oEPA
ates
environmental Protection
Agency
Environmental Monitoring and Support EPA-bOO 7-78 179
Laboratory Seph-mht-r 1978
Research Triangle Park NC 27711
Research and Development
Quality Assurance in
Support of Energy
Related Monitoring
Activities in the
Western United States
Annual Report No. 1
Interagency
Energy/Environment
R&D Program
Report
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development, U.S. Environmental
Protection Agency, have been grouped into nine series. These nine broad cate-
gories were established to facilitate further development and application of en-
vironmental technology. Elimination of traditional grouping was consciously
planned to foster technology transfer and a maximum interface in related fields.
The nine series are:
1. Environmental Health Effects Research
2. Environmental Protection Technology
3. Ecological Research
4. Environmental Monitoring
5. Socioeconomic Environmental Studies
6. Scientific and Technical Assessment Reports (STAR)
7. Interagency Energy-Environment Research and Development
8. "Special" Reports
9. Miscellaneous Reports
This report has been assigned to the INTERAGENCY ENERGY-ENVIRONMENT
RESEARCH AND DEVELOPMENT series. Reports in this series result from the
effort funded under the 17-agency Federal Energy/Environment Research and
Development Program. These studies relate to EPA's mission to protect the public
health and welfare from adverse effects of pollutants associated with energy sys-
tems. The goal of the Program is to assure the rapid development of domestic
energy supplies in an environmentally-compatible manner by providing the nec-
essary environmental data and control technology. Investigations include analy-
ses of the transport of energy-related pollutants and their health and ecological
effects; assessments of, and development of, control technologies for energy
systems; and integrated assessments of a wide range of energy-related environ-
mental issues.
This document is available to the public through the National Technical Informa-
tion Service, Springfield, Virginia 22161.
-------
QUALITY ASSURANCE IN SUPPORT OF ENERGY
RELATED MONITORING ACTIVITIES IN THE
WESTERN UNITED STATES
Annual Report No. 1
by
Mark Cher
Rockwell International
Atomics International Division
Air Monitoring Center
Newbury Park, California 91320
Contract No. 68-02-2412
Project Officer
Steven M. Bromberg
Quality Assurance Branch
Environmental Monitoring and Support Laboratory
Research Triangle Park, North Carolina 27711
ENVIRONMENTAL MONITORING AND SUPPORT LABORATORY
U. S. ENVIRONMENTAL PROTECTION AGENCY
RESEARCH TRIANGLE PARK, NORTH CAROLINA 27711
-------
DISCLAIMER
This report has been reviewed by the Environment Monitoring and Support
Laboratory, U. S. Environmental Protection Agency, and approved for publica-
tion. Mention of trade names or commercial products does not constitute
endorsement or recommendation for use.
ii
-------
ABSTRACT
The purpose of the program is to establish a quality assurance data
base for ambient air monitorinq in specified geographical areas around
present and proposed energy development projects, and to provide technical
assistance to enable existing monitoring networks to achieve a high level
of data quality. A goal of the program is to enable the government to
utilize and compare air monitoring data from diverse sources for future
study and planning purposes by providing information concerning data
quality from the individual monitoring networks.
An initial on-site review of 18 laboratories and associated field sites
was completed, and detailed evaluation reports containing a list of recommenda-
tions for eliminating observed deficiencies were submitted to EPA.
Regularly scheduled laboratory performance surveys are being carried out
for the analysis of sulfate, nitrate, SCL, NCL, and CO and for weight measure-
ment and high volume flow rate. In these surveys audit samples or devices are
submitted to participating laboratories and their results are compared with
those obtained by Rockwell.
Approximately 10% of the analyses performed by another laboratory for
metals collected in high volume filters are being repeated by Rockwell. The
causes for observed large differences in the analysis of synthetic metal
samples prepared by Rockwell are being investigated.
Quarterly field audits are being conducted at specified monitoring sites.
Known concentrations of pollutants are delivered to each continuous analyzer,
and the observed response is compared with that predicted by the agency's
calibration. Procedures have been developed to assure reliable performance
by the audit devices and standards. Additional checks on the equipment and
procedures are made by means of quarterly audits performed in the EPA/EMSL
laboratory at Research Triangle Park, North Carolina.
iii
-------
Approximately 70 site audits have been completed, and audit reports
have been submitted to EPA. Satisfactory results were found most frequently
in audits of CO, CH., and total hydrocarbon analyzers. The most unsatisfactory
results were found in audits of N0?, NO , NO, and S0? analyzers.
£ A C-
Technical assistance has been provided to participating agencies, as
requested by the Project Officer. The assistance consists mainly of instruc-
tion manuals, procedures, or literature references relating to chemical
analysis or quality assurance. Much of the technical assistance was given
informally during site visits. Assistance was also provided concerning
specific laboratory and calibration problems during visits to Rockwell AMC
by personnel from participating agencies. A brief description of these and
other types of technical assistance is given in the report.
This report was submitted in fulfillment of Contract No. 68-02-2412
by Rockwell International Air Monitoring Center under the sponsorship of
the U. S. Environmental Protection Agency. This report covers the period
July 13, 1976 to September 30, 1977. Work under this contract will continue
to July 13, 1981.
iv
-------
CONTENTS
Disclaimer ii
Abstract iii
1. Introduction 1
2. Site Evaluations 3
3. Interlaboratory Performance Surveys 9
Sulfate/Nitrate Analysis 12
S02 Analysis 13
N02 Analysis 16
CO Analysis 16
Weighing Survey 20
High Volume Flow Rate Measurement 20
Analysis of Metal Samples From Laboratory A .... 22
4. Field Audits 27
5. Technical Assistance 33
Agencies Assisted 33
U.S. Geological Survey Water Quality Field Stations. 36
General Review Meeting 36
6. Plans For Next Year 38
-------
SECTION 1
INTRODUCTION
This report describes and summarizes the activities and achievements of
Rockwell International Air Monitoring Center during the first year of the
Contract No. 68-02-2412 titled Quality Assurance in Support of Energy Related
Monitoring Activities in the Western United States. The effective date of the
contract was July 13, 1976, and the program is scheduled to be completed on
July 13, 1981.
On June 16, 1977, the contract was expanded to include additional evalu-
ations and field audits sites located in Colorado, Utah, and Montana. On
August 9, 1977, the contract was further expanded to include on-site inspection
and evaluation of seven water quality field stations of the U. S. Geological
Survey in the states of Arizona, Colorado, Montana, New Mexico, North Dakota,
Utah, and Wyoming.
The activities described in this report cover generally the period from
the beginning of the program to approximately the end of September 1977. For
convenience the title of the program is shortened to Western Quality Assurance
Program.
The purpose of the Western Quality Assurance Program is to develop and
implement a quality assurance program for use by networks monitoring air
quality around present and proposed energy development projects. These energy
projects are located in the states of Montana, South Dakota, North Dakota,
Wyoming, Utah, Colorado, Arizona, Nevada, New Mexico, and California.
A goal of the program is to enable the government to utilize and compare
air monitoring data from diverse sources for future study and planning purposes
by providing information concerning data quality from the individual monitoring
networks.
-------
This Annual Report is written as a general overview of performance
and progress thus far. As such it is not intended to treat all areas in
specifics nor will it treat each area rigorously. In order to keep this
report within manageable size, the reader will oftentimes be directed to
specific reports, available in EPA files, should greater detail be required.
The Western Quality Assurance Program is divided into four task areas
involving:
1) An initial on-site review of laboratories and field sites
2) A sample submissions audit program to the laboratories
3) On-site field calibration audits; and
4. Technical assistance to the laboratories, as required.
All the work carried out by Rockwell in connection with these tasks
is summarized in the following sections. Section 2.0 describes the on-site
evaluations; Section 3.0 discusses the results of the interlaboratory perform-
ance surveys; Section 4.0 summarizes field audit results, Section 5.0 describes
technical assistance given to the various agencies; and Section 6.0 outlines
the plans for next year.
