United States
Environmental Protection
Agency
Office of Air Quality
Planning and Standards
Research Triangle Park NC 27711
EPA-450/4-86-014
October 1986
Air
Options for
Reducing the
Costs of
Criteria Pollutant
Monitoring
-------
EPA-450/4-86-014
October 1986
Options for Reducing the Costs of
Criteria Pollutant Monitoring
SYSAPP-86/106
Prepared by
Alison K. Pollack
C. Shepherd Burton
Systems Applications, Inc.
101 Lucas Valley Road
San Rafael. CA 94903
Prepared for
Mr. David Lutz, Project Officer
Monitoring and Reports Branch
Monitoring and Data Analysis Division
Office of Air Quality Planning and Standards
Environmental Protection Agency
Research Triangle Park, NC 27711
(Contract 68-02-3889)
Under subcontract to
Radian Corporation
8501 Mo-Pac Boulevard
P.O. Box 9948
Austin, TX 78766
(Purchase Order Number K26384)
U.S. Envi; -- v, r'-r«r
Region £, :-L-l?J)
77 West j oulevard, 12th Floor
Chicago, iL b~-u04-3590
U.S. ENVIRONMENTAL PROTECTION AGENCY
Monitoring and Data Analysis Division
Office of Air Quality Planning and Standards
Research Triangle Park, NC 27711
-------
DISCLAIMER
The development of this document has been funded by the United States
Environmental Protection Agency under contract 68-02-3889. It has been
subject to the Agency's peer and administrative review, and it has been
approved for publication as an EPA document.
Mention of trade names or commercial products does not constitute
endorsement or recommendation for use.
11
-------
ACKNOWLEDGMENTS
Our sincere thanks are due to the following staff at Systems Applica-
tions: Jano Banks and Mitnra Moezzi for programming support, Tom Permutt
for his wise counsel on seasonal variability, and Howard Beckman for his
editing. We also thank David Lutz and Stan Sleva of EPA's Monitoring and
Reports Branch and Dave Armentrout of PEI for their assistance.
-------
-------
CONTENTS
1 INTRODUCTION 1
2 THE NATIONWIDE CRITERIA POLLUTANT MONITORING NETWORK:
DESCRIPTIVE DATA 5
3 PERMANENT OR TEMPORARY SHUTDOWN: MOTHBALL ING AND
ROTATION OPTIONS 21
Decision Rules for Monitor Shutdown 21
Design Value Concentration. 21
Test Period and Data Completeness 22
Historical Trends 23
Rotating Monitors as an Alternative to
Permanent Shutdown 23
Restarting Pollutant Monitoring: The Use
of Indicators 24
Example Decision Rules 25
Misclassification Probabilities After
Monitor Shutdown... 28
4 SEASONAL MONITORING 36
TSP, Sulfur Dioxide, and Nitrogen Dioxide 36
Carbon Monoxide 40
5 MONITORING COSTS AND AN EXAMPLE OF NETWORK COST SAVINGS 47
Cost Components 47
Example Savings 53
6 FURTHER STRATEGIES FOR POTENTIAL COST REDUCTIONS
AND CONCLUDING REMARKS 59
Combinations of Options: Maximizing Reduction
of Monitoring Costs 59
Adjustment of State Trend Statistics 60
Summa ry 61
References 63
IV
-------
1 INTRODUCTION
State and local air pollution control agencies currently spend approxi-
mately $55-$58 million annually on monitoring criteria pollutants. These
costs include operating and maintaining monitoring systems, new equipment
purchases, quality assurance, laboratory analysis (the portion for air
monitoring only), maintaining computerized data bases, and data summary
and reporting. Since 1980 there have been increasing efforts to hold or
reduce monitoring costs; over the same period pressures for additional
monitoring have developed (for PM-10, visibility, air toxics, short-term
NO?, and non-methane hydrocarbon monitoring). Even with new (albeit
modest) monitoring activities, budget reductions at both the State and
federal levels have been achieved by eliminating bubbler sites for SOg and
N02» by reducing the number of TSP monitors, and moving toward increased
automation.* Nevertheless, both the pressures to reduce air monitoring
costs and the necessity for additional monitoring continue. The purpose
of this document is to provide guidance to State and local agencies on how
the monitoring of criteria pollutants can become more cost-effective.
