United States
Environmental Protection
Agency
K
Atmospheric Research and
Exposure Assessment Laboratory ^«
Research Triangje Park. NC 27711 ~'f fe
Research and Development
EPA/600/S3-90/018 June 1990
&EPA Project Summary
Second Interim Report: Quality
Assurance Support for the
National Atmospheric
Deposition Program and
National Trends Network
Monitoring Activities
1987-1990
David S. Bigelow
The full report summarizes the
quality assurance activities of the
NADP/NTN Quality Assurance
Manager from mid-1988 through mid-
1989. It presents some
accomplishments and makes
recommendations for the network.
The full report outlines the
progress being made by the
NAOP/NTN monitoring program in
completing documentation of
standard operating procedures,
responding to field operation and
siting deficiencies, and discusses
data quality issues relating to the
handling of TRACE samples and the
quality coding of daily rain amounts.
It also contains an evaluation of a
second laboratory's participation in
the network and reports on a
preliminary evaluation of Nipher-
shielded Belfort rain gauge
performance in the network.
Recommendations emphasize the
Quality Assurance Manager's role in
coordinating quality assurance
activities and the need for more
support for the gathering and
interpretation of quality assurance
data.
This Project Summary was
developed by EPA's Atmospheric
Research and Exposure Assessment
Laboratory, Research Triangle Park,
NC, to announce key findings of the
research project that is fully
documented in a separate report of
the same title (see Project Report
ordering information at back).
Introduction
The National Atmospheric Deposition
Program (NADP) monitoring network
began operations in July of 1978. The
Network used and continues to use
cooperating local, state and federal
agencies as well as private industry to
operate a network of sites according to a
single set of protocols. In 1983, NADP
assumed responsibility for managing the
day-to-day operations of the National
Acid Precipitation Assessment Program's
National Trends Network (NTN). This
action resulted in the combined,
cooperative monitoring program now
known as the NADP'NTN monitoring
network. The NADP/NTN monitoring
network serves both the National
Atmospheric Deposition Program and the
National Acid Precipitation Assessment
Program.
Because the NADP/NTN monitoring
program both represents and relies on
many different public and private
agencies to accomplish its goals, it is
unique in its structure and mode of
operation. For quality assurance, this has
meant that responsibility for the
-------
development of quality assurance
programs has been, and continues to be,
a voluntary effort assumed by each of the
agencies that have responsibility for
managing specific pieces of the network
operations (i.e., Field Sites, Central
Analytical Laboratory (CAL), Coordination
Office, U.S. Geological Survey, U.S.
Environmental Protection Agency). These
quality assurance procedures are then
reviewed and accepted by the various
NADP committees that have
responsibility for overseeing each
network function. This practice has been
strongly encouraged since it would
require considerable effort and expertise
to centralize this function within a single
organization.
A decentralized approach to quality
assurance, however, creates its own
problems. Many of the network quality
assurance procedures are innovative
adaptations of recognized quality
assurance practices, and as such are
reported in the scientific literature, at
professional and technical meetings or in
various agency reports rather than in a
standard network document or operating
procedure publication. This lack of
centralized quality assurance reporting
makes documenting the achievement of
network data quality goals and objectives
difficult and also makes it difficult to
continue some key quality assurance
programs once scientific recognition for
the techniques has been achieved. The
likelihood that those quality assurance
activities that cannot be maintained by a
single agency, either because of financial
or cooperative restraints, will be
overlooked is also increased. Finally,
decentralized reporting can lead to
situations where results from one effort
may conflict with another such that no
clear picture of "true" data quality can
emerge.
Recognizing the need for a more
coordinated and thorough quality
assurance program, NADP formed a
study committee in 1982 to address this
issue. The work of this committee
resulted in the development of a Quality
Assurance Plan to cover all aspects of
network operation and the
recommendation that a full-time Quality
Assurance Manager be appointed to
administer this quality assurance
program. The desirability of a
coordinated quality assurance program
was again recognized in the NAPAP Peer
Review held in Boulder, Colorado in
1985.
