EPA-450/4-84-013
June 1984
An Update On The Use Of Paniculate Ratios
To Assess Likely PM<|Q Attainment Status
BY
A.D Thrall
And
A.B Hudischewskyj
Systems Applications, Inc.
San Rafael, CA 94903
EPA Project Officer Edwin L Meyer, Jr
EPA Contract No 68-02-3848
ENVIRONMENTAL PROTECTION AGENCY
Office Of Air And Radiation
'ce Of Air Quality Planning And Standards
Research Triangle Park. IMC 27711
-------
This report has been reviewed by the Office Of Air Quality Planning And Standards, U.S.
Environmental Protection Agency, and approved for publication as received from the contractor.
Approval does not signify that the contents necessarily reflect the views and policies of the
Agency, neither does mention of trade names or commercial products constitute endorsement
or recommendation for use.
EPA-450/4-84-013
-------
Technical Memorandum
AN UPDATE ON THE USE OF PARTICULATE RATIOS
TO ASSESS LIKELY PM 10 ATTAINMENT STATUS
13 January 1984
(Revised 4 June 1984)
SYSAPP-84/100
Prepared for
U.S. Environmental Protection Agency
Office of Air Quality Planning and Standards
Monitoring and Data Analysis Division
Research Triangle Park, NC 27711
EPA Contract No. 68-02-3848
Prepared by
A. D. Thrall
A. B. Hudischewskyj
Systems Applications, Inc,
101 Lucas Valley Road
San Rafael, CA 94903
iv
-------
Table of Contents
Page
Introduction 1
Differences In Concentration Ratios When Grouped By Time,
Location, Or Total Concentration 1
Pooling National Data To Estimate The Statistical Distribution
Of Concentration Ratios 6
Discussion Of Specific Issues 11
References 16
Supplementary Figures 17
Appendix A. Sites Used In The Analysis 51
Appendix B. Site Classifications According To Climate 55
Appendix C. 24-hour Site Cases Omitted Because At Least One
Of The Following Ratios Was Greater Than 1.05:
PMio/DS, PMiQ/TSP, DS/TSP 56
Appendix D. Annual Site Cases Omitted Because At Least One
Of The Following Annual Ratios Was Greater Than
1.05: PMio/DS, PMio/TSP, DS/TSP 57
-------
INTRODUCTION
In this memorandum the results obtained by Thrall and Burton (1983a) based
on 1980-1981 data are compared to results based on 1982 data and revised
1980-1981 data. To facilitate the comparison, the new results are
presented in the same way as in the earlier report. In addition, illus-
trations in this memo are numbered to correspond to illustrations in the
earlier report, e.g., Figure 5' here corresponds to Figure 5 in the
earlier report. For a description of the PM-10 attainment status problem
and the development of a method of solution see Pace and Frank (1982).
There are several differences between the data set analyzed here and that
used previously: (1) 1982 data, including PM 10 measurements, are now
available; (2) the first quarter of 1980 has been deleted from the 1980-
1981 data, since a change in filters nominally scheduled for the beginning
of 1980 appears not to have been implemented at a number of sites; and (3)
24-hour or annual ratios of PM 10/DS, PM 10/TSP, or DS/TSP*exceeding 1.05
have been deleted. Items 2 and 3 were suggested by EPA technical staff as
a means of improving the reliability and consistency of the data used in
the analysis.
DIFFERENCES IN CONCENTRATION RATIOS WHEN GROUPED
BY TIME, LOCATION, OR TOTAL CONCENTRATION
24-Hour Ratios Undifferentiated by TSP Concentration
The distribution of 24-hour DS/TSP ratios over time, beginning with the
second quarter of 1980 and ending with the last quarter of 1982, is shown
in Figure 5'. Jhe accompanying analysis of variance shows that differ-
ences among the average ratios for the 11 individual quarters are sta-
tistically significant.
* DS is an abbreviation for PM15 measured using a dichotomous sampler.
81003 2 1
-------
Figures 6' and 7' show the results of grouping ratios by climate and by
EPA region, respectively. Group differences are somewhat more pronounced
between regions (e.g., an average ratio of 0.49 for Region 10 versus 0.59
for Region 8) than between climates (0.52 for West Coast, 0.53 for arid,
and 0.56 for other locations).
Figures 8,1 9,' and 10' show the results of grouping ratios by state,
city, and site, respectively. In each of the figures we display the two
groups whose histograms indicate the range of ratio distributions among
all the groups. Note, however, that the analysis of variance in each
figure refers to all groups having at least two observations, and not
merely the two groups displayed. As expected, we see larger differences
between cities than between states. Since there are typically only one or
two sites per city in our sample, the variability of ratios between sites
is similar to that between cities.
24-Hour Ratios Corresponding to High TSP Concentration
Figure 11' shows the effect of grouping ratios by TSP concentration (TSP <
o O
150 pg/m versus TSP > 150 pg/m ). In most of the remainder of the report
we restrict our attention to the latter group of ratios, corresponding to
high TSP concentrations, since, as Pace and Frank point out, it is this
group of observations that would most likely produce the high PM 10
concentrations of interest. Indeed, taking the TSP measurements at face
j
value, one can argue that high PM 10 concentrations (> 150 yg/m ) can
occur only on days of high TSP concentration. We see from the results
presented in Figure 11* that screening out observations corresponding to
low TSP concentrations (< 150 ug/m3) has little effect on the average and
standard deviation of the ratios. Restricting attention to site-days
o
having high TSP concentrations (> 150 ug/m ), Figures 12' through 17' show
distributional changes in 24-hour DS/TSP ratios over time and location, in
parallel with Figures 51 through 10'.
31*003 2
-------
Comparing Figures 5' and 12', we see that one result of screening out low
TSP concentrations is that the differences among mean ratios for quarterly
periods are no longer statistically significant (the quarterly averages
are 0.50, 0.54, 0.56, and 0.56).
Comparing the results in Figures 6' through 10' with the corresponding
results in Figures 13' through 17', we see that screening 24-hour DS/TSP
ratios based on TSP concentration reduces the differences in ratios that
are associated with different site locations. The analysis of variance
given in Figure 13' shows that there is no statistically significant
difference (at the 0.01 level) between ratios measured in different
climates. Figure 14' shows that the differences between EPA regions
remain statistically significant, although the differences are reduced.
The differences among states, among cities, or among sites remain sta-
tistically significant, as shown by Figures 15', 16', and 17', respec-
tively. Nevertheless, these geographical differences are less pronounced
for high TSP concentration, as evidenced by the diminished F-statistics in
the analysis of variance tests.
Annual Ratios Undifferentiated by TSP Concentration
Following the same format for ratios of annual average DS and TSP concen-
trations as for 24-hour concentrations, we see from Figure 18' that the
differences between the 1980, 1981, and 1982 average ratios (0.57, 0.55,
and 0.54, respectively) are not statistically significant.
The analyses of variance presented in Figures 19' through 23' show that
the annual ratios are geographically more homogeneous than are the 24-hour
ratios. Partitioning site-years by climate or EPA region does not reveal
geographic differences in average ratios beyond what one might find for a
sample drawn from a geographically homogeneous population. Only the finer
partitions based on states, cities, and sites show geographic location to
have a statistically significant effect on average annual ratios.
84003 2
-------
To give an indication of the range of geographic variation in the distri-
bution of annual ratios, Figures 21', 22', and 23' contrast results for
two states, cities, and sites, respectively. Figure 21' compares the
ratios for the state of Washington (having an average of 0.44) with those
for North Carolina (having an average of 0.63). Similarly, Figure 22'
contrasts the average annual ratio for New York City (0.52, an average of
0.55 and 0.49 at two sites) with the ratio for Buffalo, New York (0.57, an
average of 0.68 and 0.49 at two sites). The case of Buffalo illustrates
that wide site variation can exist within one city. (Figure 22' also
shows that the minimum and maximum city ratios were 0.197 and 1.020,
respectively.) Finally, Figure 23' demonstrates an even wider variation
between the site-specific ratios in Will County, Illinois (with an average
of 0.36) and those in Rubidoux, California (with an average of 0.72).
Annual Ratios Corresponding to High TSP Concentration
Figure 24' shows the effect of grouping annual ratios into low and high
annual average TSP concentrations, with low and high TSP concentrations
defined as being below and above 55 yg/m , respectively. We see that the
average and standard deviation of the ratios are lower for the high TSP
concentrations (0.59 ± 0.14 versus 0.55 ± 0.11). Figures 25' through 30'
show the effect of time (year) and location for annual ratios
corresponding to high TSP concentration, in parallel with Figures 18'
through 23', which show ratio distributions undifferentiated by TSP
concentration.
In general, the result of screening out low TSP concentration has only a
slight effect on annual ratios. Statistically significant differences
between cities and sites remain, whereas the difference between states is
marginally statistically significant. The other grouping factorsyear,
climate, and regiondo not reveal statistically significant differences.
4
8^003 2
-------
The Use of Grouping Factors
Screening occurrences of high ratios (greater than 1.05) from the data
produces an effect similar to the one reported by Thrall and Burton
o
(1983a) when low TSP concentrations (less than 150 iig/m ) were excluded.
The effect is to reduce the differences between geographic regions, years,
and so on. This is not surprising; since the higher ratios correspond to
the lower TSP concentrations, the screening out of high ratios and the
exclusion of low TSP concentrations produce nearly the same set of data.
As a consequence, the TSP level (low or high) no longer appears to be a
statistically significant factor.
Thus the case for using group-specific distributions*of ratios to estimate
the probability of PM 10 attainment status, which was not strongly
supported by the previous data, has even less support in the more recent
and revised data.
Of course, some group differences, e.g., between quarters, are statisti-
cally significant. Nevertheless, the case against the use of group-
specific ratios, even in the face of statistically significant differ-
ences, is fairly strong: when group differences are great enough to be of
practical significance, e.g., for differences between sites, the sample
sizes for individual groups are too small to permit reliable estimation
and application of the group-specific distribution of ratios.
Since PM 10 measurements are now available, it is possible to examine this
issue in a more quantitative fashion, i.e., we could evaluate the efficacy
of group-specific predictions of PM 10 measurements (or predictions of
whether the PM 10 concentrations are above or below some level) using
statistical cross-validation techniques. The cross-validation analysis is
beyond the scope of the present effort, however.
By a "group-specific distribution," we mean a distribution of PM10/TSP or
DS/TSP ratio derived specifically from sites of a particular type (i.e.,
within a particular group). For example, if we only considered ratios
observed in a given city, this group of ratios would be specific for that
city and the distribution of values would be a group-specific distribution,
84003 2
-------
POOLING NATIONAL DATA TO ESTIMATE THE STATISTICAL
DISTRIBUTION OF CONCENTRATION RATIOS
Table 5' shows the sample distribution of the 24-hour ratios of PM 10/DS,
PM 10/TSP, and DS/TSP for those site-days having high TSP concentrations
(> 150 pg/m ). Because the effect of high TSP is not statistically
significant for the current data set, we include Table 5" to show the same
ratios for all site-days (undifferentiated with respect to high or low
TSP). Tables 7' and 7" give the corresponding distributions of annual
ratios. These results may be compared to the unweighted percent!les
previously reported (Tables 5 and 7 of Thrall and Burton, 1983a). Because
the use of weighting factors was previously found to have little effect on
the distribution, this procedure was not followed in the analysis reported
here.
The recently available 24-hour PM 10 measurements allow us to examine
PM 10/DS ratios that, of necessity, were previously assumed to have the
constant value 0.83. Tables 51 and 5" show that the sample median (50th
percent!le) ratio is 0.89 for high TSP concentrations and 0.85 for overall
TSP concentrations. Figures 31 and 32 display scatter diagrams of 24-hour
and annual concentrations, respectively. From the information included in
Figure 31a, we see that the coefficient (a) minimizing the sum of squared
residuals
(PM 10 - a-DS)2
is 0.82. Scatter diagrams of PM 10/TSP ratios versus 0.83 times corre-
sponding DS/TSP ratios are shown in Figures 33 and 34 for 24-hour and
annual data, respectively.
On the choice of ratio distribution to be used to estimate PM 10 attain-
ment status, our judgment is that the 24-hour PM 10/TSP ratios are
sufficiently numerous to be of use for the short-term standard, but that
the annual DS/TSP distribution multiplied by an appropriate factor (in the
84003 2
-------
TABLE 5'. 24-Hour high TSP (> 150 yg/m3), all ratios < 1.05, April 1, 1980 - December 31, 1982.
Ratio =
Percentage
Minimum
1
5
10
20
30
40
50
60
70
80
90
95
99
Maximum
Average
Standard devi
PM 10/DS
Percentile
0.40
0.40
0.68
0.73
0.77
0.83
0.85
0.89
0.91
0.93
0.95
0.98
1.02
1.04
1.05
0.86
ation 0.11
Number of cases 116
Ratio 2 = PM
Percentage
Minimum
1
5
10
20
30
40
50
60
70
80
90
95
99
Maximum
Average
Standard deviation
Number of cases
10/TSP
Percentile
0.19
-
0.30
0.36
0.40
0.43
0.46
0.50
0.51
0.54
0.55
0.57
0.58
0.63
0.64
0.48
0.09
34
Ratio 3 =
Percentage
Minimum
1
5
10
20
30
40
50
60
70
80
90
95
99
Maximum
Average
Standard deviati
Number of cases
DS/TSP
Percentile
0.16
0.18
0.26
0.36
0.44
0.48
0.51
0.55
0.58
0.61
0.66
0.71
0.77
0.88
0.98
0.54
on 0.14
283
-------
TABLE 5". ^differentiated 24-hour TSP, all ratios < 1.05, April 1, 1980 - December 31, 1982.
00
Ratio = PM
Percentage
Minimum
1
5
10
20
30
40
50
60
70
80
90
95
99
Maximum
Average
10/DS
Percentile
0.40
0.50
0.62
0.68
0.75
0.79
0.83
0.85
0.88
0.91
0.93
0.96
1.00
1.04
1.05
0.84
Standard deviation 0.11
Number of cases
287
Ratio 2 = PM
Percentage
Minimum
1
5
10
20
30
40
50
60
70
80
90
95
99
Maximum
Average
Standard deviation
Number of cases
10/TSP
Percentile
0.19
0.20
0.29
0.34
0.39
0.42
0.45
0.48
0.50
0.54
0.57
0.65
0.70
0.79
0.91
0.48
0.12
265
Ratio 3 =
Percentage
Minimum
1
5
10
20
30
40
50
60
70
80
90
95
99
Maximum
Average
DS/TSP
Percentile
0.02
0.21
0.33
0.38
0.44
0.48
0.51
0.54
0.57
0.60
0.65
0.71
0.78
0.92
1.05
0.55
Standard deviation 0.13
Number of cases
5733
-------
TABLE 7'. Annual high TSP (> 55 pg/m3), all ratios < 1.05, April 1, 1980 - December 31, 1982.