A number of options for improving the cost-effectiveness of monitoring
networks have been considered. The options presented, however, do not
apply to the NAMS network since it was established to provide the minimum
number of sites needed to provide data for national policy analyses and
trends, and for reporting to the public on air quality in major metropoli-
tan areas. The options had to be constructed so that they do not hinder
EPA air quality objectives. In view of this goal, it was determined that
changes in monitoring programs should not
Disturb the NAMS monitoring network,
Disturb designated reference and equivalence method programs,
Reduce measurement precision and accuracy for individual monitoring
sites and overall network performance,
Adversely affect the SlP-call process,
* Reduced monitoring has occurred only at sites with observations well
below the NAAQS.
-------
Adversely affect national and regional trend analysis, or
Adversely affect SASD NAAQS review requirements.
Four broad categories of options that meet these constraints have been
identified:
A. Reducing the number of operating monitoring sites
B. Streamlining and standardizing nationwide data collection,
reduction, and reporting activities
C. Reducing requirements for State and local agency reporting to
the EPA
D. Reducing instrument and system maintenance costs
Within these four broad, somewhat overlapping categories, 14 specific
options have been Identified (Table 1-1). In a previous report these
categories were described and ranked as to their potential for monitoring
cost savings (Burton and Pollack, 1985); these rankings are indicated in
Table 1-1.
This report describes the options for reducing the number of criteria pol-
lutant monitors in operation at any given time 1n a State network. We
consider reductions 1n monitoring for five of the six criteria pollu-
tants: TSP, S02» 03, N02, and CO. We do not consider reductions in moni-
toring for lead because, due to the relatively recent (1982) requirement
for lead monitoring, little lead data is available.
The options are offered as examples only, and the EPA will consider State
proposals regarding reduced monitoring on a case-by-case basis, judging
their acceptability on their individual merit, with final authority for
the SLAMS resting with the Regional Administrator, and for the NAMS, with
the Administrator.
Section 2 of the report describes the current nationwide criteria pollu-
tant monitoring network and the data base used to evaluate the monitoring
reduction options. The following two sections discuss the three options
for reducing the effective number of monitors: shutting down monitors
permanently, rotating monitors, and seasonal monitoring. In Section 5 we
demonstrate the potential cost savings for an example network of 100 moni-
toring sites after implementation of some of the suggested options.
Finally, Section 6 discusses how statistics on statewide trends can be
adjusted after one or more of the suggested options are implemented and
how cost-effectiveness can be optimized.
-------
^
00
^*
^f
u
o
0.
c
§
L.
3
m
.
41
u
L.
3
O
i2
£
c
40
3
"o
Q.
ID
41
40
^
L.
U
01
c
^
u
o
40
c
§
»-
o
VI
c
01
^
+i
U
Ol
41
1
"0
O
U
Ol
.c
Ol
c
o
L.
^
U
O
*^
VI
c
o
Ml
40
<§
2
*-
40
c
41
U
c
o
Q.
VI
4>
O
e
o
40
&
O
X
0
o>
41
40
0
C
o
40
*-»
£
ex
ov
C 41
o
O.N
E 0
a
o
ie
o -o
L. 2
41 40
41 e
O 41
40 Q.
C O
C
IO
m in
»
IO
^ v»
X C
1- O
41 VI
41 41
VI
Sc
t- o
40
Ol S
C
"£"0
O GL
40
"c 10
0 41
E a.
41 Ol
3
^ C.
01 3
ec. -D
^n
c
^
0
*J
m^
C
i
^B
C
0
in
IO
41
VI
"O
41
40
O
41
41
1/1
^
«T
O 1
fm
1- C -
A) O
"i e
3 IX)
C X
«t
41 _j in
JC l/> Ol
40 40
Ol
ene vi
c
40 Ol
0 10 C
3 t-
T3 OJ L.
41 O. O
at o w
<
^^.
"»
t |
e
4l *"* tt)
u e c.
§£"
w c m
* 0 _4»
'ojc'o
CL
C b.
Ol O X
in
41 41
O 3 IO
3
41 10 O"
JC >
40 L.
!_
Ol *- IO VI
U O 41
41 40 W IJ
j: a c.