-------
SECTION 2
SITE EVALUATIONS
Pursuant to provisions of the Western Quality Assurance Program
Contract, laboratories and field sites were visited, inspected and evaluated
by Rockwell personnel. Visits were made to the 18 agencies listed in Table 1.
Three of the agencies - - Montana, Utah, and Colorado - - were visited a
second time as a result of the expansion in June 1977 in the number of sites
audited.
A detailed report covering the site evaluations has been made for each
agency, and each agency has been given the opportunity to comment upon the
preliminary draft of the report. When appropriate, these comments were noted
and used in preparing the final report, copies of which were sent to the EPA.
The final report contains a great deal of information concerning organization,
staffing, laboratory and field equipment, operations and procedures, documen-
tation, data handling, and data validation. The individual reports should be
consulted for specific information.
The findings were summarized in each report as a table of recommendations
given to the individual agency. The table lists deficiencies found, recommended
solutions, approximate cost to implement the solutions, and an estimate of the
impact of each deficiency on the quality of the data.
Approximately 50 different specific deficiencies were observed throughout
the extensive evaluation period; however, for practicality and workability
many specific deficiencies have been combined and reduced to more general terms
for purposes of systematic and logical reporting. Table 2 lists deficiencies
generically with a brief comment pertaining to each. The deficiencies are
listed according to descending frequency of occurrence. Although many sites
were visited at some of the agencies, no differentiation as to agency or site
is attempted in this summary. The original reports must be examined for
defect particularity and specificity.
-------
TABLE 1. SUMMARY OF AIR MONITORING AGENCIES EVALUATED
Agency
Site Visited
Location
State of Arizona
Colorado Oil Shale Tract,
Colorado Oil Shale Tract,
State of Colorado
EPA/Col strip
EPA/Las Vegas
(Lockheed Electronics)
State of Montana
Montana Power Company
State of New Mexico
NOAA/Boulder
(continued)
Bureau of Air Quality
Headquarters
Field Site #1
Field Site 023
Air Quality Surveillance
Section
CARIH Station
CAMP Station
Arvada Station
Wei by Station
Greeley Station
Colorado Springs Station
Field Site
Laboratory
Air Quality Bureau
Col strip Trailer
Lame Deer
C Hill
Highway Junction
Saddle Mountain
Mi crowave
Alpine West
Field Site #3
Air Quality Bureau
Albuquerque Field Sites
Four Corners Area Sites
Headquarters
Phoenix, AZ
Denver, CO
Rio Blanco, CO
Rio Blanco, CO
Denver, CO
Denver, CO
Denver, CO
Arvada, CO
Wei by, CO
Greeley, CO
Colorado Springs, CO
Col strip, MT
Las Vegas, NE
Helena, MT
Col strip, MT
Lame Deer, MT
Anaconda, MT
Anaconda, MT
E. Helena, MT
E. Helena, MT
Butte, MT
Col strip, MT
Santa Fe, NM
Albuquerque, NM
Farmington, NM
Boulder, CO
-------
TABLE 1 (continued)
Agency
Site Visited
Location
State of North Dakota
Northern Testing Laboratory
State of South Dakota
Utah Oil Shale Tracts,
State Dept. of Health
Stanton
Beulah
Laboratory
State Laboratory
Field Site A-6
Bismarck, ND
Stanton, ND
Beulah, ND
Billings, MT
Pierre, SD
Bonanza, UT
State of Utah
Ute Research Laboratory
State of Wyoming
Yellowstone County
Bureau of Air Quality
Price
Huntington
Salt Lake City
Magna
Bountiful
Ogden
Provo
Laboratory
Laboratory Headquarters
Newcastle Field Site
Gillette Field Site
Patrick Draw Field Site
Laboratory
Billings Field Site
Laurel Field Site
Salt Lake City, UT
Price, UT
Huntington, UT
Salt Lake City, UT
Magna, UT
Bountiful, UT
Ogden, UT
Provo, UT
Fort Duchesne, UT
Cheyenne, WY
Newcastle, WY
Gillette, WY
Rock Springs, WY
Billings, MT
Billings, MT
Laurel, MT
-------
TABLE 2. EXAMPLES OF COMMON DEFICIENCIES IN THE AIR MONITORING PROGRAMS
EVALUATED
Deficiency
Comment
3.
4.
No daily permanent record of
flows, instrument parameters,
instructions, remedial steps
and action criteria kept by
the agency.
Gas cylinders, permeation tubes,
and other standards not NBS
traceable and not rechecked upon
receipt or on a scheduled basis.
Poor or poorly defined intra-
laboratory quality control
procedures in the various tests
and procedures employed.
Balances calibrated infrequently
or not at all.
Virtually all laboratories were
remiss in some aspect of record
and log book keeping. Some defects
were minor but others were of major
proportions.
Again essentially all agencies were
amiss in some form or another with
regards to standards traceability.
Only a few agencies had acceptable
internal quality assurance procedures.
A majority of the agencies were
within this category.
5.
6.
7.
Bubbler samples handled
incorrectly; flows incorrect,
no leak checks made on samplers
or apparatus.
Flow rate measurement errors;
instrument sample flows not
monitored, altitude corrections
not made.
Permeation tube oven tempera-
tures and other temperatures
not measured.
Essentially all agencies that used
bubblers had at least one suspect
handling technique among those
listed.
A majority of the agencies had
problems connected with the measure-
ment and calibration of gas flows.
Approximately half of the agencies
had problems in this area.
8.
9.
Deficient calibration
techniques, r=; :?ro and span
charts, too i-.irequent zero
and span of instruments,
improper utilization of
zero and span data.
Filters not conditioned properly,
humidity not monitored in
weiahinq room.
Half of the agencies would be
included.
Half of the agencies had improper
technique in regards to these areas.
(continued)
-------
TABLE 2 (continued)
Deficiency1*
Comment
10. NO-NO Converter
efficiency not checked.
11. Zero air not analyzed
for purity. ,
12. Substandard equipment or
laboratory facilities.
13. Ozone generators not
checked for stability.
14. Recorders not calibrated
or linearity checked.
15. Safety violations of a
serious enough nature to
mention.
16. High Volume sampling method
incorrect or flows or orifices
not Roots-meter-calibrated.
17. Too infrequent multipoint
calibrations,or calibrations
not done at the correct levels.
18. Poor permeation tube techniques;
no air flow while being stored.
19. No spare instruments for inter-
- changing.
20. N0? channel not calibrated
correctly, KI titration
errors; metal analysis
errors; manifolds dirty or
incorrectly constructed,
poor operator training,
filters handled improperly.
A deficiency was exhibited by
half of the agencies.
Fifty percent of the agencies
did not check their air.
Seven instances of such problem
areas were observed.
Five agencies did no stability
monitoring.
Five instances of this defect
were noted.
Five agencies had one or more
apparent violations of good safety
practice.
Five agencies had incorrect practices
in this category.
Deficiency noted in four agencies.
Three agencies were found to have
errors in this area.
Most agencies have no spares.
Several agencies had these or
similar defects noted in their
operations.
* For more detailed information on site instrumentation and/or methods, the
reader is directed to Site Evaluation reports available from EPA files.
7
-------
Many of the deficiencies noted are of a fundamental and critical nature;
other deficiencies are judgmental in nature and may or may not have impact on
the data. For example, a very common deficiency listed near the top of the
table is that gas cylinders, permeation tubes, and other standards are not
NBS traceable and are not checked independently by the agency. A non-traceable
standard cylinder is not necessarily or automatically bad, but the use of such
standards certainly decreases credibility and confidence in the data and raises
serious questions about data accuracy. Some of the deficiencies, such as a
lack of spare analyzers, affect data in the sense that spare instruments
minimize down time and increase the quantity of data taken. It is recognized
that the implementation of all of the recommendations, while desirable, may be
neither feasible nor practical because of budgetary or other limitations. We
are aware that some corrective procedures have been undertaken, but a complete
substantiation of the actions taken must await future reevaluation visits.