Accomplishments
Coordination of NADP/NTN
Quality Assurance Activities
The primary coordination efforts of the
Quality Assurance Manager involve
attending the frequent NADP Technical
and Subcommittee meetings; writing,
revising and reviewing NADP/NTN
documentation; reviewing agency reports
and proposals that use NADP/NTN data
or documentation; and disseminating
quality assurance information to the
various NADP committees and scientists.
Each of these activities serves as a
vehicle for obtaining and documenting
the quality assurance activities of the
network and for promoting the evaluation
and reporting of quality assurance data.
Two areas of coordination activities
received special attention during this past
year. These activities were the staffing of
a new site liaison m the NADP/NTN
Coordination Office and the drafting and
review of standard operating procedures
identified as missing during the revision
of the NADP/NTN Quality Assurance
Plan.
Other tasks undertaken by the Quality
Assurance Manager as a part of the
coordination task included: 1) the
participation by the Quality Assurance
Manager in ad hoc committees to
complete a formal procedure for
comparing analytical methods and to
review the development of a training
video for NADP/NTN Site Operators and
Supervisors (copies available through the
NADP/NTN Coordination Office), 2) the
completion of an ASTM standard for
choosing locations and sampling
strategies and 3) participation in a review
of a newly instituted collocated collector
program sponsored by the USGS. The
Quality Assurance Manager also assisted
in training USGS personnel and in
locating candidate sites for the program.
Review and Evaluation of
NADP/NTN Network Operations
Three areas of network operations were
targeted for review during the past
contract year by the Quality Assurance
Manager. The reviews covered an
evaluation and recommendation for
integrating data from a second laboratory
into the network data system, an
evaluation of a Nipher-shield modification
to the standard rain gauge used by the
network as a way of improving rain
volume measurements made by the
network, and an evaluation c
unpublished NADP/CANSAP netwoi(
collocation activities. The review
resulted in two reports, the presentatio
of one poster at the NADP Technic.
Committee Meeting held in Provincetowr
MA in October 1989, and the initiation (
two journal articles.
Evaluation of a Second
Laboratory in the NADP/NTN
Monitoring Program
Site liaison and chemical analyses <
samples from 18 NADP/NTN monitorir
sites administered by the National Pai
Service were performed by a secor
laboratory from March 24 throug
September 29, 1987. At the request
the NADP Subcommittee on Netwo
Operations, the Quality Assuranc
Manager was asked to determine tf
comparability of the second laboratory
data to that of the Central Analytic
Laboratory. The evaluation conclude
that chemistry data are not of the sanr
quality as previous and subseque
NADP/NTN data. Detection/reportir
limits are not comparable, precision is n
comparable and unresolved problen
with sulfate analyses and ion balanci
render the data questionable. Fie
measurements, however, that we
gathered and screened by the secor
laboratory and then re-evaluated by tl
Illinois State Water Survey can be us<
"transparently" with other similar da
gathered by the network.
Recommendations for the integration
this information into the NADP/NTN de
system include the coding of the secoi
laboratory's data to identify it
originating from a laboratory other th
the Central Analytical Laboratory, tl
integration of field data into the standa
data base, the entering of the laboratc
data into a special data base and t
sequestering of all laboratory chemk
analysis data from all but very spec
requests.
Nipher-Shield Evaluation
In 1987, the U.S. Environmen
Protection Agency began a comparis
of a Nipher-shielded Belfort rain gauge
the standard Belfort rain gauge used
most deposition monitoring programs
the United States. The study uses ni
NADP/NTN atmospheric depositi
monitoring sites chosen to represent
variety of snow collection conditions tl
might be found within existing natio
-------
deposition monitoring programs.
Favorable results taken from deposition
monitoring sites, primarily in northeastern
North America, have prompted the U.S.
Environmental Protection Agency and the
Electric Power Research Institute to
adopt the Nipher gauge as a standard
piece of equipment in the Acid MODES
and Operational Evaluation Network
monitoring programs, and to propose that
it be included as a standard gauge in
other North American deposition
monitoring programs such as the
NADP/NTN network.