* 0 3
Ol O
VI T3 C VI
40 L. IO
VI 40 10
U TJ
L. 41 C
O 0. 0 <-
*- in E 4)
c E o
*^* X 3
I* O VI Ot
O HI C
** V QC **^
** .0 «n
g 1/1
"*" O 01
^ C < **
c o< «o
4i ^E o
CL40
VI IO W
3 U C C
40 *- -r-
Ol
c
(_
0
*^»
c
§
o
OJ
u
^k
OJ
(A
^_
m
^3 V*
£ Ol
_o *
* VI
CM
*
o e
«
in
VI C
4) U ^A***
^ C
e o> c
o
0 0 40 40
41 40 10
in ie c L.
2|i|
VI U 41
41 --.40 C
C « 40
O U O VI
4i £
40 O. - I-
10 10 x o
U U U 40
o> o
O O Ol C
40 4j 41 O
^) O £Z
in e u
41 10 c. 10
40 u 41
»- IO
VI C Ol -
o
en c 4i
C 40 i
»» IO JZ "^
t- 40
O 3 *
£ §=».£
CO. v>
o 10
E £ vi jo
41 41 L. ^
^3 O Ol 10
*- U 40 3
> 3 40 C
0 10 C
O vi o. «
v>
0)
V)
3
41 e
otj
in O
CTI^
C C
Q.
E ^3
<0 01
in 40
U
cn oi
c
40 VI
IO
40 U
o o
ee u-
r->
I
.
L.
10
at
x
^
u.
X
L.
41
>
41
a
in
u.
41
41
40
V
c
01
40
"a
o>
c
IO
40
O
u
*
c
o
41
ID
U
41
a.
o
in
c.
41
0
o"
41
"n 41
JC 3
0 41
E jc
u
41 in
L.
ie
IO
L. O
O in
40 IO
41
C VI
i t.
J^
10
o
40
41
JD
in
in 10
c O>
0 *-
40 -a
a. c
0 10
O 01
L.
C
O m
41
IO
C Vt
IO
E f^
O *
<_> <
.
«t
u" «
4J Ol t
Ol 40
"w c "o
JP
"c to "S
10 40 N
IO
in 73
t. 10
41 t- L.
Ol O 40
OIH- c
O 41
41 O
in 40
0 10
U 4O i-
C *- 41
« x o
> i- e
a i- 41 «
JC 41
Q. HI Q. C
O 01 0
S° S 2
40 < E
C
o
u
41
^»
O
u
IO
40 01
10 C
o
^~ "O
o c
10
C JE
O
«
IO V
E ^3
40 -a
3 C
< IQ
^,
03
C
10 41 T3 vi
40 -a c 41
in 10
X 40
O c
c o c >
10 O
Ol IO 4O O
c e o
40
C 10
41 40
C 10
o -a
§^
U *->
«J
0 E
OJ O
**~ 3
c «o
u
« (A
TJ <0
C
«
4-, «.
l/t flJ
J
a m
u
c c
g
^ ^l
3 10
t -o
tsi
CO
o
41 4)
C
in ia
41 C
o oi
in c
IO
40 41
in (-
X "3
m X
X»-
<- o
E -0
41 41
IS
41
40 *
IO
J= 1-
w C
X 41
U
41
40 40
10 O,
1- O
41 «t
C
41
41 VI
41
O
IO 'O
1- C
a. 10
tj
41
40
e
c
O in
*" c o
40 O VI
u e
41 40 O
u u
i C
O 3 41
U *- JO
10 4) O
40 U VI
10 C
C L.
^ 41 X 41
41 40 10 ^
40 c E <-
0 0
4> IO Vt JS
i 41
41 ' Ol
in 41 40 40
m 10 > 40
4) X m
40 40 40
IO tb U 14-
40 O 10 0
in in
Oi X
en-o c 40
c c
§10 u c
0
41 C >
400
O »-
41 41
a. c
»- 41 O IO
O 40
40 40 O 41
C VI U
Sc .