Ratio = PM
Percentage
Minimum
1
. 5
10
20
30
40
50
60
70
80
90
, 95
99
Maximum
Average
10/DS
Percentile
0.52
-
0.52
0.55
0.60
0.61
0.80
0.81
0.83
0.86
0.86
0.91
0.93
-
0.94
0.77
Standard deviation 0.14
Number of cases
18
Ratio 2 = PM
Percentage
Minimum
1
5
10
20
30
40
50
60
70
80
90
95
99
Maximum
Average
Standard deviation
Number of cases
10/TSP
Percentile
0.25
-
-
-
0.38
0.40
0.45
0.48
0.49
0.51
0.53
0.61
-
-
0.64
0.47
0.11
13
Ratio 3 =
Percentage
Minimum
1
5
10
20
30
40
50
60
70
80
90
95
99
Maximum
Average
Standard deviati
Number of cases
DS/TSP
Percentile
0.30
0.31
0.36
0.41
0.46
0.50
0.51
0.54
0.57
0.59
0.63
0.67
0.73
0.84
0.90
0.55
on 0.11
205
-------
TABLE 7". Undifferentlated annual TSP, all ratios < 1.05, April 1, 1980 - December 31. 1982.
Ratio = PM
Percentage
Minimum
1
5
10
20
30
40
50
60
70
80
90
95
99
Maximum
Average
10/DS
Percentile
0.52
_
0.52
0.55
0.60
0.73
0.81
0.82
0.83
0.84
0.86
0.91
0.93
-
0.94
0.77
Standard deviation 0.13
Number of cases
19
Ratio 2 = PM
Percentage
Minimum
1
5
10
20
30
40
50
60
70
80
90
95
99
Maximum
Average
Standard deviation
Number of cases
10/TSP
Percentile
0.25
-
0.25
0.36
0.38
0.40
0.46
0.48
0.49
0.53
0.56
0.61
-
-
0.64
0.48
0.11
14
Ratio 3 =
Percentage
Minimum
1
5
10
20
30
40
50
60
70
80
90
95
99
Maximum
Average
DI/TSP
Percentile
0.20
0.29
0.36
0.42
0.47
0.50
0.52
0.55
0.58
0.61
0.63
0.69
0.74
0.90
1.02
0.55
Standard deviation 0.12
Number of cases
254
-------
range of 0.75 to 0.85, depending on the criterion used to guide the
choice) should be used for the long-term standard.
DISCUSSION OF SPECIFIC ISSUES
Using the above results we address the following issues raised by EPA
staff at the beginning of this updating study.
(1) What options are available to regulatory agencies in making
initial PM 10 attainment determinations based on scant or
indirect measures of PM 10?
The options considered to date consist of variations on the application of
the methodology developed by Pace and Frank (1982). Also, see Thrall and
Burton (1983b). The primary issue is the choice of ratio distribution to
be used, e.g., whether to apply a national or local sample distribution.
(2) Does the use of a single national distribution, estimated from
currently available data, still appear to be the most appropri-
ate option? If not, how should this recommendation be revised?
The use of a single national distribution still appears to be more
reliable than the use of local distributions, because of (i) the small
numbers of observations in most localities and (ii) the difficulty of
applying a geographic-specific distribution to anything other than the
identical geographical unit from which it was derived.
(3) Are the TSP cutpoints of 55 ug/nr for annual average concentra-
tions and 150 pg/m3 for 24-hour average concentrations still
appropriate thresholds for potential PM 10 nonattainment?
8^03 2
11
-------
Having screened out high DS/TSP ratios (which are almost always concurrent
with low TSP concentrations), the difference between the average ratios
for low and high TSP concentrations is no longer statistically significant,
The preferable procedure still appears to be to differentiate ratios on
the basis of the TSP (denominator) concentration rather than simply
screening out high ratios.
(4) Are the data provided by EMSL sufficient to reliably estimate
PM 10/DS and PM 10/TSP distributions? Is the 0.83 conversion
factor between DS and PM 10 still appropriate?
Sample sizes of 24-hour ratios appear to be sufficient for reliable
estimation, but few annual ratios are available. The 0.83 conversion
factor appears to be appropriate, although more precise criteria and
probably more data would be required to estimate (with two-digit accuracy)
the "best" conversion factor.
(5) Recommend a procedure to be used by regulatory agencies for
initial determination of attainment of annual and 24-hour PM 10
standards. Discuss the reliability of the procedure. Provide
all results necessary to the implementation of the procedure and
the evaluation of its reliability. These results may include
estimated distributions of DS/TSP, PM 10/TSP, and PM 10/DS as
appropriate.
We recommend that the national 24-hour PM 10/TSP distribution be used to
estimate 24-hour PM 10 attainment status and that attainment of the annual
PM 10 standard be estimated using the national DS/TSP distribution
multiplied by 0.83. For the current data, in which high ratios have been
screened out, no differentiation between low and high TSP ratios is
needed. These distributions are given in Tables 5" and 7". As indicated
by Figures 31 and 32 these procedures should yield predicted PM 10
q
concentrations typically within 10 pg/m of the measured concentration.
8*1003 2 12
-------
As discussed in the previous section, and by Thrall and Burton (1983b),
there are further methods for gauging the reliability of these procedures.
After circulation of an initial draft of this memorandum, we received
several comments from EPA reviewers to which we now respond.
(A) The exclusion of ratios greater than 1.05 causes the previously
reported relationship between DS/TSP ratios and TSP concentra-
tions to disappear. The deletion of high ratios may also induce
a bias in the estimated distribution of ratios. Two alternatives
to the deletion of high ratios are: (1) to eliminate from the
analysis those sites tending to produce high ratios (or other
anomalies) or (2) to remove extremely low ratios in addition to
the removal of extremely high ratios.
The high concentrations were deleted at the request of EPA project staff
since these ratios are regarded as not physically meaningful. At issue is
whether we wish to use the ratios to estimate the results of PM-10
monitoring, even though monitoring results may be occasionally erroneous,
or whether we wish to use the definition of PM-10, PM-15, and TSP (concen-
trations of particulate having aerodynamic diameters less than 10, 15, and
20 urn, respectively) to rule out ratios greater than unity. The latter
course was taken at the direction of EPA staff.
The deletion of high ratios does of course alter the empirical cumulative
distribution. The number of site-days or site-years in which ratios are
less than a specified value, x, with x between zero and unity, is
unaltered but the total number of site-days or site-years is reduced. The
result is that the proportion of site-days or site-years having ratios
less than x is increased. This has the effect of increasing the estimated
probability that the PM-10 concentration corresponding to a given TSP
concentration is below the PM-10 standard. That is, the effect of
deleting high ratios is to increase (or possibly leave unaltered) the
.»
number of sites deemed to be in attainment of the PM-10 standard, and to
84003 2 13
-------
reduce (or possibly leave unaltered) the number of sites deemed to have
failed to attain the standard.
The suggested alterations to the deletion of high ratios are worth
pursuing in future work. As more PM-10 measurements become available it
will be possible to compare the performance of various procedures for
estimating either PM-10 concentrations or PM-10 attainment status.
(B) Please clarify the reasoning leading to your recommendation that
a national distribution of ratios be used rather than separate
distributions for different geographic regions. Also comment on
the use of the analysis of variance to test the similarity of
distributions as opposed to a Kolmogorov-Smirnov test, for
example.
The use of computer program BMDP7D, entitled "Description of Groups
(Strata) with Histograms and Analysis of Variance," proved to be an
efficient way to make the many comparisons in which we were interested.
As noted in our previous report of 4 May 1983, an analysis of variance
finding that group differences are not statistically significant does not
imply that the distributions are identical, since the finding pertains not
to group distributions in their entirety but only to group means (and
variance in the case of BNMDP7D). In general, the "significant" factors
such as "site" remained significant under a more robust (but less power-
ful) analysis, included in BMDP7D, that used separate rather than pooled
estimates of variance. As the reviewer suggests, we did conduct nonpara-
metric (Kolmogorov-Smirnov) tests in a few instances, but these did not
alter our findings. We did not use nonparametric tests more extensively
because they require much more computation and programming time, and, more
importantly these tests are much less powerful, particularly for the
screening of approximately 100 groups with an average of about 50 observa-
tions per group.
Regarding the use of a national distribution, our judgment is that if the
choice is to apply to each group of sites either a national distribution
8i*003 2 14
-------
or the distribution obtained from that group, then the national distribu-
tion appears preferable. Our preference is based on considerations of
both practicality and statistical reliability. Comparing the magnitude of
sampling variability to the magnitude of group differences, it appears
that the use of one large-sample distribution, that admittedly represents
individual sites imperfectly, will do a better job in limiting the number
of PM-10 misclassifications. We fully agree, however, that this recom-
mendation should be further examined by comparing the performance of
various methods of classifying PM-10 attainment status.
(C) Why has the analysis focused on ratios of particulates as a means
of assessing PM-10 attainment status rather than some other means
of assessment? More specifically, the PM-10 assessments might
well be improved by taking the logarithm of particulate concen-
trations and using log TSP as a covariate of log PM-15 or log
PM-10.
EPA project staff requested that we focus our analysis on particulate
ratios, since our task was to refine the work of Pace and Frank. We did
examine the covariation of log OS and log TSP in previous work. These
results along with those of other workers indicate that a linear model
fits the log-transformed data slightly better than the untransformed data,
homogeneity of variance being an important benefit of the transforma-
tion. Our work also suggests, however, that TSP cutpoints, to be used to
classify the status of sites, derived from either particulate ratios or
logged particulates would be much the same. Again, it would be useful to
use the recently available PM-10 data to compare the rate of correct
classification of these different methods.
(D) Consideration should be given to the grouping of data into the
four quarters of the year, i.e., by season, rather than grouping
data into the quarters spanning the period under study.
Additionally, the analysis might he improved by removing sites
having an insufficient number of observations during a quarter or
year.
81*003 2 15
-------
In our previous report we did find differences between DS/TSP ratios for
the four quarters of the year. Using those pairs of 24-hour DS and TSP
concentrations reported during 1980-1981 for which the TSP concentration
exceeded 150 pg/m , the quarterly averages were 0.48, 0.54, 0.57, and
0.56, respectively. A similar analysis for the current data set was
hampered (although still possible) by the exclusion of the first quarter
of 1980. The suggestion to remove sites having an insufficient number of
observations is worth pursuing in future analyses.
REFERENCES
Pace, T. G., and N. H. Frank. 1982. "Procedure for Estimating Probabil-
ity of Nonattainment of a PM^g NAAQS using Total Suspended Particulate
or Inhalable Particulate Data." U.S. Environmental Protection Agency,
Research Triangle Park, NC.
Thrall, A. D., and C. S. Burton. 1983a. "Characterizing Ratios of
Particulate Concentrations: A Preliminary Step in Assessing Likely
Attainment Status Under a PM 10 National Ambient Air Quality Stan-
dard." Systems Applications, Inc., San Rafael, CA. (SYSAPP-83/078).
Thrall, A. D., and C. S. Burton. 1983b. "Considerations in Assessing
PM 10 Attainment Status Using TSP Data." Systems Applications, Inc.,
San Rafael, CA. (SYSAPP-83/079).
8«t003 2 16
-------
1980
01
1980
02
1980
03
1980
04
1981
Ql
1901
02
1931
03
1981
04
1982
Ql
1982
02
1982
03
1982
04
VAR 14
XCLUOED
VAL UE S
........... ..49S .......785 ""1036 """1143
TABULATIONS AND COMPUTATIONS WHICH FOLLOW EXCLUDE VALUES LISTED ABOVE
1343
******* it * ik *
1447 *«*«
**"*1254 *"«*»»*«»«**«97| ** ********MMMQgg ****M*7g3 *7g|
IDPOINTS
1.085) ~~"^
1.060)
1.0151
0.9UJ)
0.945)
0.910)
0.875)
0.8401
0.805)
0.770)
0.*7«0>
0.666)
0.630)
' 0.695)
0.62S)
0.490)
0.455)
0.420)
D.3d5)
0.350)
0.316)
0.280)
0.245)
0.210)
0.175)
0.140)
0.105)
0.070)
0.035)
ROUP MEANS ARE DENOTED
EAN 0.000
TD.OEV. 0.000
.E.S.D. ».itJ
. E. M. 0.000
AX I MUM 0.044
INIMUM 0.000
AMPLE SIZE 0
..
... ««
...
. ......... ...» «
..** .* .
..% *.**** .....
*. ..
.. ......
«
BY M'S IF THEY COINCIDE WITH "S. N'S OTHERWISE
0.638 0.663 0.670 0.647 .527 0.666
0.153 0.154 0.138 0.140 .128 0.137
0.160 0.161 0.1J2 0.132 .124 0.131
0.008 0.007 0.006 0.006 .006 0.006
1.049 1.019 1.027 1.042 .988 1.016
0.160 0.163 0.086 0.171 .131 0.136
406 606 516 600 463 669
* *
* * *
* *** ***
***** *** * *
* * **
* *
It
0.532 .636 0.520 0.662
0.118 .123 0.126 0.129
0.H2 .117 0.121 0.123
0 . 045 . 006 0 . .C*5 0 . 005
1.010 .942 1.001 1.036
0.U9 .154 0.173 0.166
669 499 637 609
«
*.
.
.
0.566
0.127
0.122
0.0;.' 7
1.017
0.022
370
THE NUMBER OF GROUPS IS NOT CONSISTENT WITH THE TWO-WAV SPECIFICATION. ONE-WAY ANALYSIS OF VARIANCE IS COMPUTED
............................. ANALYSIS OF VARIANCE TABLE «
SOURCE SUM OF SQUARES UK MEAN ROUARE F VALUr TAIL PROBABILITY
ALL CROUPS COMBINED
(EXCEPT CASES WITH UNUSED VALUES
FOR VE1R. AND
MEAN
STD.DEV.
R.E.S.D.
