vvEPA
              United States
              Environmental Protection
              Agency
             Atmospheric Research and
             Exposure Assessment Laboratory
             Research Triangle Park, NC 27711
EPA/600/3-90/004
February 1990
              Research and Development
National Performance
Audit Program
Acid Rain Audits
1988

-------
                                                EPA/600/3-90/004
                                                February 1990
          NATIONAL  PERFORMANCE AUDIT PROGRAM
                    ACID RAIN AUDITS
                        -1988-
        Robert L. Lampe and William J. Mitchell
               Quality Assurance Division
Atmospheric Research and Exposure Assessment Laboratory
     Research Triangle Park, North Carolina  27711
ATMOSPHERIC RESEARCH AND EXPOSURE ASSESSMENT LABORATORY
           OFFICE  OF RESEARCH AND  DEVELOPMENT
          U.S.  ENVIRONMENTAL PROTECTION AGENCY
     RESEARCH TRIANGLE PARK, NORTH CAROLINA  27711

-------
                                    NOTICE

      This document has  been reviewed in accordance with U.S.  Environmental
Protection Agency policy and approved for publication.   Mention of trade names
or commercial products does not constitute endorsements  or recommendation for
use.

-------
                                   ABSTRACT

      This report  presents results of  the Environmental  Protection Agency's
Semiannual U.S. Acid Rain Audit Program  for 1988.  When these audits, which are
a part of the National Performance Audit Program (NPAP), were initiated in 1981,
participation was limited to laboratories routinely analyzing acid rain samples.
However, in 1988 invitations to participate  were issued to  all NPAP participants.
As a  result,  only 52% of  the  results  received for the  1988 audits  came from
laboratories that had been regular participants in the previous years'  audits.

      Despite  the  many  new participants,  the  percentage  of participants  who
analyzed for a specific analyte was similar to that from previous audits*.  Also,
as in previous  audits, many participants had difficulty measuring accurately the
concentration of the lowest cation and anion concentrations.  Not unexpectedly,
the percentage of  the reported  results  identified as  outliers  was larger than
in the 1985-1987 audits.   The  results for the heavy metals  (Mn, Fe, Cd,  Cu,  Ni,
Pb and  Zn)  were similar  to earlier audits  in that most  of  the  participants
reported an average result (for all seven metals) close to the expected value.
                                       iii

-------
                                   CONTENTS




                                                                          Page




Abstract 	,	    iii




Tables	      v




Acknowledgments	     vi




Introduction	      1




Procedure	      2




Results and Discussion	      3




References	      5
                                      v

-------
                                    TABLES




Number                                                                     Page
  1	        6




  2	        7




  3	        8




  4	        9




  5	       10




  6	       11




  7	       12




  8	       13




  9	       14




 10	       15
                                       VI

-------
                               ACKNOWLEDGMENTS

      Recognition is due  to the technical staff of NSI Technology Services Inc. ,
Research Triangle Park, NC, who produced the high quality samples used in these
audits.  Also,  we thank the staff of Global Geochemistry Inc., Canoga Park, CA,
the referee laboratory.
                                      vii

-------
                                   SECTION  1

                                 INTRODUCTION

      The  National  Performance  Audit  Program  (NPAP)  is  conducted  by  the
Atmospheric Research and Exposure  Assessment Laboratory  (AREAL)  of  the  U.S
Environmental Protection  Agency (EPA)  at  Research Triangle  Park,  NC.   NPAP
provides EPA management a means  to  assess  the analytical proficiency of state
and local  laboratories.   The acid  rain audits,  which are a part of the NPAP
program,  are  open  to  governmental,  academic  and commercial  organizations.
Participation is voluntary.

      When our acid rain audit program began in 1975 it was restricted  to members
of the World Meteorological Organization (WHO).  In 1981, U.S. participants of
the NPAP who were known to be routinely analyzing acid  rain samples were invited
to participate  in an acid  rain audit  conducted only for  U.S.  laboratories.
Thirty-two laboratories accepted the invitation and 18 actually reported their
results.