40 O VI
JC ID OJ
in 40 13
in 10
<- cm
J3 41 IO
ie 40 ja
40 c e w
vt a/ o
LU O CJ CJ
m «
as m
41
^
40
c
0
u
-------
****
"S
3
, 01 U 4V O
0 in ^ « j=
C IB O 4V
Ol JO -0
3 U TO >,
V VI i- C
01 ou= e
*. ex o
»- t-
01 4V >o C C
t- e o ex ^
* Z o c
CO. «
««>,
K VI C U
Z Ol 01 I- 0
^ u oi oi
ai <- 4v 01
5 o «o >o
o in ex u
^j
01
U 4O
e 3
41 VI
3
cr «-
41 e
L.
Ol 01
c (-
r- f >,
4V 3
L. 0- C
0 Ol O
Q. U
ai in
* rz
>\ *- -j
u- o to
a.
C 01 H-
=> ce 0
CJ
« c
W QJ
o -a
» L.
Q 4v
C
« 0,
01
4V O
VI CX
01
o> <-
c
U O Q-
3 C UJ
O 0)
01 Ol O
ee 10 4V
e
JS
VI 3
e
4C e
4V 01
_e"£
4V Ol
4V
£g
01 O»
L.
Ol ^O
<8"3
ss
v* &.
a> uj
1^
"2
S1-
55
m
.^
8
LB
o
tfl
!
o
ai e
N a
VI
« VI
11
m.
M
W
«fl
3
U
>o
u
^^
Ul
o
*^
T3
VI
"St
u o
3 ex
c a;
o, u
VI
JAB
CL
3 01
VI U
ex
0
4V ^
C
vi *a
01
4V VI
4V
,0,
3 01
O
U
C 0
o
c
a
4V
C X
k. >~
4V 40
VI »
*- VI
o
in
T3 4V
c m
< O
u
'e oi
§u
S
L. C
4V Ol
in 4V
C C
"oil
e
u £
3 4V
"O VI
01 >,
at in
O
*r
«
VI *V
oi e
mjs 01
U 4V 4V
e
r^l
u u 6
O -D
41 ex e
o
> u e
o c o
ex"* 4V
o
o u
4V VI
*
VI 1
t- c u
01 oi ai
3 *vi
4V
0 VI C
030
». 4V
3 »O 4V
C 4V U
5 vi oi
Ol
4V U O
ecu
3 U 4V
t. o -o
4J *_ °D
in i.
C 01 O
O.4V
c
Ol 4V
oi e
19 41 VI
^§?
StSi
C 0
UJ Ul VI
4V
C
u
«v >«
in 4V
C
a
01 C
m o
Ol 4V
L. O
u c
C 3
»»-
O
VI
41
U
4V
*O
ex
-------
2 THE NATIONWIDE CRITERIA POLLUTANT MONITORING NETWORK:
DESCRIPTIVE DATA
State and local agencies are required to submit quarterly criteria
pollutant data from the NAMS and annual statistics from the SLAMS. Data
are submitted in a standard format to EPA's Storge and Retrieval of
Aerometric Data (SAROAD) system. From the raw monitoring data submitted,
EPA prepares numerous summary statistics and reports. Annual summary
statistics from EPA's "Quick Look Report" were used in the analyses in
this report.*
The Quick Look Report contains those annual summary statistics necessary
to determine whether a site achieves a National Ambient Air Quality
Standard (NAAQS) in a given year. The reported summary statistics vary by
pollutant, of course, since the forms of the standards are different.
Table 2-1 lists the annual summary statistics 1n the Quick Look Report for
each of the five criteria pollutants under consideration. We received and
processed tape copies of Quick Look Reports for these five pollutants for
the six-year period 1978 to 1983.
Using information provided by State and local air pollution agencies, the
EPA maintains a data base of site locations and characteristics; we
received a June 1985 version of this SAROAD site file. From this file we
extracted, for all sites with data in the interval 1978 to 1983, the site
address, latitude and longitude, UTM zone and coordinates, city and AQCR
population, station type (e.g., suburban versus rural, industrial versus
residential), and site type for each pollutant. The site type for each
pollutant was of particular interest in this study. Three site types are
defined in the SAROAD site file.