Unfortunately, the data collected in the
study were not being analyzed since the
conclusions of the study will have a direct
bearing on the quality of previous
precipitation measurements made by the
NADP/NTN network. The data might even
change the standard basis for
determining the representativeness of
collector volumes (catch efficiency) used
in the network. The Quality Assurance
Manager, concerned that the precipitation
measurement standard used by the
network would be changed without
benefit of the results of this important
study, thus decided to analyze the data.
Generally, all of the sites showed
higher values for Nipher-shielded rain
gauges for snow, rain and mixed types of
precipitation. The differences ranged
from negligible amounts up to 37% of the
volume for snow at one site. In weekly
data sets, statistically significant
differences occurred at only five of the
eight sites when all precipitation types
(rain, snow, mixed) are analyzed together
and in only two of the sites when snow
data were analyzed separately. Rain was
significantly different at three of the sites.
Daily data were analyzed at only two
sites but in general supported the
conclusions drawn from the weekly data.
One site showed significant differences
for rain and snow but the other site did
not. The study is scheduled to be
completed after another year of data has
been gathered.
U.S.-Canadian Intercomparisons
Beginning in 1981, the National
Oceanic and Atmospheric Administration
(NOAA) sponsored a three-year, direct
comparison of atmospheric deposition
monitoring protocols used by the National
Atmospheric Deposition Program (NADP)
and the Canadian Network for Sampling
Acid Precipitation Program (CANSAP).
The purpose of the comparison was to
determine if data from these two
networks could be used as a single data
set to characterize atmospheric
deposition in North America. Preliminary
examination of the data during the study
revealed serious contamination problems
with much of the CANSAP data. Because
no final report was ever made and
because the results of the comparison
still have value for long-term trend
analysis and network introspection, the
Quality Assurance Manager was asked to
complete the analysis of the
comparability study.
Preliminary conclusions derived from
an analysis of the complete data set are:
1) that the previously noted
contamination in CANSAP data may only
be a serious problem at unsheltered or
locally source rich sites (e.g., poorly sited
in regards to local or line sources of
emissions) and 2) that the differences at
the more rural, emission source deficient
locations exhibit quite similar results. It
was noted, however, that CANSAP values
in general are consistently higher than
values from the NADP network. The
differences appear to be significant
though at only some of the locations.
Review and Revision of the
NADP/NTN Quality Assurance
Plan
During October 1986, the NADP
Quality Assurance Steering Committee
began the task of reviewing and revising
the NADP Quality Assurance Plan. The
Quality Assurance Manager had been
given the responsibility for the revision of
the Data Management portion of the Plan.
This portion was completed in August
1988. More recently, the Quality
Assurance Manager was given
responsibility for developing another
portion of the plan detailing the integrated
assessment of quality assurance results.
No effort was put into this task pending
action by others to complete drafts of
other portions of the plan. A working
document will be available after the fall of
1990.
Future Needs of the NADP/NTN
Monitoring Network Data
Management Improvement
Data Management activities in the
Coordination Office have increased
steadily both in size and complexity ever
since the monitoring program began. As
part of ongoing efforts to reduce the lag
in data reporting and the complexity of
the data management activity, computer
programs, along with software and
hardware, have required constant tuning.
This has been especially true in the more
recent years as many of the NADP/NTN
program co-operators began developing
and improving their methods of
gathering, processing and distributing
data. When, for instance, the CAL
develops improved data screening
algorithms and improves upon the
amount and type of information passed
along to the Coordination Office, the
Coordination Office must update its data
handling programs and data base design
to accommodate the new or improved
information. These tasks are also further
complicated by the NADP policy of
implementing improvements in data
screening and coding retroactively
throughout its data set. The importance
of applying these improvements
retroactively is one of the many unique
features of the NADP program which
helps maintain the overall integrity of the
NADP data set.