S. E. N.
MAXIMUM
MINIMUM
SAMPLE SIZE
0.647
0.134
0.129
0.002
1.049
0.022
6733
BETWEEN GROUPS
WITHIN CROUPS
TOfAL
1.7284
101.9364
103.£638
10
5722
5732
0.1728
«.1/170
9.70
0.0000
U.VtNfS TEST FUK EQUAL VAIUANCLS l0,i>VZ2 6.71 0.0000
ONE-WAY ANALYSIS OF VAhlAIICt
UST STATISTICS FOR Wl I'llltl-UftOUP
VARIANCES NOT ASSUMED TO UE EQUAL
WELCH
BROWH-FORSVTHE
10.2200
10.5249
10.11
9.68
0.0000
0.0000
NOTE - ONLY THOSE GROUPS WITH NON-ZERO VARIANCE ARE USED IN THE COMPUTATIONS.
FIGURE 5'. Distribution of 24-hour DS/TSP ratios by year (1980. 1981. 1982)
and quarter.
84003
-------
00
W.COAST ARID OTHER
+ + »
VAR 14
EXCLUDED
VALUES
*..************!424 «******************.*i70e ««******«****«**«*«7765
TABULATIONS AND COMPUTATIONS WHICH FOLLOW EXCLUDE VALUES LISTED ABOVE
MIDPOINTS
1.050)*
1.0IS)*
0.980)** *« *««
0.945>*«* * *.*««*»
0.910)****** **** ******************
0.840)«******** ««.**«*«* .......................... «34
0.805)********"* *************** ********************>»«**>gi
0770 )***.*«« A**************** *****************«****.»..g7
0,7 351******************.** ********************** *»*********«***********.**]2l
.700j***************************34 *******«*******************g2 *****************».»*»»***J90
.666)***************************4Z ***********************»***5a **»****«**«***********«*»*247
.630>***************************70 **«*****************«**«**66 *******««**«***"«****«"***3Z6
.595).**.».***..*****.**.**.*.».B2 ***..*.....*.***..*....*..*94 ..*.**........**.....*..**44a
.560)****«*********«****«*******96 ***«*«««****«**««««*104 H"*********«***"«*****"»**467
.625)M*****************9s M**************************96 *«»«....«««»»«...«**.474
.490}**************************100 *****************«***«*****8g **************«********«**390
.455 I**************************102 *«****»**«***********83 *«**«**«««*«**«***«««339
.420)******«********«***********79 ««....*....*...........74 *****.«...**....*......*.*206
.385 )*««"«"«*«*«*«««.«.63 ««*******««««**«*.........64 ..........................137
.350)***«**«*****************«**43 ...........................45 .......................****8S
.315 I******************* *..«....3) *******..*«*..**...**.** .................... ........45
.200)**************** ***************** *******n*******************30
.210)******* ********** A******.********.**
.175)** ********* ********..***
.140)*** *« **
.105) *
.070) *
0.035)
GROUP MEANS ARE DENOTED BV'M'S IF THEV COINCIDE WITH "S, N'S OTHERWISE
MEAN
STD.DEV.
R.E.S.D.
S. E. M.
MAXIMUM
MINIMUM
SAMPLE SIZE
0.522
0.137
0.133
0.004
1.036
0.131
946
0.633
0.150
0.147
0.005
1.042
0.150
992
0.557
0.128
0.122
0.002
1.049
0.022
3795
ALL GROUPS COMBINED
(EXCEPT CASES WITH UNUSED VALUES
FOR CLIMATE )
******
MEAN
STD.DEV.
R.E.S.D.
S. E. M.
MAXIMUM
MINIMUM
SAMPLE SIZE
0.647
0.134
0.129
0.002
1.049
0.022
6733
SOURCE
BETWEEN GROUPS
WITHIN GROUPS
TOTAL
******* ANALYSIS OF VARIANCE TABLE **«
SUM OF SQUARES OF MEAN SQUARE
1.1900
102.4737
103.6630
2
6730
0.5950
0.0179
f VALUE TAIL PROBABILITY
33.27 0.0000
LtVENE'S TEST FOR EQUAL VARIANCES
5732
2.67J0
22.34
0.0000
ONE-WAY ANALYSIS OF VARIANCE
TEST STATISTICS FOR WITIIIN-GROUP
VARIANCES NOT ASSUMED TO BE EQUAL
WELCH
BROWN-FORSYTHE
2.1742
2,2572
32.02
29.90
0.001/0
0.0000
FIGURE 6'. Distribution of 24-hour DS/TSP by climate.
-------
REGION 1 REGION 2 REGION 3 REGION 4 REGION 6 REGION 6 REGION 7 REGION 8 REGION 9 REGION10
+ ...,... + + + ....*. + + * ...» *
VAR 14
EXCLUDED
VALUES
MIDPOINTS
*..*...557 *.*****779 *«****230s ««***1304 **««»*i68l »**«**«587 *«****«768 «**«623
TABULATIONS AND COMPUTATIONS WHICH FOLLOW EXCLUDE VALUES LISTED ABOVE
625
1.0501
1.015)*
0.9B0)*
0.945>««
0.910)**
0.875)*
0.840)**
0.805)**
0.770)**
0.735)**
0'.700)**
0.665)**
0.630)**
0.595)**
0.560)*
0.525)H
0.490)**
0.455)**
0.420)**
0.385)**
0.3S0)**
0.315)**
0.280)**
0.245)*
0.210)**
0.175)**
0.140)**
0.105)*
0.070)
0.035)
GROUP MEANS
MEAN
STD.DEV.
R.E.S.O.
S. E. M.
MAXIMUM
MINIMUM
SAMPLE SIZE
********** ***
****
** *
*
**
*****
****
*****
**«
*****
***
*
*** ********|g **«****«|4 ****«***}3
* ........28 «**««««*15 .«...«««.*
It ******** ********34 ********24 ********30
******* *«******}3 ********£3 ********4g «*«*****27
******Jg ******}} *M******Q3 ********g| ***«****g4
****«I9 .«.**..32 ********9Q *««>.**7g ***«****5|
**«***27 ********g0 M"*"gl**13a ********66 **»««*««92
******33 ********g4 *******|2g n***««**95 ********gg
......38 H...***«66 *«*..**i3| ...*»***74 H*****«i)B
******37 ******«*g9 ********9g ........54 ********95
********54 ********78 *«******4g ********72
..***...40 ********29 ********37 «««..ngf
**.****.24 ****«***24 ********|8 ..«««...39
********I4 ****«***jg ******** ***«****27
****** ***** ******* ****.***|4
**** **** *«** ****..**ii
** * **.**«*1j
* ** ** *******
* * ******
*****ll**16
**Hi*****3g
*<*****
**********
******** j
4 ****** -k
**********
*«*«13
******2i
****
*
*********
****
****
r««3l
'26
16
******* «
********|3
***«****!3
*******I0
....««..35
««***»**4g
********g7
********80
****«**j0g
M***"**101
«******l0g
*******»91
»*****£))
********gg
********29
********29
****««**I2
*******ll
******
*** *
*****
*****
*****
*****
*****
****.
*****
*****
*****
*****
*****
*****
*****
*****
13
*«*15
***30
*«*36
**39
43
40
62
«**50
*"*37
*"*38
***17
«**20
ARE DENOTED BV M'S IF THEV COINCIDE WITH "S, N'S OTHERWISE
0.529
0.146
a. 135
0.008
1.001
0.119
309
0.529
0.1)6
0.104
0.005
1.027
0.209
468
0.683
0.122
0.115
0.004
1.017
0.192
1014
0.577
0.120
0.114
0.005
0.986
0.086
698
0.534
0.130
0.124
0.004
1.049
0.022
862
0.542
0.139
0.139
0.006
0.967
0.150
494
0.572
0.133
0.129
0.007
1.026
0.226
343
0.590
0.117
0.116
0.010
0.961
0.208
149
0.530
0.146
0.137
0.006
1.042
0.154
920
********
0.489
0. 143
0.141
0.007
0.974
0.131
476
ALL GROUPS COMBINED
(EXCEPT CASES WITH UNUSED VALUES
FOR RE6ION )
MEAN
STD.DEV.
R.E.S.D.
S. E. M.
MAXIMUM
MINIMUM
SAMPLE SIZE
0.547
0.134
0.129
0.002
1.049
0.022
6733
SOURCE
BETWEEN GROUPS
WITHIN GROUPS
TOTAL
««*"** ANALYSIS OF VARIANCE TABLE *'
SUM OF SQUARES OF MEAN SQUARE
4.7168
98.9469
103.6637
9
5723
5732
0.6241
0.0173
F VALUE TAIL PROBABILITY
30.31 0.0000
LEVENE-S TEST FOR EQUAL VARIANCES
9.5723
6.85
0.0000
ONE-WAV ANALYSIS OF VARIANCE
TEST STATISTICS FOR WITHIN-GROUP
VARIANCES NOT ASSUMED TO BE EQUAL
WELCH
BROWN-FORSVTHE
9.1606
9,4048
30.83
30.32
0.0000
0.0000
84003
FIGURE 7'. Distribution of 24-hour DS/TSP ratios by EPA region.
-------
NE
ID
VAR 14
EXCLUDED
VALUES
43
ro
o
TABULATIONS AND COMPUTATIONS WHICH FOLLOW EXCLUDE VALUES LISTED ABOVE
MIDPOINTS
1.085)
1.050)
I. 015)
0.980)
0.945)"
0.910)
0.875)
0.640)*
0.805)*"
0.770)"
0.735)*
0.700)*
0.665)M
0.630)"
0.596)* *
0.660)* «" *****
0.52S)* «
0.490)** ******
0.455) **
0.4Z0) H******
0.386) *.*««*
0.360) *«**«*««J3
0.316) «*«
0.280) «**«*««
0.245) *«
0.210) *
0.176)
0.140)
0.105)
0.070)
0.035)
CROUP MEANS ARE DENOTED BY H'S IF THEV COINCIDE WITH *'S, N'S OTHERWISE
**
MEAN
STD.DEV.
R.E.S.D.
S. E. M.
MAXIMUM
MINIMUM
SAMPLE SIZE
0.668
0.111
0.115
0.021
0.942
0.602
27
.418
.126
.128
.014
.803
.194
77
ALL GROUPS COMBINED
(EXCEPT CASES WITH UNUSED VALUES
FOR STATE )
MEAN 0.647
STD.DEV. 0.134
R.E.S.D. 0.129
S. E. M. 0.032
MAXIMUM 1.049
MINIMUM 0.022
SAMPLE SIZE 6733
A********************** ANALYSIS OF VARIANCE TABLE *******
SOURCE
BETWEEN GROUPS
WITHIN GROUPS
SUM OF SQUARES DF
11.0502 37
92.6136 5695
MEAN SQUARE
0.2987
0.0163
F VALUE TAIL
18.36
!*.*****.
PROBABILITY
0.0000
*
*
TOTAL 103.6637
LEVENE'S TEST FOR EQUAL VARIANCES
5732
37,5695
4.03
0.0000
ONE-WAV ANALYSIS OF VARIANCE
TEST STATISTICS FOR WITIIIH-GROUP
VARIANCES NOT ASSUMED TO BE EQUAL
WELCH
BROWN-FORSVTHE
37, 750
37,1109
19.61
18.23
0.0000
0.0000
NOTE - ONLY THOSE GROUPS WITH NON-ZERO VARIANCE ARE USED IN THE COMPUTATIONS.
FIGURE 8'. Distribution of 24-hour DS/TSP ratios by state. The states of Nebraska
and Idaho are contrasted. Other states are not shown but are included in the analysis
of variance.
-------
St. Paul, t M1,nn.
" Pa-
VAR 14
EXCLUDED
VALUES
MIDPOINTS
TABULATIONS AND COMPUTATIONS WHICH HM.LUW EXCLUDE VALUES LtSfED ABOVE
I .050)
1.015)
0.980)
0.94S)
0.910)
0.675)
0.0401
0.805)
0.770)
0.7351 ........12
0.700) »i2
0.665) « 13
0.6301 H..»....20
0.595) 22
0.560) 23
0.525) .|7
0.490)
0.4SSI .........
0.420)
0.305) ....
0.350) «*
0.315) M
0.280)
0.246)
0.210)
0.175)
0.1401
0.105)
0.070)
0.0351
CROUP MEANS ARE DENOTED BY M'S IF THEV COINCIDE WITH *'S. N'S OTHERWISE
MEAN
STD.DEV.
R.E.S.D.
S. E. M.
MAXIMUM
MINIMUM
SAMPLE SIZE
0.320
0.115
0.115
0.021
0.680
0.173
31
0.628
0.129
0. 127
0.010
1.002
0.341
174
ALL GROUPS COMBINED
(EXCEPT CASES WITH UNUSED VALUES
FOR CITV )
MEAN 0.547
S1D.DEV. 0.134
R.E.S.D. 0.129
S. E. M. 0.002
MAXIMUM 1.049
MINIMUM 0.022
SAMPLE SIZE 6733
SOURCE
BETWEEN GROUPS
WITHIN CROUPS
....... ANALYSIS OF VARIANCE TABLE "
SUM OF SQUARES OF MEAN SQUARE
23.0828
80.5000
96
5636
0.2404
0.0143
VALUE TAIL PROBABILITY
16.82 0.0000
*
TOTAL 103.6b37
LEVENE'S TEST FOR EQUAL VARIANCES
5732
96,5636
6.80
0.0000
ONE-WAV ANALYSIS OF VARIANCE
TEST STATISTICS FOR WITHIM-GROUP
VARIANCES NOT ASSUMED TO DE EQUAL
WELCH
BROUN-FORSVTHE
96, 800
96,1428
17.74
15.69
0.0000
0.0000
NOTE - ONLY THOSE GROUPS WITH NON-ZERO VARIANCE ARE USED IN THE COMPUTATIONS.
FIGURE 9'. Distribution of 24-hour DS/TSP by city. The cities of St. Paul, Minnesota
(ratios obtained at one site) and Philadelphia, Pennsylvania (four sites) are contrast-
ed. Other cities are not shown, but are included in'the analysis of variance.
84003
-------
St.Paul.Mn.
Phil., Pa.
VAR 14
EXCLUDED
VALUES
.84
"II
rv>
ro
TABULATIONS AND COMPUTATIONS WHICH FOLLOW EXCLUDE VALUES LISTED ABOVE
MIDPOINTS
i.aasi
I. 0501
1.015) .
.980)
.945)
.9101
.B75I
.840)
.045)
.770)
.735)
.700)
.665)
.630)
.595)
.660)
.526)
.490)
.455)
.420)
.385)
.350)
.315)
.280)
.245)
.210)
.175)
.140)
.105)
.070)
.035)
GROUP MEANS ARE DENOTED 8V M'S IF THEY COINCIDE WITH $. N'S OTHERWISE
MEAN
STD.OEV.
R.E.S.D.