Such visits may be authorized at some future date at the discretion of the
Project Officer.
It must be emphasized that not all agencies and sites share in the defici-
encies mentioned in Table 2 to an equal extent. It should also be noted that
no agency had zero deficiencies. Therefore, strict numerical comparisons of
defects should be made with caution and by reference to the original reports.
The table does provide an accurate indication of the relative frequency of
types of deficiencies that are common to many of the agencies evaluated.
-------
SECTION 3
INTERLABORATORY PERFORMANCE SURVEYS
Regularly scheduled quality assurance laboratory performance surveys have
been carried out for the following analyses or measurements:
1) Sulfate/Nitrate Analysis
2) S02 Analysis
3) N02 Analysis
4) CO Analysis
5) Weighing Performance
6) High Volume Flow Rate Measurement
In addition to the laboratory surveys, a split sample program has been
carried out with one laboratory in which a large number of filters were analyzed
by the laboratory and Rockwell for trace metals and the results compared.
Table 3 lists all the surveys performed in the first year of the program
by date. Participating laboratories in each survey are indicated by a check
(/) mark.
For the chemical analysis survey (1-4 in the above list) Rockwell obtains
from commercial vendors, either directly or through EPA, multiple sets of the
appropriate samples which are then submitted for analysis to participating
laboratories. Ten replicate samples are first analyzed at Rockwell, and the
mean value of the analyses is by definition the "true" value with which the
results of all the participants are compared.
To assure the correctness of the Rockwell "true" value, several internal
and external quality control procedures are being used. Internal quality
control procedures in the Rockwell laboratory include comparison of analyses
of replicate samples, comparison of every new stock standard solution with the
old one, duplicate calibration curves before and after analysis, analysis of
quality control standards every 10 samples, and maintenance of routine quality
control charts on calibration parameters to establish laboratory control limits.
- 9 -
-------
TABLE 3. LIST OF INTERLABORATORY PERFORMANCE SURVEYS AND PARTICIPANTS
(*)
Agency
Albuquerque
Arizona State
C-a
C-b
Colorado State
Lawrence Radiation Lab
Lockheed
Montana State
Montana Power Co.
New Mexico State
North Dakota State
Northern Testing Lab
South Dakota State
Ua/Ub
Ute Research
Utah State
Wyoming State
Yellowstone County
Survey and Date by Quarter
SO^/NO"
4/76 1/77 2/77 3/77
• / / /
/ / / /
/ / / /
/ / / /
/ / / /
/ • / /
/ /
\
so2
4/76 2/77 3/77
/ /
/ / /
/
/ / /
/ / /
/ / /
/ / /
/ / •
/ / /
/ / /
N02
1/77 2/77 3/77
/ ' / /
/ / /
/ / /
/ / /
/ / /
/ / /
/ / /
/ / /
CO
4/76 1/77 3/77
/ / /
/ / /
/ / /
/ /
/ / /
^ / /
7
Weighing
1/77
/
^
/
/
^
/
/
/
•
K
/
^
/
•
"^
Hi-Vol
1/77
^
^
/
/
/
/
/
/
/
/
^
•^
•
/
^
"^
(*) A check (/) means that the agency participated in the survey.
-------
External quality control is provided by comparing the Rockwell "true" value
with the mean value obtained through replicate sample analysis either by the
QAB EMSL-RTP laboratory or by the vendor. This comparison is included in
each report.
The evaluation of laboratory performance is done in several ways. The
most direct way is to compare the concentration determined by each laboratory
with the "true" value as determined by Rockwell. Linear graphs are prepared
in which the laboratory values are plotted against the "true" values. The
slope, intercept, and correlation coefficient for each laboratory line are
then tabulated as indices of performance. Perfect agreement between a labo-
ratory and Rockwell for all analyses results in a straight line with unit slope,
zero intercept, and a correlation coefficient equal to 1.0000.
The slope of the line is a good measure of overall analytical accuracy,
provided that the intercept is small compared to the concentrations being
analyzed. For example, if the intercept is small, a slope of 1.05 implies
a level of agreement with Rockwell of 5%.
If the slope is near 1.0, a large intercept indicates a bias in the
analyses which might be caused by such errors as incorrect blank corrections
or contaminations in the water supply. The magnitude of the intercept must
be evaluated by comparison with the sample size. For example, if the sample
o 3
size is 10 yg/m , an intercept of 1 yg/m represents a 10% bias.
The correlation coefficient is a measure of laboratory precision. In
almost all cases the correlation coefficient is between 0.990 and 1.000, which
indicates good linear correlation between the laboratory and the Rockwell
analysis. Data that show a substantial amount of scatter lead to coefficients
below 0.990.
A second method used to evaluate laboratory performance is to determine
whether each laboratory analysis result falls within arbitrarily defined
concentration ranges denoted as "sample range" and "target range". The "sample
range" is intended to describe the variability in the analysis of presumably
- 11 -
-------
identical samples within one laboratory but at various times and under various
conditions; i.e., "normal laboratory operation". The "target range" is intended
to bracket an acceptable range of variability among different laboratories and
is larger than the "sample range".
Definitions of "sample range" and "target range" have not remained the same
throughout the program. For most of the surveys the "sample range" and the
"target range" were defined as R +. 3 a and R +_ 5 a, respectively, where R is the
"true" value (i.e., the mean value of Rockwell's analyses) and a is the standard
deviation of Rockwell's replicate analyses. For CO the standard deviation a is
extremely small so that a more practical definition is required. For CO the
"sample range" and "target range" were defined at various times as R +_ r where
r was either a constant ppm value (e.g., r = + 0.5 ppm), or a constant percent
of R (e.g., r = +_ 4%). Of course, the value of r was not the same for the sample
range and for the target range. In some of the earlier surveys for species other
than CO, the same arbitrary definition of range R j^ r was used, except that the
magnitude of r changed from one survey to the next.
Because the definitions of range depend on the performance of the Rockwell
laboratory through the experimentally determined value of a, great care should
be exercised in evaluating performance by means of the range criteria, partic-
ularly when comparing results from different surveys. The "sample range" and
"target range" criteria are most useful for comparing the performance of differ-
ent laboratories within the same survey.
Initially, brackets of + 3 a and +_ 5 a were arbitrarily defined. However,
since much data are now available, Rockwell is currently tabulating and analyzing
this data with the view to establishing new target range criteria based on a
statistical review of past performance by all the participating laboratories,
including Rockwell.
Each type of survey will be discussed separately below.
Sulfate/Nitrate Analysis
The sulfate/nitrate performance survey requires participating laboratories
to analyze a set of four to six filter strips spiked with varying amounts, of
sul fate and nitrate ions. At least 10 sets of filters presumed to be identical
- 12 -
-------
to those analyzed by the laboratories are analyzed by the Rockwell Chemistry
Laboratory using the methyl thymol blue procedure for SOT and the copperized
cadmium reduction method for NO^. In the interpretation of results no bias
or adjustment is made for the fact that laboratories included in the surveys
use methods different from Rockwell. All laboratories use EPA approved methods
which give presumably equivalent results, making such adjustments unnecessary.
Tables 4 and 5, taken from the 2nd quarter 1977 report, list the linear
parameters observed in the first three surveys of the contract year. To
preserve anonymity, the identities of the laboratories are not shown. (There
is no correlation between the identity codes in the various tables of this
report. Thus Agency A in one table is not necessarily the same as Agency A
in another table.) Tables 4 and 5 show that approximately one-third of the
slopes are in the range of 0.95 to 1.05 (indicating deviations of 5% or less),
and approximately one-third to one-half of the slopes are less than 0.90 or
greater than 1.10. Analytical problems are clearly indicated if the slope
falls outside the latter range.