National Air Monitoring Stations (NAMS), which are located in areas
with high pollutant concentrations and/or high population exposure;
the network was established in 1979 to provide a network of high-
* The "Quick Look Report" lists annual summary statistics for all monitor-
ing stations reporting data to the NADB. These lists are included in
the annual report by the EPA titled Air Quality Data and are available
on computer tape.
-------
TABLE 2-1. Annual summary statistics from the EPA's Quick Look Report.
Pollutant
Summary Statistics
Total Suspended
Particulate
Sulfur Dioxide
Carbon Monoxide
Ozone
Nitrogen Dioxide
Number of valid 24-nour observations
Maximum 24-hour concentration
Second highest 24-hour concentration
Number of 24-hour concentrations above 260 wg/m3
Number of 24-hour concentrations above 150 ug/nr
Annual arithmetic mean
Annual geometric mean
Annual geometric standard deviation
_
Number of valid 1-nour observations
Maximum 24-hour average
Second highest 24-hour average
Number of 24-hour averages above 365
Maximum 3-hour average
Second highest 3-hour average
Number of 3-hour averages above 1300
Maximum 1-hour concentration
Second highest 1-hour concentration
Annual arithmetic mean
Number of valid 1-hour observations
Maximum 1-hour concentration
Second highest 1-hour concentration
Number of 1-hour concentrations above 40 mg/m3
Maximum 8-hour average
Second highest 8-hour average
Number of 8-hour averages above 10 mg/m3
Number of valid daily maxima
Number of days in the ozone season
Maximum daily maximum 1-hour concentration
Second highest daily maximum 1-hour concentration
Third highest daily maximum 1-nour concentration
Number of observed daily maxima above 0.125 ppm
Number of estimated daily maxima above 0.125 ppm
Number of days assumed to have dally maxima below
0.125 ppm
Number of valid 1-hour concentrations
Maximum 1-hour concentration
Second highest 1-hour concentration
Maximum 24-hour average
Second highest 24-hour average
Annual arithmetic mean
-------
quality data with consistent siting criteria.
State and Local Air Monitoring Stations (SLAMS), -which comprise the
majority of monitors.
Special Purpose Monitors (SPMs), which are set up generally for a
short period of time only and are operated by government agencies or
private groups. SPMs were omitted from consideration in this study.
A monitoring site is identified in the SAROAD system by a unique 12-
character alphanumeric identification. This 12-character code contains
five subcodes for State, city or county local site number, operating
agency, and project classification (e.g., population oriented, background
surveillance).- With this coding, if the controlling agency for a site
changed at some point in time, there would appear to be two sites report-
ing data to the SAROAD system when actually there was only one site.
Therefore, if two sites had the same 12-character code except for the
operating agency identification, we combined the annual summary statistics
from the two separate sites. (For example, to combine two annual means a
weighted average of the two separate means was calculated, where the
weight for each year is the number of reported observations.) The
exception to this rule is the case of duplicate monitors at one location
for quality assurance checks (project classification 09); for this study
duplicate monitors were omitted-.
In the Quick Look Report a two-digit code identifies the monitoring method
for each pollutant monitored at a site. With this system, if the monitor-
ing method for a specific pollutant changes during a year, a new site is
defined in the Quick Look Report. Therefore we disregarded the method
code and combined the annual summary statistics for sites for which the
12-character SAROAD code was otherwise the same (except perhaps for
operating agency). Since the 24-hour bubbler methods for N02 and S02
monitoring are by now almost entirely phased out, we omitted from consid-
eration monitors using these methods.
The number of monitoring sites in operation in 1983 for each criteria
pollutant is shown in Table 2-2. The numbers in the table represent
actual monitoring sites, i.e., changes in operating agency code or
pollutant method codes have been disregarded. In Table 2-2 the total
number of monitors for all pollutants is greater than the total number of
monitoring stations in the country because many stations monitor more than
one pollutant. Table 2-3 shows the number of monitoring stations per
number of pollutants monitored.