S. E. H.
MAXIMUM
MINIMUM
SAMPLE SIZE
0.320
0.116
0.116
0.021
0.680
0.173
31
0.670
0.128
0.125
0.015
I .002
0.341
69
ALL CROUPS COMBINED
(EXCEPT CASES WITH UNUSED VALUES
FOR
MEAN
STD.DEV.
R.E.S.D.
S. E. M.
MAXIMUM
MINIMUM
SAMPLE SIZE
0.547
a. 134
a. 129
0.002
1.049
15.822
S7J3
* SOURCE SUM OF SQUARES
BETWEEN GROUPS 25.9093
WITHIN GROUPS 77.7543
*
. TOTAL I03.0U3G
LEVENE'S TEST FOR EQUAL VARIANCES
ONE -WAV ANALYSIS OF VARIANCE
TEST STATISTICS FOR WITIIIN-GROUP
VARIANCES NOT ASSUMED TO BE EQUAL
* WELCH
BROWN-FORSVTHE
DF MEAN SQUARE
118 0.2196
5614 0.0139
{.732
110.1,6)4
118. 840
118. 835
F VALUE
15.85
6.68
17.49
14.60
TAIL PROBABILITY
0.0000
0.0000
0.0000
0.0000
NOTE - ONLV THOSE GROUPS WITH NON-ZERO VARIANCE ARE USED IN THE COMPUTATIONS.
FIGURE 10'. Distribution ;of 24-hour DS/TSP ratios by site. Sites in St. Paul, Minnesota and
Philadelphia, Pennsylvania are contrasted. Other sites are not shown, but are included in
the analysis of variance.
-------
0 - 150
150
ro
CO
VAR 14
EXCLUDED
VALUES
***..................5294 ***«.335
TABULATIONS AND COMPUTATIONS WHICH FOLLOW EXCLUDE VALUES LISTED ABOVE
MIDPOINTS
1.085)
I.0S0)
1.015) ...........
0.980)
0.9<6)
.910)
.876) ...........................38 .*
.840) *56
.80S) 79
.770) «..........97 ..
.735) 153 ..........
.700) 250 ..*.........*«..
.665) .324 .......................
.630) 444 .................
.595) ******..SB/ ...........................37
.560) (! ................635 H..........................32
0.525) ...........«.64I .........................
0.490) ***54B
0.455) *««.*«..502
0.42*) ..........................34, ..................
0.385) ..254 *.
0.360) |64 .........
0.315) ....gg ...
.200) ...........................60
.245) ...........................36 ....
.210) ............................. .......
.176) « .*
.140)
.105)
0.070) «
0.035)
CROUP MEANS ARE DENOTED 6V M'S IF THEV COINCIDE WITH *"S, N'S OTHERWISE
MEAN
STD.DEV.
R.E.S.D.
S. E. H.
MAXIMUM
MINIMUM
SAMPLE SIZE
ALL GROUPS COMBINED
(EXCEPT CASES WITH UNUSED VALUES
FOR
0.547
0.134
0.128
0.002
1.049
0.022
6450
.544
.144
. 139
.009
.984
.I5S
283
* ANALYSIS OF VARIANCE TABLE
SOURCE SUM OF SQUARES OF MEAN SQUARE
*
MEAN
STD.DEV.
R.E.S.D.
S. E. M.
MAXIMUM
MINIMUM
SAMPLE SIZE
0.547
0.134
0.129
0.002
I. 049
0.022
6733
BETWEEN GROUPS
WITHIN GROUPS
0.0033
103.6605
I
S73I
0.0033
0.01BI
F VALUE
0.18
TAIL
PROBABILITY
0.6702
TOTAL 103.6630 5732
LEVENE'S TEST FOR EQUAL VARIANCES 1,5731 2.60"
0.1071
ONE-WAV ANALVSIS OF VARIANCE
TEST STATISTICS FOR WITHIN-UROUP
VARIANCES NOT ASSUMED TO BE EQUAL
WE1CH
BROWN-FORSVTHE
1. 308
I . 308
0.16
0.16
0.6899
0.6899
NOTE - ONLV THOSE GROUPS WITH NON-ZERO VARIANCE ARE USED IN THE COMPUTATIONS.
FIGURE 11!. Distribution of 24-hour DS/TSP ratios by 24-hour TSP concentration (low versus
high TSP, with high TSP defined as concentrations greater than 150
84003
-------
1980
01
1980
02
1980
03
1980
04
1981
01
1981
03
1981
04
1982
01
1902
02
1902
UJ
1982
04
VAR 14
EXCLUDED
VALUES
MIDPOINTS
1. 1120)
U.99J)
IB7 (*207 *29B ****
TABULATIONS AND COMPUTATIONS WHICH FOLLOW EXCLUDE VALUtS LISTED ABOVE
"626
*ft«tt*****ft**MQ?g «*«*»**««**«7qg **cge «****** *KfiQ
TABULATIONS AND COMPUTATIONS WHICH FOLLOW CXCLUOC VALUCS LICTCD ABOUC
.............56tr
0.93d I
0.84ti)
0.813)
a.xe?)
0.7t0>
0.72«>
it. 690 I
U.fi6!J)
0.63J)
a.t,aa>
U.S7S)
u.B4u)
U.SIJ)
0.41)1)!
0.4!iin
0.391H
0.360)
U.33J)
a . 30 J )
0.270)
0.240)
U.21J)
U.ID8)
0.12*1
GROUP MEANS
MEAN
STD.OEV.
R.E.S.D.
S. E. H.
MAXIMUM
MINIMUM
SAMPLE SIZE
ARE DENOTED BY M'S IF THEY COINCIDE WITH "'S. N'S OTHERWISE
NS ARE DENOTED BV M'S IF THEV COINCIDE WITH *'S, N'S OTHERWISE
.lag
,000
.000
.000
.000
.0»0
.6S9
.147
.1S6
.034
.791
.23)
19
.672
.142
.128
.024
.87*
.177
34
.671
.171
.166
.031
.984
.194
30
.497
.145
.136
.027
.893
.171
28
.622
.147
.144
.031
.739
.214
23
.644
.120
.115
.024
.827
.269
25
.648
.132
.128
.023
.020
.166
32
0.496
0.130
0.138
0.024
0.747
0.208
29
.646
.153
.144
.029
.957
.208
28
.572
.152
.161
.031
.800
.204
24
.574
.118
.17?
.H:>,
.80S
.473
II
THE NUMBER OF GROUPS IS NOT CONSISTENT WITH THE TWO-WAY SPECIFICATION. ONE-WAV ANALYSIS OF VARIANCE IS COMPUTED
AIL GROUPS COMBINED
(EXCePT CASES WITH UNUSED VALUES
FOR YiAR
MEAN
SIII.DEV.
R.E.S.D.
S. E. H.
KAX1WUII
HI III HUM
SAMPLE SIZE
AND
0.644
0.144
0.139
0.009
0.984
0.I6S
283
SOURCE
BETWEEN GROUPS
WITHIN GROUPS
TOTAL
SUM OF SQUARES
0.2218
S.6146
5.11364
OF MEAN SnUARC
10 0.0222
272 tt.HtUti
282
F VALUE TAIL PROBABILITY
1.07 0.3020
LEVEHE-S TEST FOR EQUAL VARIANCES
10. 272 0.44
a. 9,MO
ONE-WAY ANALYSIS OF VARIANCE
TEST STATISTICS FOR WUIIIN-GROUP
VARIANCES NOT AiUUHED TO iff EOUAL
WELCH
OROWN-FORSYTHE
10. 97
10. 26*
I.I*
NOTE - ONLY THOSE GROUPS WITH NON-ZERO VARIANCE ARE USED IN THE COMPUTATIONS.
3.40ft,
FIGURE 12'. Distribution of 24-hour DS/high-TSP ratios by year (1980, 1981, 1*982) and quarter
84003
-------
V. COAST
ARID
OTHER
ro
01
VAR 14
EXCLUDED
VALUES
**gB3
3948
MIDPOINTS
1.020)
TABULATIONS AND COMPUTATIONS WHICH FOLLOW EXCLUDE VALUES LISTED ABOVE
0.960)
0.930)
.0.900)
#.840)
0.810)"*
tf.760)**
I/. 720)*****
H.600)**1"'**
0.660)***»*
0.630)*
0.6/90)**
0.570)M******
0.640)***
J/.SU)*****
tf.4au)*«****«
M****
i».420)«*»"*
JJ.390)*
tf.360)
0. 33*)
*
4*
0.240)
0.210)*
0.100)
0.160) *
0.120)
GROUP MEANS ARE DENOTED 0V N'S IF THEV COINCIDE WITH *'S, N'S OTHERWISE
KEAN
STC.DEV.
R.E.S.I).
S. K. M.
MAXIMUM
MINIMUM
SAMPLE SIZE
0.581
0.149
0.1S4
0.018
0.984
0.214
72
ALL CROUPS COHBINED
(EXCEPT CASES WITH UNUSED VALUES
FOR CLIMATE >
MEAN
STD.PEV.
R.E.S.D.
S. E. M.
MAXIMUM
MINIMUM
SAMPLE SIZE
0.S44
0.144
0.139
0.009
0.984
0.155
203
0.543
0.158
0.153
0.016
0.957
0.171
94
SOURCE
BETWEEN GROUPS
WITHIN GROUPS
TOTAL
0.S22
0.124
0.122
0.012
0.747
0.155
117
SUM OF SQUARES
0. 1563
5.6802
5.8364
DF MEAN SQUARE
2 0.0781
200 0.0203
282
F VALUE TAIL PROBABILITY
3.85 0.0224
LEVENE'S TEST FOR EQUAL VARIANCES 2, 200 2.74
ONE-WAV ANALYSIS OF VARIANCE
TEST STATISTICS FOR WITHIH-GROUP
VAH I AHCF.S NOT ASSUMED TO BE EQUAL
WELCH . 2, 162
BROWN-FORSVTHE 2. 237
3.98
3.70
0.0661
0.0205
0.0261
FIGURE 13'. Distribution of 24-hour DS/high-TSP ratios by climate.
-------
REGION 1 REGION 2 REGION 3 REGION 4 REGION 5 REGION 6 REGION 7 REGION 0 REGION 9 REGION10
VAR 14
EXCLUDED
VALUES
MIDPOINTS
I .i'Z0>
0.090)
0.360)
«.93'J>
0.90JO)
0.H70)
0.84U)
0.819)
0.730)
0.7501
0./2J)
0.69)))
**40| 428 ******I264 *******820 739 209 **'
TABULATIONS AND COMPUTATIONS WHICH FOLLOW EXCLUDE VALUES LISTED ABOVE
231
0.600>M
0.070)
4.040)
0.40.0)
a.'
I'.J
U.;
J9.3JU)
0.31'.))
fl.;
M«
**
*
*
i *
H«
A * * * *
H-"
* *
*
*
>n *
|.|« **
* **
ro
cr>
K.2I4I
U.l&fl)
u.it.n
i».i z>j >
GROUP MEANS ARE DENOTED Bf H'S IF THEV COINCIDE WITH *'S. N'S OTHERWISE
MEAN
SIU.OEV.
K.F..S.D.
S. E. H.
HAKfNl'il
14 1 Mil IUII
SAMPLE SIZE
ALL
(EXCEPT CAS
P«JA KEUIQ
MEAN
STO.DEV.
P..E.S.D.
S. E. N.
IJ n 'J riJIll 1
"l\ * t llvrl
SAMPLE SIZE
0.603
0.000
0.000
0.000
0.603
0.60J
1
.477
.141
.134
.042
.690
.2*9
11
0.535 0.549
0.152 0.119
0.148 0.109
0.042 0.022
0.681 0.747
0.Z08 0.155
13 28
0.516
0.121
0.1 18
0.016
0.747
0 . 2iM
55
IS WITH UIIUSEO VALUFS
it )
0.544
B. 144
9 . i .n
a. an'}
0 . 904
283
* SOURCE
UETUEEN GROUPS
* WIIIIIN CROUPS
*
TOTAL
LEVLNE'S TESI FOR
SUM
LOUAL
0.538
0.147
0.142
0.020
0.9b7
0.171
54
OF SQUARES
0.8462
4.9902
G.B3M
VARIANCES
0.556
0.046
0.055
0.H20
0.593
0.497
5
OF VARIANCE
0.354
H . ?Bb
Il.?!i8
«.M5
W.49U
II.ZUU
2
Uf MEAM nilUARE
0
274
202
(I. 2/4
0.1058
0.01U2
0
a
u
if
H
a
F
.616
.146
. 14U
.016
. !>(M
.177
80
VALUE
5.HI
1 .t.6
0.457
£.110
0.0'J9
u. a; H
£'.Uf,'J
It. 1P4
3/1
TAIL PROBAffLITY
0
a
eeiia
1358
ONE-WAV ANALYSIS UK V.U.'IAHCE
iLsr siArisrits FOR ui IIIIH-GHOUP
VARIAHCLS NOT ASSUMED 10 HE EQUAL
WE 1C II
URUWN-FORSYTIIE
0.
U,
I!)
13
4.55
5.47
0.0032
0.0U36
NOTE - ONLY THOSE GROUPS WITH NON-ZERO VARIANCE ARC USED IN THE COMPUTATIONS.
FIGURE 14'. Distribution of 24-hour DS/high-TSP ratios by EPA region.
84003
-------
CA
VAR 14
EXCLUDED
VALUES
MIDPOINTS
I .020)
a.'JlO)
U.-.K.O)
D.'JJfl)
H . 9»IO )
M.U71O
1I.B4.Q)
U.81U)
0 . 7U.-J )
0. /.M)
fl.L'JJ)
(3T.CCO)
U.C3I)
1) . 1,10 )
TABULATIONS AHO COMPUTATIONS WHICH FOLLOW EXCLUUE VALUE!, L1SUO ABOVE
ro
0.4QJU)
H.4'.ifl)
fl.4?3)
fl.39.9)
0.3G.C)
D.JJOt
V.Z10)
0.249)
tf.104)
.
GROUI- n.-X.HS ARE DENOTED BV H'S IF THEY COINCIDE WITH «'S. N'S OTHERWISE
STD.IUV.
R.E.S.D.
S. t. M.
MAX IMUH
M Jill MUM
0.408 0.619
0.069 if. 155
0.082 1). 159
0.035 M.019
0.479 JI.984
0.326 0.177
SAN.UE SIZE 4 66
(EXCEPT CASES WITH UNUSED VALUES
FUR STATE )
MS AH
STli.DLV.
R . 1 . 6 . 0 .
