electrode is obtained, it should be
equilibrated overnight in a pH 4 or 7
buffer solution, as recommended by the
manufacturer, before it is used. The
electrode should be stored in pure
water, in a 74 micromho/cm KCI
conductivity standard, or in a 0.0001 N
acid solution; these storage media may
cause the electrode to require longer
times for stabilization than concen-
trated buffer solutions do, but they
increase the life of the electrode.
6.2.2.2 Conductivity Apparatus—The
conductivity meter and cell must have a
measurement range of 0 to 1000
micromho/cm, a precision of ±0.5% of
range, and an accuracy of ±1.0% of
range. The range most frequently used
is 10-100 micromho/cm. A tempera-
ture-compensated cell with a cell
constant of 1.0 is preferred.
6.2.2.3 Temperature—a thermistor,
thermocouple, or thermometer can be
used to measure temperature. The
temperature probe must have an
accuracy of at least 1 °C and a precision
of ±0.5°C.
6.2.3 Balance or Graduated Cylinders
The amount of precipitation sample
collected can be measured with a
balance or with graduates. Since the
density of rain samples is approximately
1.0 g/ml at 20°C, the weight of the
sample can betaken toequal its volume.
The measurement of sample volume by
graduates increases the chance of
contamination, so use of a balance is
recommended.
The precipitation volume can be
compared to that recorded by the rain
gauge (e.g., with the Aerochem Metrics
collector, 16 g of sample = 0.013 in. =
0.25 mm) to calculate the collection
efficiency of the sampler (Sect ion 8.5.4).
Differences between the field and the
laboratory weight values (container plus
sample shipped to the laboratory)
indicate either loss of sample during
shipment or a weighing error.
For weekly sampling, the balance
must have a capacity of 20 kg and a
precision of at least ±10 g. Triple beam
balances meeting these requirements
are readily available. The balance
should be kept on a sturdy level table,
and it should be zeroed before weighing.
The balance should be checked initially
in the laboratory with 1.0 and 5.0 kg test
weights. For event and sequential
sampling, a 2.6 kg capacity triple beam
balance with a sensitivity of at least ±1
g should be used.
Graduates are not recommended, but
if they are used, they should be plastic,
and they should be cleaned and dried
before use..To measure within ±10 ml,
the graduate should not be more than
1000 ml in capacity. For large samples,
several graduates or one graduate plus
a clean auxiliary container (in which to
empty the graduate) will be needed. The
graduate should be checked for accur-
acy in the laboratory by weig*
various volumes of water from
vessel.
6.3 Acceptance Testing
All precipitation collectors, rain
gauges, pH and conductivity meters,
and electrodes should be tested for
acceptance before they are used in the
field. Acceptance tests should cover the
functions of the instruments. Collectors
and rain gauges should be tested on
site. However, it is more convenient to
test meters and electrodes in a central
support laboratory, where common
standards and procedures would be
used for all instrumentation in the
monitoring network. General pro-
cedures for acceptance testing are
detailed below; procedures for carrying
out these tests are in 0 & M manual (9).
6.3.1 Precipitation Collectors and
Rain Gauges
Collector tests should include: (a)
sensor heating and actuating the lid
when the sensor is shorted with water
drops, (b) sensor cooling and return of
the lid after removal of the shorting
material (water wiped dry), (c) sen<^r
temperature attainment when lid ;
of the wet bucket (50° to 60°C),
sensor temperature when ambient
temperature is below freezing and (e) •
observation of lid cycling and sealing. If
any of these tests indicate a malfunc-
tion, contact the manufacturer. For test
c above, the temperature can be varied
by a potentiometer identified by TH in
the sensor circuit. In general, the
malfunctions can 'be rectified by re-
placing the sensor or the motor box.
Rain gauge tests should include a
check of: (a) sensitivity and accuracy, (b)
clock function, and (c) pen and recorder
functions. For test a, add an amount of
water equivalent to 0.02 in. (0.51 mm),
and measure the response. For the
Belfort recording rain gauge (5-780
series), 0.01 in. = 8.2 gm. The gauge can
be calibrated at each inch level with a
set of weights from the manufacturer or
with known weights of water.
6.3.2. pH and Conductivity Meters
All pH and conductivity meters,
electrodes, and cells should be tested in
a laboratory for acceptance before being
shipped to the field stations. The meters
usually have a serial number affixed,
but the electrodes do not, so an
identification number should.be affi
to each electrode. For each of the tv
indicated below, a total of 10 solutions
is measured, and an average value and
a standard deviation are calculated.
-------
Jan. 1981
Part l-Section 6.0
6.3.2.1 Evaluation of Conductivity
' and Cell—The conductivity meter
all are acceptable if the average
„• for a 0.0003M KCI solution is
within 2% of the accepted value, 44.6
micromho/cm at 25°C, and if the
standard deviation is less than 2%. The
results should be recorded on the
Conductivity Meter/Cell Acceptance
Test form and on the Conductivity Test
Summary form.
6.3.2.2 Evaluation of pH Meter—For
the tests presented here, a pH electrode
reference solution (Section 7.0} should
be used. The field pH meters should be
tested using a laboratory pH electrode
whose performance has been docu-
mented. The pH meter is acceptable if
the average pH and the standard
deviation are within 0.03 unit of the
documented values. The pH Meter/
Electrode Acceptance Test form, and
the pH Acceptance Test Summary form
(Section 6.8.5) should be completed.
6.3.2.3 Evaluation of pH Electrodes-
Each new electrode should be equili-
brated overnight in the storage solution
recommended by the manufacturer.
Before testing, the electrode should be
carefully rinsed with deionized water,
'hen successively placed in deion-
vater in a series of test tubes until a
c^.,slant pH reading is achieved. The
electrode is acceptable if the average is
within 0.1 pH unit of the average
historical value and if the standard
deviation is less than 0.03 pH unit.
Results should be recorded on the pH
Meter/Electrode Acceptance Test form
and on the pH Acceptance Test Sum-
mary Form (Section 6.8.5).
6.4 Sampler and Rain Gauge
Installation and Operation
For placement of neighboring col-
lectors and rain gauges of equal or
smaller height, the distance between
rain gauge and collector or between
collector and collector should at least
equal the height of the taller object to
minimize interference as well as splash
effects. To ensure that the collector dry
bucket is not an obstruction to the wet
(or precipitation sample) bucket, the
collector should be aligned either
perpendicular to the prevailing winds or
with the dry bucket downwind of the
wet bucket. The ground surface around
the collector and rain gauge should be
natural vegetation or gravel; being
paved may cause splashing into the
''ector or gauge.
e precipitation collector should be
jnted on the ground so that the rim
of the mouth or opening is level and at
least 1 m above the ground, and it
should be properly anchored against
strong winds. The collector may be
shielded from the wind, but it should not
be put in an area where there will be
excessive turbulence caused by the
shield or where there are obstructive
objects such as trees and buildings. In
the winter, loss of snow from the
collector can occur due to blow out.
The sampler installation and opera-
tion are described in the manufacturer's
instructions and in the 0 & M manual
(9). The precipitation collector requires
no calibration, but proper functioning
should be checked frequently.
The rain gauge operation is discussed
in the manufacturer's instructions. The
rain gauge should be calibrated ac-
cording to the manufacturer's instruc-
tions after installation and at least
annually (Section 6.6). A calibration
check at two points, approximately half
and full scale, should be made monthly.
The rain gauge should be mounted on
a firmly anchored support or base so
that the funnel rim is level and at about
the same height as the collector rim to
enable comparisons of collection a-
mounts between the two. The gauge
level can be checked with a carpenter's
level placed at two intersecting posi-
tions. The gauge mouth should be high
enough not to be covered by snow. In
open areas, a wind shield (e.g., swing-
leaf like the one used by the U.S.
Weather Service) should be used. For
rain gauges which contain a recorder,
the access door to the chart drive should
be on the leeward side of the prevailing
winds, and should be kept closed to
minimize dirt and moisture affecting the
chart and the mechanism.
6.4.1 Routine Checks on Collector,
Rain Gauge, and Site
The following tests should be carried
out routinely on the precipitation
collector and the rain gauge. The
detailed procedures for these tasks and
a checklist are in the O & M manual (9).
1. Collector Sensor Test - At intervals
which coincide with the sampling
schedule, the sensor should be
shorted with a piece of metal or
some water to check the lid
opening and sensor heating func-
tions. When the sensor short is
removed, the lid should close
immediately, and the sensor should
cool. The sensor should be cleaned
at least monthly, or as needed.
2. Inspection of Dry Collector Bucket -
If the collector has a dry bucket, it
should be checked after an event
or a time period which has de-
posited more than 0.25 mm (0.01
in.) of precipitation to ascertain if
the bucket contains or has con-
tained any precipitation. Precipita-
tion in the dry bucket may be
evidence of a collector malfunc-
tion. Possible causes are (a) a dirty
or faulty sensor, (b) a high sensor
heating temperature and/or a low
precipitation rate, (c) a defective
magnetic mercury switch in the
motor box, (d) a lid arm is loose
and/or has moved too far out from
the magnetic switch to actuate it.
The above causes, except for the
dirty or faulty sensor, can result in
the lid cycling. This could cause
contamination and some loss of
sample.
3. Test of Dry Sample Bucket -
Weekly, if no event has occurred,
the sample bucket should be
returned to and/or tested in the
laboratory for contamination due
to poor initial cleaning, dry de-
position, and/or handling. If con-
tamination is frequent at a site,
poor collector sealing and/or an
operator handling problem are
likely occurrences.
4. Examination of the Event Pen
Marker Trace - Weekly, the event
trace should be inspected to see if
the lid cycled. The event trace
openings and closings should
correspond to the beginning and
ending of the event, as indicated by
the slopes of the sample weight
trace. Many upanddown markings
in a short time may indicate lid
cycling. Some cycling traces may
occur during light rain events or
heavy dew when no event is
apparent. No lid movement traces
when the sample weight trace
shows an event occurred indicate
a collector malfunction.
5. Cleaning Techniques and Schedule
The collector sensor should be
washed monthly with deionized
water to remove dirt, salt, and film
buildup. If a film persists, clean the
sensor grid and plate with deter-
gent and a toothbrush. The rim of
the dry bucket should be wiped
with clean tissues (e.g., Kimwipes)
to prevent carry over of dustfall to
the sealing gasket and to the wet
bucket.
6. Adjusting the Zero Setting of the
Rain Gauge • Daily or weekly as
needed, with no precipitation in
the rain gauge, adjust the zero
setting. The zero setting fluctuates
with temperature, but generally
not more than 0.25 mm (0.01 in.).
7. Checking the Rain Gauge Pail
Level - Whenever the rain gauge
pail is removed, be sure it is
replaced correctly so that it is level.
-------
Part l-Section 6.0
Jan. 1981
8. Adjusting and Winding the Rain
Gauge Clock - Weekly, wind the
clock (or chart drum) on the
weighing gauge, and correct the
time setting if necessary. Be sure
to correct for backlash and to set
the time correctly with respect to
a.m. and p.m.
9. Rain Gauge Check - Monthly, add
several known weights or water to
the rain gauge to be sure it is
measuring correctly at about the
75mm (3 in.) and 125mm (5 in.)
levels. For the Belfort weighing
gauge, 25 mm = 1 in. = 820 gm. A
complete calibration at each inch
level should be made at least
annually.
10. Inspection of Rain Gauge Pens and
Ink - Weekly, the pens should be
inspected to see if they have ink
and are writing. If they are not
writing, clean the pens, refill them,
and be sure they are working.
11. Rain Gauge Chart Replacement -
At the appropriate interval, de-
pending on the chart range, general-
ly weekly, remove the old chart and
replace it with a new one.
12. Rain Gauge Level Check - At 6-
month intervals, measure the
gauge level to be sure it is still
horizontal.
13. Winterizing - For the precipitation
collector, encase the lid arms in
plastic and tape one end of the boot
to the table. If necessary, to
prevent the lid freezing to the
bucket, attach a heater (e.g., a 40W
light bulb) to the top of the lid.
Check the sensor when the am-
bient temperature is below freezing
to be sure the heater is working.
Allow collected samples to warm
up to room temperature before
measurement. For the weighing
rain gauge, remove the funnel
(generally in the inlet), and add
ethylene glycol antifreeze to the
pail. For a tipping bucket gauge,
turn on the heater.
14. Site Maintenance and Inspection
for Obstacles - Periodically, mow
the grass and inspect the site area
for new obstacles (e.g., a growing
tree).
6.4.2 Corrective Action
Any indication of a malfunction
should be recorded in the logbook, and
the field manager should be notified. An
attempt to diagnose and correct the
problem should be made with the aid of
the 0 & M manual (9) as soon as
possible. If the problem cannot be
corrected, the field manager or equip-
ment manufacturer should be asked for
advice and direction. The field manager
or central laboratory should maintain a
supply of spare parts. The diagnosis and
action taken should also be recorded in
the logbook.
6.5 Sampling Methodology
Since precipitation samples generally
contain microconcentrations of pol-
lutants, extreme care must be taken to
avoid contaminating the sample and to
preserve its integrity. This section gives
the methodology for sample collection,
handling, measurement and for pre-
serving sample integrity. The pro-
cedures, used to accomplish each of the
above-mentioned tasks, are in the 0 &
M manual (9).
6.5.1 Sample Collection and Schedule
Sampling schedules commonly in-
clude weekly, daily, event and subevent,
or sequential. An event can be defined
as a storm separated from a second
storm by a dry interval of at least 6 h in
the winter or at least 3 h in the summer.
The choice of sampling schedule de-
pends on the program objective and the
available funds. To correlate rain data
with aerometric and/or meteorological
data, a subevent, event, or at most a
daily schedule must be used. To mea-
sure the amount of deposition and/or
its effects, a weekly sample may be
sufficient. Sampling periods longer than
1 wk are not advised because important
changes in the sample can occur on
standing in the collector.
6.5.2 Hand/ing of Plastic Containers
Treatment of plastic containers de-
pends on the species to be measured,
the container's previous use, and its
cleanliness. In most cases, the cleaning
should be done in the laboratory; only
routine bucket rinses (when the buckets
need no.t be returned to the laboratory)
should be done in the field. The final
rinse water conductivity should be
measured, and should be below 2
micromho/cm. Procedures for cleaning
plasticware are in the 0 & M manual (9).
The container should be capped and
kept in a plastic bag until immediately
before use, and it must be resealed
immediately after use. When a bucket is
to be returned to the laboratory with or
without a sample, it should not be
sealed with its original lid; instead, the
sample should be covered with the lid
from a new bucket which is replacing
the old bucket in the collector. Thus the
chance of contamination from the lid is
minimized.
6.5.3 Sample Hand/ing
The sample container must be checked
for precipitation at the time and fre-
quency set by the schedule. If a sample
is present, the container should be
removed and weighed; the bucket
swirled to help ensure a homogeneous
sample; and a sample aliquot should be
measured after assuring that ***>
containers are correctly labele
pencil or a ball point pen should be L.. -.
to inscribe the label. If the sample is
frozen, it must be allowed to warm up to
room temperature before measure-
ment. Avoid breathing onto a sample to
prevent ammonia contamination.
6.5.3.1 Weekly Sampling—After a
week if no sample is present, the empty
bucket can be sealed and returned to the
laboratory or rinsed in the field and
reused. In either case, the bucket should
be rinsed with distilled/deionized
water, and the conductivity of the rinse
water measured to provide a blank
which reflects the previous cleaning,
operator handling, collector sealing,
and so forth. The deionized/distilled
water conductance should be checked
before it is used. After cleaning, the
container should be shaken dry, and
reused.
6.5.3.2 Event Sampling—To mini-
mize the number of buckets required as
well as storage and shipment space, the
sample should be weighed and trans-
fered from the bucket to a labeled 500
ml wide mouth polyethylene bottle.
Some sample is taken for measurement
and the bottle sealed. If suffic
sample (e.g., more than 300 ml) is ,
sent, use about 50 ml to rinse out the
shipping bottle. One 500 ml bottle per
event is sufficient for all measurements;
the rest of the sample may be discarded.
The sample transfer can be made direct-
ly from the bucket to the bottle. The
bucket should be rinsed with deionized/
distilled water before reuse.
6.5.3.3 Sequential Samples—When
samples are collected through a funnel
directly into prenumbered polyethylene
bottles, the bottles should be labeled
and sealed immediately after the
samples are measured.
6.5.4 Sample Preservation. Storage.
and Shipment
6.5.4.1 Sample Preservation—Sam-
ple degradation can occur due to
chemical interactions (e.g., with parti-
culates or gases) or to biochemical
reactions. Preservation of sample
integrity can be maximized by filtration,
sealing, and storage in the dark at about
4°C. After pH and conductivity mea-
surements, filtration should be done
with a 0.45 um organic membrane filter
(10), if inorganic species are to be
analyzed. Although biocides (e.g.,
toluene or chloroform) are effec
they can interfere in the var.
measurements or analyses (2) and thus
must not be added to the sample. If
certain species must be preserved, an
-------
Jan. 1981
Part I-Section 6,0
aliquot of the sample can be mixed with
•^servative in a separate container.
dures initiated in the field should
jntinued in the laboratory; however,
to minimize contamination, filtrations
should be done in the laboratory as soon
as possible after sample arrival.
6.5.4.2 Storage—Samples should be
stored in a refrigerator after collection.
After the laboratory analyses have been
completed and the results checked, the
sample should be transferred to a 125
ml polyethylene bottle to save space and
stored in a refrigerator or freezer for a
time period of at least six months. Thus
the samples are available for further
tests or analyses if desired. Availability
and location of storage should be
recorded in the chain of custody
documentation. It is not known whether
the best method for longterm storage is
to freeze the samples or to keep them at
4°C. Stability tests for several months
indicate that both the 4°C (10,11) and
the freezing (11) are satisfactory. Until
more evidence is available, 4°C storage
is recommended.
6.5.4.3 Shipment—If sampling week-
ly, ship samples to the laboratory
kly, on a scheduled day and by the
iod specified in the 0 & M manual
(^,. If event or subevent sampling is
being performed, ship the sample as
soon as possible after collection and
measurement; for economic reasons,
the samples can be logged in, stored in a
refrigerator, and shipped weekly. Gener-
ally, the shipment should be made early
in the week by airfUPS Blueorairparcel
post) so that samples are received in the
laboratory before the weekend. Upon
arrival, the samples should be logged in
and stored in a refrigerator to minimize
degradation. Cumulative weekly samples
should be shipped with their collecting
buckets packed into cardboard cartons
or other protective boxes.
Event and sequential samples col-
lected for special studies should be kept
cold during shipment, and should be
shipped in cardboard-enclosed Styro-
foam boxes (Polyfoam Packers Corp.,
Chicago, III.) with freeze-gel packs
which have been kept in the freezer
compartment of a refrigerator for about
24 h before shipping to ensure that they
are completely frozen. Gel-type packs
are preferred because they are less
likely to leak when unfrozen. Generally,
four packs per box are sufficient to keep
the samples cold for 4 to 5 days. The
lerature of the box interior should
measured on arrival at the labora-
tory. The central laboratory should
replace all sample containers and
shipping materials weekly.
6.6 Field Measurements
Field measurement procedures for
pH, specific conductance, and tempera-
ture are discussed in this section.
Procedures described in the 0 & M
manual (9) should be used in the central
laboratory as well as in the field. Sample
pH and conductivity are measured in
both the field and the laboratory to
detect sample changes. The field values
representing the fresh sample are the
true values if no measurement error has
occurred. A warmup of the meters may
be needed before they are used (see
manufacturer's instructions). Results of
field measurements should be recorded
on the Field Data Form (Section 6.8.5).
6.6.1 pH Determination Method
6.6.1.1 General Description—The pH
of a solution is related to free-acid
activity:
pH =-log H* 6-1
where H* is the H* activity or free H*
concentration.
Thus pH does not measure the total acid
concentration. The pH meter measures
the electrical potential difference
between a reference electrode and an
l-Tglass electrode; the glass electrode
potential varies with the activity (or
effective concentration) of hydrogen ion
(H*) in solution. Although the meter
measures electrical potential (volts), it is
calibrated to give data as pH.
6.6.1.2 Calibration—Since acid rain
samples generally have pH's between
3.0 and 6.0, the pH meter should be
calibrated with pH 3.0 and 6.0 standard
buffers. For other less acid samples, pH
4.0 and 7.0 buffers should be used. For
basic samples, pH 5.0 and 8.0 buffers
should be used. Each station should
receive the needed calibration buffer
solutions from the central laboratory.
The stations should notify the laboratory
•when the buffer supply is nearly
exhausted.
The pH 6.0 (or 7.0) buffer is used to
adjust the calibration or standardization
setting; the slope setting is adjusted
against the pH 3.0 (or 4.0) buffer. After
the slope has been adjusted, the first
setting should be rechecked with thepH
6.0 (or 7.0) buffer; if it has changed by
more than ±0.02 pH unit, the calibra-
tion should be repeated. For these
buffer solutions, a stable reading
generally occurs in 30 s to 2 min. Since
pH is temperature dependent, calibra-
tion buffers and samples should always
be measured at the same temperature.
Before calibration, unsealed electrodes
should be topped off with filling solution
(available from the electrode manu-
facturer), and the exterior should be
carefully rinsed with deionized water.
During the measurements, the pH
electrode should not touch the bottom of
the solution vessel. If a metal ring stand
is used to hold the electrode, the stand
and the pH meter should be connected
to the same electrical ground.
The pH meter should be calibrated
before and after each measurement or a
series of measurements. If the initial
and final calibrations have changed by
more than 0.02 unit, the measurements
must be repeated. If this change
reoccurs, a problem exists with the
apparatus, and it should be remedied.
6.6.1.3 Sample Analysis—Sample
measurement is performed directly
after the meter has been calibrated and
the electrode washed. Never insert the
electrode or any other object into any of
the bulk solutions; never pour a solution
back into its bulk container; and never
measure while the solution is being
stirred to eliminate errors due to
streaming potentials. For measure-
ments, a small vessel (vial or test tube)
should be used. A procedure for sample
analysis is in the 0 & M manual.
Precipitation samples usually yield a
stable potential in about 4 min. The pH
electrodes, after being rinsed in de-
ionized/distilled water, should be
stored in deionized water, in 10"4N acid,
or in a 74 micromho/cm KCI standard.
6.6.1.4 Quality Control - Accuracy
and Precision — The main problem with
pH measurements is aging of the
electrode; diagnostic tests are pre-
sented here. The first test for electrode
or procedural problems should be with
unknown test samples sent monthly to
the field station by the laboratory; these
audit samples have pH and conductance
values similar to rain, and thus should
be measured for both before returning
them to the laboratory with the results
for recheck and evaluation. If the field
pH differs from the laboratory pH by
more than ±0.15 unit, the electrode
probably needs replacing; laboratory
values are assumed to be correct
because they are measured on several
samples of the unknown and because
the electrode is checked with a second
backup electrode. Thus the field results
are indicators of the accuracy of the pH
measurements; if a number of mea-
surements are made, the precision is
also obtained and should be better than
0.05 unit. If the accuracy and/or
precision are poor, consultation with
the field operator on technique should
confirm the source of the problem, and it
should be corrected by the QA coordi-
nator as soon as possible.
-------
Part l-Section 6.0
Jan. 1981
A second test for electrode problems
should be a reference solution of known
pH to learn the precision of the station's
measurements. Each site should re-
ceive from the cent'ral laboratory a
polyethylene bottle of electrode refer-
ence solution with pH and conductivity
similar to those of rain samples, and
should check the electrode biweekly.
The measurement procedure should be
identical to that for rain samples. The
solution should be stored in the re-
frigerator and replaced when needed or
when the pH or conductivity appears to
have changed. The average value and
the standard deviation of a series of at
least five test sample measurements
taken consecutively and the time
required for electrode equilibration can
be used to evaluate the electrode
performance. The standard deviation (s)
is:
6-2
where x, = the measured pH for the ith
sample,
x = the average pH reading for the
series, and
n = the number of sample tubes mea-
sured.
For acceptable electrode behavior, s
should not be greater than 0.05 pH unit,
and the average pH should agree with
the previous month's value within
±0.10 unit. Frequently, the first mea-
surement differs by more than ±0. 1 unit
from the others, so this value can be
excluded from the average for six
sample tubes. If this behavior is
exhibited by an electrode, it is strongly
recommended that two tubes of each
precipitation sample be measured for
pH and that the second value be entered
on the data form. Performance is
determined by the time needed to attain
a stable reading— when the pH reading
becomes constant within ±0.02 unit for
1 min.; the time should be less than 5
min. for a well-behaved electrode.
Results of these tests should be guides
for the measurement technique and the
equilibration time to be used for
precipitation sample measurement. If
an electrode test at any time exhibits
out-of-control behavior (as indicated by
the above criteria), the electrode should
be replaced. If the average pH value has
changed from that of the previous
month by more than 0.10 unit, the solu-
tion conductivity should be checked. If
the conductivity has changed by more
than 10% from its original value, the
solution has degraded, and should be
replaced. Always return enough of the
solution so that it can be checked by the
laboratory.
6.6.2 Specific Conductance Deter-
mination Method
6.6.2.1 General Description—The
conductivity of a solution is the recipro-
cal of its resistance, and it is related to the
solution temperature and to the total
concentration and species of free ions
present.