Identification and Handling of
TRACE Samples
The NADP/NTN monitoring program
has agreed to code weekly precipitation
totals as TRACE when the weekly
precipitation total is the result of wet
deposition only, is not the product of
dew, frost, fog or other types of
condensate, and represents an amount of
precipitation that is below the detection
limit of the sampling system. To
implement this philosophical definition,
the network is forced to use a number of
operational definitions to identify TRACE
samples in its field, laboratory and data
management operations. The methods
are not always directly comparable and
become somewhat subjective and
confusing at best when the network
attempts to differentiate TRACE samples
from non-precipitation events and when
the network attempts to estimate "true"
sample volumes and rainfall amounts. At
the time of the original analysis of the
problem, over 21 different cases were
identified that would have to be
considered to correctly process TRACE
samples.
At the April 1989 joint Network
Operations and Data Management and
Analysis subcommittee meeting, the
Quality Assurance Manager made three
recommendations for improving the
handling of TRACE samples in the
NADP/NTN monitoring program. The
recommendations were intended to
clarify and make consistent the handling
of TRACE samples in the network. The
proposal affected approximately 3000
samples out of the entire NADP/NTN
data set collected to date.
-------
Data Quality Coding of Daily
Rain Values
The NADP/NTN monitoring network
has as a matter of policy attempted to
code rather than sequester data in
instances where network protocols have
not been followed exactly. Coding, rather
than sequestering, provides users of
network data with both a maximum
amount of information and a highly
qualified data set. For daily rain values,
however, this practice was ignored. This
was in part due to the emphasis the
network placed on weekly totals rather
than on daily amounts and the fact that
the network would accept volume
amounts from the AeroChem Metrics
Wet/Dry collector in lieu of "good" ram
gauge measurements as an acceptable
rainfall amount for a weekly total.
Common practice in the Network was to
code as missing daily rain values from
network records when problems with the
rain gauge were detected. In the mid-
80's, the practice of coding the samples
as missing was greatly curtailed;
however, this has led to the maintenance
of non-qualified daily ram records in the
network data base.
To begin the process of quality coding
daily rain values in NADP/NTN data
records, the Quality Assurance Manager
proposed a minimal set of guidelines for
the network to follow. The guidelines
were presented to the Data Management
and Analysis Subcommittee in April 1989
and are currently being reviewed by that
committee.
Recommendations
• Continue to evaluate, document and
support network data quality and
network procedures through the
publication of integrated quality
assurance reports. These reports
should be inclusive of all years of
operation through current time.
• Increased support should be made
available to expand and accelerate the
gathering and data banking of both
external and internal quality assurance
data that relate directly to the
NADP'NTN data set and to allow the
timely review and completion of
network documentation activities.
• Staffing levels should be increased to
allow for a full-time Quality Assurance
Manager and other support staff.
Increased staffing would improve the
frequency of integrated quality
assurance reports and allow for
continuity in the tracking of network
quality assurance activities. The current
three-quarter time staffing level
provides only enough resources to
manage the NADP/NTN quality
assurance program on a topical basis.
A quality assurance program
specifically designed to estimate
network precision and bias should be
developed and implemented as soon
as possible. Emphasis should be
placed on comparisons within
NADP/NTN and between peer network
data (e.g., daily vs weekly sampling,
laboratory A vs laboratory B, etc.).
The comparability of NADP/NTN data
to that of other maior networks
operating in North America should be
expanded.
David S. Bigelow is with Colorado State University, Fort Collins, CO 80523
Berne T. Bennett is the EPA Project Officer (see below).
The complete report, entitled "Second Interim Report: Quality Assurance
Support for the National Atmospheric Deposition Program and National
Trends Network Monitoring Activities: 1989-1990." (OrderMo. PB 90-187
T621 AS; Cost: $17.00, subject to change) will be available only from:
National Technical Information Service
5285 Port Royal Road
Springfield, VA 22161
Telephone: 703-487-4650
The EPA Protect Officer can be contacted at:
Atmospheric Research and Exposure Assessment Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, NC 27711
United States
Environmental Protection
Agency
Center for Environmental Research
Information
Cincinnati OH 45268
US.OFFICIAL MAIL"
'ENALTV
OW
R1VATE
•SE $300
* i*
3 MFTER
6090444
U.S.POSIAGE
s 0 .2 5 =
Official Business
Penalty for Private Use S300
EPA/600/S3-90/018
------- |