S. E. M.
M J vii Iiini
f lAA I rlUFl
MINlMUII
Q * Mb I L~ PI *JC
a APlrl. L i*l £L
0.S44
0.144
0.139
0.009
0. 934
0.155
on i
Zu J
SOURCE SUM
BETWEEN GROUPS
WITHIN GROUPS
TOTAL
Li:vLNE'S TEST FOR EQUAL
ONE-WAY ANALYSIS <;f VAIS
OF SQUARES
1.2675
4.S689
5.0364
VARIANCES
IAIII:E
OF MEAN SUUARE
17 0.0746
265 0.4172
282
17, i-1,5
F VALUE TAIL PROBABILITY
4.32 B. Bit fa
2.31 il.audl
Ttsr S.TATIS1ICS FUK VM Til I II CROUP
VARIANCES NOT ASSUMEU TO UE EQUAL
\IELCil
OROWN-FOaSVTME
17. 29
17, 66
4 . 68 .0 . 001/2
4.116 .0.00U0
FIGURE 15'.
contrasted.
NOTE - ONLY THOSE GROUPS WITH NON-ZERO VARIANCE ARE USLD IN TKE COMMIT/*! ION?.
Distribution of 24-hour DS/high-TSP ratios by state. Minnesota and California are
Other states are not shown, but are included in the analysis of variance.
84003
-------
Minn., Minn.
Rubldoux, Ca.
VAR 14
EXCLUDED
V,UlttS
MIDPOINTS
i .
TABULATIONS AND COMPUTATIONS WHICH FOLLOW EXCLUDE VALUES LISltO ABOVE
i| . Ql'.O >
0.8/0)
0.843)
U.01UI
0.700)
«.7S'J>
0.72JJ)
0 . 690 1
O.G07)
C.639)
fl.COtfl
9.570)
0 . S4.U I
0.5191
0.481)
»»
*
fl.li'J)
U.399I
o.36*>
0.33*)
PO U.ZI4)
CO 0. 1801
0!l24> '
GROUP MCANS ARE DENOTED BY M'S IF THEY
MEAN 0.418
SrO.DEV. 0.081
R.E.S.D. 0.094
S. t. H. 0.047
MAXIMUM 0.4/9
MINIMUM 0.326
COINCIDE WITH *'S. N'S OTHERWISE
0.693
0.036
0.085
0.0IS
0.878
SAMPLE SIZE 3
Al 1 £Di*lllQC t?*nU&tllFfl **
AL L GKUUr S CUff D I HE 0
(EXCEPT CASES WITH UNUSED VALUES
FOR CITY )
MEAN
STD.DEV.
R.E.S.D.
S. E. M.
MAX I MUM
MINIMUM
SAtll'LE SIZE
0.S44
0.144
0.139
0.009
0.984
0.ISS
283
Q
3)
SOURCE
BETWEEN CROUPS
WITHIN CROUPS
TOTAL
SUM OF SQUARES
2.6215
3.3149
3.8364
* LEVENE'S TEST FOR EQUAL VARIANCES
*
ONE-WAV ANALYSIS OF
TEST STATISTICS FOR
VARIANCE
WITH 114 -GROUP
OF MEAN SQUARE
31 0.0813
251 0.0132
Z8Z
31, 2S1
F VALUE TAIL PROBABILITY
6.16 . 0.0000
2.84 0.0000
* VARIANCES NOT ASSUMED TO BE EQUAL
*
*
WELCH
BROUit-FORSVTHE
31 , 31
31 . 65
11.04 0.0000
6.lb 0.0000
NOTE - ONLY THOSE GROUPS WITH NON-ZERO VARIANCE ARE USCO IN THE COMPUTATIONS.
FIGURE 16'. Distribution of 24-hour DS/high-TSP ratios by city. Minneapolis and Rubidoux are
contrasted. High TSP ratios were recorded at a single site for both cities. Other cities are
not shown, but are included in the analysis of variance.
8U003
-------
VAR 14
EX.LUDCO
VALUES
MIDPOINTS
1.020)
Rubldoux
Boise, Idaho
....«-...64 «-*.19
TABULATIONS AND COMPUTATIONS WHICH FOLLOW EXCLUDE VALUES LISILl) ABOVE
ro
vo
0. UP1))
U.U7J)
i). 84tf)
0 010)
H.70W)
a.720}
(I.60JV)
!'.(
U.(
«*.:
«.!
*
*««
U. 39.1/1 ***
IJ.36iJ) H
£.130)
0. JOJ)
0.270)
i a.210)
u.ieoi *
0. !!.#>
0.12^1 i
GRO.ii' II ANS ARE DENOTED BV M'S IF THEV COINCIDE WITH *'S. N'S OTHERWISE
ME AH
STIi.DEV.
K. F . S. D .
S. I.'. H.
MAXlMUi-t
Ml III HUH
SAI'.PLt SIZE
0.693
C.PUb
ALL GROUPS COMBINED
T CAStS WITH UNUSED VALUES
MtAII
i'ri>. DCV.
i! . L' . S . I) .
r, . L . M .
II/.X.IIUH
MINU'IUM
SAMPLE SIZE
0.544
0.144
0.139
0.009
0.904
0. 155
283
ri
*
*
M
*
*
M
«
*
*
*
H
Jff.077
B.077
v.an
0.421
0.104
0
SOURCE
OUVtEEN CROUPS
WITHIN GROUPS
TOTAL
LCVINE'S TEST FOR
OHf -WAY ANALYSIS
IKbT STATISVICS F
VAP.lAIILtS NOT ASS
WELCH
UROWH-FOUSVTIIE
SUM OF SQUARES
2.06/S
J. Ib9t)
G.Q364
EQUAL VARIANCES
OF VARIAMCb
OR Wl YHIN-GROUP
HMO 10 BE EQUAL
OF IHEAH !>OUARc
33 «.i)H9U
249 U. 1/127
282
33. 249
33. 33
33. 66
F VALUE TAIL PilOBAllILIT/
6.35 0.00911
i
2.
-------
198*
1981
1982
00
O
VAR 14
EXCLUDED
VALUES
a***********.********«****H4 ****************«**********73 **»*****«**«
TABULATIONS AND COMPUTATIONS WHICH FOLLOW EXCLUDE VALUES LISTED ABOVE
MIDPOINTS
1.085)
1.050)
1.016) *
.980)
.946)
.9)0) *
.875)
.8401* *
.80S)
.770)*** .
.736)** *
.700)***** ****** *
.666)******* ** «
.630)***«"*« *.«« .******
,696)******* **************** **********
.6601M******* M****«*« .«.*.*«..
.625)********* ********** H********
.490)**.**.** *«****..**.**** ............
.465)*** ..****** **
.420)*** ****** ***
.386)** *
.360)**
.316)
.280>*
.246)
.210) «
.175)
.140)
CROUP MEANS ARE DENOTED BV H'S IF THEV COINCIDE WITH *'S. N'S OTHERWISE
105
MEAN
STD.DEV.
R.E.S.O.
S. E. M.
MAXIMUM
MINIMUM
SAMPLE SIZE
0.673
0.112
0.111
0.014
0.841
0.277
60
ALL GROUPS COMBINED
(EXCEPT CASES WITH UNUSED VALUES
FOR VEAR >
MEAN
STD.OEV.
R.E.S.D.
S. E. M.
MAXIMUM
MINIMUM
SAMPLE SIZE
0.664
0.116
0.110
0.007
1.020
0.197
264
0.661
0.127
0.115
0.012
1.020
0.197
109
SOURCE
BETWEEN GROUPS
WITHIN GROUPS
TOTAL
LEVENE-S TEST FOR
0.541
0.101
0.102
0.012
0.768
0.316
77
SUM Of SQUARES
0.0380
3.3614
3.3994
EQUAL VARIANCES
OF MEAN SQUARE F VALUE
2 0.0190 1.42
261 0.0134
263
2, 251 0.48
TAIL PROBABILITY
0.2437
B.6210
ONE-WAV ANALYSIS OF VARIANCE
TEST STATISTICS FOR WITHIN-GROUP
VARIANCES NOT ASSUMED TO BE EQUAL
WELCH
BROWN-FORSVTHE
2, 168
2, 240
1.60
1.49
0.2049
0.2279
FIGURE 18'. Distribution of annual DS/TSP ratios by year.
-------
V. COAST
ARID
OTHER
VAR 14
EXCLUDED
VALUES
44
223
TABULATIONS AND COMPUTATIONS WHICH FOLLOW EXCLUDE VALUES LISTED ABOVE
MIDPOINTS
1.0BS)
1.060)
1.015) * "
0.980)
0.945)
0.910)*
0.875)
0.840)* "
0.805)
0.770)" "" *
0.735)* «
0.700)*
0.665)*
0.630)*
0.596)*
0.5601M
0.S25)*
0.490)*
0.466)* * **********
0.420)*****
0.386)"*
0.350)
0.316)*
0.280)* «
0.246)
0.210) *
0.176)
0.140)
CROUP MEANS ARE DENOTED BV M'S IF THEV COINCIDE WITH *'S, N'S OTHERWISE
MEAN
STD.OEV.
R.E.S.D.
S. E. M.
MAXIMUM
MINIMUM
SAMPLE SIZE
.645
.134
.129
.022
.896
.293
38
0.647
0.140
0.135
0.022
1.020
0.333
40
0.SS7
0.106
0.I0I
0.008
1.010
0.197
176
ALL GROUPS COMBINED
(EXCEPT CASES WITH UNUSED VALUES
FOR CLIMATE )
MEAN
STD.DEV.
R.E.S.D.
S. E. M.
MAXIMUM
MINIMUM
SAMPLE SIZE
0.654
0.116
0.110
0.007
1.020
0.197
264
SOURCE
BETWEEN GROUPS
WITHIN GROUPS
TOTAL
....... ANALYSIS OF VARIANCE TABLE «*
SUM OF SQUARES OF MEAN SQUARE
0.0062
3.3932
3.3994
2
261
263
0.0031
0.0135
F VALUE TAIL PROBABILITY
0.23 0.7949
LEVENE'S TEST FOR EQUAL VARIANCES 2, 251 2.78 0.0640
ONE-WAV ANALYSIS OF VARIANCE
TEST STATISTICS FOR WITHIN GROUP
VARIANCES NOT ASSUMED TO BE EQUAL
WELCH
BROWN-FORSVTHE
2, 63
2, 93
0.19
0.18
0.8287
0.8383
FIGURE 19'. Distribution of annual DS/TSP ratio by climata.
64003
-------
U)
ro
REGION I REGION 2 REGION 3 REGION 4 REGION 5 REGION 6
» » + * + H
VAR 14
EXCLUDED
VALUES
21 jg gg 45 4g *
TABULATIONS AND COMPUTATIONS WHICH FOLLOW EXCLUDE VALUES LISTED
MIDPOINTS
1.086)
1.060)
1.016)
0.980)
0.946)
0.910)
0.876)
0.840)
0.806)
0.770)
0.736)
0.700)*
0.666)*
0.630)* ]
0.696)*** « *
0.660)M* ** M***
0.626)"** M" "*"
0.490) »
0.466)
0.420) " *
0.386) * *
0.360)
0.315)
0.280) *
0.246)
0.210)*
0.176)
0.140)
GROUP MEANS ARE DENOTED BV M'S IF THEV COINCIDE WITH *'S. N'S OTHERWISE
REGION
7 REGION
. . . +
8 REGION
. . . +
REGIONI0
*»
ABOVE
16
30
*
**
*
**
13
***
H
*
*
M*
*
*
H
*
MEAN
STD.DEV.
R.E.S.D.
S. E. M.
MAN IMUM
MINIMUM
SAMPLE SIZE
.660
.124
.098
.036
.684
.197
12
,6)6
,093
094
022
,709
,356
18
0.660
0.096
0.102
0.014
0.707
0.311
49
.686
.110
,100
,020
,841
,333
30
0.646
0.094
0.090
0.014
0.749
0.312
46
0.666
0.164
0.146
0.031
1.010
0.277
24
.674
,088
,090
,024
,733
,448
14
0.496
0.109
0.119
0.046
0.617
0.352
6
0.671
0.148
0.139
0.024
1.020
0.369
38
.600
.123
.13S
.029
.700
.293
18
ALL CROUPS COMBINED
(EXCEPT CASES WITH UNUSED VALUES
FOR REGION )
****
MEAN
STD.DEV.
R.E.S.D.
S. E. M.
MAXIMUM
MINIMUM
SAMPLE SIZE
0.654
0.116
0.110
0.007
1.020
0.197
264
SOURCE
BETWEEN GROUPS
WITHIN GROUPS
TOTAL
SUM OF SQUARES
0.1356
3.2638
3.3994
ANALYSIS OF VARIANCE TABLE *
MEAN SQUARE
OF
9
244
263
0.0161
0.0134
VALUE TAIL PROBABILITY
I.13 0.3446
LEVENE'S TEST FOR EQUAL VARIANCES
9. 244
1.40
0.1893
ONE-WAV ANALYSIS OF VARIANCE
TEST STATISTICS FOR WITHIN-GROUP
VARIANCES NOT ASSUMED TO BE EQUAL
WELCH
BROWN-FORSVTHE
9.
9.
69
144
1.10
1.12
0.3779
6.3529
FIGURE 20'. Distribution of annual DS/TSP ratios by EPA region.
84003
-------
WA
NC
CO
VAR 14
EXCLUDED
VALUES
MIDPOINTS
««..««>«
TABULATIONS AND COMPUTATIONS WHICH FOLLOW EXCLUDE VALUES LISTED ABOVE
1.060)
1.015)
0.980)
0.945)
0.910)
0.875)
0.840)
0.805)
.770)
0.736)
0.700)
0.666)
0.630)
0.696)
0.560)
0.626)
0.490)
0.465)
0.420)
0.385)
0.360)
0.315)
0.280)
0.246)
0.210)
0.176)
0.140)
GROUP MEANS ARE DENOTED BV M1
MEAN
STD.DEV.
R.E.S.D.
S. E. M.
MAXIMUM
MINIMUM
SAMPLE SIZE
S IF THEY COINCIDE WITH *'S. N'S OTHERWISE
0.443 0.629
0.087 0.0B8
0.090 0.093
0.031 0.027
0.664 0.770
0.293 0.469
8 II
ALL GROUPS COMBINED
(EXCEPT CASES WITH UNUSED VALUES
FOR STATE >
MEAN
STO.DEV.
R.E.S.O.