All the options described in this report involve shutting down a pollutant
monitor either temporarily or permanently. The cost savings resulting
-------
TABLE 2-2. Primary national ambient air quality
sites.
standards and number of 1983 monitoring
Pollutant
Total Suspended
Particulate
Ozone
Sulfur Dioxideb
Carbon Monoxide
Nitrogen Dioxide*
Averaging Time
Annual geometric mean
24-hour
Maximum daily 1-hour average
Annual arithmetic mean
24-hour
8-hour
1-hour
Annual arithmetic mean
Concentration
75 ug/m3
260 ug/m3
0.12 ppm
(235 ug/m3)
80 ug/m3
(0.03 ppm)
365 ug/m3
(0.14 ppm)
10 mg/m3
(9 ppm)
40 mg/m3
(35 ppm)
100 ug/m
(0.053 ppm)
NAMS SLAMSd Total
633 1878 2511
206 375 531
216 315 531
117 322 439
57 165 222
3 Excluding NAMS monitors.
" Monitors using 24-hour bubbler methods are not included.
-------
TABLE 2-3. Distribution of the number of pollu-
tants monitored at each monitoring station.
Number of
Pollutants
Monitored
1
2
3
4
5
Total
Number of Stations
NAMSa
670
108
41
15
11
845
SLAMS6
1974
204
87
68
49
2382
Total
2644
312
128
83
60
3227
Percent of
Stations
81.9
9.7
4.0
2.6
1.9
100
a In this table a station is designated as NAMS
if it is a NAMS for at least one of the pollu-
tants monitored. Generally speaking, a NAMS
station is so designated for all pollutants
monitored.
b Excluding NAMS monitors.
-------
from these shutdowns are clearly related to the number of pollutants
monitored. For example, 1f SO? monitoring 1s discontinued at a station
where all other pollutants will continue to be monitored, then the cost
savings is less than if S02 were the only pollutant monitored at the site,
because there are certain fixed costs for the operation of the monitoring
shelter. The number of stations monitoring all possible combinations of
one, two, three, and four pollutants is provided in Table 2-4; we will
return to these tables in the cost savings discussion in Section 5. For
now, we note that of the 3,227 stations, 82 percent (2,646) monitor only
one pollutant; of these, 80 percent (2,120) monitor TSP. Of the total
3,227 stations, therefore, 66 percent (2,120) monitor TSP only.
At the vast majority of these TSP monitors, the annual geometric mean and
second-highest £4-hour maximum concentration are below the NAAQS. Figures
2-1 and 2-2 show the cumulative distribution functions (cdf) of these two
annual summary statistics for 1983. From these cdf's one can see what
percentage of the sites have annual geometric means or second highest 24-
hour maximum below any target concentration. In Figure 2-1, for example,
we see that about 95 percent of the sites are below the long-term NAAQS of
75 vg/mi: and about 82 percent have annual geometric means below
60 pg/m . Similarly, about 98 percent of the monitors show concentrations
below the short-term NAAQS of 260 yg/m , and about 94 percent have second-
highest 24 maxima below 200 yg/m . These percentages'do not change when
one considers only sites with valid data, using NADB's quarterly validity
criteria for TSP data (EPA, 1984), as can be seen in the plots.
One can see from the cdf's for 1983 annual average N02 concentration
(Figure 2-3) and annual average and second-highest 24-hour S02 concentra-
tion (Figures 2-4 and 2-5) that N02 and S02 concentrations are well below
the standard at most monitors. The same is true for the 1-hour CO NAAQS
(Figure 2-6). But the 8-hour CO NAAQS 2-7 is the controlling standard at
every monitoring site; while over 99 percent of the monitors have concen-
trations below the 1-hour NAAQS, only about 67 percent have concentrations
below the 8-hour NAAQS (Figure 2-7). Finally, the cdf for the ozone
standard (Figure 2-8) shows that only about 40 percent of the monitors
have second-highest 1983 daily maximum 1-hour concentration below the
NAAQS of 0.12 ppm.
10
-------
TABLE 2-4. Distribution of pollutants monitored
at single and multiple pollutant monitoring
stations.
Number ofStations
Pollutant Monitored NAMSa SLAMSb Total
(a) Stations where one pollutant is monitored
TSP
°3
CO
SO
NO
S02
479
78
59
51
3
1641
122
125
84
2
2120
200
184
135
5
2
(b) Stations where two pollutants are monitored
TSP, S02
TSP, 03
so2, o3
S02, CO
TSP, CO
CO, 03
03, N02
CO, N02
S02, N02
TSP, N02
38
15
18
8
8
7
7
2
4
1
70
35
26
21
15
16
12
4
2
2
108
50
44
29
24
23
19
6
6
3
(continued)
11
-------
TABLE 2-4 (concluded).