      In 1988, participation in the U.S. acid rain audit program was opened to
all NPAP participants.  One hundred  and fourteen agencies accepted the invitation
for the May 1988 (0588) audit  and  118  accepted the invitation for the October
1988 (1088) audit.  However,  only 57  of these laboratories  (50%) returned results
for the 0588 audit and  only  51  (43%)  returned results for the 1088  audit.   We
believe these low  levels  of  participation  in the 1988 audits  resulted because
many agencies that requested to participate were unfamiliar with the extensive
analyses required for precipitation samples.  When they received the samples they
decided  not to analyze  them.    Only 52%  of the  results reported  came from
laboratories who had been regular participants in the previous audits.

-------
                                  SECTION 2

                                  PROCEDURE

      The five samples used in the audit are contained in polyethylene containers
to improve sample stability.  The participating  laboratories  analyze three of
the samples for pH, conductivity, acidity and  the  cations  and anions normally
measured  in  precipitation samples.   The other  two  samples,  which  are  acid
stabilized to prevent loss of metals from the solution,  are analyzed for heavy
metals.

      The samples  are  prepared using  formulations developed by  the National
Institute of  Standards  and Technology (NIST) and  certified by  an independent
analytical (referee) laboratory using NIST reference materials.  The participants
analyze the samples using  the analytical procedures they normally employ when
analyzing their precipitation samples.   Each  laboratory is shipped  a  set of
samples consisting of one  sample  from each  of the five  series.   When diluted
1:50, each sample simulates a precipitation sample.

      The participants  dilute each sample, analyze  it  in triplicate and average
the results  for  each analyte.  They also analyze  at least one blank  of the
distilled water they used to dilute the samples to correct their results for any
contribution from the distilled water.   All  results are  then  forwarded to EPA
for summarization and evaluation.   Each participant is then sent a report that
compares his results to  the expected values for each analyte.   At the end of the
year, this summary report is prepared so that all  participants  can see how their
performance compared to that of the other laboratories.

-------
                                   SECTION 3

                            RESULTS AND DISCUSSION

      The data and statistics  in this report are descriptive of the quality of
 the  analytical data  currently being reported by agencies measuring acid rain.
 At present,  there  are no federal regulations specifying the accuracy required
 in the analysis of acid  rain  samples.  Therefore, we recommend that individual
 laboratories evaluate  their progress and  improve their analytical capabilities
 as they feel necessary based  on their own results and the averages reported in
 this report for the  0588 and  1088 acid rain audits.

      Less than 35% of the 1988 audit participants  analyzed  all samples  for all
 the  analytes.  The approximate percentage of participants who reported  results
 for  each  analyte or  class  of analyte  (minor  ions,  heavy  metals)  are  pH and
 conductivity, 94%; acidity,  26%;  sulfate and nitrate, 69%; minor anions (F, Cl),
 31%; minor cations (NH4,  Ca, Mg,  Na, K), 46%; and heavy metals (Mn,  Fe,  Cd,
 Cu, Ni, Pb, Zn),  33%.  These percentages are generally lower than those  for the
 previous audits.

      The  number of  audit  participants  is small  and   less  than  50%  of the
 participants analyze for the minor anions (Cl,  F),  cations (NH^, Ca, Mg, Na, K)
 and heavy metals (Mn,  Fe, Cd, Cu, Ni, Pb, Zn).  Further, the concentrations of
 many  of the  analytes are  near  or below  the quantification  limits  of the
 analytical  methods,  so  many participants  report  their  results  to  only one
 significant  figure  (despite  our request to report  them  to  two  significant
 figures).   For the above reasons,  it is difficult to obtain  a sufficient number
 of results on which to perform an in-depth statistical analysis.   For example,
 suppose a laboratory obtains a value of 0.14 mg per L for Na  when the true value
 is 0.15 mg per L but only reports a value of 0.1 mg per L.  The apparent error
 is -33% but in actuality it is only -7%.