S. E. M.
MAXIMUM
MINIMUM
SAMPLE SIZE
0.664
0.116
0.110
0.007
I .020
0.197
264
SOURCE
BETWEEN GROUPS
WITHIN GROUPS
TOTAL
....... ANALYSIS OF VARIANCE TABLE
SUM OF SQUARES OF MEAN SQUARE
F VALUE TAIL PROBABILITY
0.B2I0
2.6784
3.3994
36
218
263
0.0235
0.0118
1.98
0.0017
LEVENE'S TEST FOR EQUAL VARIANCES
35, 218
1.61
0.0400
ONE-WAV ANALYSIS OF VARIANCE
TEST STATISTICS FOR UITHIN-UROUP
VARIANCES NOT ASSUMED TO BE EQUAL
WELCH
BROWN-FORSVTHE
36.
36.
36
78
11.34
2.63
0.0000
0.0002
NOTE - ONLY THOSE GROUPS WITH NON-ZERO VARIANCE ARE USED IN THE COMPUTATIONS.
FIGURE 21'. Distribution of annual DS/TSP ratios by state. The states of Washington and
North Carolina are contrasted. Other states are not shown, but are included in the analysis
of variance.
84003
-------
Buffalo, M.Y.
14.Y., N.Y.
VAR 14
EXCLUDED
VALUES
U)
MIDPOINTS
1.085)
1.050)
1.015)
.980)
.945)
.91*)
.075)
.840)
.805*
.770)
.735)
.700)
.665)
.630)
.695)
.560)
.625)
.490)
.455)
.420)
.385)
.350)
.315)
.280)
.245)
.210)
. m>
.140)
CROUP MEANS ARE
MEAN
STO.OEV.
R.E.S.O.
S. E. H.
MAX IHUH
H1NIHUN
SAMPLE SIZE
TABULATIONS AND COMPUTATIONS WHICH FOLLOW EXCLUDE VALUtS LISTED ABOVE
DENOTED BV M'S IF THEY COINCIDE WITH *'S. N'S OTHERWISE
0.567
0.113
M.I49
0.0M
0.781
0.376
S
ALL CROUPS COMBINED
(EXCEPT CASES WITH UNUSED VALUES
FOR CITV I
MFAN
STD.DEV.
R.E.S.O.
S. E. H.
MAXIMUM
MINIMUM
SAMPLE SIZE
0.S54
0.116
0.110
0.007
1.020
0.197
254
0.S21
0.037
0.044
0.019
0.562
0.479
4
SOURCE
BETWEEN GROUPS
WITHIN GROUPS
TOTAL
SUM OF SQUARES OF MEAN SQUARE
1.6453 79 0.0208
1.7S4I 174 0.0101
3.3994 253
LEVCHE'S TEST FOR EQUAL VARIANCES 79. 174
ONE -WAV ANAL VS IS OF VARIANCE
TEST STATISTICS FOR WITHIN -UROUH
VARIANCES NOT ASSUMED TO BE EQUAL
WELCH 79, 37
BROWN-FORSVTHE 79. 13
F VALUE TAIL PROBABILITY
2.07 0.0000
5.86 0.0000
27.48 0.0000
1.64 0.1600
NOTE - ONLY THOSE GROUPS WITH NON-ZERO VARIANCE ARE USED IN THE COMPUTATIONS.
FIGURE 22'. Distribution of annual DS/TSP ratios by city. Buffalo, New York and New York City,
New York are contrasted. Ratios were recorded at two sites in each city. Other cities are not
shown, but are included in the analysis of variance.
81*003
-------
gubldpux, C/i.
₯111.Co., IJ1.
CO
en
VAR 14
EXCLUDED
VALUES
TABULATIONS AND COMPUTATIONS WHICH FOLLOW EXCLUDE VALUES LISTED ABOVE
MIDPOINTS
!.*>
1.050)
1.015)
.980)
.945)
.9101
.875)
.640)
.80S)
.770)
.735) N
.700)
.665)
.630)
.595)
.560) *
.525)
.490)
.455)
.420)
.385)
.350) N
.315)
.280)
.245)
.210)
.175)
.140)
WOUP MEANS ARE DENOTED BV H'S IF THEV COINCIDE WITH *'S. N'S OTHERWISE
MEAN
STO.DEV.
R.E.S.D.
S. E. M.
MAXIMUM
MINIMUM
SAMPLE SIZE
.719
.139
.149
.080
.849
.673
3
0.359
0.066
0.082
0.046
0.405
0.312
2
ALL GROUPS COMBINED
(EXCEPT CASES WITH UNUSED VALUES
FOR SITt. )
MEAN 0.SS4
STU.DEV. 0.116
R.E.S.O. 0.110
S. E. M. 0.007
MAXIMUM 1.020
MINIMUM 0.197
SAMPLE SIZE 254
SOURCE SUM OF SQUARES
BETWEEN GROUPS 1.8332
WITHIN CROUPS 1.5662
TOTAL 3.3994
LEVENE'S TEST FOR EQUAL VARIANCES
ONE-WAV ANALYSIS OF VARIANCE
TEST STATISTICS FOR WITHIN-GROUP
VARIANCES NOT ASSUMED TO BE EQUAL
WELCH
BROWN-FORSVTHE
DF MEAN SQUARE
94 0.0195
159 0.0099
253
94. 159
94. 39
94. 17
F VALUE
1.98
9.27
22.09
I.b2
TAIL PROBABILITY
0.0001
0 . 0000
0.0000
0.1626
NOTE - ONLV THOSE GROUPS WITH NON-ZERO VARIANCE ARE USED IN THE COMPUTATIONS.
FIGURE 23'. Distribution of annual DS/TSP ratios by site. Sites in Rubidoux, California and
Will County, Illinois are contrasted. Other sites are not shown, but are included in the
analysis of variance.
84003
-------
0-56
CO
CT>
VAR 14
EXCLUDED
VALUES
MIDPOINTS
l.0dS)
1.PG0)
I.015)
.980)
.945)
.910)
.875)
.040)
TABULATIONS AND COMPUTATIONS WHICH FOLLOW EXCLUDE VALUES LISftD ABOVE
79
.770)
.735)
.700)
.665)
.630)
.595)
.660)
.525)
.490)
.465)
.420)
.385)
.350)
.315) '
.280)
.246)
.210) *
.1751
.140)
GROUP MEANS ARE DENOTED BV M'S IF THEV COINCIDE WITH $. N'S OTHERWISE
*<
30
34
MEAN
STD.DEV. <
R.E.S.D.
S. E. M.
MAXIMUM
MINIMUM
SAMPLE SIZE
ALL GROUPS COMBINED
(EXCEPT CASES WITH UNUSED VALUES
FOR MHIVOL )
0.587
0.144
0.124
0.021
1.020
0.197
49
0.546
a. in?
0. 104
a.BUI
0,896
0.293
205
MEAN
STD.DEV.
R.E.S.D.
S. E. M.
MAXIMUM
MINIMUM
SAMPLE SIZE
0.664
0.116
0.110
SOURCE
BETWEEN GROUPS
WITHIN GROUPS
TOTAL
ANALYSIS OF VARIANCE TABLE ««
SUM OF SQUARES OF MEAN SQUARE
O.J688
3.3306
3.3994
I
252
253
0.0688
0.0132
F VALUE TAIL PROBABILITY
6.21 0.0233
1.020
0.197
254
LEVENE'S TEST FOR EQUAL VARIANCES
I. 2S2
1.66
0.1985
ONE-WAV ANALYSIS OF VARIANCE
TEST STATISTICS FOR W1THIH-GROUP
VARIANCES NOT ASSUMED TO BE EQUAL
WELCH
BROWN-FORSVTHE
I. 61
I. 61
3.62
3.62
0.0619
0.0619
NOTE - ONLV THOSE GROUPS WITH NON-ZERO VARIANCE ARE USED IN THE COMPUTATIONS.
FIGURE 24'. Distribution of annual DS/TSP ratios by annual TSP concentrations (low versus
high TSP, with high TSP defined as concentrations greater than 55 yg/m-3).
84003
-------
1980
1981
1982
VAR 14
EXCLUDED
VALUES
MIDPOINTS
.925)
it.....*.........*.*.*.*....95 »»»»»»»»»»«»»»»«««««»»»56 »» »*»»»»*»**»»»*»»»«94
TABULATIONS AND COMPUTATIONS WHICH FOLLOW EXCLUDE VALUES LISTED ABOVE
.876)
.8601*
.825)
.800)
.775)"
.750)*
.725)
.700)"
.675)"
.650)*
**
.625)*
.600)*
.676 )H
.550)*
.525)*
,600)***«**** «*
,475). ««
.460)* .*««
.425)
.400)*** ***
.376)* *
.360)"
.326)*
.300) ***
.275)
CROUP MEANS ARE DENOTED BY M'S IF THEY COINCIDE WITH
M****
**
**
'S, N'S OTHERWISE
MEAN
STD.DEV.
R.E.S.D.
S. E. M.
MAXIMUM
MINIMUM
SAMPLE SIZE
0.664
0.106
0.103
0.014
0.841
0.333
63
,539
,111
,107
,012
,896
,293
90
0.539
0.102
a.me
0.013
0.768
0.315
62
ALL CROUPS COMBINED
(EXCEPT CASES WITH UNUSED VALUES
FOR VEAR )
MEAN 0.546
STD.DEV. 0.107
R.E.S.D. 0.104
S. E. M. 0.007
MAXIMUM 0.896
MINIMUM 0.293
SAMPLE SIZE 20S
SOURCE
....... ANALVSIS OF VARIANCE TABLE **
SUM OF SQUARES OF MEAN SQUARE
F VALUE TAIL PROBABILITY
*
*
*
*
BETWEEN GROUPS
WITHIN GROUPS
TOTAL
LEVEME'S TEST FOR
0.0230
2.3071
2.3301
EQUAL VARIANCES
2
202
204
2, 202
0.0116
0.0114
1.01
0.12
0.3675
0.8B93
ONE-WAY ANALVSIS OF VARIANCE
TEST STATISTICS FOR WiTHIN-GROUP
VARIANCES NOT ASSUMED TO BE EQUAL
WELCH
BROWN-FORSVTHE
2, 123
2, 185
1.02
1.03
0.3643
0.3606
FIGURE 25'. Distribution of annual DS/high-TSP ratios by year.
8U003
-------
W.COAST
ARID
OTHER
VAR 14
EXCLUDED
VALUES
36
190
CO
00
TABULATIONS AND COMPUTATIONS WHICH FOLLOW EXCLUDE VALUES LISTED ABOVE
MIDPOINTS
.925)
.900)*
.875)
.850)" "
.826)
.800)
.776)" """
.750) **
.726)*
.700)
.676) *» *
.660)" * «
.626)*"* *" »*«««
.600, .........to.
.575)**«
.550)M*" M M*
.626) * «**
.500)*"*
.476)*"
.460)"" «
.426)"** «
.400)**
.376)
.350) **
.326)"
.300)*
.276)
GROUP MEANS ARE DENOTED BY M'S IF THEV COINCIDE WITH *'S, N'S OTHERWISE
***«
*
MEAN
STD.DEV.
R.E.S.D.
S. E. M.
MAXIMUM
MINIMUM
SAMPLE SIZE
.646
.143
.137
.027
.896
.293
29
.642
.120
.121
.020
.769
.333
36
.647
.095
.093
. ana
.841
.311
.141
ALL GROUPS COMBINED
(EXCEPT CASES WITH UNUSED VALUES
FOR CLIMATE )
SOURCE
...*.«« ANALYSIS OF VARIANCE TABLE **«"*««
SUM OF SQUARES OF MEAN SQUARE F VALUE TAIL PROBABILITY
MEAN
STD.DEV.
R.E.S.O.
S. E. H.
MAXIMUM
MINIMUM
SAMPLE SIZE
.646
.107
.104
.007
.896
.293
206
BETWEEN GROUPS
WITHIN GROUPS
TOTAL
LEVENE'S TEST FOR EQUAL
0.0005
2.3296
2.3301
VARIANCES
2
202
204
2, 202
0.0003
0.0116
0.02
3.77
0.9768
0.0248
ONE-WAV ANALYSIS OF VARIANCE
TEST STATISTICS FOR W1THIN-GROUP
VARIANCES NOT ASSUMED TO BE EQUAL
WELCH
BROWN-FORSVTHE
2. 50
2. 68
0.02
0.02
0.9800
0.9834
FIGURE 26'. Distribution of annual DS/high-TSP ratios by climate.
84003
-------
CO
UD
**«***
REGION 1 REGION 2 REGION 3 REGION 4 REGION 5 REGION 6 REGION 7 REGION 8 REGION 9 REGION10
»... » » * * * * * *
VAR 14
EXCLUDED
VALUES
********I7 ********j7 *******«49 »*42 ********40 »****« ««*«»«
TABULATIONS AND COMPUTATIONS WHICH FOLLOW EXCLUDE VALUES LISTED ABOVE
MIDPOINTS
0.925)
0.900)
0.875)
0.850) *
0.426)
0.800)
0.775)
0.760)
0.725)
0.700) *
0.676)
0.650) *
0.626)
0.600)
0.576)"
0.560)H* **
0.525>"
*
H*
«*
**
**
*****
M
**«
***
0.500)
0.476)
0.460)
0.425)
0.400)
0.375)
0,350)
0.325)
0.300) * *
0.276)
CROUP MEANS ARE DENOTED BV H'S IF THEV COINCIDE WITH "'S, N'S OTHERWISE
N
*
*
*
MEAN
STD.DEV.
R.E.S.D.
S. E. M.
MAXIMUM
MINIMUM
SAMPLE SIZE
0.546
0.024
0.026
0.011
0.681
0.621
,511
,100
,099
,027
,709
,366
14
ALL CROUPS COMBINED
(EXCEPT CASES WITH UNUSED VALUES
FOR REGION )
MEAN
STD.DEV.
R.E.S.D.
S. E. M.
MAXIMUM
MINIMUM
SAMPLE SIZE
,546
107
,104
.007
,896
,293
206
0.548 .671 0.545 0.547
0.099 .106 0.097 0.111
0.100 .090 0.092 0.115
0.016 .021 0.016 0.024
0.707 .841 0.749 0.768
0.311 .333 0.312 0.333
41 26 40 22
H
* SOURCE SUM OF SQUARES
*
* BETWEEN GROUPS 0.1326
* WITHIN GROUPS 2.1976
*
* TOTAL 2.3301
* LEVENE'S TEST FOR EQUAL VARIANCES
* ONE-WAV ANALYSIS OF VARIANCE
TEST STATISTICS FOR WITHIN-GROUP
* VARIANCES NOT ASSUMED TO BE EQUAL
* WELCH
* BROWN-FORSVTHE
0.562
0.090
0.090
0.028
0.733
0.448
0.518
0.104
0.098
0.046
0.617
0.362
10 6
DF MEAN SQUARE
9
195
204
9. 195
9, 44
9, 113
0.0147
0.0113
0.571
0.137
0.133
0.025
0.896
0.369
30
F VALUE
1.31
1.21
1.04
1.45
.470
.113
.117
.031
.654
.293
13
TAIL PROBABILITY
0.2361
0.2899
0.4234
0.1763
FIGURE 27'. Distribution of annual DS/high-TSP ratios by EPA region.