6.6.2.2 Calibration—The conductiv-
ity (or resistance) varies with the
electrode area and spacing as well as
with the temperature and the ion
concentration; therefore, the mea-
suring apparatus has to be calibrated to
obtain the cell constant or to adjust the
meter. For calibration, a KCI solution of
known conductivity should be used, and
the temperature of the KCI standard and
the sample should be the same. For rain
samples, a 0.0005M KCI solution with a
specific conductance of 74 micromho/
cm at 25°C should be used in the
calibration with the measurement
procedure recommended by the instru-
ment manufacturer. All conductances
should be reported in micromho/cm or
in microSiemen/cm corrected to 25°C.
If the apparatus has automatic tempera-
ture compensation or if the standard
and the unknown are at the same
temperature, no temperature correc-
tions are generally necessary if the
25°C KCI value is used for calibration.
(Temperature coefficients of the two
solutions are assumed comparable.) If
the apparatus does not have tempera-
ture cdmpensation or if the KCI standard
and the unknown are at different
temperatures, corrections must be
applied; KCI values near 25°C, based on
a temperature coefficient of 2% per
degree, are in Table 6-2.
The conductivity apparatus should be
calibrated before and after each mea-
surement or series of measurements. If
a change of more than 5% occurs, the
measurements should be repeated; if
the drift reoccurs, a problem exists with
the apparatus, and it must be corrected.
In general, stable values (for minutes)
occur in about 30 s.
Table 6-2.
Temperature Specific
Conductance of
0.0005M KCI
T°C
20
21
22
23
24
25
26
27
28
Micromho/cm
66.8
68.2
69.5
71.0
72.4
73.9
75.4
76.9
78.4
6.6.2.3 Sample Analysis—Conductiv-
ity of the samples can be measured "n
the same aliquot used for pH; i
conductivity must be measured be
pH to avoid error due to salt contamina-
tion from the electrode. The conductivity
cell should be washed with distilled/
deionized water after calibration and
rinsed with sample solution. The
procedure is in the O & M manual (9). If
the temperature of the sample is not the
same as that of the standard and if the
apparatus does not have automatic
temperature compensation, the mea-
sured conductance should be corrected
to 25°C by adding 2%/°C if below 25°C
or by subtracting 2%/°C if above 25°C.
6.6.2.4 Quality Control - Accuracy
and Precision—Unlike the pH electrode,
which has a limited life, the conductivity
cell generally has few problems. How-
ever, in contrast to the pH buffers,
which are concentrated and quite
stable, the working conductivity stand-
ard is a very dilute, 0.0005M KCI, and
may either degrade slowly or become
contaminated. To minimize errors due
to changes in the calibration standard,
the working 74 micromho/cm solution
should be replaced approximately
monthly. When a new working standard
is received, it should be checked aga>' *
the old working standard, and the
values should agree within 10%. If t...,
do not, notify the laboratory. Always
return enough of the old standard to the
central laboratory for it to be remea-
sured. Never return the old working
standard before receiving a new one.
Conductivity standards should be
sealed and stored in a refrigerator to
minimize changes. Generally, changes
of less than 5% monthly may be ignored;
if greater than 5%, the field values can
be corrected for the large changes by
prorating with time in a linear manner.
Such corrections should be duly noted.
If the conductivity meter has its own
builtin standardization circuit, it can be
used to check the KCI standard. If the
KCI standard has changed from its
original value by more than 5%, the
laboratory should be informed. The cells
should be stored as recommended by
the manufacturer.
Another means of evaluating the
working conductivity standard is to use
unknown test samples submitted month-
ly from the laboratory to determine the
accuracy and precision of the station's
specific conductance measurements.
These test samples must be returned to
the central laboratory with the next
sample shipment for remeasuremer
ensure that they have not change
value. If the laboratory finds that tne
station's conductivity differs from the
laboratory's by more than 10%, the
-------
Jan. 1981
Part l-Section 6.0
laboratory should inform the field and
••lity assurance personnel, and should
ice the old conductivity standard.
v,.o.3 Temperature Measurements
6.6.3.1 General—Each field ther-
mometer and temperature probe should
be assigned an identification number so
that it will be possible to trace its
certification and to document it proper-
ly. To minimize any chance of con-
tamination, the probeshould be washed
and dried before the temperature mea-
surement of a solution, and the probe
should never be placed in a solution
before pH and conductivity measure-
ments unless there are duplicate
vessels for each sample. The probe
should be placed only in the solution
used for .rinsing, not for pH and
conductivity determinations.
6.6.3.2 Calibration andTraceability—
The central support laboratory should
maintain and store an NBS-calibrated
thermometer, as a primary standard,
and one field thermometer should be
certified against this as a secondary
standard. All field thermometers or
temperature probes should be cali-
brated against the certified (secondary)
thermometer in a circulating water bath
'' *he 0° to 25°C range; the procedure
ild be similar to that used for
.ifying the secondary standard. The
Thermometer Calibration Log (Section
6.8) should be completed, one copy
should be filed in the laboratory, and
another should be sent to the field with
the temperature probe. If the water bath
does not have a cooling coil, an ice-
water mixture can be used to achieve
the low temperature. Generally, calibra-
tion at two temperatures, near 0° and
25?C, is sufficient. Linear temperature
behavior may be assumed.
After initial calibration, the tempera-
ture behavior probes should be cali-
brated at least once a year in the labora-
tory or in the field by using a certified
thermometer and two temperatures
(i.e., 0°C and ambient).
6.6.4 Gravimetric Measurements
6.6.4.1 General—The volume of rain
is measured as mass of rain with a
density of 1 g/ml. The mass of rain is
measured in the field to determine the
rain collector efficiency (compared to
the rain gauge), and the mass of the
sample sent to the laboratory is mea-
sured to determine if leakage occurred
in shipment.
The field site should have a 20 kg
—>acity balance for weighing rain
ets; the balance should be in a
n free from drafts andon a table that
minimizes vibrations; and the legs of the
balance should be adjusted to level the
balance.
Before each weighing, the balance
pan should be brushed off with a soft
brush and the balance zeroed. After
each weighing, the balance should be
cleaned of all (potentially corrosive)
chemicals.
Periodic maintenance should be
according to the manufacturer's recom-
mendations; factory maintenance is
usually once a year, and a record of
maintenance should be kept for each
balance by completing the Balance
Factory Service form (Section 6.8.5).
6.6.4.2 Balance Calibration—Each
balance, before being shipped to the
field, should be calibrated with NBS
traceable weights in the central support
laboratory. In the field the balance
should be zeroed before each rain
sample is weighed. Annually, a full
calibration as part of the network
evaluation (Section 6.7) should be
performed by weighing two NBS trace-
able weights (1.0 and 5.Okg) on the field
balance. The auditor-recorded actual
reference weight, measured weight,
and weight difference should be re-
corded on the Audit Record form (Sec-
tion 6.8.5).
6.6.4.3 Rain Gauge Calibration—To
calibrate the weighing gauge, a set of
weights is generally available from the
manufacturer. With a dual traverse(0-6
and 6-12 in.) pen recorder such as the
Belfort, the range 5-7 in. has been
difficult to calibrate and to keep cali-
brated; this range is generally not
needed if the rain gauge bucket can be
emptied after each event or week of
events. However, antifreeze must be
added to the weighing gauge bucket in
the winter to melt captured snow, so
severe or prolonged storms can bring
the gauge to 5-7 in.; accordingly, it is
recommended that the bucket be
emptied when the 5 in. level is ap-
proached before adding new antifreeze.
The tipping bucket gauge is calibrated
by using a slow drip technique to add a
controlled volume of water. The set
screws may need to be adjusted to limit
the travel of the tilting bucket.
6.7 Network Evaluation
Establishing a schedule for audits and
independent checks to evaluate the
quality of data provided by the total
measurement system is an important
part of an overall QA program. An audit
for a precipitation chemistry network
should include both qualitive and
quantitative evaluations. The quantita-
tive evaluation of precision and ac-
curacy is discussed in Section 9; the
remainder of this section discusses a
qualitative audit of performance.
6.7.1 System Audit
A system audit is an on-site inspec-
tion and review of the QA efforts used
for monitoring (sample collection,
sample analysis, data processing, etc.).
The QA project plan (Section 3) should-
be the basis for a system audit.
For each monitoring network, a
system audit should be conducted as
soon as possible after the start of
monitoring; subsequent system audits
should be scheduled at least once per
year. About 4 to 6 weeks before the
audit, the QA coordinator should send a
questionnaire to the field operations
manager, and the questionnaire, (Sec-
tion 6.7.3) should be returned before the
auditor's visit to allow the auditor to
review and evaluate the recorded
information. The system audit should be
conducted by (a) interviewing the field
manager and the station operators, (b)
visiting the monitoring site, and (c)
writing a summary report with results,
observations, and recommendations.
6.7.2 System Audit Questionnaire
The system audit questionnaire
(provided herein) is in the recommended
format, but it can be modified and
supplemented to fit the needs of each
specific monitoring network. In the
questionnaire, the R & G denote
"recommended strongly" and "guide-
line", respectively.
-------
Part l-Section 6.0 8 Jan. 1981
Field Agency Questionnaire
Forms
General Information
Precipitation Monitoring Resources/Staff Size, Organization,
Qualification and Utilization
Precipitation Monitoring Resources/Staff Training
Precipitation Monitoring Resources/Equipment
Precipitation Monitoring Resources/Facilities
Monitoring Network/Network Design
Monitor Network/Network Status
Monitor Network/Network Operation
Monitor Network/Network Maintenance
Monitor Network/Network Calibration
Quality Assurance: Data Quality Assessment Requirements/General .
Quality Assurance: Data Quality Assessment Requirements/Methods.
Quality Assurance: Data Quality Assessment Requirements/
Required Calculations For Data Quality Assessment
Data Handling and Reduction/General
-------
Jan. 1981 9 Part (-Section 6.0
General Information
C -(ionnaire completion date:
:y visit date:
Agency name and address:
Agency mailing address (if different from above)
Telephone number: FTS: Commercial:! )-
Agency Director:
Monitoring Supervisor:
Quality Assurance Coordinator:
Survey Conducted by:
Affiliation of Auditor:
Persons present during entrance interview:
Persons present during exit interview:
-------
Part l-Section 6.0 10 Jan. 1981
Precipitation Monitoring Resources
A. Staff Size. Organization. Qualifications and Utilization
Yes No Comments
1. Provide a current organizational chart
showing precipitation monitoring and
data handling personnel. . ( )
2. Are the following items adequate to
demonstrate that the equipment can
be operated, calibrated and maintained:
a. Staff size? (R)
b. Program organization? (R)
c. Staff qualifications? (R)
d. Staff utilization? (R)
B. Staff Training
1. Do staff members receive periodic training
to upgrade employee's skills? IG)
a. At least once a year? (G)
2. Are the manufacturer instrument man-
uals, and quality assurance guideline
documents for precipitation monitoring
available to the operators? (R)
3. Are the following references available?
a. Atmospheric Environment (G)
b. Journal of the Air Pollution Control
Association (G)
c. Environmental Science and Technology(G)
-------
Jan. 1981 11 Part l-Section 6.0
Precipitation Monitoring Resources
tuipment
Yes No Comments
/. Does the agency have the necessary hand
tools, electrical testing, and calibration
weights to operate, and maintain the in-
struments, to calibrate the rain gauges
and assure the data quality in the
network? IP) .
2. Are the following types of equipment
available or on hand?
a. Hand tools-screwdrivers, wrenches.
etc. (G)
b. Multimeter 61 analog (G)
62 digital (G)
c. Soldering iron, gun, and accessories (G)
d. Electric drill, saws. etc. (G)
e. Tubing cutters and accessories (G)
3. For precipitation collection are the
following types of equipment used?
a. Automatic precipitation collectors (R) .
Recording rain gauge sensitive to
0.07 in (0.25 mm). (R) .
c. Meteorological gear:
(J) wind speed 1C)
(2) wind direction (G)
4. Are the means to calibrate the rain
gauges available? (R) -
D. Facilities
7. Is adequate space available to operate
and maintain the network? (Rl
2. Is the work space used for the sample
measurements maintained at 25°C ±5°C?(R)
3. Are the instruments operated at a
normal line voltage between 105
and 125V? (G)
-------
Part l-Section 6.0
12
Jan. 1981
A. Network Design
Monitoring Network
Yes
No
Comments
1. Is the network designed in accordance
with the program objective? (R)
a. Is there a written plan describ-
ing the network prescribing:
(1) The basis for design of the
network, selection of instru-
ments and siting? (R)
(2) The locations of the instru-
ments (site locations) by UTM? IR)
(3) The sampling schedules? (R)
(4) The methods of sampling
and analysis? IR)
(5) The method of data handling
and analysis procedures? (R)
b. Have the site description ques-
tionnaires been submitted? (R)
2. Are instruments installed at a site
in accordance with:
a. Manufacturer's specifications? (R)
b. EPA guidelines? (R)
c. Sound scientific principles? (R)
3. At the sites:
a. Are there any obstacles the
height of which subtends an
angle greater than 30°? (R)
b. Are the precipitation collec-
tor and rain gauge at least 10
feet apart? (R)
c. Are the rain gauge and collec-
tor placed in a line perpendi-
cular to the prevailing winds? (G).
d. Is the rain gauge level? (ft)
e. Is the access door of the rain
gauge on the leeward side of
the wind path? (G)
f. Can the rain gauge measure 0.01
inches of precipitation? (R)
-------
Jan. 1981
13
Part l-Section 6.0
itwork Design
Monitoring Network
4. Does network design consider:
a. Access? (R)
b. Power availability? (R)
c. Localized interferences? IR)
5. Is the precipitation fall to the
sites unobstructed? (R)
6. Are sites located in accordance with
current EPA site location criteria
and ambient precipitation monitor-
ing siting guidelines? (R)
7. a. Does each site have a written
identifying purpose? (R)
b. Are the samplers located at the
optimum site to meet the purpose? (R)
Yes
No
Comments
B. Network Status
•es the agency have the following
ords identifying the history and
-.aft/5 of each monitoring site?
(R)
a. Completed site identification
form? IR)
b. Photographs or slides. One
photograph or slide toward each
of the four compass directions
and one closeup photograph of the
instrumentation at the site? (R)
c. Date site was started up and date
site was shut down, as appropriate? (Ft)
d. Model, manufacturer and serial
number of instruments at the site
and dates each instrument operate? (R)
e. Reasons ior periods of missing date? (R)
C. Network Operation
1. Are the instruments in the network
operated in strict accordance with the
agency's written Standard Operating
•ocedures? (R)
-------
Part l-Section 6.0
14
Jan. 1981
Monitor Network
C. Network Operation
Yes
No
Comments
2. Are these procedures compatible with:
a. The Federal reference or equivalent
method if available? (R) .
b. The manufacturer's instruction manual?(R)
c. The Agency Quality Assurance plan? (R)
3. Is a formal schedule used for operating
the instruments? (Attach a copy of the
schedule.!
(R)
4. Does the agency maintain an adequate
supply of expendables necessary to
service the instruments? (G)
5. Precipitation samples are picked up:
a. Every 24 hours? (G)
b. Every 7 days? (G)
c. Other (specify) hours (G)
6. In the winter:
a. Are the necessary precautions taken
with the rain gauge (antifreeze,
funnel removal, heater)? (R)
b. Are precautions taken with the collec-
tor (boots for lid arms, seal removal)? (G)
D. Network Maintenance
1. Is preventive maintenance performed in
strict accordance with the agency's
Standard Operating Procedure? (R)
2. Is a formal written schedule used for
performing preventive maintenance on
the instruments? (Attach a copy.)
3. Does the agency have access to a qualified
instrument repairman?
a. Is it a staff position?
b. Is it a manufacturer's serviceman?
c. Does the agency have a service con-
tract on any instruments? (Specify
on back of this page.)
4. Are the recommended spare parts ade-
quate and available to maintain instru-
ment downtime?
5. Are sample buckets, containers and lids
kept clean?
6. Is the collector sensor cleaned period-
ically?
(R)
(R)
(Rl
(R) -
(G)
(R) -
(Rl -
-------
Jan. 1981
15
Part l-Section 6.0
Monitor Network
D. Mvtwork Maintenance
Yes
7. Is the collector rim of the dry bucket
wiped clean with Kimwipes weekly? (G)
8. Is the rain sensor tested at every site visit ?(R)
9. Is the collector dry bucket inspected
for moisture whenever a sample is
removed? (R) .
10. Are the rain gauge pens checked for
ink weekly? (G)
J1. Is the rain gauge clock wound at the
prescribed intervals? (R)
12. Is the time clock on the rain gauge
recorder accurate (to 'A hr/week)? (G)
13. If not. is it corrected weekly? (R)
E. Network Calibration
Comments
1. a. Does the agency conduct frequent rain
gauge. pH and conductivity apparatus
calibrations? (R) _
; the frequency adequate to demonstrate
the accuracy, precision, and completeness
for all data submitted? (R) _
c. Are the rain gauges calibrated:
ID Upon installation? (R) .
(2) On a semi-annual basis? (R) -
(3) When major maintenance is
performed? (R) _
14) Before removal from operation? (R) _
(5) When the daily service and periodic
audits fail to meet the prescribed
performance specifications? (R) _
d. Are new calibration or standard
solutions for pH and conductivity
checked against the old ones? (R) _
e. Is there a statistically significant
quantity of calibration data routinely
generated to determine the accuracy
and precision, at the 95% confidence
level, of all data? (R) .
he conductivity working standard
laced monthly? (R)
-------
Part l-Sectlon 6.0 16 Jan. 1981
Monitor Network
E. Network Calibration
Yes No Comments
3. Does the agency own or have access to
calibration capability for each:
a. Rain gauge (R)
b. Recorder (R)
c. pH and conductivity apparatus (R)
4. a. Do all calibration standards pur-
chased have a maximum analytical
error of ±2%? (R) .
b. Is the accuracy of all calibration
standards specified and documented? (R)
5. Are all instruments calibrated under the
same conditions as they are operated? (R) .
6. Are instruments calibrated at the same
site as they are operated? (R) .
7. Are the pH and conductivity meters cali-
brated daily before sample measurement?(R) .
8. Are records kept documenting all:
a. Audits? (R) .
b. Calibrations per (R)
-------
Jan. 1981 17 Part l-Section 6.0
Quality Assurance: Data Quality Assessment Requirements
•>neral
Yes No Comments
1. Are data quality "reporting organiza-
tions" properly defined? (R)
a. Is operation by common team of field
operators? (G)
b. Does organization have common
calibration facilities? (G)
c. Is organization supported by a
common laboratory or headquarters? (R)
2. Does the agency have a formal quality
assurance program, developed and
implemented, as evidenced by a
written plan? (Rj
3. Are confidence limits assigned to a/I
data collected? (R)
4. Are all the data submitted documented
to show that they are acceptable? (R)
5. Have written Standard Operating Pro-
cedures been developed and imple-
mented for each routine task performed:
i.e. operation, maintenance, calibration,
etc.? (R)
6. Are all quality assurance data docu-
mented and available for inspection? (R)
7. Does thf agency ha ve a quality assurance
coordinator to insure that a maximum
quantity and quality are generated? (R)
8. Does the agency participate in the EPA
quality assurance audit program? (R)
the chain of custody maintained on
II samples? (R)
10. Is complete traceability maintained
on all data? (R)
11. Does the agency maintain sufficient
quality assurance information related
to data collection and analysis to
demonstrate that the data are accept-
able? (R)
-------
Part l-Section 6.0 18 Jan. 1981
Quality Assurance: Data Quality Assessment Requirements
B. Methods
Yes No Comments
1. Is a one point (minimal) check on the rain
gauge performed monthly? (R)
2. Is the balance zeroed before each
weighing? (Rj
3. Is the outside of the wet bucket wiped
dry before weighing? (G)
4. Does the pH electrode have an identifi-
cation number? (G)
5. Is the pH electrode stored in the
proper pH buffer? (R)
6. Is the pH electrode rinsed well with
distilled/deionized water after removal
from the buffer? (R)
7. Are the samples allowed to come to room
temperature before they are measured
for pH and conductivity? (R)
8. For the pH measurement is 4 M KCI added
to the sample aliquot to yield a 0.04 M
solution? (R)
9. Are the sample aliquots measured for pH
and conductivity discarded after meas-
urement? (R)
10. Is the electrode test run monthly? (G)
/ 7. Are the electrode test results sub-
mitted to the laboratory? (Rj
12. Are the conductivity standards and
electrode test solution refrigerated? (G)
13. Is the conductivity of the rinse water
measured and recorded? (R) :
/ 4. Are the data properly recorded in a log
book and on the data? (R)
a. Are the container labels made out
in pencit.or indelible ink? (R)
b. Are the samples sealed securely
before shipment? (R)
c. Are shipments made on the
prescribed schedule? (R)
15. Are the stations audited semi-annually?(R)
16. As a check on precision, have at least
two pairs of collocated samplers been
operated during the calendar quarter? (R)
7 7. Do the sites meet the Class A require-
ments? (R)
18. Are calibration, sampling, & analysis
the same for collocated samplers? IR)
C. Required Calculations for Data Quality Assessment
1. Are the average deviations ofthe pH elec-
trode from the test solution calculated
monthly? (R)
2. Have the following quality assurance
reporting requirements been met? (Ask
to see copy of a typical report.)
a. A list of all monitoring sites and
their identification codes in each
reporting organization (Regional
Office and EPA-EMSL/RTP)? ( j
b. Changes in assignments of monitoring
site (Region and EPA/EMSL)? ( )
c. Quarterly reports? ( )
d. Quality assurance portions of the
annual report? ( )
-------
Jan. 1981 19 Part l-Section 6.0
Data Hand/ing and Reduction
icral
Yes No Comments
1. Have standard data forms (bound or
loose} been developed and implemented
to document the results as contained in
the Standard Operating Procedures-' (R)
2. Do the standard data forms document
complete data traceability? (R)
3. Are all standard data forms completed.
dated, and signed by the person per-
forming the task? (R)
4. Are the station data record sheets for
each sample made out in duplicate? (R)
5. Is one copy of the data sheet sent with
each sample? (R)
6. Is the rain gauge chart sent weekly
with the samples? (R)
7. Is a log book maintained? (R)
8. Are problems, equipment changes, new
standards, new operators, etc. docu-
mented in the log book and reported to
the central laboratory? (R) .
9. Is the sample container labelled with
station identification, weight, date.
sample number? (R) .
10. Are duplicate or sequential samples
identified properly? (R) .
11. Do standard data forms show a com-
plete chain of custody and traceability
~r all data from generation to sub-
ssion to the data bank? (R) .
-------
Part l-Section 6.0
20
Jan. 1981
6.7.3 Auditor Site Visit
The site visitation should cover all
aspects of site operation, so the auditor
can evaluate the operator's overall
capabilities. The auditor should bring a
test sample and add it to a clean bucket
at the station; observe the operator's
weighing and measurements of its pH
and conductivity; record the data on the
Audit Record form (Section 6.8.5);
watch the operator transfer the sample
back to the test sample bottle, rinse the
bucket, and test the rinse water con-
ductivity to ensure that the bucket is
clean; and finally inspect all the equip-
ment, calibrate the rain gauge, offer
advice, and ask questions while the
operator goes through his rounds and
tests. If there are any problems, the
auditor must attempt to correct them or
bring them to the attention of the field
manager.
6.8 Documentation
All data, observations, changes, or
modifications must be dated and docu-
mented on data forms and/or in
logbooks in triplicate and duplicate,
respectively (carbon paper may be
used). One copy of each should be kept
in the station records, and another
shipped with the sample; the third copy
(the data form) should be mailed
separately from the sample to the
laboratory to help trace a missing
sample.
6.8.1 Logbook
A bound notebook with perforated
pages that can be easily torn out should
be used. All problems and actions, dates
and times of visits, equipment changes,
procedural changes, or modifications,
standard solution changes, electrode
replacement, operator changes, and so
forth should be recorded in the logbook,
and a copy should be sent to the central
laboratory.
6.8.2 Rain Gauge Chart
The rain gauge charts, marked with
station identification, date and nota-
tions for tests, problems, and so forth
should be submitted weekly to the
central laboratory.
6.8.3 Field Data Form
Field Data forms should contain all
information required for identification
of the sample—station, date, weight,
sampling times, pH, conductivity, prob-
lems, and supply orders. Triplicate
forms should be made out for each
sample.
A sample of a data form used by NADP
for event sampling is in Section 6.8.5.
Instructions for filling out the form are in
the 0 & M manual(9). Under item 11
(Remarks) should be included unusual
occurrences—plowing, harvesting,
burning, increased atmospheric pollu-
tion or dust, and so forth—and any other
information, comments, or observa-
tions—power outage, smell, moisture in
the dry bucket, new electrode, new
conductance cell or standard.
The importance of the information
requested in item 11 cannot be over-
emphasized. Careful observations of the
sample and the occurrences in the
surrounding environment can aid in
evaluating the validity of the sample and
the collected data.
6.8.4 Sample Identification
Samples must be labeled so that they
can be readily and correctly matched
with their data forms. The label should
contain station, date, and sample
weight marked with a pencil or a ball
point pen so that it is legible if it gets
wet.
For duplicate (collocated) samplers or
sequential samplers, each sample
container should be coded and logged in
separately so they can be easily identi-
fied in the data base. The station can
distinguish between duplicate samples
by adding a -1 and -2 in the site number
space on the data form; for sequential
samples, add -11, -12, -13, and so forth
to denote chronological order of collec-
tion for each event.