According to Tables 4 and 5, approximately 70% of the analyses show
3 = 3 -
intercepts of less than 1 yg/m for SO. and less than 0.33 yg/m for NO-.
These intercepts should be evaluated with respect to the sample sizes, which
3 =3
were 30 yg/m or less for SO. and 10 yg/m or less for NO.,. A potential
analytical problem is indicated if the intercepts are significantly greater
than the values mentioned above.
The correlation coefficient was less than 0.990 for three SOT entries
and three NOl entries. Three of the six linearity problems involve one
laboratory, and two involve another laboratory.
SOpAnalysis
£
SOy samples consist of standard ampules containing freeze-dried solutions
of sodium sulfite and tetrachloromercurate. The "true" value for each sample
was obtained by the Rockwell laboratory by analysis of at least 10 replicate
samples. The usual quality control procedures were used for each analysis.
- 13 -
-------
TABLE 4, 'SUMMARY OF RESULTS OF S04 SURVEYS
Agency
Code Quarter/Year
A 4/76
1/77
2/77
B 4/76
1/77
2/77
C 4/76
1/77
2/77
D 4/76
1/77
2/77
E 4/76
1/77
2/77
F 4/76
1/77
2/77
G 4/76
1/77
2/77
EPA/QAB 4/76
1/77
2/77
Slooe
_
-
1.0054
0.8097
1.3508
1.1020
1.0538
1.0871
0.8692
0.82S2
1.0300
1.0458
1.0355
0.8952
0.9127
0.6431
1.0747
1.8603
0.9260
1.0313
0.7555
1.0327
1.0528
0.9457
Interceot
_
-
0.3327
1 . 0663
-2.9730
-1.2856
-0.0123
0.1471
0.5557
1.5705
0.0747
0.4330
-0.1941
0.7181
-0.5646
4.1915
0.8746
2.2721
0.0751
0.4593
0.9818
-0.1157
-0.1898
0.2899
Correlation
Coefficient
_
-
0.9994
0.9963
0.9939
0.9922
0.9964
0.9930
0.9972
0.9980
0.9843
0.9984
0.9995
0.9985
0.9989
0.9549
0.9876
0.9923
0.9951
0.9946
0.9942
0.9999
0.9998
0.9972
14
-------
TABLE 5. SUMMARY OF RESULTS FOR N0~ SURVEYS
Agency
Code
A
B
C
D
E
F
EPA /QAB
Quarter/Year
4/76
1/77
l]ll
4/76
1/77
2/77
4/76
1/77
2/77
4/76
1/77
2/77
4/76
1/77
2/77
4/76
1/77
2/77
4/76
1/77
2/77
SI one
.
-
1.3440
1.0267
1.0118
0.8835
1.0511
1.0250
0.7795
0.9411
0.8447
0.8778
1.2492
0.8781
0.9136
-2.1799
1.0497
2.0848
1.0178
1.0504
0.9479
Interceot
-
-0.06C4
0.3102
0.9563
0.3492
-0.0590
-0.3482
.2920
0.5791
0.2034
-0.0697
-0.2033
0.0175
-0.0054
44.6805
1.0345
0.0130
-0.0032
0.1670
-0.0378
Correlation
Coefficient
_
1.000
0.9981
0.9991
0.9996
0.9998
0.9770
.9589
0.9957
0.9968
0.9990
0.9952
0.9977
0.9905
-0.2331
0.9994
0.9998
1.0000
0.9999
0.9994
15
-------
Table 6, taken from the 3rd quarter 1977 report, summarizes the S02
results for the first three S02 surveys. In all but two entries the slope
was between 0.85 and 1.11, and the correlation coefficient was between
0.990 and 1.000.
The S00 sample concentration ranged from 16 to 227 yg/m . Intercepts of
o
greater than 10 yg/m are of concern, and they were found in about 23% of the
analyses.
NOp Analysis
N02 samples consist of standard ampules containing solutions of sodium
nitrite. The "true" value for each sample was obtained by the Rockwell labo-
ratory by analysis of at least 10 replicate samples. The usual quality control
procedures were used for each analysis.
Table 7, taken from the 3rd quarter 1977 report, summarizes the ML
results for the first three N02 surveys. In all but four entries the slope
was between 0.87 and 1.09. Samples in all surveys were 0.9 yg/ml or less;
therefore, intercepts greater than +0.04 yg/ml are of concern. Approximately
25% of the entries show intercepts in excess of these limits. The correlation
coefficient was below 0.990 in only three entries.
CO Analysis
3 3
CO samples are contained in 0.85 m (30 ft ) aluminum cylinders. Three
different concentrations were submitted for analysis in each survey. Replicate
analyses were performed in the Rockwell Quality Assurance laboratory using a
Bendix nondispersive infrared analyzer, and an NBS SRM cylinder containing
95.0 ppm CO in air was used as the standard. The "true" value was established
by analysis of at least five cylinders of each concentration.
The results of the first three surveys were for the most part very
satisfactory. Table 8 shows slopes and intercepts for each survey. Multiple
entries under each agency signify that the unknown cylinders were analyzed at
more than one site. As shown in the table the majority of the indicated slopes
are in the range 0.95 to 1.05, and the intercepts are usually less than 1 ppm.
16
-------
TABLE 6. SUMMARY OF RESULTS OF S02 SURVEYS
Agency
A
B
C
0
E
F
G
H
I
EPA/QAB
Survey
Quarter
4th '76
2nd '77
3rd '77
4th '76
2nd '77
3rd '77
4th '76
2nd '77
3rd '77
4th '76
2nd '77
3rd '77
4th '76
2nd '77
3rd '77
4th '76
2nd '77
3rd '77
4th '76
2nd '77
3rd '77
4th '76
2nd '77
3rd '77
4th '76
2nd '77
3rd '77
4th '76
2nd '77
3rd '77
Slope
1.3483
1 . 0973
1.0178
0.9627
0.9769
0.9226
1.0618
0.9311
0.9947
0.8757
0.6677
0.8883
0.9804
0.9708
0.9373
1.1123
1.0052
1.0316
0.9489
1.0130
0.9103
1.0995
0.8618
1.0114
1.0000
0.9781
0.9741
0. 9447
0.9798
Intercept
-10.9745
-8.2229
-10.4733
1.6888
0.2874
-0.6433
-5.9543
3.0098
-12.0752
-0.1165
7.4901
4.7308
2.4875
-2.0890
1.2077
-0.4824
-1.4080
0.7284
-8.9719
-14.0940
-5.9870
-15.6153
10.9976
0.5971
-2.4845
-4.8377
2.854
3.3206
-3.3785
Correlation
Coefficient
0.9605
0.9990
0.9935
0.9990
0.9997
0.9987
0.9963
0.9994
0.9999
0.9993
0.9590
0.9986
0.9995
0.9992
0.9989
0.9978
0.9997
0.9986
0.9999
0.9989
0.9967
0.9975
0.9911
0.9997
0.9990
0.9999
0.9994
0.9996
0.9S98
17
-------
TABLE 7. SUMMARYOF RESULTS OF NO,, SURVEYS
Agency
A
B
C
D
E
F
G
H
EPA/QAB
Quarter/
Year
1/77
2/77
3/77
1/77
2/77
3/77
1/77
2/77
3/77
1/77
2/77
3/77
1/77
2/77
3/77
1/77
2/77
3/77
1/77
2/77
3/77
1/77
2/77
3/77
1/77
2/77
3/77
Slope
1.0276
1.0000
1.0367
0.9457
0.9669
0.9326
1.2608
0.8787
0.9406
1.0085
0.9872
0.8198
1.0862
0.9955
0.9309
0.9176
0.9319
0.8409
0.9606
0.9757
0.9645
0.9807
0,7632
0.9093
0.9319
0.9237
0.9263
Mean Slope
and Rel .