84003
-------
OH
WA
VAR 14
EXCLUDED
VALUES
TABULATIONS AND COMPUTATIONS WHICH FOLLOW EXCLUDE VALUES LISTED ABOVE
MIDPOINTS
0.925)
.900)
.876)
.860)
.826)
.800)
.776)
.750)
.726)
.700)
.676)
.650)
.626)
.600)
.676) M«*«
.650)
.626) **
.500) »* *
.475)
.450) *
.425) M
.400) ««
.376)
.360)
.325)
.300) *
.276)
CROUP MEANS ARE DENOTED BV M'S IF THEV COINCIDE WITH *'S, N'S OTHERWISE
MEAN
STD.DEV.
R.E.S.D.
S. E. M.
MAXIMUM
MINIMUM
SAMPLE SIZE
.574
.083
.089
.017
.749
.481
23
ALL CROUPS COMBINED
(EXCEPT CASES WITH UNUSED VALUES
FOR STATE )
MEAN
STD.DEV.
R.E.S.D.
S. E. M.
MAXIMUM
MINIMUM
SAMPLE SIZE
.546
.107
.104
.007
.896
.293
206
0.428
0.080
0.079
0.030
0.642
0.293
7
SOURCE SUM OF SQUARES OF MEAN SQUARE F VALUE TAIL
*"*
BETWEEN CROUPS 0.5846
WITHIN CROUPS 1.7464
TOTAL 2.3301
LEVENE'S TEST FOR EQUAL VARIANCES
ONE-WAV ANALYSIS OF VARIANCE
TEST STATISTICS FOR WiTHIN-GROUP
VARIANCES NOT ASSUMED TO BE EQUAL
WELCH
BROWN-FORSVTHE
NOTE - ONLV THOSE GROUPS WITH NON-ZERO.
Of
annual DS/high-TSP ratios
33 0.0177 1.74
171 0.0102
204
33. 171 1.69
33, 26 11.81
33, 53 2.37
VARIANCE ARE USED IN THE COMPUTATIONS.
by state. The states of Ohio
PROBABILITY
0.0128
0.0313
0.0000
0.0025
and
FIGURE 28'. Distributi .
Washington are contrasted. Other states are not shown, but are included in the analysis
of variance.
-------
W11J Co., 111.
El Cajon, Ca.
VAR 14
EXCLUDED
VALUES
TABULATIONS AND COMPUTATIONS WHICH FOLLOW EXCLUDE VALUES LISTED ABOVE
MIDPOINTS
0.02E)
a.900)
0.87S)
0.860)
0.826)
0.800)
0.776>
0.750)
0.725)
0.700) H
0.675)
0.650)
0.625) *
0.600)
0.575)
0.550)
0.525)
0.500)
0.475)
0.450)
0.425)
0.400)
0.375)
,0.360) N
0.326)
0.300)
GROUPZMEANS ARE DENOTED BV M'S IF THEV COINCIDE WITH $. N'S OTHERWISE
MEAN
STD.DEV.
R.E.S.D.
S. E. M.
MAXIMUM
MINIMUM
SAMPLE SIZE
.359
.066
.082
.046
.405
.312
2
0.710
0.103
0.129
0.073
0.783
0.637
2
ALL GROUPS COMBINED
(EXCEPT CASES WITH UNUSED VALUES
FOR CITV >
MEAN 0.646
STD.OEV. 0.107
R.E.S.D. 0.104
S. E. M. 0.007
MAXIMUM 0.896
MINIMUM 0.293
SAMPLE SIZE 206
NOTE - ONLV THOSE GROUPS WITH NON-ZERO VARIANCE ARE USED IN THE COMPUTATIONS.
FIGURE 29'. Distribution of annual DS/high-TSP ratios by city. Will County, Illinois and El
Cajon, California are contrasted. High-TSP ratios were recorded at one site in each city.
SOURCE SUM OF SQUARES
BETWEEN GROUPS 1.2603
WITHIN GROUPS 1.0698
TOTAL 2.3301
LEVENE'S TEST FOR EQUAL VARIANCES
ONE-WAV ANALYSIS OF VARIANCE
TEST STATISTICS FOR W1THIH-GROUP
VARIANCES NOT ASSUMED TO BE EQUAL
WELCH
BROWN-FORSVTHE
DF MEAN SQUARE
64
140
204
64, 140
64. 29
64, 20
0.0197
0.0076
F VALUE
2.68
3.94
18.77
2.24
TAIL PROBABILITY
0.0000
0.0000
0.0000
0.0234
84003
-------
VAR 14
EXCLUDED
VALUES
MIDPOINTS
.926)
.900)
.876)
.860)
.825)
.800)
.776)
.760)
.725)
.700)
.675)
.660)
.625)
.600)
.676)
.560)
.626)
.600)
.475)
.450)
.425)
.400)
.376)
.350)
.326)
.300)
.275)
KJ.1.1! Co.,,111.
Cajon,
TABULATIONS AND COMPUTATIONS WHICH FOLLOW EXCLUDE VALUES LISTED ABOVE
GROUP MEANS ARE DENOTED BY N'S IF THEY COINCIDE WITH 'S, N'S OTHERWISE
MEAN
STD.DEV.
R.E.S.D.
S. E. H.
MAXIMUM
MINIMUM
SAMPLE SIZE
0.359
0.066
0.082
0.046
0.405
0.312
2
0.710
0.103
0.129
0.073
0.783
0.637
2
ALL GROUPS COMBINED
(EXCEPT CASES WITH UNUSED VALUES
FOR SlTC )
MEAN 0.646
STD.DEV. 0.107
R.E.S.D. 0.104
S. E. M. 0.007
MAXIMUM 0.896
MINIMUM 0.293
SAMPLE SIZE 205
***************«****«*«*«
FU
SOURCE
BETWEEN GROUPS
WITHIN GROUPS
ANALYSIS OF VARIANCE TABLE **'
SUM OF SQUARES DF MEAN SQUARE
1.4309
0.8992
73
131
0.0196
0.0069
F VALUE TAIL PROBABILITY
2.86 0.0000
TOTAL 2.3301 204
****«***»«*«*«******
LEVENE'S TEST FOR EQUAL VARIANCES 73, 131 8.36 0.0000
0«K««»««*«««««*««««««»
ONE-WAY ANALYSIS OF VARIANCE
TEST STATISTICS FOR WITHIN-GROUP
VARIANCES NOT ASSUMED TO BE EQUAL
WELCH
BROWN-FORSYTHE
73.
73.
29
22
16.83
2.07
0.0000
0.0287
NOTE - ONLY THOSE GROUPS WITH NON-ZERO VARIANCE ARE USED IN THE COMPUTATIONS.
D!s*!f!but!on of annual DS/high-TSP ratios by site. Sites in Will County, Illinois
C°ntraSted- «*" '»« -e not shown but are included in tne
-------
N- 237
COR= .9647
MlAN
X 54.263
Y 44.G32
ST.DEV. REGRESSION LINE RES. MS.
32.669 X- 1.1504*Y+ 2.2429 74.319
27.158 Y- . 801 94*X-t- 1.3123 51.359
ZERO INTERCEPT REGRESSION LINE RES. MS.
Y = 0.81971*X 51.6382
*************
'.***+ X. .Y.+ '
200
175
150
125
100
75
25
0.0
1
2
11 111
11 1
2 1
1 22 1
111 1
221 2 1
1 1 1
11 4
112321 1
154 1
232 1
7331 1
23776 1 11
286313
17D921
CF122
BF934 1
383
8932
2
*
*
*
*
*
*
*
*
Y +
0.0
25.
+
50.
75.
. * .
100
125
150
175
200
225
250
. . * . .
275
300
..+....
325
3
350
DS
FIGURE 31a.. Regression analyses for 24-hour concentrations: PM 10 versus DS
81+003
43
-------
N = 265
COR= .H916
MEAN ST.DEV. REGRESSION LINE RES. MS.
X 92.468 49.215 X- 1.6673*Y+ 18.292 498.56
Y 44.490 26.318 Y- ,47680*X+ .40157 142.58
ZERO INTERCEPT REGRESSION LINE RES. MS.
Y = 0.48018*X 142.0728
* * xx***x*«*«*w* *** ******XXK ******************* + * *y + * * + Y + * +*
»«o»~«A~ "?""
2ZZ + *
175
IStf
1?5
l l l l l
100 + 111 +
11 1
1 31 *
2 1 *
1 111 1 11 2 1
75. + 2 11 1 +
1 12 11 1 *
1 12112 111 *
112 1 1 1 *
2 1 12 11 1 *
5-3. + 42333211 *
1114214 211 *
13351232 1 1 *
133 533411 1 1 *
1365341113 *
25. + 13AD2421 21 1 *
14122 *
14554 21 *
. 1131 *
35. 105 175 245 315 385 455
0.0 70. 140 210 280 350 420 490
TSP
FIGURE 31b. PM 10 versus TSP.
8*t003
44
-------
co
o
320
280
24JET
2J0
160
80.
40.
0.0
.
;
*
+
*
,
.
f
.
+
.
.
,
.
+
.
.
.
»
+
.
.
.
.
+
,
,
. 3
. 4*
Y 51
.X
0.0
N= 5733
COR= .SS46
MEAN ST.DEV.
X 72.696 39.553
Y 39.894 24.213
ZERO INTERCEPT
Y = 0.54709*X
***************************************
i
1
i 11
l i i
i i
11 11 l
1 1 21 1
1 1 1
1 212 21313 1 21
2111 1 2112 4421 1 1
21 12221514 11 21
1 121 53221312 1 1
1 12 36 5312 221 11
443328379642341 1 11
112286D79A96235211 1
44488GGEBCA44422 1 11 1
1167GILKPBE836522
46HLV***YCEB634 1 11
3FU*****USI5212131 1 1
7E******OPA45621 2
1CX******QGA1 2 13 1 11 1
50*****«*RD4332 21 11
4U******MB31111111 1
*****OKA42 311 1
**YE7241
21 1
+ + + + + + + +
35. 105 175 245
70. 140 210 280
REGRESSION LINE RES. MS.
X- 1.4453*Y+ 15.036 340.60
Y= .54136*X-f .53966 127.57
REGRESSION LINE RES. MS.
127.6199
* * .+....**. ..+....+*. . . +*
X
*
*
*
Y
*
1 *
*
*
.
1 1
1
*
*
1 *
*
*
1 *
*
*
*
*
1 *
*
*
*
*
*
*
*
*
*
*
*
315 385 455
350 420 490
TSP
FIGURE 31c. DS versus TSP.
8^003
-------
N- 19
COR= .8985
MEAN ST.DEV. REGRESSION LINE RES.MS.
X 47.6?8 18.444 X» 1.2179*Y+ 3.2985 69.398
Y 36.448 13.607 Y- .66289*X+ 4.8359 37.773
ZERO INTERCEPT REGRESSION LINE RES.MS.
Y = 0.75171*X 38.739
* *+ * y ** y + + 4. + ft + *
^ A aT****'****^** *»'» ~
75.J0T + 1 +
67.5
52.5
0
. 1 1
1
37.5 +1 +
1 2
11
1
1
30. J? + +
1
! i *
22.5 + 1 +
15.0
12.50 37.50 62.50 87.50 112.5 137.5 162.5
25.00 50.00 75.00 100.0 125.0 150.0 175.0
DS
FIGURE 32a. Regression analysis for annual concentrations: PM 10 versus DS.
84003 46
-------
N- 14
COR = .7618
MEAN ST.DEV. REGRESSION LINE RES.MS.
X 85.418 27.227 X= 1.4938*Y+ 25.906 337.00
Y 39.339 13.886 Y- .38853-X+ 6.6516 87.650
ZERO INTERCEPT REGRESSION LINE RES.MS.
Y = 0.45969*X 85.0163
X.*.* *. . *Y + +..*.+.«..+*
75.0 * 1 +
57.5
60.0
?2.5
0
Q.
45.0
1 1
1
37.5 + 1 +
11 1
1 *
. *
1
30.0 + «.
. *
22.5 + 1 I
15.0 + *
"** **«X«*» ** + ****«*,« + + » + »^* * + + ». + .
17.50 52.50 87.50 122.5 157.5 192.5 227.5
0.000 35.00 70.00 105.0 140.0 175.0 210.0 245.0
TSP
FIGURE 32b. PM 10 versus TSP.
81*003 47
-------
160
140
BZ.
6.7.
4*.
23.
.*. .«****.
+
f
+
1
2
"17.50
0.000
N= 25<;
COR* .3*44
MEAN ST.DEV. REGRESSION LINE RES. MS.
X 74.475 27.592 X- 1.3208*/* 20.529 219.29
Y 4,C'.645 17.641 Y= .53991-X+ .63493 89.645
ZERO INTERCEPT REGRESSION LINE RES. MS.
Y = 0.54741*X 89.3398
**w*******************«* * + ** «* + + +. + 4. * +*
X
i
+
Y
»
*
1 1
1
1
1
1 1
1 +
11
1 1
1111
11 1 1
112 11 21 1
1 11 1
12 1 22 1 1 1 *
1 1121211112 2 1 *
1 41355222 121 1
131425233 21 2 *
126A4274122 11 *
1654465 11 1 *
261143112 12 *
11213 311 2 *
1 2 11 11 *
111 *
1 1
1
* « ^ » **^« ** ^^« »***» * * »*«**
52.50 87.50 122.5 157.5 192.5 227.5
35.00 70.00 105.0 140.0 175.0 210.0 245.0
TSP
FIGURE 32c. DS versus TSP.
8«f003
48
-------
*
9
^
.750 +
*
*
*
,
.675 +
m
f
.
^
.600 +
a
»
O.
-------
.65 +
o
.60 +
.55 +
.
.50 «
CO
|HMK »
O
«-« .45 +
S:
«x
.40 +
1
.35 +
Y
.30 +
.25 + 1
A*» * * * *«
N- 14
COR- .8221
MEAN ST.DEV. REGRESSION LINE RES. MS.
X .49571 .09212 X» .70584-Y+ .15672 .00298
Y .48027 .10729 Y» .95753*X+ .00561 .00404
...-.... + .... + .... + .... + .... + .... + .... + ....+X... +
Y
1
1 '.