      Also,  in previous audits,  we had  observed  that  many  participants had
 relatively large positive and negative percent differences  for analytes in the
 same general class  (anions,  cations,  metals) even  though  the same analytical
method was  used  to  determine all  the  analytes  in that  class.   It  was  also
 observed that participants who had participated in many of  the previous audits
 tended  to  show  large variations  in  the magnitude  and  size of  the   percent
 difference for the same analyte from one audit  to the next.  This variation was
 quite randomly distributed.

      To remove as much  of  these  random variations  as  possible from the data,
 it was decided to calculate a percent  difference by lavel for each laboratory
 for each class of analyte rather  than  calculate  a  percent difference for each
analyte within each  class of analyte (the approach used in earlier reports).
This was accomplished by determining the  sum of the concentrations for  all the
analytes in a specific class for  the expected and for the reported values.  The
difference between the two sums was then determined and the percent difference
 --(reported-expected) 100/expected--was calculated.

-------
      In  the  case  of  the  minor  anions  and  the  minor  cations  almost  all
participants who  reported results for one  analyte  in these  classes  reported
results for all  the analytes in these classes.  Thus, the percent difference for
each participant for each analyte class was used to calculate the average percent
difference for that analyte class, i.e.,  cation,  anion.

      However,  because  a  significant number of the participants who  reported
results for the metals did not  analyze  for all  seven  of the metals,  it was
decided to  report each participant's  average  percent  difference  and  not to
combine them to obtain a grand average.

      Tables 1 and 3 present the  results reported (n) for all analytes except
the heavy metals  for  the  0588 and  1088  audits,  respectively. Tables  2 and 4
present the results for thes^e two audits after outliers  were  removed  using an
outlier rejection method  that is  based on  the  standard  deviation historically
obtained for these audits.   This procedure considers all percent differences that
exceed 3 times the historical standard deviation  (10% for these  audits) to be
outliers (1).   For information purposes, the concentrations of the individual
anions and cations in the  1988 samples are  presented in  Table  5.

      Tables 6  through 9 present the results for the  0588  and 1088 audit samples
that contained the heavy  metals.  Tables 6  and  8  present the  results  reported
by all participants for the 0588  and 1088  audits, respectively.   Tables 7 and
9 present the results  after outliers were  removed.   For information purposes,
the concentrations of the  individual metals in each  0588 and 1088 audit sample
is presented in Table 10.

      A comparison  of  Tables  1  and 2  and a comparison of  Tables  3  and 4
demonstrate that  quite a few participants encountered  difficulty  measuring
acidity and the cations and anions.   The number of outliers reported is larger
than those reported for 1985,  1986  and 1987.   It is quite possible  that this
increase resulted because many laboratories were analyzing this  type  of synthetic
precipitation sample for the first time.   Also,  the  lower the concentration of
the analyte, the higher percentage of the results that were rejected  as outliers.
This  is  consistent  with  the  previous  audits.    Unfortunately, a  majority of
precipitation samples contain ions at these levels.

      Inspection of the average results for  1985, 1986 and 1987  (2,3,4) and those
in this report  for the  analytes most frequently measured show no trend in the
performance obtained from audit to audit.

      Inspection  of  the  individual  laboratory results  for   the heavy metals
(Tables 6,7,8  and 9) shows that most laboratories obtained results that agreed
very  well  with the expected  value  in both 1988 audits.   When  a laboratory
reported results that  differed markedly  from the  expected value  in one audit,
it generally showed good agreement between the expected and reported values in
the other 1988 audit.