Std.Dev.
All Surveys
1.0214
1.9%
0.9484
1.8%
1.0267
20,0%
0.9385
11.0%
1.0042
7.8%
0.8968
5.5%
0.9669
0.8%
0.8844
12.5%
0.9273
0.5%
Intercept
-0.0329
-0.0282
-0.0516
0.0230
-0.0290
0.0141
-0.0336
-0.0201
-0.0207
-0.0041
-0.0179
0.0475
0.1362
0.0250
0.0169
0.0070
0.0135
0.0498
0.0308
-0.0285
0.0190
0.0178
0.0515
0.0506
0.0101
-0.0071
0.0081
Correlation
Coefficient
0.9978
0.9998
0.9937
0.9992
0.9973
0.9996
0.9818
.9996
0.9904
0.9999
0.9979
0.9990
0.9699
0.9999
0.9997
1.0000
0.9990
0.9970
0.9993
0.9975
0.9988
0.9997
0.9260
0.9990
0.9992
0.9979
0.9995
18
-------
TABLE 8. SUMMARY OF RESULTS FOR CO SURVEYS
Quarter/Year
4/76
1/77
3/77
Agency
A
*•
B
C
D
E
F
G
Slope
1.0257
1.0257
1.0324
1.0324
1.0257
1.0394
1.0652
1.2002
1.0567
1.0453
0.9976
1.0433
0.9948
1.0389
0.9839
0.9696
0.9885
0.9860
0.9976
0.9745
0.9977
0.9855
1.0024
1.0234
0.9562
Intercept
0.9354
0.9354
-0.1100
-0.1100
0.9354
0.4687
-0.2943
-1.1073
0.0575
0.1744
-0.1519
-0.8187
-0.4227
-0.3520
0.5149
0.1609
-0.1849
0.6367
-0.9519
0.6536
-0.2198
-0.0198
0.1767
-1.1968
0.9692
Slope
1.0342
1.0471
1.0633
1.0445
1.0474
1.0467
1.0578
1.0603
1.0847
1.0921
0.9676
0.9676
0.9934
1.0020
0.9462
0.9571
0.9805
0.9656
1.0026
0.9941
0.9669
0.8871
0.9908
Intercept : Slope
0.4461
0.1439
0.0302
0.2044
-0.1975
-0.0146
-0.1075
0.0004
-0.4361
-1.0442
0.0054
0.0054
0.4011
-0.1996
-0.0036
-0.1308
0.2032
0.0026
-0.3653
0.7182
0.6882
0.7766
-0.1385
1.0607
1.0068
1.0272
1.0875
1.0332
1.0310
0.9665
1.1229
1.1290
1.0757
1.0945
1.0073
1.0672
0.9855
1.0017
1.0155
0.9858
0.9799
0.9909
Intercept
0.1377
-1.1833
1.0887
0.0086
-0.0318
0.0005
0.4786
-0.3214
-0.8475
-2.213
-1.405
0.5545
-1.466
0.2675
-1.0465
0.7633
0.2261
0.2982
-0.1266
19
-------
The good agreement between the various agencies and Rockwell is probably due
to the fact that agencies use cylinders for calibration that contain CO at
ambient levels and require no dilution. The CO concentrations supplied by
vendors are apparently sufficiently reliable for this particular analysis.
Weighing Survey
For this survey 12 sets of 3 "weights" each were purchased, each contain-
ing a 1, 2, and 5 g mass. All weights were modified slightly by filing away
some of the mass. After thorough cleaning, each weight was weighed several
times on two different balances (one capable of weighing to 0.01 mg and the
other a microbalance) over a period of several days. Each balance was cali-
brated just prior to use with an NBS certified set of Class S weights. One
set was sent to each participating agency, which then weighed the objects and
reported results to Rockwell. The objects were reweighed after return to
insure that no significant change had occurred during transit.
The results of the weighing survey are summarized in Table 9, which is
taken from the first weighing survey report. Table 9 shows differences between
the weight reported by the agency and the weight determined by Rockwell. Except
for three agencies, all measurements agree with Rockwell to within 1 mg or better.
The average deviation, excluding the three agencies showing large discrepancies,
is less than 0.3 mg.
At least one of the three agencies with large deviations has determined
that the problem was faulty balance operation and has taken steps to correct
the problem. It is not known what the other two agencies have done in regard
to this problem. A second weighing survey is scheduled to begin late in 1977-
Other weighing surveys will be repeated semi annually if they are deemed
necessary.
High Volume Flow Rate Measurement
One of the required tasks in the program is to conduct semi annual audits
of the flow rate calibration of high volume samplers as performed by the partic-
ipating agencies. For this purpose Rockwell submitted audit devices to partici-
pating agencies together with a test protocol. Agencies used the devices and
at the same time performed a normal calibration. Agencies reported their
measured flow rates as well as additional pressure and temperature data from
20
-------
TABLE 9. SUMMARY OF WEIGHING SURVEY
Agency
A
B
C
D
E
F
G
H
H (reweigh)
I (balance A)
I (balance B)
J
K
L
M
N
N (reweigh)
0
QAB EMSL/RTP
Deviations
ig
-0.3
-0.2
0.2
-0.4
-0.5
0.2
0.9
1.9
0.0
-0.1
-0.1
2.6
-0.3
-0.1
-0.2
-1.8
-1.7
0.1
-0.4
from Rockwell
2g
0.0
-0.2
0.1
0.5
-0.6
0.1
0.3
1.9
2.0
-0.1
0.0
2.8
-0.3
-1.0
0.0
-1.9
-1.6
-0.1
-0.3
in mg
5g
0.4
-0.2
-0.1
-0.5
-0.1
0.2
0.2
1.9
1.8
0.0
0.2
1.9
-0.2
-0.8
-0.3
-2.2
0.0
-0.1
-0.2
21
-------
which Rockwell calculated "true" flow rates. The results of the comparison
between values measured by each agency and calculated by Rockwell are summari-
zed in Table 10, This table gives each agency's calibration error, which is
defined as the percent difference between the indicated flow and the "true"
flow, measured at 1.132 m3/min (40 ft /min).
Significant discrepancies were found in several of the sites, and
specific recommendations were proposed in the report to study the causes
for the discrepancies. A basic problem appears to be a lack of definite,
uniform set of instructions, approved by EPA, which is specific for the
various types of calibration and measuring equipment used by the agencies.
One evidence of this is that some agencies report flows converted to standard
conditions while others report flow at ambient conditions. In many cases the
basis for the flow measurement is ambiguous, especially if no temperature or
pressure corrections are made. Further evidence of a lack of uniform instruc-
tions is the large number of errors that were discovered after the preliminary
survey results were sent to the agencies. Although it is possible that some
of the errors were caused by instructions from Rockwell which were not abso-
lutely clear, it is also apparent that operators are not all thoroughly
familiar with calibration procedures and the means for converting scale
reading to actual flow.
A specific problem in the high-volume audit occurred in two sites which
use flow controllers. In these sites the flow appeared to change drastically
when the different resistance plates were attached to the audit fixture.
Since the report was written, it was determined that this anomaly was an
artifact generated by the audit fixture, and an investigation is currently
under way to eliminate this problem in the next survey, which is scheduled
for late in 1977.
Analysis of Metal Samples From Laboratory A
One of the requirements of the Western Quality Assurance Program is to
duplicate approximately 10% of the analyses performed by a designated labora-
tory (denoted here as Laboratory A) for metals collected with particulates in
high-volume samplers. The method of analysis is atomic absorption, and the
elements that were originally listed included Be, Ca, B, Zn, Cd, Cr, Co, Cu,
Fe, Pb, Mn, Mo, and Ni.