6.8.5 Data Forms
For the convenience of the manual
user, blank data forms are listed and
then provided on the following pages;
many of these were adapted from EPA
and other forms. No documentation is
on these forms. The titles are at the top,
as customary for a data form. To relate
the form to the text, a nur .ber is in the
lower right-hand cornt. —for example,
1.1/6.3.2 indicates form 1/version 1 as
discussed in Section 6.3.2. A revision of
the form may be documented with
1.2/6.3.2, form 1/version 2 and so
forth.
Form Number Title
1.1/6.3.2 Conductivity Meter/Cell
Acceptance Test Form
2.1/6.3.2 Conductivity Test Sum-
mary Form
3.1/6.3.2 pH Meter/Electrode Ac-
ceptance Test Form
4.1 /6.3.2 pH Acceptance Test Sum-
mary Form
5.1/6.6.4 Balance Factory Service
Form
6.1/6.6.4 Thermometer Calibration
Log
7.1/6.7.2 Audit Record Form
8.1/6.8.3 Field Data Form
6.9 References
1. Galloway, J.N., and G.E. Likens,
Water, Air and Soil Pollut. 6, 241
(1976).
2. Galloway, J.N., and G.E. Likens,
Tellus30, 71 (1978).
3. Cooper, Jr., H.B.H.,J.A. Lop d
J.M. Demo, Water, Air Soil. ^t.
6, 351 (1976).
4. Gatz, D.F., R.F. Selman, R.K.
Langs, and R.B. Holtzman, J. Appl.
Meteorol. 10, 341 (1971).
5. Morgan, J.J., and H.M. Liljestrand,
The Measurement and Interpreta-
tion of Acid Rainfall in the Los
Angeles Basin, California Institute
of Technology Report, No. AC-2-
80, February 20, 1980.
6. Robertson, J.K., T.W. Dolzine and
R.C. Graham, Chemistry and Pre-
cipitation from Sequentially Sam-
pled Storms, EPA report to be
published.
7. Raynor, G.S., and J.P. McNeil, The
Brookhaven Automatic Sequential
Precipitation Sampler, BNL-50818,
Brookhaven National Laboratory,
January 1978; Atmos. Environ.
73,149 (1979).
8. P.B.S.K. Associates, P.O. Box 131,
State College, PA. 16801, Bulletin
177.6801.
9. Operations and Maintenance Man-
ual for Precipitation Measure-
ment Systems, U.S. Environ-
mental Protection Agency, Re-
search Triangle Park, N/ 'n
press).
10. Peden, M.E., and L.M. Skowron,
Atmos. Environ. 12. 2343 (1978).
11. Rothert, J.E., Battelle Pacific
Northwest Laboratories, Richland,
Washington, MAP3S Program,
telephone communication.
-------
Jan. 1981
21
Part l-Section 6.0
Conductivity Meter/Cell Acceptance Test Form
Date of Test:
Preparation date of a 0.0003 M KCI reference solution:.
Meter type/serial no. f_
Conductivity cell type/serial no..
Indicate whether test of
_meter or
cell
(Analyst Signature)
Conductivity Values Obtained:
Aliquot 1:
Aliquot 2:
Aliquot 3:
Aliquot 4:
Aliquot 5:
Aliquot 6:
Aliquot 7:
Aliquot 8:
Aliquot 9:
Aliquot 10:
Test Solution
Average conductivity
± standard deviation:
Accepted _
Rejected.
QA Manual for Precipitation
1.1/6/3/2
-------
Part l-Section 6.0
22
Jan. 1981
Conductivity Acceptance Test Summary Form
Meter Type/
Serial ti
Cell Type/SerieItt
Date of Ref
Soln. Prep.
Date of
Check
Conductivity Value
Average ± Standard
Deviation (umho/cm)
Number
of
Values
Ana.
Initials
QA Manual for Precipitation
2.1/6.3.2
-------
Jan. 1981
23
Part l-Section 6.0
pH Meter/Electrode Acceptance Test Form
Date of Test:
Preparation date of pH electrode reference solution:
Meter type/serial no Z
pH electrode type/serial No.
Indicate whether test of.
.meter or
electrode
(Analyst Signature)
pH Values Obtained:
4.0 (3.0) Buffer before:
7.0(6.0) Buffer before:
Aliquot 1:
Aliquot 2:
Aliquot 3:
Aliquot 4:
Aliquot 5:
Aliquot 6:
Aliquot 7:
Aliquot 8:
Aliquot 9:
Aliquot 10:
4.0 (3.0) Buffer after:
7.0 (6.0) Buffer after:
pH Electrode Test Solution
Average pH ± standard deviation:
Accepted.
Rejected.
QA Manual for Precipitation Measurement
3.1/6.3.2
-------
Part l-Section 6.0
24
Jan. 1981
pH Acceptance Test Summary Form
Meter Type/
Serial #
Electrode Type/
Serial #
Date of Ref.
Soln. Prep.
Date of
Check
pH Value
Average ±Standard
Deviation
Number
of
Values
Anai,
Initials
QA Manual for Precipitation
4.1/6.3.2
-------
Jan. 1981
25
Part l-Section 6.0
Balance Factory Service Form
Balance
Identification
Number
Date of
Factory
Service
Name of Factory Service
Representative
Date Next
Service
Due
Manual for Precipitation
5.1/6.6.4
-------
Part l-Section 6.0
26
Jan. 1981
Thermometer Calibration Log
Date
Therm.
Serial #
Temperature tt1
NBS
Value
Lab
Value
Temperature #2
NBS
Value
Lab
Value
Temperature #3
NBS
Value
Lab
Value
Slopea
Intercept*
Anal.
Init.
"Calculated from a linear least squares fit with the lab value on the X-axis and the NBS value on the Y-axis. The
laboratory thermometer correction to NBS is, then: NBS value - slope xflab value) + intercept.
QA Manual for Precipitation
6.1/6.6.4
-------
Jan. 1981
27
Part l-Section 6.0
Audit Record Sheet
sue Identification:
Operator:
Date:
Auditor:
Lab Anal. Before
pH Cond
A udit Sample No:-
Field Anal.
pH Cond
Lab Anal. After
pH Cond
Difference Between Initial and Final Lab Values:
Difference Between Lab and Field Values:
initial final
Cond =
Cond
initial final
Comments:
QA Manual for Precipitation Measurement
7.1/6.7.2
-------
Field Data Form
CAL/NREL USE ONLY
BULK
DA
QA
NS/Exclude
LD
NO
NN
1 Station
Name
ID .II I I I I I
2 Observer
Initials •
Signature
3 Sample Bucket
Check ||
One
Dry-Side
Wei-Side
Check Yes or No for each item for wet-side samples only
if No, explain in remarks
t. Collecinr nil/wars to have operated properly and sampled all precipitation
events dunny enure sampling period
2 Rain gauge appears to have operated properly during the meek
3 Collector opened and closed at least once during the week
4 Bucket
On
Bucket
Off
EOT/-II
ESJ/CDT 101
CSJ/MOJ til
Circle Time Zone
MS J/POT 121
PST 131
AKDT II)
AKS!. MSI 151
American 161
Samoa
5 Site Operations
6 Sample Condition Complete for all samples containing precipitation
1. Bird Droppings
2. Cloudy or Discolored
3. Lots of Soot or Dirt
Describe all else in remarks
7 Sample Weight - Grams
Only lor Buckets with water, ice. or snow
Bucket &
Sample & Lid
~ I I I I I 1*1 I Suc*e' * Lid
I I | Ul Samplt
-1 I I I I 1*1 I weight
8 Precipitation Record
For wet-side samples
only
Type
Circle one for each
day of precipitation
Amount In inches
or circle
one
R-Rain only s-Snow only m-Mixlure u-Unknown
z-zero t-Trace mm-Missing
Bucket On To Bucket Off
Tues
R S M U
Z T MM
R S M U
Z T MM
Thur
R S M U
T MM
Fn
R S
Z T MM
Sai
R S M U
Z T MM
Sun
R S M U
Z T MM
Mon
R S M U
Z T MM
Total sampling period precipitation from rain gauge
Total precipitation from sampler = sample weight x 0,00058
inches/gram
Tues
R S M U
Z T MM
inches
inches
9 Sample Chemistry
Only for wet-side buckets with precipitation
Conductance
Distilled water
"j //S/um
OH-
Aliquot Removed Standard Certified Standard Measured Correction Factor
rruixi 111 ui i=rrmn
Correction Factor Sample Measured Sample Corrected
pH
pH 4 Observed
Sample
10 Supplies
Circle if needed
pH 4
pH 7
75 AiS/cm
Field Forms
11 Remarks
For example: Contamination by operator, equipment malfunction, harvesting in
area
QA Manual for Precipitation
Revised 10/1/80
8.1/6.8.3
-------
Jan. 1981
Part l-Section 7.0
7.0 Laboratory Operations
Quality data generated by an analyt-
ical laboratory requires detailed infor-
mation on needed equipment, control of
instrument performance, and a formal
QC program.
7.1 Facilities
Minimum facilities which must be
available to laboratories analyzing
rainwater samples include adequate
benchspace, lighting, electrical power,
temperature control, compressed gases,
exhaust hoods, reagent storage space,
refrigerated sample storage, glassware,
analytical equipment, and an area for
washing glassware (1).
7.1.1 Benchspace
Benchspace should be adequate for
equipment, and should allow each
analyst a minimum of 4 ft2 for prepara-
tion of reagents. A minimum of 8 ft2
each is needed for the atomic absorp-
tion spectrophotometer, ion chroma-
tograph, and automated colorimeter
system; 4 ft2 each is needed for the
analytical balance, pH meter, and
ductivity meter. Bench tops should
5 to 38 in. above the floor.
/.1.2 Lighting
Lighting should be sufficient for
easy reading of glassware gradations,
balance verniers, and other measuring
lines. It is recommended that the
lighting of benchtops be at least 100
footcandles.
7.1.3 Electrical
The electrical system should provide
both 115 v and 230 v, and should have
the capacity to accommodate the
wattage requirements of all the electri-
cal devices in the laboratory. Voltage
must be stable to protect instrument
components and to avoid instrumental
response changes. Most modern instru-
ments have builtin voltage regulators,
so that the only requirement of the
voltage source is that "spikes" greater
than ±10% do not exist; if necessary,
voltage regulators can be readily
obtained. All electrical equipment
should be carefully grounded, prefer-
ably using three-pronged plugs.
7.1.4 Laboratory Temperature
The laboratory temperature should be
maintained at 22±3°C, without exces-
sive drafts or temperature changes. It is
°<5oecially important to isolate the ion
•jmatograph, which has a conductiv-
detector, from drafts and tempera-
ture fluctuations. Dilute solutions have
conductivity temperature coefficients of
approximately 2%/°C.
7.1.5 Compressed Gases
Compressed gases are needed for
the ion chromatograph and for atomic
absorption analyzers. Compressed air
for the ion chromatograph provides a
constant 80 psi to the pneumatically
actuated valves; any inert gas is
acceptable, and purity is not critical
although particulates should not be
present.
Since gases used for atomic absorp-
tion spectroscopy eventually pass the
light path of the atomic absorption
detection system, standards of quality
must be met. For the analysis of calcium,
magnesium, sodium, and potassium,
acetylene is the fuel for the atomic
absorption burner and air is used as the
oxidant; both gases should be free of oil,
water droplets, and particulates. A filter
in the line will remove these impurities.
Air flow to the atomic absorption
burner is approximately 20 liter/min.;
a standard #1 air cylinder lasts about
5 h at this rate. If a cylinder is used, it
should be compressed air, not "breath-
ing" air or any other type. Instead of
cylinder air if a compressor is used, the
flow should be at least 30 liter/min. at
40 psi. A surge tank will help provide a
smooth flow.
Acetylene for atomic absorption is
burned at 5 liter/min. when air is the
oxidant; at this rate, a standard sized 1 B
cylinder will last about 30 h. Acetylene
in cylinders is dissolved in acetone, and
the ratio of acetone vapor to acetylene
vapor changes as the acetylene is
depleted. Because atomic absorption of
some elements is affected by this ratio,
it is good practice to renew acetylene
cylinders when the pressure is below
100 psi.
7.1.6 Exhaust Vent for Atomic Absorp-
tion Burner Gases
An exhaust vent is necessary to
protect personnel from toxic vapors, to
minimize the effects of room drafts on
the flame, and to protect the atomic
absorption instrument from corrosive
vapors. The venting system should
provide a flow of 5400 to 8400 liter/
min. The vent system has an "inverted
fun nel" intake over the burner, tubing to
the blower intake, the blower itself, and
tubing from the blower exhaust. Stain-
less steel sheets and flexible stainless
steel tubing are recommended for
construction of the vent system. All
connections should be made with
screws or rivets because temperatures
up to 310°C may occur at the inlet.
Atomic absorption instrument manu-
facturers provide details of construction
for vent systems.
7.1.7 Exhaust Hood
Acids are used to stabilize samples
for metal analysis and to wash glass-
ware and plasticware. It is recom-
mended that these operations be done
in an exhaust hood. The air flow through
the hood should achieve a linear face
velocity of 100 ft/min with the sash
fully open.
7.1.8 Vacuum
Rainwater samples are filtered to
remove particulates. Vacuum or inert
gas pressure filtration should be used to
minimize the time of exposure of the
sample to laboratory air and dust.
7.1.9 Sink
A sink with both hot and cold water
should be provided for washing glass-
ware and plasticware. An area (about
16 ft2) should be provided for air drying
glassware and plasticware.
7.1.10 Storage Space
Adequate shelf space should be
available for storing reagents, glass-
ware, and plasticware. Closed cabinets
are preferred to minimize chances of
dust settling in containers.
7.1.11 Distilled or Deionized Water
Distilled or deionized water is used
in an analytical laboratory to prepare
reagents, to make dilutions, and to
rinse glassware and plasticware. Table
7-1 offers criteria for water purity (2,3).
Water having a conductivity less
than 1.0 micromho/cm (resistivity
greater than 1.0 megohm/cm) is
acceptable for analysis of major constit-
uents in rainwater. In the past, high
purity has been obtained by distilling
water; however, distillation systems
have several drawbacks. Even water
double or triple distilled contains easily
detectable impurities (3). Stills require
periodic shutdown and careful cleaning
and water production is relatively low.
Ion exchange systems, on the other
hand, provide high quality water, are
relatively maintenance free, and pro-
vide water on demand. The only main-
tenance required is to change cartridges
periodically. It is preferable to pretreat
the feed water with a reverse osmosis
system to remove a high percentage of
ionic impurities and to prolong the life of
-------
Part l-Section 7.0
Jan. 1981
the ion exchange beds. In the last stage
of treatment, a 0.2-micron filter should
be used to remove microorganisms and
particles. A meter to monitor the
conductivity of the water should be
installed inline directly before the
spigot, and the system should be
checked if the conductivity becomes
greater than 1.0 micromho/cm. If trace
organics are to be determined, an
activated charcoal filter should also be
used for purification.
Water treatment equipment can be
leased from companies such as Culli-
gan; complete systems, such as those
listed in Table 7-2, can be purchased.
High-purity water should be pur-
chased in polyethylene bottles except
when organics are to be analyzed; then
the storage container should be glass.
High-purity water should be periodically
checked with a conductivity meter
independent of the inline meter. If the
laboratory conductivity meter is cali-
brated (Section 6.3), the conductivity of
the laboratory deionized water should
be checked with each set of rainwater
samples analyzed.
7.1.12 Equipment
Major equipment needs for rain-
water analysis include an analytical
balance, a pH meter, a conductivity
meter, an atomic absorption spectro-
photometer, an ion chromatograph, and
an automated colorimetric system. In
addition .the laboratory should have a
desiccator, a drying oven that heats to
200°C, an autoclave capable of main-
taining 121±0.5°C, an NBS-certified
thermometer, a thermometer which can
be calibrated against the NBS-certified
thermometer, a set of Class-S weights,
and other weights which can be
calibrated against the Class-S weights.
7.2 Analytical Reagents
7.2.1 Reagent Quality
Reagents used for analyses must
meet standards of quality denoted by
Table 7-1. Water Purity
the terms "analytical reagent grade,"
"reagent grade," and "ACS analytical
reagent grade." All of these grades are
equivalent, and they identify reagents
which conform to current specifications
of the Committee on Analytical Rea-
gents of the American Chemical Society
(4).
It may not be possible to obtain dyes of
analytical reagent grade for automated
colorimetric ammonium and phosphate.
For these, a statement of purity should
be obtained from the manufacturer and,
if necessary, the weights of dye used in
reagent preparation should be adjusted.
The lanthanum nitrate used as a flame
buffer in atomic absorption should be
"atomic absorption grade."
7.2.2 Drying of Reagents
Hydrated salts should be stored in a
desiccator overnight before weighing.
Anhydrous reagent chemicals should
be dried overnight in an oven at 105° to
110°C, allowed to cool in a desiccator,
and promptly weighed before dissolu-
tion.
7.2.3 Traceability to NBS
Traceability to NBS is usually pro-
vided by analysis of an NBS standard
reference material (SRM). As of May
1980, NBS had not provided a rainwater
SRM, so the only way to obtain NBS
traceability is to reference calibration
standards to NBS-SRM's. Table 7-3 lists
NBS SRM's applicable to rainwater.
These materials may be ordered from
NBS:
Office of Standard Reference Materials
Room B311, Chemistry Building
National Bureau of Standards
Washington, D.C. 20234
Telephone: 301/921-2045
This NBS traceability information is
provided mainly as a service. Labora-
tories are not required to trace all
calibration standards to NBS. Calibra-
tion standards and independently
Degree of Purity
Pure
Very Pure
Ultrapure
Theoretically Pure
Maximum
Conductivity
(micromho /cm)
10
1
0.1
0.055
Approximate
Concentration
of Electrolyte
(mg/l)
2-5
0.2-0.5
0.01-0.02
0.00
Table 7.2. Commercially Available Water Purification Systems
Manufacturer Reverse Osmosis System Ion Exchange System
Barnstead
Culligan
Millipore
RO pure
Milli-RO
NANO pure
Milli-Q. Super-Q
prepared "analyst spikes" should be
prepared from ACS reagent graH°
chemicals. When a new stock calif
tion standard is prepared, it should
checked against the old standard.
7.2.4 Storage of Reagents and Re-
agent Solutions
Analytical reagents have a finite shelf
life, so all chemicals received by the
laboratory should be dated by the
receipt clerk and labeled "Do not use
after . . ." Shelf lives of chemicals vary
with manufacturer, so the manufac-
turer should be consulted. Unless
otherwise specified by the manufac-
turer, inorganic chemicals have a shelf
life of 5 yr at room temperature. The
analyst should become familiar with
publications (5,6) related to preparation,
standardization, and storage of rea-
gents.
Concentrations of reagents in solu-
tion may change due to: (1) biological
action, (2) chemical reaction (e.g., oxida-
tion), (3) evaporation, and (4) adsorp-
tion-desorption phenomena on solution
container surfaces. All of these effects
can be slowed by refrigeration. Guide-
lines for reagent storage are in applic-
able analytical methods and in the
publications (5,6,7) noted above. As
discussed below (Section 7.3), pr
ethylene or Teflon bottles are
preferred storage vessels for all rea-
gents for inorganic analyses. Chemical-
ly resistant borosilicate glass such as
Kimax or Pyrex are acceptable for
storage except for reagents to be used
for alkali and alkaline earth metal
analyses.
7.3 Glassware and
Plasticware
This section discusses recommended
types of and cleaning procedures for
glassware and plasticware, and Federal
specifications for volumetric glassware.
7.3.1 Choice of Vessels for Prepara-
tion and Storage of Reagents and
Samples
Potential problems include: (1) con-
tamination of the solution by leaching of
species from container surfaces, (2) loss
of species from solution due to adsorp-
tion on container walls, and (3) break-
age of containers. The second is most
important for trace metal analysis,
which is not addressed in this docu-
ment; several plastic materials are
acceptable for.containers to be used in
trace metal analysis if the solution is
acid stabilized. To minimize the other
two problems, plastic vessels should
used when possible. Most labora
vessels are available in plastic.
The type of plastic depends on the
application. A recent article(7) dis-
-------
Jan. 1981
Part l-Section 7.0
cusses trace contamination, durability,
and recommended used and cleaning
edures for plastics. Polystyrene
„ Teflon (TFE) and conventional or
low-density polyethylene (CPE) were
found to be the cleanest plastics,
presumably due to the relatively simple
manufacturing processes. Polypro-
pylene (PP) and linear polyethylene
(LPE) were much more contaminated.
The Teflon CPE and FEP fluorocarbon
resins were recommended because
they are relatively clean chemically
resistant tough materials suitable for
field applications. Plastics such as LPE
are acceptable for major constituent
analysis of rainwater.
All glassware used in the laboratory
should be borosilicate, and should be
thoroughly cleaned as detailed in the 0
& M manual(8).
7.3.2 Volumetric Glassware
All glassware calibrated to contain
(TC) or to deliver (TD)—a precise mea-
sure of volume—must meet the NBS
specifications for Class A volumetrics(9,
10), as given in Table 7-4. The volume of
solution and the internal volume of the
glass container itself change with
temperature. The temperature (usually
20°C) at which the volumetric glass-
ware was calibrated is indicated on the
glassware; solutions should be ±5°C of
Table 7-3. NBS-SRM's Applicable to Rainwater Analysis
Measurement SRMtt Type Value
Unit
pH
pH
Conductivity.
185e Potassium Acid 4.004
Phthalate
186 Ic Potassium Dihydrogen (6.863)"
Phosphate j \
1 86 He Disodium Hydrogen (7.415)
Phosphate
999 Potassium Chloride
pH (25°C)
pH (25°C)
pH (25°C)
K". Cl
Ca+* 915 Calcium Carbonate
929 Magnesium Gluconate
i. . ft/Os 193 Potassium Nitrate
NHt\ P0<~3 194 Ammonium Dihydrogen
Phosphate
Na\Cr 919 Sodium Chloride
SOS None Available
8 Instructions are included for preparation of each pH value using different
ratios of 186 Ic and 186 lie.
Table 7-4. Tolerances for Class A Volumetric Glassware"
Type of
Glassware
Graduated flask
Transfer pipet
Capacity
(ml)
25
50
100
200
250
300
500
1000
2000
2
5
10
25
30
50
100
200
Limit of Error
(ml)
±0.03
±0.05
±0.08
±0.10
±0.11
±0.12
±0.15
±0.30
±0.50
±0.006
±0.01
±0.02
±0.025
±0.03
±0.05
±0.08
±0.10
'Abridged reference 10
Vess than and including:
the calibration temperature for accurate
volume measurements.
Disposable glassware (e.g., Pasteur
pipets; culture tubes) should not be used
because they may be of questionable
cleanliness and because sodium may
readily leach from the soft glass used in
the manufacture of such disposable
items.
Disposable plasticware (e.g., poly-
propylene test tubes) may be used, but
they should be thoroughly cleaned
before use.
7.3.3 Cleaning of Glassware and
Plasticware
Glassware and plasticware include
the borosilicate volumetrics, test tubes,
and flasks and the polyethylene con-
tainers. All should be rigorously cleaned,
segregated, and dedicated to various
analyses. Glassware for trace metals
analyses should be routinely acid
cleaned; other glassware should be
rinsed in deionized/distilled water.
Detailed cleaning procedures are in the
0 & M manual(8).
7.4 Laboratory Support for
the Field
The laboratory must prepare stan-
dards for calibrating field instruments
and for field testing the quality control
samples. Clean sample containers and
shipping materials should be supplied
weekly. This section discusses refer-
ence solutions, laboratory evaluation of
field equipment, and routine materials
supplies. Detailed procedures for prepara-
tions of solutions are in the 0 & M
manual(8).
7.4.1 Conductivity Standards
Field calibration for conductivity mea-
surement is a single point calibration
•with the standard 0.0005M KCI. Each
month, sufficient quantity is prepared
for each field site to receive 1 liter of this
dilute standard. It is recommended that
this standard be prepared in a large
polyethylene carboy dedicated to this
purpose.
The conductivity of the field standard
is established by comparison with a
calibration curve, then the Field Con-
ductivity Standard form (Section 7.8) is
filled out, and a 1-liter bottle for each
site is filled and labeled for shipment.
After 1 mo, a new standard is sent to the
field, and the old field conductivity
standard is returned to the laboratory,
and rechecked, and the data recorded
again on the Field Conductivity Stan-
dard form so that the before and after
values can be compared.
-------
Part l-Section 7.0
Jan. 1981
7.4.2 pH Electrode Reference Solu-
tion
A pH electrode may perform well in a
pH buffer, but it may be inaccurate for a
dilute sample; therefore, it is important
to check pH electrodes with -a dilute
solution. The pH electrode reference
solution is a dilute solution used to
check all laboratory and field pH
electrodes before the latter are sent to
the field. The pH electrode reference
solution should be protected from air in
a sealed container, and should be
refrigerated. The data should be re-
corded on the Field pH Electrode Test
Solution form (Section 7.8).
7.4.4 Evaluation of Equipment
All meters and electrodes are tested
in the laboratory before they are shipped
to the field. Meters usually have a serial
number affixed, but electrodes do not,
so an identification number should be
given to each electrode. Acceptance
tests are in Section 6.3 and detailed
laboratory procedures are in the 0 & M
manual(8)."