-------
                                  REFERENCES

1.    Duncan,  A.J.   Quality Control and Industrial Statistics.   R.D.  Erwin and
      Co.,  Inc.,  1967.   P.  89.

2.    Parr, B.F., R.L.  Lampe, G. Pratt, E.T. Hunike ndW.J. Mitchell.   National
      Performance Audit Program.  Ambient Air Audits of Analytical Proficiency,
      1987.   U.S.   Environmental  Protection Agency  Report EPA  600/3-89-006.
      Research Triangle Park,  North Carolina 27711.   January 1989.

3.    Parr, B.F. , R.L.  Lampe, G. Pratt, O.L.  Dowler andW.J. Mitchell.  National
      Performance Audit Program.  Ambient Air Audits of Analytical Proficiency,
      1986.   U.S.   Environmental  Protection Agency  Report EPA  600/4-87-038.
      Research Triangle Park,  North Carolina 27711.   December 1987.

4.    Parr, B.F., R.L.  Lampe, G. Pratt, O.L.  Dowler andW.J. Mitchell.   National
      Performance Audit Program.  Ambient Air Audits of Analytical Proficiency,
      1985.   U.S.   Environmental  Protection Agency  Report EPA  600/4-87-002.
      Research Triangle Park,  North Carolina 27711.   January 1987.

-------
Table 1.  Results Reported For All Analytes Except
             Heavy Metals—0588 Audit
Audit Analyte
0588 pH

Conductivity
,i S/cm
Acidity
ueq/L
804
mg S/L
N03
mg N/L
Anions (Gl, F)
mg/L
Cations
(NH4, Na, Ca,
Mg, K)
mg/L
Sample
1
2
3
1
2
3
1
2
3
1
2
3
1
2
3
1
2
3
1
2
3
N
52
52
52
48
48
48
17
17
17
38
40
40
38
40
40
19
21
20
22
25
22
Average
Reported
Value
4.37
3.53
3.55
20.98
148.00
133.00
56.82
262.00
268.00
0.80
4.48
3.37
0.38
2.24
1.89
0.43
3.03
1.45
0.50
4.00
1.15
Expected
Value
4.40
3.50
3.52
16.50
156.00
130.00
41.10
312.00
292.00
0.66
3.91
2.89
0.11
2.14
1.81
0.33
3.21
1.42
0.42
3.86
1.02
% Diff.
-0.68
0.86
0.85
27.15
-5.13
2.30
38.24
-16.02
-8.22
21.21
14.58
16.61
245.45
4.67
4.42
30.30
-5.61
2.11
19.05
3.63
12.75

-------
Table 2.  Table 1 Results After Outliers Removed
Audit Analyte
0588 pH


Conductivity
}i S/cm

Acidity
Meq/L

S04
mg S/L

N03
tag N/L

Anions (Cl, F)
mg/L

Cations
(NH4, Na, Ca,
Mg, K)
mg/L
Sample
1
2
3
1
2
3
1
2
3
1
2
3
1
2
3
1
2
3
1
2
3

N
51
52
52
42
43
46
5
12
11
31
36
35
34
38
38
11
19
19
18
24
18

Average
Reported
Value
4.40
3.53
3.55
18.44
147.00
127.00
40.14
342.00
319.00
0.66
4.02
2.96
0.11
2.13
1.81
0.33
3.19
1.49
0.45
3.93
1.10

Expected
Value
4.40
3.50
3.52
16.50
156.00
130.00
41.10
312.00
292.00
0.66
3.91
2.89
0.11
2.14
1.81
0.33
3.21
1.42
0.42
3.96
1.02

% Diff.
0.00
0.86
0.85
11.76
-5.76
-2.31
-2.34
9.62
9.25
0.00
2.81
2.42
0.00
-0.47
0.00
0.00
-0.62
4.93
7.14
1.81
7.84


-------
Table 3.  Results Reported For All Analytes  Except
             Heavy Metals—1088 Audit
Audit Analyte
1088 pH


Conductivity
iiS/cm

Acidity
ueq/L

S04
ng S/L

N03
mg N/L

Anions (Cl, F)
mg/L

Cations
(NH4, Na, Ca,
Mg, K)
mg/L
Sample
1
2
3
1
2
3
1
2
3
1
2
3
1
2
3
1
2
3
1
2
3