Number of Stations
Pollutant Monitored NAMS SLAMS Total
(c) Stations where three pollutants are monitored
TSP, 03, S02 11 34 45
TSP, 03, CO 8 10 18
03, S02, N02 4 14 18
03, CO, N02 3 8 11
CO, S02, N02 4 6 10
TSP, 03, N02 358
TSP, S02, N02 426
03, CO, S02 156
TSP, CO, S02 123
TSP, CO, N02 213
(d) Stations where four pollutants are monitored
TSP, 03, CO, N02 3 19 22
TSP, 03, S02, N02 4 17 21
03, CO.* S02, N02 0 20 20
TSP, 03, CO, S02 6 10 16
TSP, CO, S02, N02 2 24
(e) Stations where all five pollutants are monitored
There are 11 NAMS and 49 SLAMS, or 60 total,
stations where all five pollutants are monitored.
a In this table a multiple pollutant station is
designated as NAMS it if is a NAMS for at least
one of the pollutants monitored. Generally
speaking, a NAMS station is so designated for
all pollutants monitored.
b Excluding NAMS monitors.
12
-------
uj in
0£ t)
o
0)
^ u
« in
UJ
I
_iin
m
-ICNJ
t i i i i
SOWN
J 1-
1 (. L
O
,o
ro
CM
cr
O
. c
CM O
4J
t)
O C
.CD V
U
c.
o
o
O oo
- *~
c
T5
o;
OJ
e
o
OJ
t)
C
C
CD
OOi
08
09
OS - Of OE OZ
soj?>o 3 A 1 1 e [ n um }
OJ
E
U
!_
^j
cu
e
o
c
fQ o
D e
c
c
IDC/5
3"
C
O C
u ^
c
D U
i«- C
o
oo
u t_
w o
o
mm
h ^
C
O
c
I V
C\J O
c
LU O
cr u
13
-------
UJW
52
(/>
>-
2
_!*>
I I I I I
. -4_o
CO
M
£
o
.o
vD c
O
o c
- I t-> V
o
o
o
O
^
I
TT
CM
ID .
JC W
cnt_
o
JZ*J
n c
c o
O E
O
Q_
00
o
__0
rr
s_
O
I
^~
csi
X
t- <
O "3^
C CP
O C
U'O
CJ D
C
C
o
- co
u z
D -c
i2 _l
C/0
t_
U I-
tn o
14
-------
t i I I I I I I I I I I
001
15
-------
(M
O)
en
"0
t_
0)
t3 O
C -
c c
t) O
E
it_
O
w
cz
o<:
-2
jJ
O (T-
C C
3
3
C
O O
- c
CD
C
O
CM 1_
4-J
LU C
ac v
3 O
CD C
!_ ^
U '.'
OCI
16
-------
a: in
=>u
00 *J
o
en
I V
a>
ft
ir>in
001
17
-------
tu
IU
§91
«
O
o
u in
in rr>
rxvo
m
A *»
SOWN
o
in
o
.LO
CP
E
O
.in c
CD O
O *J
c
.o u
rn o
c
o
o
CM
o
V
2 4,
c
o
CJ
u
O C/5
.o ro
co
o
"in
O
JZ
(U
o -
c c
o o
u e.
<0
CO
c/o
c
O CT>
- C
JJ
CJ T3
C D
3
*t- U
C
c -
o
4J OO
l_ C/5
*J
W !_
O
93 .
> n
- CD
3 -
E C.
CJ -
I C
CM CD
CJ
UJ C
& o
1- 1 I 1 |___J
18
-------
UJ~
§m
o
»>
«
u in
o
m
(0
!;
men
O C
O Z>
C
u u
M- C
i.
*-> 1_
v> o
<*-
o
a; m
> CD
D C
e o
o -
i C
(NJ OJ
o
LU C
ce o
Z3 CJ
19
-------
I/I
UJ
K in
a u
o
o
_) U
_ICD
_lin
tfl
in co
IV ID
I I I I I I I I I I I
pjepue;s SOWN " ~-:>I=T-
in
on
X
E
>» w
'E ~ °
Q. rV-*J
-£" c
*J O
c WE
2 ^-
X
c
o
o
OJ
in c
fsl O
. N
O O
O
.c
o
t\J
X
e
>.