7.4.5 Supplier of Material
All sample containers sent to the field
must be washed and tested in the
laboratory for cleanliness. All sample
containers and shipping materials sent
to the laboratory from the field sites
must be replaced routinely by the
central laboratory as soon as possible
after arrival to avoid exhausting the
field's supply, thus minimizing down-
time. All other supplies required by the
field should be shipped when requested
on the data form (Sections 6.8.3 and
6.8.5; 0 & M manual, Section 2.8.3).
7.4.3 Preparation of Audit or Test
Samples
Accuracies of field conductivity and
pH measurements are evaluated with
field quality control audit samples.
Monthly, an audit sample should be
prepared using the procedures in the 0
& M manual(8). Aliquots(about60ml)of
the diluted audit sample should be sent
to each site. Three aliquots in bottles
should be kept by the laboratory; when
the samples are sent to the field, the
laboratory should determine the initial
pH and conductivity of the three
aliquots, and then the aliquots should
be refrigerated.
When the field audit samples from all
sites have been returned to the labora-
tory, the samples are reanalyzed along
with the three aliquots. The laboratory's
pH electrode is checked for accuracy
against a backup electrode..The data are
reported to the QA coordinator (Section
9.2).
7.5 Analytical Methodology
To obtain acceptable data, one must
choose an analytical procedure and
range appropriate for the sample. The
next step is to gain experience with the
instrument and its performance charac-
teristics—its precision and accuracy.
One way to gain such experience is to
document performance before analyz-
ing samples.
7.5.1 Selection of Analytical Methods
Discussed here are state-of-the-art
analytical procedures applicable for
rainwater analysis, in Table 7-5; others
may be used if similar sensitivities are
achieved.
Existing data(12) demonstrate that
procedures are not sensitive enough to
provide meaningful data for most
phosphate (PO-T3) analyses and that
results of potassium (rO, calcium(Ca**),
and magnesium (Mg++) analyses are
frequently at the detection limits so
careful analytical procedures must be
used. Recommendations for optimal
instrument sensitivity and performance
are in Sections 7.5.3 and 7.6. Apossible
alternate is to obtain greater sensitivi-
ties for potassium (K+), calcium (Ca*+),
and magnesium (Mg*+) with graphite
furnace atomic absorption.
7.5.1.1 Gravimetric Measurements—
In the field and the laboratory, the
volume of rain is determined by mea-
suring the mass and by assuming a
density of 1 gm/cm3. Mass of rain is
measured in the field to compare
collector efficiency with the rain gauge;
the sample mass sent to the laboratory
is checked to determine if leakage
occurred in shipment.
In addition to a 20 kg capacity balance
for weighing rain buckets, the analytical
laboratory should have an analytical
balance which achieves the precision
necessary for the preparation of rea-
gents and salts for calibration stan-
dards. Balances should be in a tempera-
ture-controlled room free from drafts
and on a rigid table to minimize vibra-
tions. The legs of the balance should be
adjusted so that it is level. Electronic
balances should be electrically ground'"'
Before weighing, the balance p;
brushed off with a soft brush, and ti....
the balance zero is set. After the
weighing, the balance should be cleaned
of all (potentially corrosive) chemicals. If
analytical balances are not in use, the
beam should be raised above the knife
edges.
Analytical balances should be cali-
brated daily using either NBS Class-S
weights or weights traceable to NBS
Class-S; NBS Class-S weights should
be stored as primary references in the
laboratory, and daily calibration weights
should be certified against these; and
the Certification of Weights to NBS form
(Section 7.8) should be completed. All
certified weights should be recertified
every 6 mo.
The procedure used to certify weights
should be repeated five times: the
balance should be zeroed according to
manufacturer's recommendations, the
NBS-certified 1.0- and 5.0-g weights
should measured, and the test 1.0- and
5.0-g weights should be weighed. Each
balance should be calibrated using
weights close to those actually mea-
sured. High-capacity balances should
1-3 calibrated similarly, with 1.0-and
5.0-kg weights. The Balance Calibre
Log (Section 7.8) should be compk
each time the balance is calibrated.
The 1.0- and 5.0-kg weights should
be shipped to the field in wooden
storage boxes suitable to protect the
weights when they are not being used.
Explicit instructions on the care of
calibration weights should be given by
the QA coordinator or field manager.
Balances should be maintained peri-
odically according to manufacturer's
recommendations. For the laboratory
analytical balances, factory mainten-
ance is usually once a year. A record of
maintenance should be kept for each
balance; (complete the Balance Factory
Service form (Section 7.8)).
7.5.1.2 pH Measurement—Labora-
tory pH measurement is a check on field
measurement and sample degradation.
Table 7-5. Procedures Recommended for Rainwater Major Constituent
Analysis
Analysis Instrumental Method
Volume (weight)
pH
Conductivity
Strong acid
Total acid
Chloride, phosphate,
nitrate & sulfate
Ammonium, phosphate
Sodium, potassium,
calcium, magnesium
Analytical balance
pH meter
Conductivity meter
Gran plot using pH meter
Potentiometric titration with pH meter
Ion chromatography (preferably automated)
Technicon automated co/orimetry
Atomic absorption spectrophotometry (flame)
-------
Jan. 1981
Part l-Section 7.0
The procedure is given in the 0 & M
ie pH meter should be calibrated
..ore and after each measurement or
each series of 20 measurements. If
initial and final calibrations have
changed more than 0.02 unit, the mea-
surements should be repeated; if
changes reoccur, a problem with the
apparatus should be remedied.
Each day the meter is used, it should
be calibrated, and real-time quality
control procedures (Section 7.6) should
be applied. The quality control sample
(Section 7.6) should be included with
each set of samples analyzed.
Samples are measured directly after
the meter has been calibrated and the
electrode has been washed. Neither the
electrode nor any other object should be
inserted into any of the bulk solutions,
and no solution should ever be poured
back into a bulk container. For measure-
ment, use a small vessel (vial or test
tube) rinsed first with distilled/deion-
ized water and then (if sufficient solu-
tion permits) with sample.
The working pH electrode should be
checked weekly; if performance is
unacceptable, replace it with a new
tested electrode.
".1.3 Strong Acid (by the Gran
iod)—Data from a recent study(11)
...uicate that 95% of rainwater samples
from the eastern United States contain
strong acid at a concentration at 1.6 x
10~"N or less (pH greater than 3.8) and
that the median rainwater concentra-
tion of strong acid is 4.6 x 10~5N (pH of
4.3). Strong acid is determined using
the Gran(12) plot, based on the Nerns-
tian response of the pH electrode. The
procedure is in the 0 & M manual(8).
At setup time, conditioning solution
intercepts and spike recovery data
should be evaluated. Real-time evalua-
tion of this information along with
quality control samples to be analyzed
are discussed in Section 7.6.
Key performance indicators for Gran
titration are the solution temperature,
the conditioning solution's initial po-
tential, time for potential stabilization
after each addition of base, the correla-
tion coefficient of the linear least
squares fit of a function (Jj vs volume V of
base added, and the spike recovery. All
potentials should be measured at the
same temperature. During the initial
setup and between analyses, the titra-
tion vessel should be rinsed with condi-
tioning solution; 7.0 ml of conditioning
solution should be pipetted into the
sel; and the potentials of the two
Jtions should be measured. The
rinsing and the 7.0 ml addition should
be repeated until the measured potentials
of two solutions agree to ±1.2 mv. If this
cannot be achieved, the temperature is
probably not stable or there is an
electrode drift problem. If the potential
readings are not stable within 30 s, the
electrode drift is excessive, so the meter
and pH electrode should be reevaluated
(Section 6.3).
7.5.1.4 Acidity—Acidity is deter-
mined potentiometrically by titrating to
a pH of 8.3. The procedure is in the 0 &
M manual(8). Care must be taken during
the titration to C02-free water and to
protect the solution from air. The
temperature is crucial because pH is
temperature dependent. The meter and
pH electrode should be evaluated (Sec-
tion 6.3).
Spikes should be analyzed each day
before and after the samples are
analyzed, and the data evaluated
immediately. The real-time evaluation
of this information along with the
quality control samples to be analyzed
are discussed in Section 7.6.
7.5.1.5 Conductivity—Conductivity
measurements are made both in the
field and in the laboratory. Laboratory
measurements serve to check for
sample degradation and to evaluate
field measurements. The procedures for
calibration of the apparatus and deter-
mination of conductivity are in the 0 &
M manual(8).
The conductivity apparatus is cali-
brated using KCI solutions of known
conductivity before and after each mea-
surement or series of measurements. If
a change of more than 5% occurs,
repeat the measurement; if drift reoc-
curs, a problem with the apparatus must
be corrected. In general, stable values
occur in 30 s.
If conductivity of the sample is to be
measured on the aliquot poured for pH,
the conductivity must be measured
before the pH. The conductivity cell
should be washed after calibration—
first with distilled/deionized water and
then with a rinse of sample, using the
same two vials or test tubes of sample to
be used for the pH measurement
(discussed above in Section 7.5.1.2).
Dip the conductivity cell three times in
the rinse test tube and three times in the
measurement solution, and then take
the reading; between measurements,
rinse the cell thoroughly with deionized
water, and shake off excess water.
Store the cell as recommended by the
manufacturer.
Real-time quality control procedures
as well as performance evaluations
using the results of the quality control
samples are in Section 7.6. New
conductivity cells should be checked
upon receipt, using the conductivity cell
acceptance tests in Section 6.3.
7.5.1.6 Automated Colorimetric Mea-
surements— Ammonium (NH/) and
orthophosphate (PCU~3) ions are mea-
sured using automated colorimetric
procedures. Data from a recent study
(11) indicated that 95% of rainwater
samples collected in the eastern United
States contain phosphate concentra-
tions less than 0.04 microgram/ml, and
95% contain ammonium concentra-
tions of 0.05 to 1.6 microgram/ml. The
median concentration was 0.28 micro-
gram/ml for ammonium and 0.008
microgram/ml for phosphate. Am-
monium concentrations are generally at
or below the detection limit of the auto-
mated colorimetric procedure.
The automated ammonium analysis
(13) uses the Berthelot reaction—a blue
compound is formed after addition of an
ammonium salt and sodium phenoxide
to sodium hypochlorite. The ammonium
concentration is determined spectropho-
tometrically at 630 nm. The automated
phosphate analysis (14) uses a phospho-
molybdenum blue complex formed and
measured spectrophotometrically at 880
nm. Both procedures are in the 0 & M
manual (8).
Key performance indicators for auto-
mated colorimetric measurements are
baseline noise, calibration standard
response, and calibration curve linear-
ity. After instrument performance is
documented, baseline noise and cali-
bration standard response are mea-
sured (scale expansion) to identify the
most sensitive analysis and to be used
as daily guides for evaluating perform-
ance.
Real-time quality control procedures
as well as information on quality control
samples to be analyzed are in Section
7.6.
7.5.1.7 Ion Chromatopraphic Mea-
surements—Chloride (Cl~), nitrate
(N03~), and sulfate (S0<=) anions are
analyzed by ion chromatography. A
recent study(11) showed that 90% of
rainwater samples from the eastern
United States contain CI" concentra-
tions from 0.02 to 2.0 microgram/ml,
99% contain N03" in the range of 0.1 to
10.0 microgram/ml, and 96% contain
S04= in the range of 0.2 to 10.0
microgram/ml. The median concentra-
tions are 0.26 microgram/ml for CI",
1.44 microgram/ml for N03", and 2.39
microgram/ml for S04=. The procedures
in the 0 & M manual(8) recommend
instrument setup to achieve maximum
sensitivity for these analyses.
Key performance indicators for the
ion chromatograph are column back-
pressure, resolution, and baseline noise
and drift. After instrument performance
is documented, baseline noise and
calibration standard response are
-------
Parti-Section 7.0
Jan. 1981
defined (scale expansion) to indicate the
most sensitive analysis and to serve as
guides to evaluate daily performance.
As part of real-time QC, the first calibra-
tion curve linear least squares fit should
be determined, and the analyst spike
should be calculated and compared with
the known value. After the analytical
data are measured, performance is
evaluated using results of QC samples
and curve parameters (Section 7.6).
7.5.1.8 Atomic Absorption Measure-
ments—Atomic absorption is used to
determine sodium (Na+), potassium (K+),
calcium (Ca*+), magnesium (Mg**) ca-
tions. Recent data(11) show that 96% of
rainwater samples from the eastern
United States contain Na* in the range
of 0.02 to 1.0 microgram/ml; 94%
contain K+at less than0.25 microgram/
ml; 68% contain Ca** at less than 0.25
microgram/ml, and 29% contain Ca+* in
range of 0.25 to 1.25 microgram/ml;
and 66% contain Mg+* at less than 0.04
microgram/ml, and 30% contain Mg*+
in the range of 0.04 toO.16 microgram/
ml. The median concentrations are 0.25
microgram/ml for Na*, 0.06 micro-
gram/ml for K+, 0.13 microgram/ml for
Ca*+, and 0.03 microgram/ml for Mg+*.
These low concentrations are detection
limit analyses for K*, Ca+t, and Mg++, so
very sensitive procedures must be used.
Procedures for the analyses are in the 0
& M manual(8), with recommendations
for maximum atomic absorption sensi-
tivity. Since the detection limit is
determined to a large extent by light
source noise, the most stable light
source available should be used.
Electrodeless discharge lamps (EDL's)
are more intense than the conventional
hollow cathodes; an EDL is not available
for K* analysis.
Key performance indicators for atom-
ic absorption are the baseline noise, the
calibration standard response, and the
curve linearity and reproducibility. After
instrument performance is documented,
baseline noise and calibration standard
response are defined (scale expansion)
to give the most sensitive analysis as
guides to evaluate daily performance.
Real-time QC control procedures as
well as information on analysis of QC
samples are in Section 7.6.
7.5.2 Choice of Analytical Ranges
For pH and conductivity measure-
ments, one specifies the instrument
precision. Instruments such as the ion
chromatograph, the Technicon auto-
mated colorimetric analyzer, and atomic
absorption analyzers have wide analyt-
ical ranges. Because very dilute sam-
ples are analyzed, the lowest range
should be the instrument's most sensi-
tive range; higher ranges are also used
so that a high off-scale response on the
lowest range can be measured to better
than ±10% on the higher range.
7.5.3 Documentation of Instrumental
Performance
Instrument performance should be
studied before rainwater samples are
analyzed to give the analyst experience,
to provide useful QC information, and to
evaluate laboratory analytical tech-
niques used at the concentrations of
rainwater samples.
The instrument performance study
for each constituent to be determined
should consist of repetitive analyses of
calibration standards at different con-
centrations. For pH, the Gran strong
acid, the acidity and conductivity, and
the instrument performance proce-
dures are different from those for the
other measurements. For pH, strong/
total acid and conductivity, solutions of
10"3, 10~". 10"5, 10"6N HCI are prepared
in COa-free water, and carefully pro-
tected from air until they are analyzed;
five aliquots of each of these solutions
are poured, and five aliquots of the pH
electrode reference solution are poured.
The meter is calibrated for each mea-
surement, as indicated by the specific
procedure and the samples analyzed.
The samples should be analyzed in the
following order—electrode reference
solution, 10"6N HCI, 10"5N HCI, 10"4N
HCI, 10"3N HCI—repeated for five
sample aliquots. For each analysis, the
pH, Conductivity, and Gran Strong Acid
Instrument Performance form (Section
7.8) is completed.
For all other analytes, the instrument
is set up for the most sensitive analysis,
and the calibration standards contain-
ing concentrations to yield responses of
5%, 10%, 20%, 40%, 60% and 80% of
full scale are analyzed. A synthetic
rainwater sample (Section 7.6) con-
taining concentrations close to the
median rainwater concentrations should
also be analyzed. The six-point calibra-
tion curve and the synthetic rainwater
sample should be analyzed consecu-
tively five times. A linear least squares
fit of expected concentration versus
response is calculated for the calibra-
tion standards; the curve parameters
from this fit are used to calculate
regression concentrations of each data
point. The average or standard deviation
of the regression concentrations should
be calculated at each concentration,
and the percentage relative standard
deviation and the percentage accuracy
should be compared to the known
concentration.
Data for each calibration curve should
be considered separately to determine
detection limits. A linear least squares
fit of each calibration curve should be
calculated; one of the curve's param-
eters which can be calculated from this
fit is the Hubaux and Vos( 15) detection
limit (Appendix A), which is mi
higher than the one calculated as ,
times the baseline noise; thus it couia
be considered the lowest level for which
a precise analysis can be done. After a
sufficient number of values are avail-
able, an upper control limit can be used
to reject poor calibration curves. Until
control limits are established using
range chart techniques (Section 8.0),
the average detection limit can be taken
as the control limit for evaluation of
daily instrument performance.
All data from the instrument per-
formance study are tabulated and
plotted as QC data on the Precision-
Accuracy Instrument Performance form
and its corresponding plot (Section 7.8),
and the key data should be summarized
on the Instrument Performance Sum-
mary form.
The strip charts, the reduced data and
the Precision-Accuracy forms from this
study are filed to be readily available to
the analysts and the laboratory director.
If, at any time, degradation of instru-
ment performance is suspected, then
the original study can be reviewed to
evaluate instrument performance. An
example of a instrument performance
study is in Appendix B.
7.6 Quality Control Progiv
for Chemical Analysis of
Precipitation Samples
When analytical data are reported, it
is essential to specify the quality. State-
ments about quality should refer to the
particular data set being reported, not to
laboratory analyses in general, so it is
necessary to implement a formal QC
program which indicates the accuracy
and precision of each data point
reported.
Both an internal and an external QC
program should be implemented. In-
ternal QC includes calibration and real-
time control by the analyst, analysis of
special QC samples by the analyst,
review of the data by the laboratory
supervisor and QA coordinator, sched-
uled data checking for transcription
errors by data processing personnel,
and a final review of all QC data by the
QA coordinator before reporting. The
external QA includes analysis of blind
samples received from an agency
external to-the central laboratory.
This section specifies QC samples to
be analyzed, and discusses responsibil-
ities for evaluation of the QC <"
Control charts for both analyst re
and managerial review are stress>x,-..
Procedures require real-time review of
analytical performance by the analyst.
-------
Jan. 1981
Part l-Section 7.0
and QC review of all data directly after
;"'it to the computer.
1 Sample Handling in the Labora-
tory
7.6.1.1 Sample Logistics—All samples
received by the laboratory should be
checked in by a receiving clerk who
records the site, date, and other
identification; checks the field data form
against sample labels to identify dis-
crepancies; assigns a laboratory identi-
fication number to the sample; records
the number and the date of arrival on
the data form and in the logbook;
examines the data form and the sample
for certain conditions; and codes the
information on the data form. The
codes, which are basically footnotes
that may be useful later in interpreting
the data, should be stored with the
sample data in the computer (Section
8.0). Table 7-6 suggests information to
be coded. After logging in the samples,
the receiving clerk should refrigerate
them as soon as possible. Data forms
received with the samples are kept by
the data clerk.
The receiving clerk should replace the
old sample bucket or containers with
clean, sealed, bagged ones, and should
ship the new ones in shipping cartons
other required materials to the field
, as soon as possible. These can be
sent by ground transport since each site
should have a 3-week supply on hand.
As soon as possible after receipt of
samples, aliquots should be poured into
clean 35 ml vials or 10 ml plastic test
tubes for pH and conductivity analyses,
and the vessels should be sealed with
paraffin film. Then the remainder of the
sample should be filtered, as indicated
below (Section 7.6.1.2), and the filtrate
should be sealed in a labeled plastic
Table 7-6. Sample Information to
be Coded
Snow/ice
Mixed: snow/rain: hail/rain
Sample contaminated
Possible sample leakage in shipping
Sampler inoperative — no sample
Insufficient sample for complete
measurement
Rain gauge inoperative
Noticeable suspended particu/ates
Lid cycling
Field pH and conductivity measured x
days after scheduled sample removal
or end of event
pH/'conductivity/temperature meter
inoperative
tie partially frozen
iual condition in area
Duplicate samples
Sequential samples
bottle of appropriate size and refriger-
ated until it is analyzed. Each analyst
should take an aliquot for analysis, and
return the sample to the refrigerator as
soon as possible. If a volume of a sample
is low, the analysis priorities should be
in the following order: pH, conductivity,
SQ4=. NQ3~. CI", NH/, Na*, K*, Ca*2,
Mg*2, PCV3, and acidity. If there is
adequate sample after pH and conduc-
tivity are determined, the sample should
be analyzed for NH/ before the other
constituents.
After all analyses have been com-
pleted and the results checked, the
sample may be transfered into a 125 ml
polyethylene bottle for storage in a
refrigerator or freezer for 6 mo to 1 yr for
other tests or analyses. The best method
for long-term storage may be to freeze
the samples or to keep them at 4°C.
Stability tests for several months
indicate that both the 4°C(16,17) and
the freezef 17) methods are satisfactory.
7.6.1.2 Filtering the Samples—Par-
ticulates in rainwater can contaminate
the sample over a period of time(18);
thus it is necessary to remove them by
filtration as soon as possible after the
sample is received. All parameters,
except pH and conductivity, should be
determined on the filtered sample. A
procedure for sample filtering is in the 0
& M manual(8).
7.6.1.3 Scheduling the Analyses—
Once a week, a list of all received
samples for which analysis results have
not been reported should be prepared
from the receiving clerk's logbook, and
copies given to laboratory personnel to
alert everyone to the analyses that are
to be done and to the data yet to be
reported.
7.6.2 Laboratory Documentation
The following documents should be
available to the analyst and the super-
visor, and should be constantly updated.
1. Laboratory Standard Operational
Procedure - detailed instructions
on laboratory and instrument
operations.
2. The Laboratory Quality Assurance
Plan • clearly defined laboratory
QA protocol, including personnel
responsibilities and use of QC
samples.
3. List of In-House Samples - dates
for completion of analysis to allow
the analyst to schedule further
analyses.
4. Instrument Performance Study
Information - information on base-
line noise, calibration standard
response, precision as a function
of concentration, and detection
limits used by analyst and super-
visor to evaluate daily instrument
performance (Section 7.5.3).
5. Quality Control Charts - most
recent QC charts with control
limits for all calibration curve QC
parameters and for all QC samples;
once a month, update all control
limits to include data from analy-
ses of the previous month; gener-
ate plots of all QC samples and
curve parameters.
6. Data Sheet Quality Control Report
- Generate a QC report after data
for each analysis are input to the
computer (preferably within 1 day
of analysis); present information
for all QC parameters, flag all data
which exceed the statistically
established QC limits; have the
supervisor review this report to
decide what is to be done for out-
of-control samples; if necessary,
the supervisor will schedule repeat
analysis of samples.
7. The Analyst's Spike Plot - daily
when the analysis is set up, the
first sample analyzed should be the
analyst spike; percentage recovery
for this sample should be calcu-
lated and plotted by the analyst in
real time (Section 7.6.5).
7.6.3 Traceability of Calibration
Standards
7.6.3.1 NBS Traceability —As com-
mon reference points for laboratories,
NBS traceabilities of chemical gravi-
metric, and thermometric calibration
standards are desirable. As of May
1980, NBS had not provided a rainwater
SRM; however, preparation of an
analyst spike or an audit spike (Section
7.6.4) from NBS-SRM's is not required,
but is recommended.
For chemical traceability all calibra-
tion standards must be prepared from
ASC reagent grade salts, and the
accuracy of calibration standard prep-
arations must be checked. With proce-
dures proposed here, accuracy is
checked by running an independently
prepared analyst spike (Section 7.6.4)
with each analysis and by checking
each new set of stock standards against
the old.
For gravimetric measurements, NBS
traceability is provided by daily balance
checks with weights traceable to NBS-
certified weights, so each laboratory
should maintain a set of NBS-certified
weights (Section 7.5.1.1). The recom-
mended procedure is to purchase
weights traceable to NBS from a
commercial supplier and to have them
certified by an NBS-approved labora-
tory; it is much more expensive to have
NBS calibrate a set of weights directly.
-------
Part l-Section 7.0
Jan. 1981
For temperature measurements, NBS
traceability is provided by thermometers
calibrated against an NBS-calibrated
thermometer; so each laboratory should
maintain an NBS-calibrated thermom-
eter, and should calibrate thermom-
eters for daily use against the NBS
thermometer (Section 6.3). It is time
consuming but not expensive to have
NBS calibrate a thermometer; NBS will
calibrate at $30.00/point (calibration
needed at two points only, 0°C and
25°C); and NBS will visually inspect for
flaws and reject if flaws are found. For
NBS calibration, thermometers should
be shipped (not mailed) to:
National Bureau of Standards
Route 270
Quince Orchard Road
Gaithersburg, MD 20760
7.6.3.2 Reference Water Samples—
The QC chemist may dilute and mix
reference water samples to prepare
synthetic rainwater samples for in-
dependent internal QC for an audit spike
sample. The two sources of reference
water/wastewater samples are(19):
1. EPA Cincinnati - (Quality Assur-
ance Branch, EMSL-Cincinnati
EPA, Cincinnati, Ohio 45268,
51 3/684-7327.) QC water/waste-
water samples without charge to
be used as "secondary checks . . .
within laboratory quality control"
prepared solely for internal QC
samples; data not reported to EPA.