N
49
49
48
43
43
43
11
11
11
34
33
35
35
34
34
16
13
15
27
26
26

Average
Reported
Value
3.94
4.25
3.71
62.37
23.94
86.64
116.00
64.78
192.00
2.55
1.07
4.80
1.04
0.14
0.14
1.29
0.38
0.66
2.46
0.47
1.09

Expected
Value
3.91
4.27
3.68
66.14
24.08
96.43
124.00
53.67
209.00
2.11
0.90
4.01
0.87
0.13
0.11
1.27
0.32
0.43
2.38
0.41
0.99

% Diff.
0.77
-0.47
0.82
-5.70
-0.58
-10.15
-6.45
20.70
-8.13
20.85
18.89
19.70
19.54
7.69
27.27
1.58
18.75
53.48
3.36
14.63
10.10


-------
Table 4.  Table 3 Results After Outliers Removed
Audit Analyte
1088 pH


Conductivity
,j S/cm

Acidity
lieq/L

S04
mg S/L

N03
mg N/L

Anions (Cl, F)
mg/L

Cations
(NH4> Na, Ca,
Mg, K)
mg/L
Sample
1
2
3
1
2
3
1
2
3
1
2
3
1
2
3
1
2
3
1
2
3

N
49
48
48
42
41
39
8
6
8
29
30
30
32
29
31
13
8
11
27
22
21

Average
Reported
Value
3.94
4.30
3.71
62.93
24.85
93.53
132.00
53.43
235.00
2.12
0.89
4.05
0.87
0.12
0.11
1.32
0.32
0.45
2.46
0.43
1.04

Expected
Value
3.91
4.27
3.68
66.14
24.08
96.43
124.00
53.67
209.00
2.11
0.90
4.01
0.87
0.13
0.11
1.27
0.32
0.43
2.38
0.41
0.99

% Diff.
0.77
0.70
0.82
-4.85
3.20
-3.00
6.45
-0.45
12.44
0.47
-1.11
1.00
0.00
-7.69
0.00
3.44
0.00
4.65
3.36
4.88
5.05


-------
      Table 5.  Concentrations of the Anions and Cations in Each
                           1988 Audit Sample8


Cl
F
NH4
Na
Ca
Mg
K

Sample 1
0.280
0.051
0.078
0.186
0.047
0.021
0.084
0588
Sample 2
2.738
0.474
0.835
1.771
0.388
0.063
0.798

Sample 3
1.254
0.167
0.471
0.393
0.038
0.042
0.078

Sample 1
1.078
0.191
0.335
1.340
0.115
0.068
0.524
1088
Sample 2
0.283
0.039
0.080
0.179
0.047
0.028
0.073

Sample 3
0.349
0.084
0.610
0.235
0.046
0.017
0.077
Expressed as mg analyte per L except for NH^ which  is mg N/L.
                                  10

-------
             Table 6.   Heavy Metal Results for Sample  4 of 0588 Audit
Laboratory
317003
320001
334001
335002
336001
339001
342002
345001
347003
406001
412001
436007
501001
506012
508016
517008
525009
529005
542010
548005
Na
7
7
7
7
7
7
7
7
7
4
7
7
7
5
7
7
6
7
7
5
Reported
Valueb
0.97
0.94
0.96
0.92
0.97
0.96
1.02
1.04
0.92
0.78
0.40
1.00
1.01
0.85
0.99
0.93
0.82
0.95
0.99
0.80
Expected
Value
0.96
0.96
0.96
0.96
0.96
0.96
0.96
0.96
0.96
0.82
0.96
0.96
0.96
0.83
0.96
0.96
0.85
0.96
0.96
0.83
% Diff.
1.04
-2.08
0.00
-4.17
1.04
0.00
6.25
8.33
-4.17
4.88
-58.33
4.17
4.99
2.41
3.13
-3.13
-3.53
-1.04
3.13
-3.61
reported.
     aN - the number of metals for which results were reported.