- 22 -
-------
TABLE 10. SUMMARY OF RESULTS IN THE HIGH VOLUME SURVEY
Aqency
A
B
C
D
E
F
G
H
I
J
K-l to
K-5
K-6 to
K-10
. \
No.
'Reporting
Sites
13
1
1
1
1
1
3
1
1
1
5
5
Error %
Min. Max.
-21.2 9.2
-4.6
-2.2
-1.0
-11.4
XXXX
6.5 25.0
0.4
-0.9
-1.7
-12.8 -8.8
6.7 26.8
Remarks
Scatter between sites is of concern and
should be investigated further.
Agency's .calibration ~is satisfactory.
Agency 'srcalibrati on is good.
Agency Is^calTbrati on is excellent.
Calibration is fair.
Agency uses flow controller. Flow appeared
to change drastically when resistance plates
were changed. Anomalous effect subsequently
traced to unsuitable audit procedure.
Several errors in the measurement had to be
corrected. High intercept should be investi
gated further.
Agreement is excellent.
3
Agreement at 1.132m /mln is excellent.
However, slope and intercept are far from
ideal. May have a calibration problem.
Aqency ''^calibration- is excellent.
Agency 's calibration data are- internally
consistent among the" sites, -but all are
low by -abou1rlO% compared to Rockwell.
AgeiTcyhs-caTi-bration data are "internally
consistent among the siteSj but all "are ~~
high .compared td~ Rockwell. .An error in
reading of manometer is suspected.. Agency
concurs with this assessment.
(continued)
-------
TABLE 10. (continued)
Agency
L
M
N
0
P
No.
Reportinn
Sites
1
2
3
1
2
Error %
Min. Max.
-1.4
-8.4 -7.9
25.9 38.8
-0.1
XXX
<
Remarks
Agency's calibration is excellent.
Agency's calibration is satisfactory.
There is an apparent small constant
negative bias in the calibration data.
All agency's calibrations show a large,
positive bias. Problem is serious and
requires investigation.
Agreement i? .excellent.
Agency uses flow controller. F.l-ow appearec
to change drastically when resistance plate
were changed. Anomalous effect subsequent'
traced to unsuitable audit procedure.
24
-------
At least 3 sets gf samples were received from Laboratory A during the
first year of the contract. The first set was analyzed in December 1976, but
some difficulty was encountered in comparing results with those of Laboratory
A because of uncertainty concerning the history of the filters. Of all the
metals listed above, Rockwell could obtain valid data only for Cu, Fe, and Mn,
since the rest of the trace elements were below the detection limit. For the
three elements that were measured, plots of Laboratory A data vs Rockwell data
showed very large scatter.
Analysis of a second and third batch of filter samples disclosed the
following problems:
1. High and variable blanks for Cu, Fe, Mn, and Ni, suggesting
contamination in handling.
2. Concentration levels for Cd, Co, Cr, Mo, and Ni were either
below the Rockwell detection limit or at such low values that
analysis could not be performed with better than +_ 50% accuracy;
thus no comparisons could be made. In many cases the values
reported by Laboratory A were also very low.
3. Comparison plots for Cu, Fe, Pb^ and Mn show a large amount
of scatter.
The cause for the discrepancies between the Laboratory A and the Rockwell
analyses is not known. In order to investigate the origin(s) of the differ-
ences, a test program was designed, and, with the concurrence of the Project
Officer and Laboratory A, was initiated during September 1977. The test
consists of a sample exchange program in which synthetic metal samples prepared
by Rockwell will be analyzed by both Laboratory A and the Rockwell Chemistry
Laboratory. Three types of samples were chosen in order to investigate three
general sources of discrepancies: contamination, extraction, and analytical
procedures. The samples included are as follows:
25
-------
1. Uncut blank filters.
2. Spiked filters containing known concentrations of Cd, Co, Cr,
Cu, Fe, Mn, Mo, Ni, and Ph. The concentrations are chosen to
be significantly above detection limits.
3. Solutions of the above metals dissolved in 10% HNCL.
In addition to the above analyses, all of the liquid extracts analyzed by
Laboratory A will be returned to Rockwell for reanalysis.
Extensive reporting forms have been designed to obtain a large amount of
information from Laboratory A concerning its analytical and quality control
procedures. From an evaluation of the Laboratory A procedures and the results
of the analyses of all the samples, it should be possible to isolate and
perhaps identify problems that are traceable to contamination, extraction,
or analytical procedures. Results from this task will be made available in
the next report together with all pertinent forms and procedures.
26
-------
SECTION 4
FIELD AUDITS
As an integral part of the Western Quality Assurance Program, regularly
scheduled audit trips are conducted by Rockwell personnel to various sites
throughout the country; most of the sites are in the western states where
various energy sources are being developed. Table 11 shows a list of all the
sites audited and the dates of the audits.
The audit procedure consists of delivering to each analyzer five differ-
ent concentrations of analyte, including zero, using a portable audit device.
The output of the analyzer is measured with a digital voltmeter or with the
measuring device used in the station. The result of each audit is a five-
point response curve which is graphed and compared with the latest calibration
curve in use at the station. To distinguish between the audit curve and the
station calibration curve, we sometimes refer to the audit results as the
"calibration audit". It should be understood that the constants derived from
the audit curve are used by Rockwell, EPA, and the agency for evaluating the
station calibration. The "calibration audit" is not intended to give a substi-
tute calibration for use by the agency for data reduction. Differences between
the station calibration and audit results are given either as a percent calcu-
lated at full scale, or as differences in calibration and audit constants.
Where the differences are unusually large, an attempt is made to locate the
source of the problem.
The audit devices used in the audits are portable calibrators designed
and built at the Rockwell Air Monitoring Center. These audit devices are
capable of generating and delivering mixtures of various gaseous pollutants
in air at precisely known concentrations. Since flows are reproducible and
accurate to +1%, dilution of pollutants is known to +. 2% over a 3000 to 30
dilution range. Mixtures are produced either by dilution of gases contained
by cylinders, by means of an ozone generator, by gas phase titration, or by
dilution of gases diffusing from permeation devices. Standards are traceable
either directly or secondarily to the National Bureau of Standards (NBS).
(For further detailed information the reader is directed to the Standard
Appendix attached to any Audit Report.)
27
-------
TABLE 11. SUMMARY OF AUDITS PERFORMED BY ROCKWELL DURING THE PERIOD
AUGUST 1976 TO SEPTEMBER 1977-
Agency
Site
Date
C-a
C-b
State of Colorado
EPA/EMSL
EPA/Corvallis
L.L.L.
State of Montana
1
023
CARIH
CAMP
Greeley
Arvada
Colo. Springs
RTP
Col strip
Imperial Valley
Col strip
Billings
Laurel
C Hill
Highway Junction
Microwave
Saddle Mountain
Butte
Dec. 1976; April 1977; Aug. 1977
Dec. 1976; April 1977; Aug. 1977
Dec. 1976; April 1977; July 1977*
July 1977*
July 1977*
July 1977*
July 1977*
May 1977; Aug. 1977
Sept. 1977
March 1977; May 1977
Sept. 1976; June 1977; Sept. 1977
Sept. 1976; March 1977; June 1977;
Sept. 1977
Sept. 1976; March 1977; June 1977;
Sept. 1977
June 1977; Sept. 1977
June 1977; Sept. 1977
June 1977; Sept. 1977
June 1977
Sept. 1977
State of New Mexico
NOAA
State of North Dakota
Ua/Ub
State of Utah
State of Wyoming
Farmington
Water Tank
Shiprock
Reservation
Boulder
Stanton
Bismarck
A6
Price
Huntington
Salt Lake City
Magna
Bountiful
Ogden
Provo
Patrick Draw
Feb.