Samples prepared from ACS rea-
gent grade chemicals are sent as
concentrates in sealed-glass am-
pules; when diluted according to
instructions, should give thecalcu-
lated EPA values which are sent in
a separate envelope. Available
samples for rainwater analyses:
MINERAL/PHYSICAL ANALYSES
Na+, K+. Ca*+, Mg+\ pH. S0«=, Cf,
F", alkalinity/acidity, total hard-
ness, total dissolved solids, and
specific conductance, two concen-
trations
NUTRIENTS
nitrate-N, ammonia-N, Kjeldahl-N,
orthophosphate, and total phos-
phorus, two concentrations
2. Environmental Resource Associ-
ates (ERA/ - (120 East Sauk Trail,
Suite 150, South Chicago Heights,
Illinois 60411. 312/755-6060.)
Commercially prepared reference
water/wastewater samples of
known composition from reagent
grade salts; selected samples from
the batch analyzed by ERA and by
three independent laboratories.
7.6.4 Preparation of Analyst's Spikes
and QC Audit Spikes
7.6.4.1 Analyst's Spike—When pre-
paring calibration standards, the analyst
should prepare an analyst's spike from a
different stock solution; concentration
of the spike should be approximately at
the midpoint of the calibration curve;
however, if the majority of samples have
concentrations below the midstandard,
the spike should be prepared within that
range.
7.6.4.2 Quality Control Audit Spikes—
For management of laboratory analyses,
the QC chemist should prepare a set of
audit spike samples once a month;
these are simulated rainwater samples
containing all of the major constituents
of rainwater. Table 7-7 is a suggested
scheme for the preparation of an audit
spikS sample. In this scheme, stock
standards are prepared by weighing
reagent grade chemicals, and diluting
them to 1 liter. Each stock solution
undergoes a preliminary dilution, dilu-
tion A; each aliquot of dilution A is
diluted to 1 liter (in a volumetric flask) to
obtain dilution B. The resulting rain-
water concentrations (last column.
Table 7-7), are approximately the
medians obtained in a recent study of
Eastern United States precipitation
samples(11). Each month, the QC may
vary the final rainwater concentrations
of dilution B by varying the volumes of
dilution A delivered. Dilution A should
be kept refrigerated, especially if nitrate
and ammonium are present.
If a QC chemist weighs several
reagents into the same volumetric flask
for the stock standard, the amount of
pipetting would be decreased, but the
possible ratios of the mixed species
would be limited. The level of pipetting
could be reduced by mixing several
species at the dilution A level without
loss of possible ratios of the mixed
species.
7.6.5 Real-Time Quality Control Pro-
cedures
Real-time QC procedures, which
stress analyst evaluation of the calibra-
tion curve during analysis and plotting
of one QC data point when each
analysis is set up, are designed to spot
problems during the analysis so that
corrections can be made immediately.
QC data obtained by analysis of special
QC samples (Section 7.6.6) are not
evaluated in real time. Data from
analyses with the automated colori-
meter, ion chromatograph, and atomic
absorption spectrophotometer yield
calibration curve correlation coeffi-
cients (linearities) and detection limits
which should be evaluated in real time.
The control limits for detection limits
and correlation coefficients are to be
statistically established (Section '
but until sufficient data are availab
calculate these limits, the limits estab-
lished in the instrument performance
study may be used (Section 7.5.3).
7.6.5. / Real-Time Plotting of Analyst
Spike Data—After each instrument is
calibrated, the analyst should immedi-
ately run an analyst spike (Section
7.6.4) as the first QC sample to ensure
that calibration standards were cor-
rectly prepared and that no degradation
of the standards has occurred. After the
analyst spike sample has been run, its
value should be calculated by using the
first calibration curve of the day. The
percentage recovery should be calcu-
lated and plotted as indicated in Figure
7-1. (For pH and conductivity, the
absolute magnitude, not the percentage
recovery, should be plotted.) The
horizontal average recovery and control
limit lines (Figure 7-1) are those
calculated in the most recent monthly
QC report. If an out-of-limits data point
is noted, an explanation should be
sought. If eight successive values fall on
one side of the average line, the
indicated bias should be evaluated.
7.6.5.2 Balance—The analytical r-
ance should be calibrated daily agt
Class-S weights, and the Balance
Calibration Log (Section 6.8.5) should
be completed. The balance should be
zeroed before each use.
7.6.5.3 pH Measurement—The pH
meter should be calibrated as indicated
in the O & M manual(8). After the first
calibration, the first sample analyzed
should be the pH electrode reference
solution. The analyst should plot and
evaluate the pH value (Figure 7-1).
Backup electrodes should always be in
the laboratory to check the first elec-
trode(s) if the reading differs from the
previous one by more than ±0.03 unit.
The pH calibration drift should be
evaluated after 20 samples are analyzed;
if the drift is more than ±0.02 pH units,
the analysis should be stopped, and the
meter and electrodes should be checked.
7.6.5.4 Strong Acid and Acidity
Measurements—For strong acid deter-
mination, each day when sample
measurements are begun, three condi-
tioning solutions and an analyst spike
should be measured and calculated
using the linear least squares fit, as
described in the procedure in the O & M
manual(8). If correlation coefficients
from the calculation are less f"
0.9990, the indicated problem shi
be eliminated. The value of Ve (tne
equivalent volume of base added) for
each conditioning solution and the
-------
Jan. 1981
Part l-Section 7.0
Table 7-7. Preparation of a Synthetic Rainwater Sample
Stock"
Dilution A
Primary
Salt
NaCI
KH2PO*
KN03
CaSO* • 2H20
MgSO* • 7H20
NH*N03
NaN03
HzSO*(0.1Nf
Weight g
0.1648
0.0904
0.2586
0. 1074
0.2536
0.4437
0.3697
—
Species
Cl
PO*
K
Ca
Mg
NH*
Na
H
fjg/ml
100.0
25.0 J
100.0
25.00
25.00
100.00
100.0
101.0
Secondary
Species
Na
K
N03
SO*
SO*
N03
N03
SO*
fig/ml
64.83
10.27
58.80
59.92
98.83
343.8
269.7
9606.0
ML"
20
4
10
—
12
20
10
10
Primary
Species
Cl
PO*
K
—
Mg
NH*
Na
H
Ijg/ml
20.00
1.000
10.00
—
3.000
20.00
10.00
10.10
Secondary
Species
Na
K
NO3
—
SO*
N03
N03
SO,
fjg/ml
12.97
0.4108
5.880
—
11.86
68.76
26.97
480.3
Dilution B°
Salt
NaCI
KH2PO*
KN03
CaSO* • 2H20
MgSO* • 7H20
NH*NO3
NaN03
H,SO*/0. INf
Vol. Oil. A
(ml)
13
8
6
5C
10
14
10
5
Primary
Species vg/ml
Cl
PO*
K
Ca
Mg
NH*
Na
H
0.260
0.008
0.060
0.125
0.030
0.280
0.100
0.0505
Secondary
Species fjg/ml
Na
K
NO3
SO*
SO*
NO3
NO3
SO*
0.169
0.004
0.035
0.300
0.119
0.963
0.270
2.402
Final Ion Concentrations
"Dilution B"
Species Cone (^/ml)
H
Cl
PO*
N03
SO*
Na
K
Ca
NH*
0.050
.260
.008
1.27
2.82
0.269
0.064
.125
.280
"Concentrations are based on a final dilution volume of 1 liter
"ML gives the volume to take to dilute to 100 ml
c Volume (ml/ of stock
"Either standardized in the laboratory or purchased from a commercial supplier
analyst spike percentage recovery
ild be plotted and obtained as
Hed in the 0 & M manual(8). At the
of the day, an analyst spike and a
conditioning solution sample should be
analyzed. The initial conditioning solu-
tion potential for each sample should be
within 1.2 mv of the potential for the
conditioning solution. According to the
0 & M manual(8), an analyst spike
should be analyzed before and after
samples are determined and these
values should be plotted daily (Section
7.6.5.1).
7.6.5.5 Conductivity Measurement-
Each day, the conductivity apparatus
should be calibrated before and after
samples are analyzed. The first sample
should be an aliquot of a 3x10~4M KCIor
the pH electrode reference solution. The
analyst should calculate the conductiv-
ity using the first calibration curve, and
then plot and evaluate the value of the
conductivity of this reference sample
(Figure 7-1).
7.6.5.6 Automated Colorimetric Anal-
ysis (Technicon)—This instrument should
be set up, and the baseline noise and
instrument response should be evalu-
ated by comparison with data from the
instrument performance study. Any
problem noted should be investigated.
:nstrument should be calibrated as
ibed in the 0 & M manual(8). For
real-time QC, the first calibration curve
should be checked for linear response
and adequate detection limit by using a
least squares fit of the first calibration
curve and by calculating a detection
limit. Linearity should not be less than
0.9990; the detection limit should be
within the statistically established
control limits.
The first sample analyzed should be
the analyst spike. Concentrations of
this sample should be calculated from
the first calibration curve, and the
value obtained should be plotted and
evaluated (Figure 7-1). In addition, the
calibration response during analysis,
should be checked to see that it is
changing less than 5% from one
calibration curve to the next; if a greater
change is noted, the analysis should be
stopped, and an explanation sought.
7.6.5.7 Ion Chromatographic Analy-
sis (Dionex)—Calibration procedures
are in the 0 & M manual(8). For real-
time QC the baseline noise and the
response of the first standard at setup
time should be monitored. The first
sample analyzed should be the analyst
spike; it should be calculated from the
first calibration curve, and the value
should be plotted and evaluated (Figure
7-1).
7.6.5.8 Atomic Absorption Analy-
sis—Atomic absorption calibration
procedures are in the 0 & M manual(8).
For real-time QC, the first calibration
curve should be analyzed, and the linear
least square fit of response vs concen-
tration should be calculated. The
correlation coefficient should beO.9995
or greater, and the detection limit
should be within the statistically
established limits. The first sample
analyzed should be the analyst spike; it
should be plotted in real time.
7.6.6 Analysis and Evaluation of
Quality Control Samples
Each day (each analytical run) or at
least after every 50 samples, a reagent
blank, an old sample, an analyst spike,
and an audit spike should be analyzed.
For most procedures discussed (except
the ion Chromatographic, the Gran
strong acid and the acidity analyses), an
analytical run may include more than 50
samples so QC is about 8%. For ion
Chromatographic analysis, one analyt-
ical run may include 20 samples; 4 QC
samples and 20 samples make the QC
16.7%. The Gran strong acid procedure
is relatively slow, so fewer than 25
samples/day can be determined; recom-
mended are only two analyst spikes and
one audit spike as QC samples for a QC
of 1 2%.
7.6.6.1 Reagent Blank—This deion-
ized water QC sample, which is sub-
jected to the same preparation procedure
as the routine samples being analyzed,
should be analyzed to check for random
contamination which may have oc-
curred in sample preparation or analysis.
7.6.6.2 Old Sample—This randomly
chosen, previously analyzed QC sample
(if no sample degradation has occurred)
-------
Part l-Section 7.0
10
Jan. 1981
700
Percent
Recovery
90
80
Upper Control Limit
Average % Recovery Monthly QC
Report Mar 80
Lower Control Limit
1 2 3
6 7 8 9 10 11 12 13 14 15
Order ol Analysis
Figure 7-1. A nalyst spike plot for S04°analysis spike stock solution prepared
4 Jan 1980; expected cone. 2.00 fjg/ml.
provides information on analytical
precision for different days of analysis; it
may provide information on sample
stability, but this is not its primary
purpose. Sample degradation can be
corroborated by a repeat measurement
at another time and the value in the QC
data noted accordingly.
7.6.6.3 Analyst Spike—This QC sam-
ple prepared by the analyst from a stock
solution independent of that used to
prepare the calibration standards
provides information on the accuracy of
the calibration standard and the preci-
sion of analysis. The analyst spike
should be analyzed at the beginning of
the run; results should be calculated
and plotted in real time (Section 7.6.5).
7.6.6.4 Audit Spike—This QC sample
is prepared by the quality control
chemist (Section 7.6.8) and it is ana-
lyzed as an unknown (blind) by the
analyst. The data are reported to the QC
chemist who calculates percentage
recovery and reports the value to the
analyst and the laboratory director. The
purpose of this QC sample is to assess
data quality independently of analyst
judgment.
7.6.7 Evaluation of QC Data
7.6.7.1 QC Data Handling—If a com-
puter is available, data for QC samples
are calculated as they are being input
and compared to control limits estab-
lished in the most recent monthly QC
chart (Section 7.6.2). After input, the
Data Sheet QC Report (Section 7.6.2) is
printed by the computer. If no computer
is available, this procedure can be done
manually. The report on the perform-
ance of the QC samples should be given
to the laboratory supervisor for evalua-
tion (Section 7.6.7.2) to see if the data
are acceptable for reporting or if re-
analysis is necessary.
When the data form is input to the
computer, the audit spike data are also
calculated; these spike data should be
given to the QC chemist, who tabulates
the data and calculates the percentage
recoveries. The QC chemist routinely
gives the audit spike data to the
laboratory director, who reviews this
independent audit data before reporting
the analytical data to the project
manager. Once a month, the QC
chemist performs a QC audit to review
the audit data with the analyst.
Once a month, all QC sample data and
calibration curve parameters obtained
during the month are combined with all
previous data for the same parameters;
the data are plotted, and the new control
limits are calculated. A copy of the
monthly QC plot should be given to the
laboratory supervisor, the QC chemist,
and each analyst.
7.6.7.2 Quality Control Reports and
Guidelines for Initiating Corrective
Action—After input of data to the
computer, the Data Form Quality
Control Report (Section 7.6.2) is given to
the laboratory supervisor; this report
flags any out-of-limits conditions for the
analysis. The analyst or the supervisor
should check to see that any out-of-
limits conditions are not results of
transcription errors from stripchart to
data form or from data form to computer;
if a transcription error is found, the
transcription of all data on the data form
should be checked, a corrected data
form should be input, and a new QC
report issued. The problem should be
carefully documented and the old r
form and QC report should be filed v.
the new. The new QC should replace the
old in the computer QC data base. If the
out-of-limits conditions are not due to
transcription errors, another explana-
tion should be sought.
The most critical parameters are the
calibration curve parameters, because
any problem with them directly affects
the data. If an explanation cannot be
found for out-of-limits calibration
parameters, all samples analyzed be-
tween the questionable calibration
curves should be reanalyzed.
If only one of several QC samples is
out-of-limits, an explanation should be
sought but if one cannot be found, no
action is needed. The supervisor may
assume that the problem was with the
particular QC sample itself, but may
retain the out-of-limits data in the QC
data base.
If several QC samples are out-of-
limits but an explanation is not found, all
samples analyzed with the QC samples
(between bracketing calibration curves)
should be reanalyzed.
In any case, an out-of-limits QC
sample requires evaluation and an
explanation by the supervisor. T'
explanation may be notes on the
report. If samples are to be reanalyze^,
the supervisor should note this on the
data form and in the QC report. At the
time of the reanalysis, it should be noted
on the data form that the samples are
being reanalyzed, and the date of the
first and the repeat analyses should be
given.
The supervisor also evaluates the QC
plots when they are periodically gener-
ated. Since the supervisor has already
evaluated out-of-limits conditions in the
QC reports, all out-of-limits conditions
should have been explained or elimi-
nated (by reanalysis) before the monthly
plots are generated. Accordingly, the
monthly plots should be examined
primarily for systematic bias. If no
systematic bias is present, all plotted
values should be evenly distributed
about the average-value line. If at any
time, eight successive values appear on
one side of the line, a bias in the data
exists; the analyses should be stopped,
an explanation sought, and any action
taken should be noted on the plot.
7.7 Evaluation of Laboratory
Performance
The QC procedures (Section ^
stress the supervisor's roles in eval
ing QC and in scheduling reanalyses
until data are acceptable for reporting.
This section discusses the QC chemists'
-------
Jan. 1981
11
Part l-Section 7.0
roles in evaluating laboratory perform-
by independent QC checks and by
al audits.
7.7.1 Independent Internal Quality
Control
The QC chemist is to ensure that QC
procedures are implemented and to
provide independent judgement on the
quality of the data generated in the
laboratory. Quality of data is checked by
introducing audit spikes and by synthet-
ic rainwater samples inconspicuously
added to the sample stream (Section
7.6.6.4). Audit spikes may be prepared
by the QC chemist from NBS-SRM's
(Section 7.2.3), from reagent grade salts
or from reference waste/wastewater
samples (Section 7.6.3.2). Whatever the
source, a reference rainwater with a
known composition in the range of
typical rainwater samples is submitted
to the laboratory as a blind audit sample
to be analyzed routinely, and the data
reduction is done by the individual who
handles the data. Finally, the data are
flagged in the data set by the QC
chemist who checks for out-of-limits
conditions and then reports the data to
the laboratory director.
7.7.2 External Performance Audit
Laboratories analyzing rainwater
r 'es are recommended to partici-
(. .1 an interlaboratory comparison at
least two times a year. An external
agency (e.g., EPA or USGS) should
provide to each laboratory audit samples
which contain constituents of interest
in concentrations typically found in
rainwater samples; these samples
should arrive in the laboratory as blind
samples, and exactly the same proce-
dures should be applied to them as in
routine analysis. When audit data are
reported with regular sample data, the
external agency informs the laboratory
of the blind sample identities. These
audits can be used by the QA coordina-
tor to assess analytical accuracy (Sec-
tion 9.2).
7.7.3 System Audit
Once a year, the project manager or a
designee should document current
standard operating and QA procedures
by completing the Laboratory Ques-
tionnaire (Section 7.7A) in cooperation
with the laboratory supervisor and the
QC chemist.
The questionnaire will be submitted
to the laboratory director 4 to 6 weeks
before the interview is conducted to
amplify responses to various questions.
During the interview, written proce-
dures, QC charts, and audit spike
recovery data may be examined. The
auditor summarizes all findings in a
System Audit Report.
7.7A Laboratory Questionnaire
In the laboratory questionnaires, the
R and G denote "recommended strong-
ly" and "guidelines", respectively.
-------
Part l-Section 7.0 12 Jan. 1981
Preliminary Questionnaire
Questionnaire Completion Date
Laboratory:
Street Address:
Mailing Address (if different from above):
City:
State: Zip:
Laboratory Telephone No.:Area Code: No.:
Agency Director:
Quality Assurance Officer:
(Quality Control Chemist)
Questionnaire Completed By (If more than one. please indicate which section(s) of the questionnaire completed):
-------
Jan. 1981 13 Part l-Section 7.0
Introduction-Laboratory Operations
5" -iard Operating Procedures I SOP) yes
1. Has an offical agency Standard Operating Procedures Manual been written?
2. Is the SOP Manual followed in detail?
3. Does it contain alt quality control steps practiced?
4. Does each analyst have a copy at his/her disposal?
5. Has a methods validation study been completed for each analysis?
6. Are plots of instrumental accuracy and precision available for every analysis?
7. Are detection limit data tabulated for each analysis?
-------
Part l-Section 7.0
14
Jan. 1981
Laboratory Personnel
Position
Name
Academic Training
BS MS
HS BA MA Ph.D
Special Training
Years
experience
in rainwater
analysis
% time
presentl
spent
in rainwater
activities
-------
Jan. 1981
16
Part l-Section 7.0
Laboratory Staff Training
1. j formal training program used? Yes No
If yes. is it:
Agency-wide Yes_
In-house Yes.
On-the-job training Yes_
.No.
.No_
2. Training outside local agency (courses attended).
Course
description
or title
Who attended
Course
length
Course
type'
Year of
attendance
'State. Federal. College, University, Other.
3. Publications routinely received and used by agency.
4. Do you feel that your agency training is adequate?
If not. what would be required to make it adequate?
Yes
No
-------
Part l-Section 7.0
16
Jan. 1981
Laboratory Facilities
Item
A vailable
Yes No
Comments
(where applicable, cite system.
QC check, adequacy of space)
1. Filter room or desiccator
(either required for TSP)
maintained at 15°-35° C
and 50% relative humidity
2. Gas
3. Lighting
4. Compressed air
5. Vacuum system
6. Electrical services
7. Hot and cold water
8. Laboratory sink
9. Ventilation system
10. Hood space
11. Cabinet space
12. Bench-top area
(cite linear ft.)
13. Lab space
(cite linear ft.)
14. Lab space utilized for
offices (cite sq. ft.)
15. Office space
(cite sq. ft.)
16. Storage space
(cite sq. ft.)
17. Shared space
-------
Jan. 1981 17 Part l-Section 7.0
Laboratory Equipment
Items
n of Equipment
units Make Model
Condition/age
Good-fair-poor
Ownership
Air Water
% of use time
used in rainwater
programs
Balance
analytical
Vacuum Filtration
apparatus
NBS Calibrated
thermometer
Dessicator
Ion
Chromatograph
Technicon
Atomic Absorption
Balance, top
loader
Class "S"
weights
Balance table
Distilled water
c 'onized
Conductivity
meter
Glassware
pH meter
Drying oven
Hot plates
Refrigerator
-------
Part l-Section 7.0 18 Jan. 1981
Laboratory Operation
A. RECEIVING CLERK
(Name)
Yes No
1. Are all chemicals dated on receipt and thrown away when shelf life is exceeded? (G)
2. Are all samples received by the laboratory logged into a bound notebook? (R)
3. Are all samples filtered before analysis? (ft)
4. Are all samples stored in the refrigerator between analyses? (R)
B. GRAVIMETRIC MEASUREMENTS
7. Is the analytical balance calibrated daily with weights traceable to NBS? (R)
2. Is the "Balance Calibration Log" kept up to date? (R)
3. Is factory service scheduled and the "Balance Factory Service Form" completed? (R)
Date next service is due
C. ANALYST - pH.
(Name)
Yes No
1. Does the analyst have his/her own copy of the standard operating procedures? (R)
2. Does the analyst have his/her own copy of instrument performance data? (R)
3. Does the analyst have his/her own copy of safety instructions? (Ftj
4. Does the analyst have his/her own copy of the latest monthly QC plots? (R)
5. Is the analyst aware of the most recent control limits? (R)
6. Does the analyst have a copy of the most recent list of samples in-house to be analyzed? -fR)
Date of list
7. Are all solutions properly labelled? (R)
5. Has a "pH Meter/Electrode Acceptance Test Form" been completed for the meter and electrode currently
in use? (R)
9. Is the "pH Acceptance Test Summary Form" kept up-to-date? (R)
10. Is the "Field Quality Control Audit Sample Report" completed on a monthly basis? (R)
71. Are rinse and measurement tubes poured for buffers and samples? (R)
12. Is the pH meter calibrated before and after samples are analyzed? (R)
13. Is the pH meter recalibrated after every set of 20 samples? (R)
14. After the initial calibration of the day. when the meter is recalibrated after a series of measurements, is the
old calibration information written down before the meter settings are changed? (G)
75. Is the pH electrode reference solution analyzed first and are the results plotted real time? (G)
16. Are the following control samples analyzed with each run?
Blanks (R)
Old Samples fR)
Audit Spike (R)
17. Does the analyst review the quality control data sheet output by the data clerk, and then decide whether
or not to release data for reporting? (G)
18. Are electrodes stored as recommended by the manufacturer? (R)
19. Are electrodes checked and filled, if necessary, before each analysis? fR)
-------
Jan. 1981 19 Part l-Section 7.0
Laboratory Operation
''ALYST - GRAN STRONG/TOTAL ACID
(Name)
Yes No
1. Does the analyst have his/her own copy of the standard operating procedures? (R)
2. Does the analyst have his/her own copy of instrument performance data? (R)
3. Does the analyst have his/her own copy of safety instructions/' (R)
4. Does the analyst have his/her own copy of the latest monthly QC plots? (R)
5. Is the analyst aware of the most recent control limits? (R)
6. Does the analyst have a copy of the most recent list of samples in-house to be analyzed? (R)
Date of list
7. Are all solutions properly labelled? (R)
8. Has a "pH Meter/Electrode Acceptance Test Form" been completed for the meter and electrode currently
in use? (R)
9. Are micropipets calibrated on at least a weekly basis or whenever the tip breaks? (Rl
10. Are repipets calibrated on a weekly basis? (R)
//. Is the stock 1.0 N NaOH standardized each month against potassium acid phthalate? (R)
12. Is solution temperature carefully monitored during analysis to see that it changes by less than 0.1°C? (Rl
13. Are conditioning solution data and analyst spike data calculated and plotted real time? (G)
14. Are the V function correlation coefficients of these data examined to ensure that they are greater than
0.9990? (G)
/ re the following analyzed each day?