          average concentration (mg per L) for  all metals  for which results were
                                       11

-------
            Table  7.  Heavy Metal Results for Sample 5 of 0588 Audit
Laboratory
317003
320001
334001
335002
336001
339001
342002
345001
347003
406001
412001
436007
501001
506012
508016
517008
525009
529005
542010
548005
Na
7
7
7
7
7
7
7
7
7
2
7
7
7
5
7
7
6
7
7
5
Reported
Valueb
0.33
0.34
0.33
0.32
0.33
0.34
0.36
0.40
0.32
0.17
0.32
0.39
0.35
0.26
0.35
0.32
0.26
0.33
0.36
0.26
Expected
Value
0.33
0.33
0.33
0.33
0.33
0.33
0.33
0.33
0.33
0.17
0.33
0.33
0.33
0.26
0.33
0.33
0.27
0.33
0.33
0.26
% Diff.
0.00
3.03
0.00
-3.03
0.00
3.03
9.09
21.21
-3.03
0.00
-3.03
18.18
6.06
0.00
6.06
-3.03
-3.70
0.00
9.09
0.00
    an - the number of metals for which results were reported.

    ''The average concentration  (mg per L)  for all metals  for  which results were
reported.
                                        12

-------
            Table 8.   Heavy Metal Results for Sample 4 for 1088 Audit
Laboratory
334001
335002
336001
342002
347003
353001
406001
412001
436007
501001
508016
521004
527002
529005
542010
Na
7
7
6
6
7
1
6
7
7
7
7
6
7
7
6
Reported
Value6
0.85
0.80
0.81
0.83
0.82
0.13
0.79
0.86
0.85
0.87
0.84
0.79
0.91
0.83
0.77
Expected
Value6
0.79
0.79
0.76
0.76
0.79
0.09
0.71
0.79
0.79
0.79
0.79
0.76
0.79
0.79
0.76
% Diff.
7.59
1.27
6.58
9.21
3.80
44.44
11.27
8.86
7.59
10.13
6.33
3.95
15.19
5.06
1.32
     aN = the  number of metals for which results were reported.

     ''The average concentration (mg per L)  for  all  metals for which results were
reported.
                                       13

-------
            Table 9.  Heavy Metal Results for Sample 5 of 1088 Audit
Laboratory
334001
335002
336001
342002
347003
353001
406001
412001
436007
501001
508016
521004
527002
529005
542010
Na
7
7
6
5
7
1
6
5
7
7
7
5
7
6
5
Reported
Valueb
0.43
0.40
0.39
0.40
0.41
0.41
0.40
0.45
0.43
0.43
0.40
0.38
0.41
0.42
0.40
Expected
Value6
0.41
0.41
0.38
0.38
0.48
0.07
0.40
0.38
0.41
0.41
0.41
0.38
0.41
0.40
0.38
% Diff.
4.88
-2.44
2.63
5.26
-14.58
14.28
0.00
18.42
4.88
4.88
-2.44
0.00
0.00
5.00
5.26
     8N - the number of metals  for which  results were  reported.

     The average concentration (mg per L) for all metals  for which results were
reported.
                                        14

-------
             Table  10.  Concentrations  of the Heavy Metals  in Each  1988
                              Audit Sample in mg Analyte/L
                         0588                              1088
                Sample  4    Sample 5          Sample 4         Sample 5
Mn
Fe
Cd
Cu
Ni
Pb
Zn
0.034
0.082
0.034
0.070
0.022
0.110
0.610
0.019
0.049
0.019
0.029
0.014
0.054
0.142
0.030
0.081
0.030
0.056
0.022
0.089
0.482
0.019
0.007
0.018
0.041
0.018
0.067
0.238
t. U S. Government Printing Office 1990-768-159/00396
                                           15

-------