Feb.
Feb.
Feb.
Dec.
Dec.
Sept
Nov.
Nov.
Nov.
Sept
June
June
June
June
June
Nov.
1977
1977
1977
1977
1976
1976
5
9
April
April
1977;
1977;
July
Sept.
1977
1977
. 1977
1976
1976
1976
9
9
9
April
April
April
1977;
1977,
1977,
Aug.
Sept.
June
1977
1977
1977
. 1977
1977
1977
1977
1977
1977
1976
J
.
9
9
9
9
Sept.
Sept.
Sept.
Sept.
Apri 1
1977
1977
1977
1977
1 977 ;
Aug.
1977
* No audit report was issued for this audit because of technical problems.
28
-------
The portable audit devices are checked in the laboratory prior to and
at the conclusion of each audit trip." Tests include a multipoint flow cali-
bration check for the flow controller and a full performance test in which
the portable audit device is compared with a fixed laboratory calibrator.
The latter test is carried out as follows. The response of an analyzer in
ppm/volt is determined twice by performing consecutive one point calibrations
with the same calibration standards, using first the fixed calibrator and then
the portable device. Performance checks are made for NO, N02, 03, and S02-
In this way all functional components of the calibrator are checked. Ideally
the ratio of analyzer response measured with the two devices should be 1.00.
Since four independently calibrated flow controllers are involved, each having
a maximum uncertainty of +_ 1%, the portable calibrator is judged to be in good
order if the response ratio is between 0.96 and 1.04. Usually the observed
response ratio is between 0.98 and 1.02. A faulty solenoid valve, a leak in
the system, contamination in the lines, an erroneous flowmeter calibration,
or any other conceivable failure mode would be expected to cause the experi-
mentally determined response ratio to deviate significantly from 1.00, and
corrective action would then be necessary.
In addition to the audit of the continuous analyzers, checks are also
made on the high volume flow rate and on some of the meteorological instru-
ments. The high volume flow rate is checked using a calibrated orifice plate.
Each instrument audit results in a comparison between the station cali-
bration curve and the audit calibration curve. Normally, response is linear.
Where single point audits are made (e.g., Hi-vol sampler), the response is
assumed linear. If deviations from linearity or zero differences are observed,
they are specifically mentioned in the audit report.
Criteria for evaluating differences between station measurements and
audit results are not rigidly defined and have never been established. The
definition of "satisfactory" or "unsatisfactory" agreement is necessarily
arbitrary and depends on the intended use of the data as defined by both the
agencies involved and the EPA.
For the purposes of the Rockwell audits, the following working guidelines
have been adopted:
29
-------
Discrepancies of 10% or less at full scale require no specific
corrective action and may be considered "satisfactory".
Discrepancies in excess of 10% at full scale are of immediate
concern. In such cases, Rockwell will make a specific attempt
to locate the source of the problem and, if this is not practical,
likely sources of error will be identified and recommendations will
be made to the agencies for further study. A typical recommendation
of this type is to request the agency to recalibrate (or calibrate
for the first time) the analyzer in question.
An additional check on the Rockwell equipment and procedures is made by
means of the quarterly audits performed in the EPA/EMSL laboratory at Research
Triangle Park (RTP), N.C. These special audits are designed to compare
standards and procedures between Rockwell and EPA. During the first year
of the program, two audits of the EPA laboratory were made. Checks were
made for NO, NO^, 03, CO, CH., and SO^. Except for one ozone audit, in
which an analyzer operational problem was identified, differences at full
scale were always less than 5%.
During the first year of the program approximately 70 audits were
performed; hence, a very large amount of data has been collected. To evaluate
the performance of each agency, the individual audit available in EPA files
should be consulted. Table 12 gives a summary of overall performance by
instrument type. Table 12 shows the spread of the results in the column
labeled "Range of Discrepancies in %". Although the largest positive and
largest negative differences are quite dramatic, another measure of the
agencies' accuracy relative to Rockwell can be seen by examining the average
discrepancy calculated without regard to sign. Column 4 shows the percentage
of the audits that are considered satisfactory by virtue of the fact that the
discrepancy was 10% or less.
Table 12 demonstrates that of all continuous analyzers agencies are most
successful in calibrating CO, CH , and THC analyzers. In the case of CO, 86%
of all audits results showed full-scale differences of 10% or less. This
conclusion is consistent with the results obtained in the interlaboratory
performance survey for CO analysis. As indicated previously in Section 3
30
-------
TABLE 12. SUMMARY OF AUDIT RESULTS
**
(1)
Parameter
measured
NO
NOX
N02
°3
CO
so2
CH4
THC
(2)
Number
of audits
31
28
23
35
21
65
8
14
(3) (4)
Range of Percent of audits with
Differences in % 10% or less discrepancy
Max
78
97
36
27
57
84
8
103
Min.
-28
-34
-78
-30
-11
-70
-12
-6
. *
Ave
16.7
16.5
18.2
9.8
8.6
14.5
6.9
13.9
55
50
43
66
86
57
75
71
* Average discrepancy without regard to sign
** Further breakdown of these data by site, agency or station is given in
the audit reports for various agencies. All reports are available in
EPA files.
31
-------
and Table 8, the majority of CO cylinder analyses were accurate to + 5%. The
obvious explanation for this observation is that most agencies calibrate these
instruments by means of span cylinders which require no dilution and a minimum
of handling. In most cases each site has its own span cylinder; hence, cali-
brations are done frequently.
The most unsatisfactory results are found in the audit of N09, NO , NO,
£ X
and SO- analyzers. To calibrate these instruments, portable calibrators are
normally used to prepare dilutions from concentrated mixtures or from permea-
tion devices. Portable calibrators are generally in short supply, and their
use requires personnel with some degree of skill and experience. Conclusions
that could be drawn from the results in Table 12 are as follows:
(1) Greater emphasis should be given to checking the integrity
of the portable calibrator.
(2) Calibration of these analyzers should be performed more
frequently.
(3) Greater care should be taken to insure that any given operator
is familiar with calibrator operation and its correct use.
32
-------
SFCTIQN 5
TECHNICAL ASSISTANCE
In accordance with the provisions of the contract, Rockwell has provided
technical assistance to the various participating agencies, as requested by
the Project Officer. Some of the assistance has been relatively minor, con-
sisting of instruction manuals, procedures, or literature references relating
to chemical analysis and quality assurance sent to agencies on request. Much
of the technical assistance in air pollution monitoring was given informally
during site evaluation and field audit trips when extensive discussions and
information exchange between Rockwell and agency personnel took place.
The following is a list of major interactions with various agencies during
which Rockwell provided technical assistance. The identity code in this section
is not necessarily the same as that in other sections (Agency A here is not
necessarily the same agency as Agency A mentioned in other sections).
AGENCIES ASSISTED
Agency A
Assistance was given Agency A in modifying its trailer and monitoring
procedures. (October 1976. )
Agency B
Personnel from Agency B visited Rockwell to discuss quality assurance
procedures, AA analysis procedures, and use of the Technicon Analyzer for
analysis (November 1976 and September 1977).
Agency C
Information was supplied to Agency C concerning QA procedures in air
monitoring programs, including calibration procedures and calibration standards.
Assistance was also provided to establish a quality assurance program in the
analytical chemistry laboratory (November 1976).
33
-------
The site evaluation report and the first audit results were discussed
with legal personnel from the Agency. Additional discussions were held with
air monitoring personnel concerning quality assurance procedures (January 1977).
Written procedures for the analysis of SOT, N0~, and arsenic were sent to
the agency's laboratory upon request (April and August 1977).
Split sample filters were received from Agency C and analyzed for trace
metals. The comparison with the agency's analysis disclosed good agreement
for Cd, Cu, Fe, Pb, and Zn but not for Al (June and July 1977).