Three conditioning solutions and an analyst spike initially. (R)
An analyst spike and a conditioning solution at the end of the analysis. (R)
An audit spike. (R)
16. Does the analyst review the quality control data sheet output by the data clerk, and then decide whether
or not to release data for reporting? (G)
17. Are electrodes stored as recommended by the manufacturer? (R)
18. Are electrodes checked and filled if necessary before each analysis? (R)
-------
Part l-Section 7.0 20 Jan. 1981
Laboratory Operation
F. ANALYST - TECHNICON :
(Name)
Yes No
1. Does the analyst have his/her own copy of the standard operating procedures? (R)
2. Does the analyst have his/her own copy of instrument performance data? (R)
3. Does the analyst have his/her own copy of safety instructions? (R)
4. Does the analyst have his/her own copy of the latest monthly QC plots? (R)
5. Is the analyst aware of the most recent control limits? (R)
6. Does the analyst have a copy of the most recent list of samples in-house to be analyzed? (R)
Date of list
7. Are all solutions properly labelled? (R)
8. Is the "Standard Preparation Form" completed when new stock standards are prepared? fR)
3. Are dilute calibration standards prepared fresh daily? (R)
tO. Is the analyst spike prepared fresh daily from an independent stock? fR)
1 7. Is the calibration curve at least a five point curve? fR}
12. Is the first calibration curve of the day checked for detection limit and linearity? fR)
13. Are the analyst spike data calculated and plotted real time? (G)
14. Is each new calibration curve checked to see that instrumental response changed less than 5%? (R/
15. Are the following control samples analyzed with each run?
Blanks (R) —
Old Samples (R/ —
Analyst Spikes fR)
Audit Spikes (R)
16. Does the analyst review the quality control data sheet output by the data clerk, and then decide whether
or not to release the data for reporting? (G)
7 7. Is water pumped through all lines daily before and after analysis? (G)
18. Are pump tubes changed at least once per three days? (G)
19. Is the pump cleaned when the pump tubes are changed? (G)
20. Is soap solution pumped through all lines once per week? (G)
21. Is the flowcell cleaned with a sulfuric acid-potassium dichromate solution once per month? (G)
22. Is the pump oiled once per three months? (G)
Date of last service _
23. Is the colorimeter mirror assembly and color filter cleaned and the alignment optimized once per
three months? fG)
Date of last service
-------
Jan. 1981 21 Part l-Section 7.0
Laboratory Operation
NALYST - DIONEX
(Name)
Yes No
1. Does the analyst have his/her own copy of the standard operating procedures? (R)
2. Does the analyst have his/her own copy of instrument performance data? (Ft)
3. Does the analyst have his/her own copy of safety instructions? (R)
4. Does the analyst have his/her own copy of the latest monthly QC plots? (R)
5. Is the analyst aware of the most recent control limits? (R)
6. Does the analyst have a copy of the most recent list of samples in-house to be analyzed? (R)
Date of list
7. Are all solutions properly labelled? (R)
5. Is the "Standard Preparation Form" completed when new stock standards are prepared? (R)
3. Are dilute calibration standards prepared fresh weekly? ' (R)
10. If manual techniques are used, are samples and eluent prepared fresh daily from the same
concentrated stock buffer? W
/1. Is the analyst spike prepared from an independent stock? (R)
12. Is the calibration curve at least a four point curve for each analytical range? (R)
13. Is the first calibration curve of the day checked for detection limit and linearity? (R)
14. Are the analyst spike data calculated and plotted real time? (G)
*" Are the following control samples analyzed with each run?
Blanks (R)
Old Samples (R)
Analyst Spikes (R)
Audit Spikes (R)
16. Does the analyst review the quality control data sheet output by the data clerk, and then decide whether
or not to release the data for reporting? (G)
7 7. Is the drip tray examined daily for reagent spills, and are spills cleaned up daily? (G)
18. Are pumps oiled once per week? (G)
19. Is the anion precolumn cleaned once per month with 0.1 M NaiCOi? (G)
20. Is the Br~. NOy resolution checked once a month and documented with a "Dionex Resolution Test Form"?(R)
-------
Part l-Section 7.0 22 Jan. 1981
Laboratory Operation
H. ANALYST-AA
(Name)
Yes No
1. Does the analyst have his/her own copy of the standard operating procedures? (R)
2. Does the analyst have his/her own copy of instrument performance data? (R)
3. Does the analyst have his/her own copy of safety instructions? (R)
4. Does the anal-jit have his/her own copy of the latest monthly QC plots? (R)
5. Is the analyst aware of the most recent control limits? (R)
6. Does the analyst have a copy of the most recent list of samples in-house to be analyzed? (R)
Date of list
7. Are all solutions properly labelled? (R)
8. Is the "Standard Preparation Form" completed when new standards are prepared? (R)
3. Are dilute calibration standards prepared fresh monthly? (R)
JO. Is the analyst spike prepared from an independent stock? (R)
11. Is the instrument allowed to warm up at least 15 minutes with the flame on before the final
wavelength adjustment is made? (R)
12. Is the calibration curve at least a five point curve? (R)
13. Is the first calibration curve of the day checked for detection limit and linearity? (R)
14. Are the analyst spike data calculated and plotted real time? (G)
15. Is each new calibration curve checked to see that instrumental response changed less than 5%? (R)
16. Are the following control samples analyzed with each run?
Blanks {R)
Old Samples (R)
Analyst Spikes (R)
Audit Spikes (R)
/ 7. Does the analyst review the quality control data sheet output by the data clerk, and then decide whether
or not to release the data for reporting? (G)
-------
Jan. 1981 23 Part l-Section 7.0
Laboratory Operation
' TA CLERK
(Name)
Yes No
1. Does the data clerk do a 100% QC check for accuracy of data input to the computer? (R)
2. Does the data clerk routinely report quality control data sheet information to the analyst? (R)
3. Does the data clerk submit quality control data sheet information to the lab manager along with the analytical
data to be reported? (R)
4. Does strip chart reduction by on-line electronic digitization receive at least 5% manual spot checking? (R)
5. Are control charts or equivalent checks (e.g., computer calculated range limits or regression charts) current
and available for inspection? (R)
6. Are provisions made for data storage of at least 3 years for all raw data, calculations, quality control data
and reports? (R)
7. Do laboratory records include the following information:
a. Sample identification number (R)
b. Station identification (R)
c. Sample type (R)
d. Date sample received in laboratory (R)
e. Time, date and volume of collection (R)
f. Date of analysis (R)
g. Analyst (R)
h. Results of analysis (including raw analytical data) (R)
i. Receptor of the analytical data (R)
8. Are rain gauge chart data for event times and amount checked? (G)
5. Does laboratory follow chain-of-custody procedures from sample receival to discard? (G)
10. Are computer printouts and reports routinely spot checked against laboratory records before data
are released? (R)
11. Are manually interpreted strip chart data spot checked after initial entry? (R)
/ 2. Are minimum detection limits calculated by an approved method such as either Hubaux- Vox (Appendix A) or
baseline standard deviation? (R)
13. Are calibration curve coefficients tabulated and regularly reviewed as evidence for instrumental control?(An
alternative is to use Regression-Hypothesis testing in lieu of control charting.) (R)
14. Are control charts, regression charts or computer QC data bases up-to-date and accessible? (R)
15. Has the data handling and reduction system been examined by legal counsel to determine the soundness of
the system in possible litigation? (G)
-------
Part l-Section 7.0 24 Jan. 1981
Laboratory Operation
J. QUALITY CONTROL CHEMISTRY
(Name)
K. LABORATORY TECHNICIAN
Yes No
J. Does the QC chemist have his/her own copy of the standard operating procedures? (R)
2. Does the QC chemist have his/her own copy of instrument performance data? (R)
3. Does the QC chemist have his/her own copy of safety instructions? (R)
4. Does the QC chemist have his/her own copy of the latest monthly QC plots? (R)
5. Is the QC chemist aware of the most recent control limits? (R)
6. Does the QC chemist prepare a blind audit spike once per month? (R)
7. Does the QC chemist routinely review and report blind audit spike data to the laboratory manager? (R)
8. Does the QC chemist update control limits and obtain new control chart plots once per month? (R)
(Name)
1. Are all containers washed before they are sent to the field? (R)
2. Is the conductivity of the last rinse water measured for 10% of the washed containers? (R)
3. // the conductivity of the rinse is greater than 2 iimhos/cm. is the container rinsed further? (R)
4. After the containers and lids are dried are the containers capped immediately? (R)
5. Are precautions taken not to touch the inside of the containers and lids? (R)
6. Are all samples stored in a refrigerator when not being analyzed? (R)
7. Are precautions taken not to breath on sample? (R)
8. After completion of the analyses, are the samples stored in a refrigerator for a time period of at least
six months? (R)
LABORA TORY MANAGER
(Name)
1. Does the laboratory manager have his/her own copy of the standard operating procedures? (R) —
2. Does the laboratory manager have his/her own copy of instrument performance data? (R)
3. Does the laboratory manager have his her/own copy of safety instructions? (R) —
4. Does the laboratory manager have his/her own copy of the latest monthly QC plots? (R) —
5. Is the laboratory manager aware of the most recent control limits? (R) —
6. Does the laboratory manager review the following before reporting data:
a. The data itself? (R)
b. The quality control data sheet with analyst notes? (G) —
c. The quality control chemist blind audit data report? (R) —
d. The ion summation ratios for the data? (R) —
e. The calculated vs measured sample conductivity? (G) —
-------
Jan. 1981
26
Part l-Section 7.0
7.8 Data Forms
•ns for recording laboratory
i as, including calibrations and
Qv, procedures, are mentioned
throughout Section 7. On the
following pages, blank data forms are
provided for the convenience of the
manual user. Many of these forms are
taken or adapted from EPA forms and
from other sources. The title is at the
top of each, as is customary for a data
form. To relate the form to the text, a
form number (e.g., 1.1/7.3.2) is in the
lower right-hand corner; the 1.1 means
form number 1, version 1, and the 7.3.2
means Section 7.3.2. A revision of the
form would change the number to
1.2/7.3.2, or form 1, version 2, and so
forth. The numbers and titles of the
forms listed herein are listed below:
Form Number
1.1/7.4.1
2.1/7.4.2
3.1/7.5.1.1
4.1/7.5.1.1
5.1/7.5.3
6.1 X7.5.3
7.1/7.5.3
8.1 /7.5.3
9.1/7.6
10.1/7.6
Title
Field Conductivity
Standard Form
Field pH Electrode
Test Solution
Certification of
Weights to NBS Form
Balance Calibration
Log
pH, Conductivity and
Gran Strong Acid
Instrument Perform-
ance Form
Precision Accuracy
Instrument Perform-
ance Form
Precision Accuracy
Instrument Perform-
ance Plot
Instrument Perform-
ance Summary Form
Most Recent Monthly
Control Limit
Audit Spike Recovery
Data
7.9 References
1. Quality Assurance Handbook for
Air Pollution Measurement Sys-
tems. Vol. II - Ambient Air Specific
Methods. EPA - 600/4-77-027a,
May 1977,. p 3-4 of 2.0.
2. Handbook for Analytical Quality
Control in Water and Wastewater
Laboratories. EPA - 600/4-79-
019, March 1979, p 2-1 to 2-6.
3. Applebaum, S.B., and G.J. Crits,
"Producing High Purity Water,"
Industrial Water Engineering (Sept./
Oct. 1964); Karamina, N.A., Amer.
ib. 8(3). 24(1976).
tagent Chemicals. American
Chemical Society Specifications,
5th Edition, American Chemical
Society, Washington, DC (1974).
5. "Standard Methods for Prepara-
tion, Standardization, and Storage
of Standard Solutions for Chemical
Analysis," from Part 31 of 1976
Book of ASTM Standards, American
Society for Testing and Materials,
Philadelphia (1977).
6. Standard Methods for the Exam-
ination of Water and Wastewater,
13th Edition, American Public
Health Association, New York
(1971).
7. Reichgott, M., "Organic Coatings
and Plastic Chemistry," Vol. 41,
1979, Paper presented at 178th
National Meeting of the American
Chemical Society, Washington,
D.C. Sept 9-14, 1979.
8. Operations and Maintenance
Manual for Precipitation Mea-
surement Systems. United States
Environmental Protection Agency,
Research Triangle Park, N.C., (in
press).
9. United States Pharmacopeia. United
States Pharmacopeial Convention
Inc., Rockville, MD., XIX, 644
(1975).
10. Peffer, E.L. and G.C. Mulligan,
"Testing of Glass Volumetric
Apparatus," NBS Circular 434,
National Bureau of Standards
(1941).
11. Electric Power Research Institute,
"Acid Precipitation in the Eastern
United States," Contract Nos.
RP1376-1 and RP1630-2, Rock-
well International Environmental
Monitoring and Services Center
(1978-1980).
12. Gran, G., Analyst 77. 661 (1952).
13. Method 350.1 from Methods for
Chemical Analysis of Water and
Wastes. EPA-600/4-79-020, March
1979.
14. Method 365.4 from Methods for
Chemical Analysis of Water and
Wastes. EPA-600/4-79-020, March
1979.
15. Hubaux, A., and G. Vos, Anal.
Chem., 42. 849(1970).
16. Peden, M.E., and LM. Skowron,
"Ionic Stability of Precipitation
Samples, Atmos. Environ., 12.
2343(1978).
17. Rothert, J.E., Battelle Pacific
Northwest Laboratories, Richland,
Washington, MAP3S Program,
private communication.
18. Peden, M.E., LM. Skowron and
F.F. McGurk, "Precipitation Sample
Handling, Analysis, and Storage
Procedures," United States DOE,
Pollutant Characterization and
Safety Research Division, Contract
No. EY-76-S-02-1199, Research
Report 4, COO-1199-57 (1979).
19. Manual on Industrial Water and
Industrial Wastewater, 2nd Ed..
ASTM Special Publication 148-H,
American Society for Testing and
Materials (1965), p869.
-------
Part l-Section 7.0 26 Jan. 1981
Field Conductivity Standard Form
Date of Preparation of
O.IM KCI Stock Solution:
Date of Preparation of
Dilute Field Standard:
(Analyst Signature)
Laboratory Analysis Before Shipment to the Field
1.
2.
3.
A verage Cond.
± Std. Dev.
Laboratory Values After Use In The Field:
Date of Lab Lab Value
Field Site ft Analysis (ftmhos/cm)
QA Manual for Precipitation Measurement 1.1/7.4.1
-------
Jan. 1981 27 Part l-Sectlon 7.0
Field pH Electrode Test Solution
L ' Preparation of
0../C* Solution
(Analyst Signature)
Volume of Stock Taken
Final Dilution Volume of
Field pH Electrode Test Solution
Laboratory Analysis Before Shipment to the Field
Conductivity pH Ave. ± Std. Dev.
1.
2.
3.
Laboratory Analysis of Aliquots Returned From the Field
Field pH Values Lab Values After Return
Site ft Ave. ± Std. Dev. (data) Date of Analysis pH Cond. Anal. Init.
QA Manual for Precipitation Measurement 2.1/7.4.2
-------
Part l-Section 7.0
28
Jan. 1981
Certification of Working Weights to NBS Form
Date of Certification:
Weight Set Serial #:
(Analyst Signature)
Balance 0
NBS 1kg
NBS 5kg
NBS 1kg
NBS ,
Test 1kg
Test 5kg
Test 1kg
Test 5kg
Balance 0
NBS 1kg
NBS 5kg
NBS 1kg
Test 1kg
Test 5kg
Test 1kg
5kg
Balance 0
NBS 1kg
NBS 5kg
NBS 1kg
5kg
Test 1kg
Test 5kg
Summary:
1kg
Balance 0
NBS 1kg
NBS 5kg
Test 1kg
Test 5kg
Average = Standard Deviation
QA Manual for Precipitation Measurement
3.1/: 1
-------
Jan. 1981
29
Part l-Section 7.0
Balance Calibration Log
ilance
ID
Number
Date
Test Weight #/
Known
Mass
Mass
Found
Test Weight #2
Known
Mass-
Mass
Found
Test Weight #3
Known
Mass
Mass
Found
Analyst
Initials
QA Manual for Precipitation Measurement
4.1/7.5.1.1
-------
Part l-Section 7.0
30
Jan. 1981
pH. Conductivity, and Gran Strong/Total Acid Methods Validation Form
Conductivity:
Date of Test
Concentration
HCI(M)
HzCr
W6
;o~5
10'*
w-3
Expected
Conductivity 1
umhos/cm
1
0.43"
4.26"
42.44"
42/.0"
Conductivity Found c>^
-------
Jan. 1981
31
Part l-Section 7.0
Precision-Accuracy Instrument Performance
one.
Taken
(ug/ml)
Curve"
Parameters
Slope
Intercept
Error
Corr. Coef
Limit
Instrumental Response
Curve 1
Curve 2
Curve 3
Curve 4
Curve 5
Sr>f>rif><: •
Date
.
-------
Part l-Section 7.0
32
Jan. 1981
Precision-Accuracy Instrument Performance Plot
Species Plot
Percentage
Recovery
Concentration (fig/ml)
Figure 1. Percentage recovery as a function of concentration.
Percentage
Standard
Deviation
Concentration (fjg/'ml)
Figure 2. Percentage standard deviation as a function of concentration.
QA Manual for Precipitation Measurement
7.1,
-------
Jan. 1981
33
Part l-Section 7.0
Instrument Performance Summary Form
Species
Date of
Test
Concentration Above
Which 10% Precision
is Obtained ffjg/ml)
Average Detection
Limit ± Standard
Deviation f/jg/m/J
Average Detection
Limit x 3x (Standard
Deviation) f/jg/ml)
Analyst
Initials
QA Manual for Precipitation Measurement
8.1/7.5.3
-------
Part l-Section 7.0
34
Jan. 1981
Date
Most Recent Monthly Control Limits
Full Date Curve Dup. Old Analyst Sp/X
Anal. Scale Control Curve Det. Sample" Sample Blank
Species Rangeti Std. Cone. Limits Est. Error Limit Diff. Diff. Magnitude LCL UCL
Instrument
Dionex
Technicon
cr
NHS
1
Atomic
Absorption
Gran Strong
Acid
Gran Total
Acid
PH
Conductivity
/Va*
r
Ca" /
2
Mg"
H*
H*
aif duplicate field samples are taken.
QA Manual for Precipitation Measurement
9.1/7.C
-------
Jan. 1981 35 Part l-Section 7.0
B. Audit Spike Recovery Data* Month:
Cone. Date of Cone. % Date of Cone. %
ument Species Taken Analysis Found Rec. Analysis Found Pec. Analysis Found Pec.
Dionex Cl~
NOi
Dionex S0«
Technicon NHt'
Technicon Pol
"" indicate data points out of limits by * in % Rec. Column.
QA Manual for Precipitation Measurement 10.1/7.6
-------
Parti-Section 7.0 36 Jan. 1981
Audit Spike Recovery Data" (Cont'd) Month:
Cone. Date of Cone. % Date of Cone. % Date of Cone.
Instrument Species Taken Analysis Found Rec. Analysis Found Rec. Analysis Found
Atomic Na*
Absorption
Atomic K*
Absorption
Ca*
* Indicate data points out of limits by * in % Rec. Column.
QA Manual for Precipitation Measurement 10 1/7.6
-------
Jan. 1981 37 Part t-Section 7.0
Audit Spike Recovery Data" (Cont'd) Month:
Cone. Date of Cone. % Date of Cone. % Date of Cone. %
rument Species Taken Analysis Found Rec. Analysis Found Rec. Analysis Found Rec.
Atomic Mg"
Absorption
pH Meter pH
"Indicate data points out of limits by" in % Rec. Column.
PA Manual for Precipitation Measurement 10.1/7.6
ph Meter Gran
Strong
Acid
-------
Part l-Saction 7.0 38 Jan. 1981
Audit Spike Recovery Data* (Cont'dl Month:
Cone. Date of Cone. % Date of Cone. % Date of Cone. %
Instrument Species Taken Analysis Found Rec. Analysis Found Rec. Analysis Found Re
Gran .
Total
Acid
Conductivity
"Indicate data points out of limits by* in % Rec. Column.
QA Manual for Precipitation Measurement 10.1/7.6
-------
Jan. 1981
Part l-Section 8.0
8.0 Handling, Validation, and Reporting
Quality control of data handling and
other aspects of analysis and reporting
are part of data management activities,
and require discussions of the follow-
ing: calculating detection limits, spot-
checking and screening for accuracy of
data entry, reporting formats, calculat-
ing of averages and statistics for
cumulative reports, and classifying and
handling outliers.
Data obtained in the field and in the
analytical laboratory must be validated
and quality controlled to assure that
they accurately represent the concen-
trations of the the sample constituents.
Validity depends on control of error and
bias during all phases of data handling
from analytical results to final reporting.
8.1 Data Logistics
Involved with data acquisition, reduc-
tion, and reporting are the field operator,
the analyst, the data entry staff, the
laboratory director, the program man-
(or project engineer), and the
tated QA coordinator.
.^eally, laboratory data should go
directly from an instrument to a machine
readable raw data base to avoid human
transcription errors. Because not all
laboratories have computer facilities,
manual data-recording techniques are
discussed, but all calculations and data-
processing steps in this section can be
performed by automated processing.
If voltages from the instrument are
not recorded automatically, a data form
must be prepared. Each analyst should
keep a bound notebook to record all
analytical data; this notebook should
have carbons so copies can be pulled for
data reduction. Typical data forms are in
the 0 & M manual(1) for each analytical
procedure and blank forms are in
Section 7.8 and 8.8.
Whenever data forms are prepared
from strip charts, transcriptions should
be checked by reprocessing 10% of all
values; if errors are found, all data
should be reprocessed.
The sections below briefly describe
manual data-recording practices for
several typical techniques used in
precipitation analysis.
r " 1 pH and Conductivity
pH and conductivity measure-
j read directly from meters are
recorded on a prepared data form. Meter
readings are recorded instead of "peak
height." Baseline readings are ignored.
8.1.2 Anion Chromatography for
Chloride. Phosphate. Nitrate, and
Sulfate
Peak heights on strip charts are a
measure of response. Draw baselines,
read each peak height from the baseline
with a clear plastic ruler, and record
each peak height on the strip chart and
also on the data form. Because this is a
chromatographic technique, care must
be exercised in drawing the baseline.
8.1.3 Automated Colorimetry for
Ammonium and Phosphate
Read the data on the strip charts as
steady-state voltage peaks; draw a
straight line between baseline points on
the chart, read each peak height from
the baseline using a clear plastic ruler,
and record the peak heights on the strip
chart and on the data form.
8.1.4 Atomic Absorption for Sodium.
Potassium, Calcium, and Magnesium
Process the data the same as for
automated colorimeter (Section 8.1.3).
8.1.5 Strong Acid by Gran Method
In the microtitration, electrical poten-
tial (mv) or pH is recorded as a function
of the volume of base added to the
sample. As a QC procedure, the initial
potential reading of the conditioning
solution, before addition of sample, is
recorded on the data form (Section 8.8)
and the final temperature before
titration should also be recorded.
8.1.6 Acidity
The sample is titrated potentio-
metrically with a basic solution to an
end point of pH 8.3. The normality (N) of
the base and the volume (ml) required
are recorded on the data form.
8.1.7 Volume
In determining the amount of sample
precipitation, assume a density of 1.0
gm/ml; thus the ratio of mass (gm) to
volume (ml) is 1.0. Record the mass
directly on the data form.
8.2 Field Observation Coding
Encode field observations of weather,
sample condition, sampler, and other
relevant data into computer files or onto
summary forms in a standard format for
ease of interpretation. A six-digit code is
most convenient to summarize weather
conditions, sample conditions, and
equipment conditions:
Weather codes aa include:
00 no event
01 rain
02 snow
03 mixed snow/rain
Samples codes bb include:
00 sample intact
01 sample contaminated (unspeci-
fied cause)
02 Noticeable suspended
particles
03 sample leakage in shipment
04 insufficient sample for complete
analysis
Equipment codes cc include:
00 no equipment problems
01 sampler inoperative or
malfunctioned (no/or negligible
sample)
02 rain gauge inoperative or mal-
functioned
03 field pH and conductivity
measured late (x days after
scheduled sample removal)
04 pH/conductivity meters
inoperative
For example, a mixed rain/snow sample
with noticeable particles and no equip-
ment problems would be coded 030200.'
8.3 Software Requirements
Data handling from raw data input
through finished report, should be
computerized as much as possible.
Figure 8-1 shows typical data flow from
chemical analysis through final QC
reporting. The software has several
basic functions:
Data input (Section 8.3.1)
Calculation of concentrations from raw
data and calibrations (Section 8.4.2)
Data storage and indexing (Section
8.3.2)
Control charts and related tests (Section
8.4.4)
Data output and reporting (Section 8.6)
8.3.1 Data Input
Data can be input in two ways,
manually or on line. For manual data
entry, the data clerk should screen all of
the terminal input by manual compari-
sons of the computer printout with the
original data forms or by duplicate
entries. For automated on line entries,
errors can usually be detected by
software reasonableness checks, by
monitoring data display while data are
being taken, and by occasional spot-
checks or audits of data acquisition
apparatus.
-------
Part l-Section 8.0
Jan. 1981
f Start J
I
Initial
Calibration
(Analyst)
I
Analysis
of Samples
with Spikes, etc.
f Analyst}
f
Final
Calibration
(Analyst)
I
Manual
Data Entry
(Data Clerk)
\ ._ _
f
1
Initial QC
Checks
(Immediate Results
to Analyst)
1
f
^
Error
Recovery
_,
Calibration
Constants
Flag Out-
of-Range
QC Analyses
'
ifAftei all
Analyses
for Sample)
Cumulative QC Report
1. Calib. constant
2. Graphs of accumulated
QC data on range
charts.
3. Summary of duplicate
analyses.
4. For all samples:
a) Calculated vs meas.
conductivity.
b) Charge balance of
ions.
c) Lab vs field
pH & conductivity.
ions: Na/CI, etc.
e) Collected sample
amount vs rain gauge.
Figure 8-1. Typical scheme of analysis and QC functions.