I
1
+
1
*
1
*
1
*
1
I +
*
*
1 +
i !
+
+
*
+
.325 .375 .425 .475 .525 .575 .625
.350 .400 .450 .500 .550 .600 .650
0.83(DS/TSP)
FIGURE 34. Annual PM 10/TSP versus 0.83(DS/TSP).
84003
50
-------
Appendix A
SITES USED IN THE ANALYSIS
Codes are from SAROAD (Storage and Retrieval of Aerometric Data) in
AEROS Manual Series, Vol. V: AEROS Manual of Codes (1976). EPA-450/2-76-005,
Research Triangle Park, NC. Asterisks before names indicate a county.
010380003A07
010380023A07
010380023A57
010380026A07
010570001A07
012380029A07
012540001A07
013200001A07
020040003A07
030050110A07
030440006A07
030600002A07
030600002A57
030600004A07
030600004A57
041440001A07
050500002A07
050520003A07
051260002A07
052220003A07
052800005A07
052820002A07
054020002A07
054020003A07
054080002A07
054180103A07
055760004A07
056300003A07
056535001A07
056860003A07
056980004A07
060080003A07
060580001A07
061260001A07
061260001A57
Birmingham, Alabama
Center Point, Alabama
Mobile, Alabama
Mountian Brook, Alabama
Tarrant City, Alabama
Anchorage, Alaska
Apache Sitegreaves National Forest, Arizona
*Maricopa County, Arizona
Phoenix, Arizona
Little Bock, Arizona
Azusa, California
Bakersfield, California
Chico, California
El Cajon, California
Fresno, California
*Fresno County, California
Livermore, California
Lompoc, California
Los Angeles, California
Pasadena, California
Richmond, California
Rubidoux, California
San Francisco, California
San Jose, California
*Arapahoe County, Colorado
Denver, Colorado
Lakewood, Colorado
84003
51
-------
81*003
061820001A07 Pueblo, Colorado
061820001A57
062220101A07 *Weld County, Colorado
070420003A07 Hartford, Connecticut
070478001A07 *Litchfield County, Connecticut
080020001A07 Dover, Delaware
080180001A07 *New Castle County, Delaware
090020017A07 Washington, District of Columbia
090020019A07
104360035A07 Tampa, Florida
110200001A07 Atlanta, Georgia
110200039A07
114500017A07 Savannah, Georgia
120370004A07 Pearl City, Hawaii
130220003A07 Boise City, Idaho
141220014A07 Chicago, Illinois
141220022A07
142360010A07 Evahston, Illinois
148320007A07 *Will County, Illinois
151520016A07 Gary, Indiana
152040021A07 Indianapolis, Indiana
152160002AQ7 Oeffersonville, Indiana
162500003A07 Harshalltown, Iowa
162500004A07
171800011A07 Kansas City, Kansas
173560007A07 Topeka, Kansas
173740012A07 Wichita, Kansas
180080002A07 Ashland, Kentucky
183090001A07 Okolona, Kentucky
191490101A07 Kisatchie National Forest, Louisiana
200010001A07 Acadia National Forest, Maine
210120001A07 Baltimore, Maryland
210120008A07
210120009A07
211380002A07 Rockville, Maryland
211380007A07
220240012A07 Boston, Massachusetts
220240013A07
222160011A07 Springfield, Massachusetts
222160011A57
222640016A07 Worcester, Massachusetts
222640016A57
231180015A07 Detroit, Michigan
231180020A07
241040025A07 Duluth, Minnesota
241620007A07 International Falls, Minnesota
242260049A07 Minneapolis, Minnesota
242260051A07
243300003A07 St. Paul, Minnesota
251260003A07 Jackson, Mississippi
52
-------
260030001A07
262380002A07
262950001A07
264280007A07
270160005A07
270310101A07
271100020A07
281880028A07
290480001A07
290580001A07
310720005A07
311380001A07
312320005A07
320040001A07
320090001A07
330660003A07
330660010A07
330660010A57
332000003A07
333520001A07
333520001A57
334680005A07
334680011A07
334680079A07
340700010A07
340945101A07
341160006A07
341160006A57
341160101A07
341160102A07
341160102A57
360060014A07
361220020A07
361300013A07
361300021A07
361300041A07
361460001A07
361660014A07
363080010A07
364140002A07
364340005A07
364340005A57
366420012A07
366420012A57
367760002A07
372200035A07
380420111A07
380500104A07
380560013A07
381460D15A07
390100064A07
390100068A07
Affton, Missouri
Kansas City, Missouri
Mark IWain National Forest, Missouri
St. Louis, Missouri
Butte - Silver Bow, Montana
Custer National Forest, Montana
*Missoula County, Montana
Omaha, Nebraska
Beno, Nevada
Winneroucca, Nevada
Camden, New Jersey
*Essex County, New Jersey
Jersey City, New Jersey
Albuquerque, New Mexico
Bayard, New Mexico
Buffalo, New York
*Erie County, New York
Lackawanna, New York
New York City, New York
Charlotte, North Carolina
Croatan National Forest, North Carolina
Durham, North Carolina
Akron, Ohio
Cincinnati, Ohio
Cleveland, Ohio
Columbus, Ohio
Dayton, Ohio
Ironton, Ohio
Medina, Ohio
Middletown, Ohio
Steubenville, Ohio
Youngstown, Ohio
Oklahoma City, Oklahoma
*Crook County, Oregon
Deschutes, Oregon
Eugene, Oregon
Portland, Oregon
*Allegheny County, Pennsylvania
84003
53
-------
390400002A07 Avalon, Pennsylvania
390780725A07 Bethlehem, Pennsylvania
390780725A57
396620001A07 North Braddock, Pennsylvania
397140003A07 Philadelphia, Pennsylvania
397140003A57
397140019A07
397140020A07
397140023A07
397140024A07
397140032A07
397140036A07
397140037A07
397140037A57
397140038A07
397260021A07 Pittsburgh, Pennsylvania
410300012A07 Providence, Rhode Island
420560003A07 *Charleston County, South Carolina
440380006A07 Chattanooga, Tennessee
442540006A07 Hurfreesboro, Tennessee
451310050A07 Dallas, Texas
451700002A07 El Paso, Texas
451710004A07 *E1 Paso County, Texas
452330024A07 'Harris County, Texas
452560034A07 Houston, Texas
454715001A07 Seabrook, Texas
460520001A07 Hagna, Utah
460920001A07 Salt Lake City, Utah
470265101A07 Green Mountain National Forest, Vermont
480200020A07 'Arlington County, Virginia
481440005A07 'Hampton County, Virginia
481560002A07 Hopewell, Virginia
482140007A07 Norfolk, Virginia
482630001A07 Reston, Virginia
482660002A07 Richmond, Virginia
491840057A07 Seattle, Washington
491840057A57
491840073A07
492040013A07 Spokane, Washington
500280004A07 Charleston, West Virginia
502000002A07 Weirton, West Virginia
502120002A07 Wheeling, West Virginia
510240002A07 Beloit, Wisconsin
510490001A07 Chequaroegon National Forest, Wisconsin
511180009A07 Green Bay, Wisconsin
54
-------
Appendix B
SITE CLASSIFICATIONS ACCORDING TO CLIMATE
ARID
030050110A07 Apache Sitegreaves National Forest, Arizona
030600004A07 Phoenix, Arizona
030600004A57
052800005A07 Fresno, California
061820001A07 Pueblo, Colorado
061820001A57
290580001A07 Winnemucca, Nevada
451710004A07 El Paso County, Texas
WEST COAST
051260002A07 Chico, California
052220003A07 El Cajon, California
054020002A07 Livermore, California
054180103A07 Los Angeles, California
012380029A07
104360035A07
114500017A07
142360010A07
152160002A07
173740012A07
211380007A07
222160011A07
222160011A57
222640016A07
222640016A57
251260003A07
270160005A07
270310101A07
271100020A07
281880028A07
334680079A07
367760002A07
390100064AQ7
410300012A07
420560003A07
500280004A07
502000002A07
510240002A07
OTHER
Mobile, Alabama
Tampa, Florida
Savannah, Georgia
Evanston, Illinois
Jeffersonville, Indiana
Wichita, Kansas
Rockville, Maryland
Springfield, Massachusetts
Worcester, Massachusetts
Jackson, Mississippi
Butte - Silver Bow, Montana
Custer National Forest, Montana
Missoula County, Montana
Omaha, Nebraska
New York City, New York
Youngstown, Ohio
Allegheny County, Pennsylvania
Providence, Rhode Island
Charleston County, South Carolina
Charleston, West Virginia
Weirton, West Virginia
Beloit, Wisconsin
55
-------
Appendix C
24-HOUR SITE CASES OMITTED BECAUSE AT LEAST ONE OF THE FOLLOWING
RATIOS WAS GREATER THAN 1.05: PM 10/DS, PM 10/TSP, DS/TSP
010380023A0782 529
010380023A0782 8 3
010570001A0780 5 3
010570001A0780 5 9
010570001A0780 515
010570001A0780 521
010570001A0780 527
010570001A0780 626
010570001A0780 7 8
010570001A0780 714
010570001A0780 720
010570001A07801117
010570001A0781 8 2
030600002A0781 6 3
030600002A0781 621
030600002A0782 312
052220003A0782 926
052820002A0780 527
056535001A0782 411
056860003A0782 821
062220101A0782 6 4
062220101A0782 616
062220101A0782 8 9
062220101A0782 926
062220101A078210U
062220101A07821020
062220101A0782H13
062220101A07821119
062220101A078212 1
062220101A07821213
062220101A07821225
070478001A0782 628
090020017A0781 913
090020017A0781 919
090020017A078110 1
090020017A07811013
090020017A078111 6
090020017A07811118
130220003A07821231
141220022A0782 616
152040021A07821119
162500003A0782 312
162500004A0780 924
162500004A07801111
162500004A07801117
162500004A07801223
200010001A0782 8 3
200010001A0782 827
210120009A07801123
210120009A078110 7
211380007A0782 920
220240012A0780 626
220240013A0780 8 1
220240013A0781 9 7
242260051A078012 5
260030001A0780 831
290580001A0780 7 2
310720005A0782 8 3
311380001A0781 329
311380001A07811230
311380001A0782 228
311380001A0782 4 5
311380001A0782 8 3
320040001A0782 216
320090001A07801217
32.9.99.?.?91A.97Q1 4 4
320090001A0781 8 8
320090001A0781 820
320090001A0781 9 7
330660003A078010 6
330660003A07801111
330660003A078012 5
330660003A0781 1 4
330660003A0781 122
330660003A0781 2 3
330660010A07821231
332000003A0780 8 7
332000003A0780 813
332000003A0780 819
332000003A0780 825
332000003A0780 831
332000003A0780 9 6
332000003A0780 918
332000003A0780 930
332000003A078010 6
332000003A07801012
332000003A07801024
332000003A078011 5
332000003A07801111
332000003A07801117
332000003A07801123
332000003A07801211
3 3 2000003A07 801223
332000003A0781 221
333520001A0781 522
334680005A0781 9 7
334680079A0781 9 7
340700010A07821113
340700010A078212 1
340700010A078212 7
340700010A07821213
340700010A07821219
340700010A07821231
341160006A07821231
341160006A578212 1'
341160101A0780 7 2
341160102A0781 919
341160102A0781 925
341160102A0782 4 5
361300013A0781 826
361300021A0780 7 8
363080010A07811112
363080010A0782 2 4
366420012A0782 5 5'
366420012A0782 8 3'
366420012A07821219
367760002A078212 7
367760002A07821225
380500104A0780 415
380500104A0780 5 9
380500104A0780 521
380500104A07801129
380560013A0780 9 6
380560013A0780 918
381460015A0780 4 3
381460015A0780 415
381460015A0780 427
381460015A0780 620
390780725A0782 210
396620001A0782 210
396620001A0782 3 6
396620001A0782 5 5
396620001A0782 722
397140003A0782 722
397140003A07821125
397140003A5782 3 6
397140003A5782 716
397260021A0782 7 4
397260021A0782 716
451700002A0780 421
451700002A0780 614
454715001A0781 8 2
460520001A0781 721
460520001A0782 1 5
481S60002A0781 416
491840057A07821231
491840073A0780 4 9
492040013A0782 7 4
510240002A0782 216
511180009A0782 4 5
84003
56
-------
Appendix D
ANNUAL SITE CASES OMITTED BECAUSE AT LEAST ONE OF THE FOLLOWING
ANNUAL RATIOS WAS GREATER THAN 1.05: PM 10/DS, PM 10/TSP, DS/TSP
332000003AQ781
050520003A0782
052800005ft0782
061820001A0782
062220101A0782
130220003A0782
290480001A0782
381460015A0782
460520001A0782
460920001A0782
84003
57
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
1. REPORT NO.
2.
3. RECIPIENT'S ACCESSION NO.
4. TITLE AND SUBTITLE
An Update On The Use of Particulate Ratios To
Assess Likely PM^p Attainment Status
5. REPORT DATE
June 1984
6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
A. D. Thrall, A. B. Hudischewskyj
8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Systems Applications, Incorporated
101 Lucas Valley Road
San Rafael, California 94903
10. PROGRAM ELEMENT NO.
11. CONTRACT/GRANT NO.
12. SPONSORING AGENCY NAME AND ADDRESS
U.S. Environmental Protection Agency
Office of Air Quality Planning and Standards
Research Triangle Park, North Carolina 27711
13. TYPE OF REPORT AND PERIOD COVERED
14. SPONSORING AGENCY CODE
15. SUPPLEMENTARY NOTES
EPA Project Officer: Edwin L. Meyer, Jr.
16. ABSTRACT
This report describes results of statistical analyses of PM15', TSP and PM10
data collected during 1980-82 at collocated sites in the National Inhalable
Particulate Monitoring Network. Results from this report have been used to support
the use of national distributions for PM10/TSP ratios. Selection of a distribution
of PMio/TSP ratios is an inherent assumption made in order to estimate nonattainment
probabilities for PM10 at sites collecting only TSP data.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.lDENTIFIERS/OPEN ENDED TERMS
c. COSATI Field/Group
PM
TSP
IP
Monitoring
PM1Q/TSP Ratios
Nonattainment probability
8. DISTRIBUTION STATEMENT
19. SECURITY CLASS (This Report)
21. NO. OF PAGES
63
20. SECURITY CLASS (This page)
22. PRICE
EPA Form 22201 (Rev. 4-77) PREVIOUS EDITION is OBSOLETE
------- |