Information was given to the agency's chemistry laboratory to assist it
in setting up a Technicon analyzer and to develop the analytical procedures
for arsenic (June and July 1977).
A set of class S weights was purchased and delivered to the agency's
laboratory (July 1977).
Agency D
A cylinder containing approximately 15900 ppm CO was analyzed by compari-
son with an NBS standard. The CO cylinder is being used as the agency's CO
standard (April 1977).
A detailed set of instructions for performing ozone calibrations by means
of the gas phase titration technique was written and sent to Agency D (May
1977).
Agency E
A Bendix portable calibrator belonging to Agency E was checked in the
Rockwell quality control laboratory. Several major plumbing deficiencies were
discovered, and specific recommendations were made to rectify the problems
(May 1977).
Written instructions describing analysis of arsenic and selenium, as well
as filter extraction procedures, were sent to Agency E upon request (June 1977),
34
-------
Agency F
A detailed set of instructions for performing ozone calibrations by
means of the gas phase titration technique was written and sent to Agency F
(May 1977).
A study was performed in the Rockwell quality control laboratory to
develop the procedures for calibrating ozone analyzers by means of the
reverse gas phase titration technique. These procedures are needed because
many of the stations operated by this agency do not have a NO-NO analyzer;
/\
thus, the normal gas phase titration technique cannot be used (August 1977).
Two cylinders containing CH. in air were analyzed by comparison with an
NBS standard and sent back to Agency F. The cylinders are used by the agency
as hydrocarbon standards (July 1977).
A plumbing modification was designed and plans were sent to Agency F
to help it improve its CO and HC sampling system. The modification is
desirable to reduce waste of calibration gas and to improve audit capability
(August 1977).
Agency G
Agency G requested information on SOp analyzers, calibration equipment,
data reduction hardware, and meteorological instrumentation. The requested
information was sent to the agency (July 1977).
A set of class S weights was purchased and delivered to the agency's
laboratory (July 1977).
Agency H
Agency H requested assistance in setting up a new Technicon analyzer.
Assistance was given in telephone conversations between Agency H and Rockwell
laboratory personnel (August 1977).
Agency I
A special study was conducted in August 1977 in the Rockwell quality
control laboratory with personnel from Agency I to determine the cause for
the large discrepancies in S02 calibrations found in field audits of stations
35
-------
operated by Agency I, A procedural error in the calibration method used
by Agency I was found which largely accounted for the audit discrepancies.
Agency I has subsequently modified the procedure to eliminate the error
(August 1977), as evidenced by subsequent field audit results.
EPA/EMSL, RTP
A special joint study was conducted at EPA's request in the Rockwell
quality control laboratory with Mr. Ken Rehme of the EPA/EMSL/RTP laboratory
to evaluate a new procedure for calibrating ozone analyzers based on a boric
acid/KI absorption solution. The boric acid/KI procedure was compared with
UV photometry and gas phase titration. Good correlations were obtained
among the three methods (September 1977). It is understood that further
work is underway under Mr. Rehme's direction to align these results with
current agency-EPA policy regarding the gas phase titration technique.
U.S. GEOLOGICAL SURVEY WATER QUALITY FIELD STATIONS
One activity was added-on to the contract toward the end of the first
year, an on-site inspection and evaluation of seven water quality field
stations of the U.S.G.S. in the states of Arizona, Colorado, Montana,
New Mexico, North Dakota, Utah, and Wyoming. The visits have been completed,
and reports are currently being written (September 1977).
GENERAL REVIEW MEETING
A meeting of representatives from the participating state and private
agencies, EPA, and Rockwell was held in Denver on April 26-27, 1977, to
review the Western QA program. The main speakers and topics were as follows:
S. Bromberg, EPA/EMSL/RTP. Introduction and Purpose of Meeting
R. Williams, EPA/Region VIII. General Remarks about the Purpose
of the Program
G. D'Alessio, EPA/Washington. Energy Overview -- National
R. Snelling, EPA/EMSL/Las Vegas. Energy Overview — Western States
M. Rinaldi, State of New Mexico. Energy Overview — Regional Office
T. Thoem, EPA/Region VIII. Energy Overview -- Regional Office
E. Parry, Rockwell International. Laboratory Evaluation Discussion
G. Colovos, Rockwell International. Laboratory Audit Discussion
M. Cher, Rockwell International. Field and Laboratory Audit Discussion
36
-------
In addition to hearing the above speakers, the entire group participated
in general discussions on all subjects concerning the program.
37
-------
SECTION 6
PLANS FOR NEXT YEAR
The general plans for next year are to continue the current program of
laboratory performance surveys, field audits and technical assistance. Since
the ultimate purpose of the contract is to help the participating agencies in
improving the quality of their data base, some modifications in the program
will undoubtedly become necessary. For example, it is anticipated that techni
cal assistance efforts may be increased in order to help correct difficulties
found during performance surveys and field audits. If a history of good per-
formance warrants the phasing out of certain tasks, this ought to be done in
order to concentrate efforts on specific trouble spots.
To accomplish the desired goal of the program, constant reevaluation of
the program is required by Rockwell, the Project Officer, the EPA regional
offices, and the participating agencies. Changes in the program will be
effected whenever it seems appropriate, given the consent and cooperation
of the Project Officer and all interested parties.
38
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
REPORT NO.
2.
3. RECIPIENT'S ACCESSION-NO.
TITLE ANDSUBTITLE
QUALITY ASSURANCE IN SUPPORT OF ENERGY RELATED
MONITORING ACTIVITIES IN THE WESTERN UNITED STATES
Annual Report No. 1
5. REPORT DATE
July, 1978
6. PERFORMING ORGANIZATION CODE
AUTHOR(S)
Mark Cher
8. PERFORMING ORGANIZATION REPORT NO.
AMC8303.135AR
PERFORMING ORGANIZATION NAME AND ADDRESS
Rockwell International
Atomics International Division
Air Monitoring Center
Newbury Park, California 91320
10. PROGRAM ELEMENT NO.
11. CONTRACT/GRANT NO.
Contract No. 68-02-2412 .
12. SPONSORING AGENCY NAME AND ADDRESS
U.S. Environmental Protection Agency
Quality Assurance Branch
Environmental Monitoring and Support Laboratory
Research Triangle Park, North Carolina 27711
13. TYPE OF REPORT AND PERIOD COVERED
14. SPONSORING AGENCY CODE
EPA/ORD
15. SUPPLEMENTARY NOTES
16. ABSTRACT
This report describes and summarizes the activities during the first year of
the program named above. The activities described are part of a continuing 5-year
program.
The purpose of the program is to establish a quality assurance data base for
ambient air monitoring in specified geographical areas around present and proposed
energy development projects, and to provide technical assistance to enable existing
monitoring networks to achieve a high level of data quality. An initial on-site
review of 18 laboratories and associated field sites was completed. Regularly
scheduled laboratory performance surveys are being carried out for the analysis
of sulfate, nitrate, SOoj NO?, and CO and for weight measurement and high volume
flow rate. Approximately 10% of the analysis performed by a specified laboratory
for metals collected in high volume filters are being repeated in the Rockwell
laboratory. Quarterly field audits are being conducted at specified monitoring
sites. Technical assistance has been provided to participating monitoring groups,
as requested by the Project Officer.
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.lDENTIFIERS/OPEN ENDED TERMS
COSATI Field/Group
Quality Assurance
Ambient Air Monitoring
Site Evaluations
Laboratory Audits
Field Audits
Technical Assistance
8. DISTRIBUTION STATEMENT
RELEASE TO PUBLIC
19. SECURITY CLASS (ThisReport)
UNCLASSIFIED
21. NO. OF PAGES
44
20. SECURITY CLASS (Thispage)
?ITY CLASS (Thispag
UNCLASSIFIED
22. PRICE
EPA Form 2220-1 (9-73)
------- |