8.3.2 Data Storage and Indexing
At a minimum, large data bases
should be stored on a computer read-
able medium (e.g., disk, tape or cassette)
which can be accessed efficiently. A
duplicate backup file stored in a
different location should be maintained.
Special file attributes (e.g., random
access, keys, and indexing) can be
useful for efficient data management.
Where elaborate file organization with
high capacity on-line storage devices
are available, the following data organ-
ization is recommended:
Main Data File(s) - raw (concentration)
data with appropriate retrieval keys.
Duplicate File - for backup.
Index File(s) - data for cross-referencing
site, date of sampling, etc., with data in
main data file.
QC File - calibration constants, control
limits, and QC sample data (spikes,
blanks and duplicates); indexing, by site
and date with cross-reference to the
raw data to which the QC data pertain, is
desirable.
8.3.3 QC Functions
A computer enables extensive auto-
mation of QC functions. One valuable
function is immediate flagging of error
conditions: spike data which are out of
control; data which are below detection
limits; and calibration constants which
are out of tolerance. Software avoids
tedious control chart plotting and much
manual mathematical labor. Equations
and criteria for data flagging have been
described (4, Appendixes E & H).
8.4 Data Handling and Sta-
tistical Analysis of Data
In an acid precipitation monitoring
network, data handled both in the field
and in the laboratory must be critically
reviewed to identify and isolate errors.
Data should be validated at each step of
the measurement process—beginning
with sample validation in the field and
followed by a preliminary physical
screening process when the sample is
received from the field. After data enter
the storage-retrieval system, a more
detailed screening process is taken. All
procedures should be well docum 1
so that new personnel can t. /
understand them.
In establishing statistical screening
procedures, it is necessary to recognize
characteristics of the chemical analy-
ses. For the most part, analyses are
done in a batch mode with multipoint
calibrations run before and after the
samples, which may total several dozen.
In an efficiently run laboratory, most
analyses are automatically sequenced
with data recorded continuously on strip
charts or by computer. As discussed in
Sections 7.5 and 7.6 on chemical
analyses, ion chromatography mea-
sures several constituents in the same
analysis.
8.4.1 Quality Control of Data Hand-
ling
Table 8-1 summarizes the data
handling steps and the QC to be applied
at each step. A range check is some-
times effective as an additional screen
against keypunch errors.
8.4.2 Calculations
Calibration standards should be run
as spikes and blanks at the beginning
and end of the analysis and periodically
during the analysis. Sample dar 'e
calculated from linear least squ<
parameters of the bracketing calibi. ^n
standards; equations used for these
calculations are available in most
elementary statistics books(2).
The linear least squares fit yields the
following parameters: slope (m), inter-
cept (b), error of fit (e), correlation
coefficient (r), and the detection limit
(dl). The slope and intercept define a
relationship between concentration
standards and instrument response:
xi + b
8-1
where Ypi = predicted instrument re-
sponse (not Yai, actual
instrument response),
and
xi concentration of stan-
dard i.
Equation 8-1 is the preferred fit where
major components of random variance
are assumed to be in instrument
response. Rearrangement of Equation
8-1 yields concentration corresponding
to an instrumental measurement:
Xj = (yBi - b)/m 8-2
where Xj calculated concentratr r
a sample,
YBj = actual instrument response
for a sample, and
-------
Jan. 1981
Part l-Section 8.0
Table 8-1. Suggested QC Spotcheck of Data Handling
Data Handling Step QC Procedure
/>._. iual reading of strip chart
Transfer of analysis results to data sheet.
Input of data (field or analytical form)
into computer
Electronic digitization of strip chart
Field report records of event time and
amount
Check 10% of data
Check 100% of data
Check data printout 100% vs raw
data form
Check 5 to 8% for baseline and
timing
Check 100% vs rain gauge strip
chart
mandb=calculated slope and inter-
cept from the latest cali-
bration standards run.
The error term is calculated from the
difference between the predicted in-
strument response yp, and the actual
instrument YBi response for a given
calibration standard.
e = [(Yai- mxi- b)2/(n-2)]1/2
8-3
where n = number of calibration stan-
dards.
This term, which indicates how much
random scatter is in the calibration, has
the same units as the y variable
(instrument response), and thus should
be directly compared between calibra-
only when all setup parameters
, factor, concentration range, etc.)
are identical.
The correlation coefficient, a measure
of how much of the variation in
instrument response is explained by
differences in concentration (as opposed
to pure random variation), is discussed
in elementary statistics texts(2).
8.4.3 Statistical Evaluation Tech-
niques
Sampling is assumed to collect single
(not duplicate) samples for most pre-
cipitation events so 100% QC of each
sample cannot be attained; however,
with QC procedures addressed to each
analysis batch, good screening of proper
instrument functioning can be achieved.
The following discussions present
statistical checks available under the
constraints of non-duplicate sample
collection. Table 8-2 summarizes avail-
able QC statistics.
Table 8-2. 'Quality Control Data
Type of Information
8.4.3.1 Calibrations—In addition to a
mathematical estimate of the relation-
ship between instrument output signal
and concentration (by regression curve),
calibration provides statistics for eval-
uating the analytical method. Duplicate
calibrations, before and after analysis of
field samples, yield data on instrument
reproducibility and drift. Statistics from
routine calibration data include:
Mimimum Detection Limit - A conser-
vative method (e.g., Hubaux and Vos,
Appendix B (3)) yields the detection limit
as a function of the quality of the linear
least squares fit of the calibration data.
Comparison of the detection limit
calculated from each day's standardiza-
tions with the one for the technique and
range determined during instrument
documentation may indicate problems
with technique.
Slope and Intercept - The slope and
intercept of the least squares fit of the
data for a technique and instrument
range are fairly constant. Visual inspec-
tion of values from successive days or
comparison with values obtained during
the method validation can be used for
quality control.
Correlation Coefficient - This easily
calculated statistic r is often used in a
semi-quantitative way to evaluate
goodness of fit of the relation of one
variable with respect to another. Values
near +1.0 and -1.0 are good and values
near 0 are poor. The actual range of
good fit values will depend on the
particular measurement or test. The
value can be used as a more quantita-
tive measure in two ways. First, r2 is
equal to the fraction of total variance s2
(s = standard deviation) due to the
Data Evaluated
1. Calibration Curve
Slope (m)
Intercept (b)
Error (e)
Correlation coefficient (r)
Section limit (dl)
£.. dlank
3. Spike sample
4. Old sample
Not charted
Not charted
Calculated magnitude
Non-Gaussian, not used for QC
Calculated magnitude
Calculated magnitude
Percent recovery
Difference between old and new values
correlation of the tested variables; the
balance is due to random error. Second,
r together with sample size n can be
used to calculate the 95% confidence
limits of the derived slope of the linear
relation between the tested variables
(2).
Residual Error - The scatter of data
points off the regression line is a
measure of "noise" in the calibration,
and it is related to expected precision of
the analysis.
8.4.3.2 Spikes and Blanks—If the
experimental design calls for only one
analysis of nonduplicate field samples,
the primary tool for monitoring the
integrity of the analysis is the occa-
sional insertion of spikes of known
concentrations. If calibrations are done
before and after each batch of samples,
the QC samples of known value should
be near the middle of the batch for
optimum control. If the final calibration
is omitted, a QC spike and blank should
be the last samples analyzed. The
concentration values of the spikes
should be entered into the computer
before analysis, so that control limits
can be checked as the analysis results
are keyed into the computer. If control
limits are exceeded, there should be an
immediate report by the QC coordinator
to the laboratory director. If no computer
is available, the laboratory director, QC
coordinator, or analyst should check the
spike results immediatley to determine
if an out-of-control condition exists.
8.4.3.3 Old Samples—Another con-
trol check is reanalysis of old samples to
obtain information on sample stability,
which varies from sample to sample the
H+, IMH/, and N03" are susceptible to
degradation; if degradation is observed,
the sample handling and preservation
techniques should be examined. If
degradation is not observed (i.e. agree-
ment is good), old sample data can be
used to calculate the precision of
analysis. The precision may vary from
day to day.
8.4.4 Control Charts and Related
Tests
With control charts, QC data are
tabulated and plotted as functions of
time; the abscissa is the chronological
order of analysis, and the ordinate may
be either the range, absolute magni-
tude, value of the difference of replicates,
or percentage recovery. All data are
plotted. An expected value and control
limits are calculated, and these are
plotted as the average-value line and
the control limit lines. Table 8-3 shows
types of data which are quality controlled
using the control chart. Other EPA
publications (4-Appendix H,5) give
details of chart construction.
-------
Part (-Section 8.0
Jan. 1981
Table 8-3. Types of Data Controlled by Controlled Chart Methods
Type of Data Value Plotted Expectation Value Control Limits
Duplicate Difference Absolute Value
of Difference
If/,1
QC spike
QC Blank
Percentage
Recovery. X
Calculated
Magnitude
Mean Difference
from method vali-
dation or other
historical data
base. l7l
700%
0
Lower limit - 0
Upper limit - \d\ + ZS
lower limit - 0
or 100-d-ZS)
Upper limit
(J+ZS) • 100
Lower limit - none
Upper limit - dl
d, - Individual duplicate difference
\3\- Mean of absolute value of duplicate differences
Z - Coefficient from distribution tables for any specified level of precision
((2) and Appendix H of (4)).
S - Estimate of standard deviation from the historical data base
dl- Detection limit
Computer calculation and data evalu-
ation with or without plotting may be
substituted for manually plotting the
data. Because the purpose of control
charts is to flag potential problems as
soon as possible, a computer can be
easily programmed to recognize control
criteria without having to plot the data
manually. In either case, the control
chart results should be correlated as
soon as practical after an analysis to
expedite corrective action.
Any control situation unlikely tooccur
statistically should be examined with-
out delay. Many of these occurrences
may be due to chance rather than to
technical problems with the analysis.
Two types of limits are frequently
defined:
Control Limits • the 3s deviation from
the mean value, where s is estimated
standard deviation for technique and
concentration range; thus the upper
control limit (UCL) is u + 3s (u is mean or
expected value of the charted parameter),
and the lower control limit (LCL) is u - 3s
(or 0 if the calculated value is negative).
Setting limits at ±3s implies that less
than 2% of valid data are flagged due to
random error alone and that other
flagged data may be assumed to
indicate nonrandom error, i.e., malfunc-
tion or contamination.
Warning Limits - The ±2s deviations
from the expected value of the control
parameter include about 95% of the
expected random variation about the
mean. However, simple probability
would predict less than 1 % probability of
two independently chosen values for
the charted parameter to exceed the
upper (UWL) or the lower (LWL) warning
limits due to chance alone; thus two
successive values which exceed one or
the other (not both) warning limits
should be reason for investigation of
analytical control.
A final control applied in control
charting is testing for a number of
successive points which lie uniformly
on one side of the expected value.
Probability predicts less than 1 % chance
of eight independent consecutive values
occurring on the same side of the mean
value line, so eight or more would
indicate a small systematic bias in the
technique that may not be observed
until after many days of analysis
because of the relatively large number
of points required.
8.5 Data Validation Criteria
Data validation based on a set of
criteria is the process of filtering the
data and either accepting or rejecting
them. All procedures in this section are
applied after data are first obtained to
identify and flag questionable data for
subsequent investigation. Validation
includes investigation of apparent
anomalies.
8.5.1 Least Squares Fit of Calibration
Curves
As a QC function, calibrations are
analyzed for excessive scatter from the
fit line and for anomalous intercepts.
Cumulative records on each analyte
concentration range and method allow
the laboratory director to examine
trends or discrepancies in coefficients
and standard deviations of fit.
8.5.2 Control Limit Checks
To know if inputs from an analysis
batch are valid, it is necessary to know
as soon as possible after analysis if
control limits were exceeded on the
known QC samples that were analyzed
with the unknown samples.
8.5.3 Detection Limit Checks
Of immediate value to the analyst is
notification that a sample is below the
detection limit of the measurement
range. Ideally, this knowledge should be
available immediately after raw data
entry, then the analyst has two options:
1. Reanalyze the sample on an in-
strument scale of greater sensitivity.
2. Improve the sensitivity (as calculated
by Hubaux-Vos(3)) by multiple
measurements of the unknown, by
judicious choices of concentration
values for the standard curve, or by
use of another analytical tech-
nique. If a sample is below all
detection limits and cannot be
reanalyzed because of lack of
material or cost of reanalysis, a
flag should be entered into the data
base indicating that the datum is
below the detection limit. With a
fully automated system, this flag-
ging could serve as an additional
QA procedure because the analyst
cannot mistakenly enter data
below the detection limit if the
computer program is written to
reject all such data. Statistical
handling and reporting of bp'~-v-
detection-limit data are disc
in Section 8.6.4.
8.5.4 Inconsistencies in Precipitation
Collection Results
Stations equipped with duplicate
samplers or with a sampler and a rain
gauge can test data quality by checking
the quantity of precipitation measured
by the two instruments. Amounts of rain
which differ by more than 15% from a
duplicate should be flagged for investi-
gation of faulty function. To compare the
amounts collected by samplers and by
rain gauges, convert the sampler weight
to inches by multiplying the weight
(grams) by 0.00058 in./gm. However,
compare carefully because light rain-
falls generally yield high weighing rain
gauge capture vs the sampler, and
heavy rainfalls (and wind) yield high
sampler capture vs. tipping bucket
gauge. Differences of 30% between the
sampler and the rain gauge are fre-
quent; for snow, much greater differ-
ences can occur. It has been suggesed
by MAP3S personnel, if the ratio (rain
gauge - sampler volume) /rain gauge is
greater than 0.5, reject the sample; if
the sampler volume is greater than the
rain gauge volume, use the sampler
volume for calculations. The reasr
the difference in sample amount
tured by the rain gauge and sampler
must be resolved before a recom-
mendation is made.
-------
Jan. 1981
Part l-Section 8.0
As an additional check on data
';ty, the stripchart record of the
itation gauges should be com-
,d with the field data form. If
discrepancies in time of event or
i precipitation amount are noted, they
must be resolved before the data can be
reported.
8.5.5 Unusual Ion Ratios
Another check on data validity is
subjective analysis of the ratios of ions
in individual samples. Table 8-4 shows
average and typical ion weight ratios for
terrains and locations. Average values
for sea water(6) should apply to most
areas(7), but those for the earth's
crust(6) cannot be assumed to represent
any specific region. The range data (last
column) were taken from a 1968
national study(8). Significant depar-
tures from these or from typical ratios
for the area are reasons for laboratory
director to investigate an individual
sample or an entire analytical proce-
dure. The seawater ratio of SO^VNA* is
in the table so that measured SO/
values can be corrected for seawater
contribution; this ratio is preferred so
S04/CI because there are nonsea
sources of Cl~ and because loss of
atmospheric Cl~ can occur by oxidation.
Mrast, the chief source of Na+ is the
except in arid areas (e.g., south-
v. _.ern United States).
8.5.6 Comparison of An/on and Cation
Equivalents
The principle of electroneutrality
requires that total anion equivalents
equal total cation equivalents. For the
EPRI Acid Precipitation Study(9), the
term [cations - anions]/0.5[anions +
cations] averaged 0.21 ±0.30. The
greatest spread in data occurred at low
concentrations; for concentrations
above 100 microequivalents, the average
was 0.18±0.22. Discrepancy from zero
and the variation suggest that errors
exist in the data and/or that important
constituents (e.g., HCOs", organic
anions) have not been analyzed. (For
solutions with pH below 5.0 at normal
ambient conditions, the HCOs" concen-
tration is negligible.) The anion/cation
ratio criterion is in Table 8-4.
8.5.7 Comparison of Measured and
Calculated Conductances
For dilute solutions (e.g., below 10~3
M)of known composition, the equivalent
conductance is the sum of the equivalent
ionic conductances i in the solution at
infinite dilution (Table 8-5). From the
relations between the equivalent A and
lecific conductance
Table 8-4. Ion Ratios for Various Sources
Ratio
I Anions/
I Cations0
Primary
Source
Average
ValuefB)
Rainwater
Source
Acceptable
Range (8)
(ug/ug)
0.50 -J. 20"
C/V/Va+b
/VaV/C+b
M0+VCa++b
SO4=/Na+c
seawater
earth's crust
seawater
earth's crust
seawater
earth's crust
seawater
1.8
0.01
27.8
1.1
3.2
0.6
0.25
industrial area
seacoast
arid region
(soil particles)
seacoast
inland
seacoast
inlar. '
seacoast
1.8 -
1.5 -
0.8 -
6
1.2 •
0.1 -
0.03 -
0.25
3.5
1.8
1.0
13
4
1.0
0.3
"Do not test if analyses of chief constituents incomplete.
"Do not test if either number is negative (i.e.. if data less than Hubaux-Vos d.l.(3))
c To correct S0t° for seawater SOt° contribution.
d microeq/microeq
Table 8-5. Equivalent Conductance at Infinite Dilution. 25°C
Ion
H*
NHS
/Va+
A, (mho/cm)
350.0
74.5
50.9
Ion
Ndl'
cr
X •, (mho/ cm)
79.0
70.6
75.5
1 '/2 = value for 1 g. eq.
= 1000/f/N
Thus
8-4
8-5
where N
= g. equivs. of ion i/liter,
Mi = g. moles of ion i/liter, and
Zi = valence or charge of ion i.
With Equation 8-5 and Table 8-5, the
calculated specific conductance of a
solution containing
H+ = 7x10~5 mol/liter,
NH/ = 4x10"5mol/liter,
S04° = 4x10~5 mol/liter, and
N03~ = 3x10'5 mol/liter.
is
1000 K = 7x10'5 (350^4x10'5
(74.5)+2(4x1Q-5)
(79.0)+3x10~5(70.5)
= 3592x10"5 mho/cm
or
K = 35.9 micromho/cm.
The calculated specific conductance
value can be compared with the mea-
sured values for precipitation samples,
which are generally 10"4M or less.
For the EPRI study(8), the difference
between the calculated and measured
conductances averaged about 22.5%,
with the measured values being lower.
These results are consistent with the
anion/cation results, and probably are
due to measurement error, to concen-
tration changes in the time interval
between the conductivity measurement
and the analysis, to analysis error or
nonmeasurement of a constituent, or to
a combination of these. The H* ion is the
chief contributor to the conductance of
the solution, and any significant error in
the (-T concentrations will generally be
evident in a comparison of the con-
ductances and of the anion/cation
equivalents.
8.6 Data Reporting
Data reporting should include a
hardcopy printout, a computer tape, and
a calendar of events (precipitation) at
each site. This section does not address
data interpretation (Section 8.7) or
advanced computation.
8.6.1 Calendar of Events
A typical calendar of events for nine
stations is shown in Figure 8-2. The
codes R, S, M, and X denote rain, snow,
mixed (liquid and solid phases), and
missed samples.
8.6.2 Entities and Units Reported
Generally, the entities measured
directly and reported are pH, specific
conductance, concentrations of major
constituents, and the amount of pre-
cipitation. Cumulative amounts of
constituents deposited and average
quantities or concentrations for various
-------
Part l-Section 8.0
Jan. 1981
.JANUARY 19*0
•?ITE tt 2 4 6 8 10 12 14 16 18 20 22 24 26 26 3'"
1 3 9 ? 9 11 - 13 15 17 1? 21 23 25 27 29
11 R R
21 R R X
41
52
61
72
M
R
MS R
S
R
R
R
R R
R R
R
R
R
R
R
R
M
R R
R ri R
R R R
R
R R
R
S
S
M S
R
91
R
M
SITE X
11
21
FEBRUARY 19*0
4 6 y 10 12 14 16 13 20 22 24 26 28
5 7 9 11 13 15 17 19 21 23 25 27 29
X S S
S R
41
R
M
R (1 M
R
61
72
11
h M
M
R
R
R
R
M
R R
P.
M
M
91
R R R
fl = Rain
S = Snow
M - Mixed (rain & snow)
X = Missed or lost sample
Figure 8-2. Calendar of events.
time periods, two other important
terms, are discussed in Section 8.6.3.
Reporting units are typically:
pH - pH units,
conductivity - microS/cm or micromho/cm,
concentration - micromol/liter, and
precipitation - mm (1 in. = 25.4 mm).
Transformation from microgram/ml
(ppm), the generally used analytical
term, to micromol/liter is:
micromoles/liter = microgram/ml
x conversion factor
8-6
The conversion factor for each ion is in
Table 8-6.
Table 8-6. Factors to Convert
Microgram/ml to Micro-
mol/liter (micromol/l =
microgram/ml x factor)
Analyte
cr
NOi
S04=
POi*
hT
NHt
Na+
/r
Mg"
Ca++
Molecule
Weight
35.46
62.01
96.07
94.98
1.01
18.04
22.99
39.10
24.31
40.08
Factor
28.2
16.1
10.4
10.5
990.1
55.4
43.5
25.6
41.1
25.0
8.6.3 Weighted Mean Values and
Deposition
Generally reported on a cumulative
basis for a given time interval are
precipitation weighted mean concen-
tration Ci for various ions.
8-7
where P\ = amount of precipitation
in event j (mm or ml),
Cij= concentration of COP
stituent i for
event j (micromol/l).
For pH or 1-T concentration, the cumula-
tive value is calculated, and the final pH
-------
Jan. 1981
Part l-Section 8.0
pH = -log
-PHJ.
8-8
"IMC cumulative average concentration
at each site can be used in studies of
distribution of the constituents as a
function of time (month or season) and
geographic location.
Total deposition of an analyte per unit
area (mg/m2) in a precipitation event or
time interval is calculated as:
mg/m -microgram/ml x 25.4xP
8-9
where P = precipitation (in.)
and 25.4 = 2.54 cm/in.
x 10~3 mg/microgram
x 10"cm2/m2.
Treatment of missing data in calculating
the above values is discussed in Section
8.7.3.
8.6.4 Reporting and Treating of Be-
low-Detection-Limit Data
Data below the detection limit (BDL)
for the analytical method used should
be flagged with a code both in the
printouts and in all computer readable.
data forms. Because it is important for
data interpreters to know the detection
' '• a suggested code is the negative
he treatment of these data in data
(,. _-essing programs depends on the
analysis being performed; examples are
replacing all BOL data with 0 if an
arithmetic mean is being taken or with
Vi DL(10) if a geometric mean is being
computed. Any computed summary
value which has 25% or more BDL
contributing data points should be
coded with a unique flag.
8.6.5 Reporting of Out-of-Control
Data
Analytical data in the same batch with
an out-of-control QC sample must be
flagged in the data base as suspect. All
such data must be either reanalyzed or
reported as invalid, unless the labora-
tory supervisor decides that the cause of
the out-of-control condition could not
have affected the analytical results.
Data reported as invalid should not be
used in statistics, analysis, or other
interpretation. For data stored or
reported on computer readable media, it
is often best to use a character or code to
indicate out-of-control, missing, BDL,
and so forth.
8.6.6 Treatment of Outliers
Any data point that lies far beyond the
normal range of values can be considered
utlier. Responsibilities should be
ed between invalidations by experi-
mental and data interpretation person-
nel. Experimental personnel should
invalidate data only for experimental
reasons such as known contamination
or violation of QC criteria for a spike,
blank, or duplicate; they should not
invalidate outliers. Outliers in data
interpretation is discussed in Section
8.7.2.
8.6.7 EPA Rain Water Data Base
The EPA is establishing a precipita-
tion data base for all United States and
Canadian deposition programs. This
System of Atmospheric Deposition
(SAD) data will be coordinated through
the National Computer Center (NCC).
Prospective contributors/users should
contact Gardner Evans, United States
EPA, Environmental Research Center
(ERC), (MD 075), Research Triangle Park,
NC 27711.
8.7 Statistics in Data Inter-
pretation
This section discusses briefly the
types of statistics that should be used in
precipitation data interpretation and the
subjects of outliers and missing data.
8.7.1 Parametric and Nonparametric
Methods and Tests
Many procedures routinely used in
statistical analysis involve the assump-
tion that the parent population from
which a sample is drawn has a normal
(gaussian) distribution. Normal distribu-
tion is not needed to compute means,
weighted means, and variances; the
most common formulas for computing
means and variances converge to
values for the parent population for
large sample sizes, regardless of
distribution. Many common hypothesis
tests normal distribution; these are
called parametric tests because they are
concerned with parameters (mean and
standard deviations) of a single func-
tional form (the gaussian curve). For
example, the Student t-test for deter-
mining whether two samples come
from populations with the same mean
assumes that the samples come from
normally distributed populations with
the same variance; thus it is imporant to
know whether the data set one is
working with comes from a normally
distributed population.
Precipitation data frequently show
skewed distributions sometimes inter-
preted as log-normal(11). The distribu-
tions 'are inevitably nonnormal (for all
analytes except pH) because negative
values of concentration and rainfall are
not allowed. The resulting distortions
could be small if the mean is two or
more times the standard deviation, but
this has not been the case in precipita-
tion for analytes other than pH.
To test whether an analyte has a
normal distribution, use an optimized
chi-square test(12) for sample popula-
tions of over 100 with the number of
class intervals chosen according to the
algorithm of Mann and Wald(13). If the
number of data points inthe set isbelow
100, use the Lilliefors(14) test for
normality; this test may also be used for
sample sizes greater than 100. If an
analyte fails the test for normality,
repeat the test with the logarithm of
each data point to determine if the
parent distribution is log-normal.
All data belowthe DLof theanalytical
procedure must be set to positive
nonzero values. An acceptable value is
'/2 DL. A significant number of BDL data
points results in a distribution which is
neither normal nor log-normal. If the
distribution is log-normal and the
interpreter wishes to proceed with
parametric hypothesis testing, log-
transformed data must be used. It is
unacceptable to use ordinary para-
metric statistics on a distribution which
is skewed enough toappear log-normal.
If the distribution fails both the
normal and the log-normal tests, do the
hypothesis testing with nonparametric
tests, also known as distribution-free
statistics; these tests do not assume
that the parent distribution is normal so
they work whether it is or not. If distribu-
tion is normal, these tests are only
slightly less powerful than the para-
metrics, so they can serve as a viable
alternative if the data analyst does not
wish to test for normality. Another
advantage of nonparametrics is its
being less sensitive to outliers than
parametrics; this insensitivity to outliers
is due to dealing with ranks and
categories of the data rather than
original data points. A good reference
on nonparametric hypothesis testing is
Conover(15).
8.7.2 Outliers
Data interpreters may delete outliers,
depending on the computations being
performed. For example, computations
of mean values are generally not
sensitive to deletion of the relatively few
data from both ends of the distribution,
but computations of skewness may be
very sensitive to the loss of data at the
extremes of the normal distribution.
A large concentration event which is
an outlier in the concentration distribu-
tion may not be an outlier in the deposi-
tion distribution if the total rainfall
amount is small. The decision to delete
data from the ends of the distribution
must be made with the full understand-
ing of the interpretations being con-
sidered; this understanding is not
possible if data are deleted by experi-
mental personnel before interpreters
see the data.
-------
Part l-Section 8.0
Jan. 1981
In precipitation chemistry, selection
of outliers is difficult due to the
following considerations:
1. Spatial distributions of pollutants
vary, so one site or more than one
site in similar locales must be
considered separately from sites in
different locales.
2. Seasonal amounts of pollutants at
each site generally vary.
3. Concentrations of pollutants can
reach unusually high or low values
in precipitation events.
Considering the above, no data that
have met the QC criteria (discussed in
previous sections) should be discarded.
Because relatively few nonnormal data
points or outliers are expected, the
effect on the mean values of any
constituent should be small.
Nonparametric hypothesis testing
provides no way to check for outliers. (A
point can be identified as isolated from
its parent population only if a form for
the parent population is assumed.)
Fortunately, nonparametric tests are
comparatively insensitive to outliers so
the outliers may be left in the data base.
Parametric tests for detecting outliers
are in Reference 4, Appendix F; these
are designed to eliminate only one or
two outliers from a data set, but it is
absolutely invalid to drop detected
outliers and then to pass through the
data again to look for other outliers.
8.7.3 Missing Data
If a small percentage of samples are
lost (e.g., due to downtime of equip-
ment), the effect of neglecting their
contributions to the precipitation
weighted average, the pH or the
concentration of a constituent is ex-
pected to be small. However, the effect
of the lost samples on the total amount
of material deposited is more important;
the deposition contributions of lost
samples should be calculated by using
the precipitation weighted mean sea-
sonal concentration values for that site
and the measured or reported (National
Weather Service) precipitation amounts
for the missed events.
8.8 Data Forms
Blank data forms on the following
pages can be used for manual transcrip-
tion of data from digital meters or from
strip chart records to prepare for data
processing. Details on preparing data
forms for each laboratory analysis are in
the respective sections of the 0 & M
manual(1).
8.9 References
1. Operations and Maintenance
Manual for Precipitation Mea-
surement Systems, United States
Environmental Protection Agency,
Research Triangle Park, N.C., (in
press).
2. An Introduction to Mathematical
Statistics. H.D. Brunk, Blaisdell
Publishing Co., Waltham, MA,
1976, p 210.
3. HubauxA.andG.Vos,Anal.Chem.,
42,849(1970)
4. Quality Assurance Handbook for
Air Pollution Measurement Sys-
tems - Vol. I- Principles, United
States Environmental Protection
Agency, Research Triangle Park,
N.C., EPA-600/9-76-005 (1976).
5. Handbook for Analytical Quality
Control in Water and Wastewater
Laboratories. United States En-
vironmental Protection Agency,
Cincinnati, OH, EPA-600/4-79-
019(1979).
6. Handbook of Chemistry and Phy-
sics. 48th ed. Edited by R.C. Weast,
The Chemical Rubber Co., Cleve-
land, OH 44128, 1967-68.
7. Mero, J.L., The Mineral Resources
of the Sea, Elsevier Publ., New
York, N.Y. p.25(1964).
8. Lodge, Jr., J.P., J.B. Pate, W.
Basbergill, G.S. Swanson, K.C. Hill,
E. LorangeandA.L. Lazrus, "Chem-
istry of United States Precipitation,"
Final Report on the National Pre-
cipitation Sampling Network, Na-
tional Center for Atmospheric
Research, Boulder CO, August
(1968).
9. Electric Power Research Institute,
"Acid Precipitation in the North-
eastern United States," Contract
Nos. RF1376-1 and RF1630-2,
Rockwell International Environ-
mental Monitoring and Services
Center (1978-1980).
10. Nehls, G. and G. Akland, J. Air
Pollut. Control. Assoc., 23, 180
(1973).
11. "The MAP3S/RAINE Precipitation
Chemistry Network: Statistical
Overview for the Period 1976-
1980," Atmos. Environ., submitted
for publication.
12. Kendall, M.G. and A. Stuart, The
Advanced Theory of Statistics, Vol.
II, Chapter 30, Hafner Publishing,
New York, (1967).
13. Mann. H.B. and A. Wald, "On the
Choice of the Number of Class
Intervals in the Application of the
Chi-squared Test," Annals of
Math. Statistics. 13. 306-7 (1942).
14. Lilliefors, H.W., "On the Kolmogorov -
Smirnov Test for Normality with
Mean and Variance Unknown," J.
Amer. Statistical Assoc., 62, 399-
402(1967).
15. Conover, W.J., Practical Nonpara-
metric Statistics. John Wiley and
Sons, New York, 2nd Ed. (1980).
-------
Jan. 1981
Part l-Section 8.0
Gran Strong Acid Data Sheet
Sample #
Initial mv
Temp. °C
pi NaOH
Injected
mv
reading
Sample ti
Initial mv
Temp °C
fjl NaOH
Injected
mv
reading
Date:
Cone
Micr
Ca
NaOH N
opipette
libration
15 /il = /jl
Total volume
-------
Part (-Section 8.0
10
Jan. 1981
Data Sheet
N
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
ID
Baseline 1
Std 1
Std 2
Std 3
Std 4
Std 5
Sample 1
Sample 2
Sample 3
Sample 4
Sample 5
QC 1
Sample 6
Sample 7
Sample 8
Sample 9
Baseline 2
Sample 10
Sample 1 1
Sample 12
Std 1
Std 2
Std 3
Std 4
Std 5
Baseline 3
Peak Ht.
B1
B2
B3
-------
Jan. 1981
Part l-Section 9.0
9.0 Data Quality Assessment
Precision and accuracy of precipita-
tion data provide quantitative assess-
ments of data quality. Procedures for
determining precision and accuracy
were designed for specific monitoring
and analysis methods. Measurement
methods used in precipitation chemistry
monitoring systems for integrated
sampling follow a three-step process;
the first two are functions of field
operations, and the third is part of the
laboratory operations.
1. Collection of the sample in a
suitable sampler.
2. Initial field analysis for pH. specific
conductance, and weight.
3. Expanded chemical analyses in the
laboratory.
At each step, QA should be assessed.
Precision of sampling is determined
by using collocated samplers; field
measurement accuracy is determined
by test sample audit. Precision of
laboratory analytical methods is deter-
mined with duplicate samples; accuracy
is determined by an internal and
vnal blind sample audit.
i».1 Evaluation of Field Oper-
ations
9.1.1 Sampling Precision
Precision is estimated by duplicate
samplings with collocated precipitation
samplers. Each network of sites operated
by the same agency should have a
duplicate sampler of the type used for
routine monitoring. The collocated
sampler operated during routine samp-
ling should be set up consistent with
siting criteria (Section 5.0).
Data from the collocated sampler are
compared to data from the station
sampler. The measured difference in pH
(pH units), conductivity (micromho/cm),
and total mass captured between the
samplers are used to calculate preci-
sion. For each pair of measurements,
differences can be calculated using:
Ai = yi-xi 9-1
where Ai = difference in measure-
ments for ith precision
check (appropriate units),
yi = pH, conductivity, or
weight measured by the
duplicate collocated
sampler for ith precision
check, and
xi = pH, conductivity or
weight measured by the
the corresponding station
sampler for the ith pre-
cision check.
Quarterly, standard deviation s is
estimated for each site with collocated
samplers:
where Sj=quarterly standard devia-
tion of jth instrument
during ith precision
check,
AJ = difference in measure-
ments during ith pre-
cision check, and
n = number of precision
checks on the instru-
ment during calendar
quarter.
If the network contains more than one
collocated sampler, an overall precision
estimate sa for each measurement
method can be computed quarterly:
sa=f^ S'j*- 9.3
Lki=' ' J
where sa = overall deviations
for a specific measure-
ment method,
Sj =standard deviations of
jth sampler for measure-
ment method, and
k = number of collocated
measurements within
the network.
Data from both the station sampler and
the collocated sampler are recorded oh
standard Field Data forms (Section
6.8.5) with corresponding identifica-
tions. Copies of these forms should be
submitted to the QA coordinator, who
summarizes the data on the Precision
Check form (Section 9.5), which is used
in preparing QA reports (Section 9.3).
9.1.2 Accuracy of pH and Conductiv-
ity Measurements
To assess the accuracies of field
measurements of pH and conductivity,
audits should be conducted using test
samples prepared and sent from the
central laboratory each month (Section
7.6). The samples should be measured
as soon as possible after receipt, and the
results should be returned to the
laboratory with the remainder of the
sample. The audit results should
include two laboratory analyses—one
before the sample is sent to the f itid and
one after it is returned and the field
analysis results. The QA coordinator
should record all the data on the
Monthly Field Audit Report (Section
9.4).
Accuracy for each measured variable
is estimated by computing value differ-
ences:
Aj = f| - (lij+ In) /2 9-4
where A|=difference in measure-
ments for jth site (appro-
priate units),
f,-field analysis of variable
for jth site.
In = initial laboratory
analysis of variable
before shipment to the
jth site, and
l(j=final laboratory analysis
of variable after return
from jth site.
This equation should be used only if |ln -
If!| is less than the control limits.
Data should be summarized in the
Monthly Field Audit Report by the QA
coordinator, who also computes an
average monthly network difference Am
for the variables pH and conductivity
Am=(1/nJZ,1,
9-5
and for the monthly standard deviation
Sm
Sm ~
9-6
whereA'm=average monthly net-
work difference for a
given variable,
A|=difference between lab l|
and field fj at the jth site,
and
n = number of sites audited
during the month.
Monthly reports should be summarized
.in a quarterly report (Section 9.3).
9.1.3 Method Precision
Each network should maintain at
least one pair of collocated samplers
(Section 9.1.1). The collected samples
should be tagged so their respective
measurements in the. field and the
-------
Parti-Section 9.0
Jan. 1981
laboratory can be used to assess
quantitatively the monitoring (sampling,
handling, and measurement etc.) pre-
cision by using Equations 9-1 through
9-3 above, where y and x also represent
concentrations of species. Data should
be reported on the Precision Checks
forms (Section 9.4).
9.2 Evaluation of Laboratory
Operations
The QA coordinator must implement
routine activities to assess the precision
and accuracy of laboratory chemical
analyses of precipitation samples.
Laboratory data quality is assessed by
determining the precision, variability,
accuracy of chemical analyses.
9.2.1 Analytical Precision
To estimate the contribution of
analytical variability to total variability,
duplicate sample analysis should be
performed on about 10%of the routinely
analyzed samples. Samples randomly
selected for replicate analysis by the QC
chemist should contain a large quantity
of precipitation. The chosen samples
should be split—one-half for analysis
immediately, the other half for re-
frigeration at about 4°C to be analyzed
within a week by a different analyst. The
sample should be properly identified,
and the results of the duplicate analyses
should be recorded by the QA coordina-
tor in the Report of Duplicate Analyses
(Section 9.4).
9.2.2 Accuracy of Chemical Analysis
Accuracy of chemical analysis should
be determined monthly on blind samples
submitted to the laboratory by randomly
selected field sites. These samples,
prepared by diluting various water
standards, should be shipped in sealed
plastic bottles to the field sites. Each
sample should be identified, and ac-
companied by two postcards. On receipt
at the field site, the sample should be
refrigerated at 4°C until it is forwarded
to the laboratory. At the end of the first
week in which no event has occurred,
the sample should be transferred to a
clean weighed bucket or to the usual
container for shipment. The sample
should be weighed, and an aliquot
should be measured for pH and for
specific conductance. A field data form
should be filled out with measured
values and the other required data; to do
this, a non-existing precipitation record
will have to be used. Sample and site
identification, sampling date(s), pH,
conductivity and weight should be
recorded on postcards. The sample and
the data form should be sent to the
laboratory (as for any sample), and the
postcards mailed to the QA coordinators
for both the monitoring network and for
the blind sample preparation laboratory;
thus, if one card is lost, the information
will be available from the second.
At regularly scheduled intervals, the
analytical laboratory should send a
printout of sample results to the QA
coordinator. The QA coordinator identi-
fies the blind samples, has the data
transferred to the QA data file, notifies
the analytical laboratory of sample
results that must be deleted from the
precipitation data file; and sends the
laboratory the expected QA sample
values. If there is a difference between
any analyzed and expected values that
is greater than the experimental error,
that analysis must be repeated. (Samples
must be stored in a laboratory refrigera-
tor until results have been approved by
the QA coordinator.) Data should be
summarized as indicated in Equations
9-4 and 9-5 (Section 9.1.2). The QA
coordinator should obtain samples from
EPA's NAP and from the USGS semi-
annually for overview purposes.
9.3 Reporting of Data
9.3.1 Laboratory Results
When analyses are complete and the
analyst has released the data for
reporting, a printout or summary of the
following must be prepared:
Analytical data
Anion-cation comparison data
Measured vs. calculated conductance
data
Quality control data
The anion-cation and the measured vs.
calculated conductance comparison
data were discussed in Sections 8.5.6
and 8.5.7. The QA data should be a
listing of the results of the QC param-
eters which pertain to the particular
data being reported and to the audit
results. Data which exceed previously
established QC limits should be flagged.
The QA coordinator should provide
the following reports to the project
manager:
Precision - Quarterly reports, sum-
marizing all data on collocated samples
and duplicate analyses performed
during the 3 mos, should include
differences in analytical results for the
split samples and the average differ-
ence for each analyte; the report should
compare average differences with the
QC ranges typical for the laboratory and
the measurement method.
Accuracy - Quarterly reports, summa-
rizing all data on three blind QA samples
in the 3 mos, should include the different
constituent concentrations and an aver-
age difference and standard deviation
computed for each. Data for the com-
parison should be obtained for each
blind sample by the QA coordinator who
summarizes the differences (mg/liter)
for each constituent.
9.3.2 Audit Data Basis for F
Measurements
The pH and the conductivity are
generally measured in the field because
of simplicity, importance, and the
possibility of chemical change before a
sample would arrive in the laboratory.
To assess the quality of measurement
by the field operator, audit or test
samples sent from the laboratory should
be run. (To assess sampling precision,
collocated samplers should be run
(Section 9.1).) Results should include
the two laboratory analyses (before the
sample is sent to the field and after it is
returned) and the field analysis. Be-
cause the amount of data is generally
not large, computer data management
is optional. The steps involved in using
these data for checking the quality of
field measurements are in Figure 9-1.
Control limits for differences between
field and laboratory measurements
must be set during the preliminary
instrument performance documenta-
tion. Factors contributing to random
variability should include differences
between well-calibrated instruments;
storage time (several days between
measurements); storage conditions
(temperature, agitation, etc.);
differences in field and labora
solution temperatures during mea-
surement (effect can be negligible if
instruments have temperature com-
pensation). After a sufficient data base
with real measurements is established,
new control limits can be calculated
from actual data. The data base should
give the standard or average deviation
for each station and for the network as a
whole. Typical control limits for 95%
confidence are approximately 0.15 unit
for pH and 10% for conductivity. Data
should also indicate when a bad pH
electrode or conductivity standard
needs replacing.
The laboratory manager should eval-
uate the data carefully if an electrode is
malfunctioning because it might be
necessary to invalidate all data taken by
that electrode since the last acceptable
audit; thus monthly test samples are
suggested.
9.4 Data Forms
Blank forms on the following pages
were taken or adapted from EPA forms
and from other references. No page-by-
page documentation is given in the top
right-hand corners on these forms. The
titles are at the top of the figures -
customary. To relate each form t-
text, a form number is given mthelov.^r
right-hand corner: 1.1/9.1 indicates
form 1, version 1 (Section 9.1). A
-------
Jan. 1981
Part (-Section 9.0
Stan
Prepare
Blind
Sample
(Laboratory)
Analysis
L2
(Laboratory)
Analysis
L,
(Laboratory)
Ship
to
Field
Analysis
F
(Field)
Investigate
Field
Operations.
c
No Problems \
-End- J
Return Sample to
Laboratory
Figure 9-1. Audit procedure for field measurements of pH and conductivity.
revision of the form would be changed
to 1.2/9.1 three forms included here for
user convenience are: form 1, version 2,
and so forth.
Form Number Title
1.1/9.1 Precision Checks
2.1/9.1 Monthly Field Audit Report
3.1/9.2 Report of Duplicate Analyses
-------
Part (-Section 9.0
Jan. 1981
Site ID
Month/Year
Precision Checks*
Station Sampler #
Collocated Sampler ft
Date
Weight C
(gm) D
S
pH C
(pH units) D
S
Conductivity C
fumho/cm) D
* S - Station Sampler
C - Collocated Sampler
D- c r
— o • \*
n
i,= ll: D.
n l=''
«.
or-L
HI
1/2
Weight
(gm)
pH
Conductivity - A,
(fimho/cm) "!-,
QA Coordinator
Date
QA Manual for Precipitation Measurement
1.1/9.2
-------
Jan. 1981
Part l-Section 9.0
Monthly Field Audit Report
.iple #:.
Date of Preparation of Field Audit Sample:.
(Analyst Signature)
Laboratory Analysis Before Shipment
To The Field
Laboratory Analysis After Return a
From The Field
Date:
Conductivity
Date:
Conductivity
pH
1.
2.
1.
2.
3.
±Average
Std. Dev L
3. .
±Average
. Std. Dev .
Site tt
Field
Date
Laboratory Analysis of Audit Samples Vs. Field Analysis
Conductivity (fjmho/cm)
Field
Analysis
Lab"
Analysis
Dili.
Field
Analysis
Lab"
Analysis
Diff.
These data are for three laboratory aliquots which are analyzed before shipment of samples, are then refrigerated, and are
reanalyzed with the samples returned from the field.
"Values after return from the field.
QA Manual for Precipitation Measurement
2.1/9.2
-------
Part l-Section 9.0
Jan. 1981
Analyte
Analysis
Technique
Report of Duplicate Analysis
Routine*
(ID ;
Date Result
Duplicate*1
(ID-. ;
Date Result
Diff.'
pH
Conductivity
Sulfate
Nitrate
Chloride
Phosphate
Carbonate
Bi-Carbonate
Acidity
Strong Acid
Ammonium
Sodium
Potassium
Magnesium
Calcium
* Value reported as routine sample
** Duplicate sample, might have different ID
•"Diff. = Duplicate - Routine
QA Manual for Precipitation Measurement
3.1/9.2
-------
Jan. 1981
Part I - Appendix A
Appendix A
Hubaux-Vos Detection Limit Calculation
According to Hubaux and Vos(1),
DL(Yo) is the limit "at which a given
analytical procedure may be relied upon
to lead to detection." The Hubaux-Vos
detection limit (DL) is significantly
greater than the smallest detectable
signal because the DL definition re-
quired a high level of certainty that a
signal represents the presence of the
analyte. An example calculation - To
establish analytical precision and
accuracy as a function of concentration,
repetitively analyze five times an
approximately 7-point calibration curve
from the lowest standard to the highest.
Then perform a linear least squares fit of
all (~35) data points. Use the slope and
the intercept values to calculate con-
centrations for all standard solutions.
Calculate the average concentration
and the standard deviation for each
concentration. Express the accuracy as
a percentage recovery found, as com-
pared to that expected. Express the
precision as percentage standard devia-
tion. This procedure was applied to Pb,
Cd, and Cr analyses by atomic absorp-
tion spectrometry. Linear least squares
fit of each calibration curve produced a
Hubaux-Vos DL for each curve. The
average detection limit was calculated.
The data in Table A-1 indicate, at the
Hubaux-Vos DL for a single calibration
curve, 10% to 20% analytical precision
and 80% to 100% analytical accuracy.
Table A-1. Hubaux- Vos Detection Limit Data
Highest Lowest
Cone. Std. Cone. Std.
Analyte lug /ml) (ug/ml)
Pb 5.0 0. 1
Pb 10.0 0.3
V 0.5 0.01
r 1.0 0.03
An Example
1. X=A + By
where X = instrument response,
A_(ZX)(Zy2)-(Zy)(ZXy).
n(Zy2 - (Zy)Vn)
ZXy-(ZX)(Zy)/n,
Iyz - (Zy)Vn
Overall DL.
(ug/ml)
0.13
0.17
0.006
0.019
Avg DL,
(ug/ml)
0.224
0.325
0.011
0.036
Recovery
at Avg
DL. %
92
90
81
97
Std. Dev.
at Avg
DL. %
17
16
12
12
A-1
n = total number of data points, and
y = concentration (generally inl/L/g/ml or ng/ml).
S=
wnere S*
n-2
A-2
T = Studentsj-test for n-2 [degrees of freedom, 95% confidence
(Table A-2).
Xc = A + ST V 1 /n, + 1 /N + y-z/l(Zv2 - (Zyp/n).
YD = 222
>32
12.706
4.303
3.182
2.776
2.571
2.447
2.365
2.306
2.262
2.228
2.201
2.179
2.160
2.145
2.131
2.120
2.110
2.101
2.093
2.086
2.06
2.04
If delta = less than some tolerance
level, stop; A-9
DL = YD; if not, replace Y0 by new YD
new YD = YD + delta/B. A-10
If number of iterations = less than
50, go to Equation A-7.
Reference
1. Hubaux, A., and G. Vos, Decision and
Detection Limits for Linear Calibration
Curves, Anal. Chem., 42(8), 849 (1970).
-------
Jan. 1981
Part I - Appendix B
Appendix B
Methods Validation Study for Lead Analysis by Flame Atomic Absorption
First the flame atomic absorption
instrumental conditions were optimized,
and then a 7 point calibration curve for
lead (Pb) at concentration of 0.3 to 10.0
microgram/ml was plotted. Each con-
centration was analyzed sequentially
five times.
The strip chart recording of the raw
data is in Figure B-1. Peak heights were
measures of responses. A scale expan-
sion of 10 times was used, so instru-
mental response was approximate
0.0039 absorbance units/cm. Data
were tabulated and calculated (Section
4.2.3 of this manual), and summarized
in Table B-1. The data would be useful in
documenting instrument performance
for future reference and for communi-
eating accuracy and precision to people
outside of the laboratory. The accuracy
(% recovery) and precision (% std. dev.)
as functions of concentration are
plotted in Figure B-2. The "3s value"
would be useful as an upper control
limit in real-time quality control (Section
4.3.5).
> a
J[
. j. . . . i
OAIAMAHK
I I I I I I I I I I
i i
i i
i i
i i
figure B-1.
» /ro
§roo!
§ 95!
I 50
1 ss!
> *0\
jr • • •
'
2\ 3\ 4\ 5\ 6> 7\ 8< 9 10
Pb Concentration pg/ml
I 18
5 16
| 14
Q 72;
"S 10
If
5 4
|2
1
2 3 4 5. 6 7 8 9 1O\
Pb- Concentration ftg/.ml
Figure B-2. Accuracy and precision\of Pb analysis as a fundtion of concentration.
-------
Part I - Appendix B
Jan. 1981
Table B-1
Precision-Accuracy Methods Validation Form
Cone.
Taken
(ug/ml)
Curve*
Parameters
Slope
Intercept
Error
Corr. Coef
Del. Limit
Instrumental Response (cm)
Curve 1
Curve 2
Curve 3
Curve 4
Curve 5
.Sp/a rif c-
Date
o*/*,"-
(Analyst Signature)
Curve Parameters Overall*
Slope:
Intercept.
Error:
Correlati
Detectior
nn C.nef:
^ Limit:
Average ±
Standard Deviation
3 a Value"
Regression Concentration"
Cone.
Taken
fag/ml)
Curve
1
Curve
2
Curve
3
Curve
4
Curve
5
A verage
Cone.
Found
Standard
Deviation
%
Recovery
% Std.
Dev.
"•for Dionex. indicate approximate umho/cm full scale ffjmho/cm scale x volts on recorder); ForAA, indicate scale expansion;
For Technicon. indicate method and flow/cell length.
"The curve parameters are calculated using the linear least squares fit of Appendix D
cFrom a linear least squares fit of all data points.
"'This is the average value+ 3x(standard deviation).
"This calculated from the instrumental responses given above, using the "Curve Parameters Overall".
'•/Vof included in linear least squares fit.
------- |