1}
o
C
O
U
O
.
o
o
o
i i i
00 i Ob 09
I Z.
o
c cr>
o c
o -
wo
W D
>»- o
o c
o
C/1
c o
DOO
J3CD
- c
T3 O
- i.
> 4_>
T3 C
V
D U
E C
D O
(_> O
OJ
c
CO O
I rsi
(NJ O
CO
m w t.
co S r
20
-------
3 PERMANENT OR TEMPORARY SHUTDOWN:
MOTHBALLIN6 AND ROTATION OPTIONS
Pollutant concentrations are below the NAAQS at many monitoring stations
across the country and have been for many years. At these stations,
shutting down the monitoring of low concentration pollutants merits
consideration. Pollutant monitors could be permanently shutdown or
shutdown for a year or two at a time in tandem with other monitors. We
refer to permanent shutdown as the mothball ing option and temporary
shutdown as the rotation option. Complete monitoring should be reinsti-
tuted, however, if there are any indications that the pollutant concentra-
tions might be increasing. In this section we describe decision rules
that could be used in considering the permanent or temporary shutdown of
monitors and the use of indicators to determine whether monitoring should
begin again. We then examine the number of monitors nationally that could
be shut down under the example decision rules, and show how the probabi-
lity of misclassification can be determined under given decision rules.
DECISION RULES FOR MONITOR SHUTDOWN
Four major decision rules must be established in order to determine which
pollutant monitors might be permanently or temporarily turned off. The
first of these concerns how high the pollutant concentrations are relative
to the appropriate NAAQS. The remaining three decision rules are con-
cerned with data quality and completeness and trends in historical data.
Design Value Concentration
Pollutant monitors should be eligible for shutdown only if the design
value concentration (DVC) is below some level or percentage of the
corresponding NAAQS. This cutoff concentration must be chosen to be low
enough so that the probability of attainment misclassification is very
low, provided that other criteria discussed below are also met. At the
same time the DVC for each pollutant cannot be so high that so many
monitors are turned off, thus leaving insufficient data for the SlP-call
process and for NAAQS-review data analysis. Since increasing the DVC
cutoff provides a potential for greater misclassification, States must
21
-------
demonstrate that the probability of such misclassification, based on
historical data, is low. An example of how these probabilities may be
assessed is provided at the end of this section.
The DVC cutoff value, expressed as a percent of the NAAQS, need not be the
same for all five pollutants. In fact, it need not be the same for all
States for a given pollutant. What should be constant, though, is the low
probability of misclassification. The selection of a misclassification
probability and cutoff concentrations is the responsibility of the EPA;
the States must provide substantive analyses of historical air quality
data to demonstrate that EPA objectives may be met.
For TSP and 502 there are two primary NAAQS, a long-term annual mean NAAQS
and a short-term 24-hour NAAQS. .For these two pollutants a monitor would
be eligible for shutdown only if the DVC cutoffs for both the short- and
long-term NAAQS were met.
Test Period and Data Completeness
A pollutant monitor should be eligible for shutdown only if specific data
quality and data completeness tests are met. The second decision rule is
therefore the number of years for which the DVC has been below the cutoff
concentration; this period of time is referred to as the test period.
During the test period, quality assurance performance standards must be
achieved for the pollutant monitors being considered for shutdown. The
third decision rule is the level of data completeness required during the
test period. A monitor at which the DVC is less than the cutoff concen-
tration only for the most recent year and where only half of the days have
valid measurements should not be eligible for shut down. The longer the
concentrations are less than the DVC, and the higher the level of histori-
cal data completeness, the less likely is the chance of future attainment
misclassification.
It might be desirable to relate data completeness criteria to the DVC;
i.e., the lower the DVC, the less historical data would be required (above
a predefined minimum and as long as misclassification probabilities are
still low). Also, in the case of pollutants showing seasonal concentra-
tion patterns, e.g., CO at some urban sites, data completeness criteria
should be developed for each season within a year. Applying seasonal
criteria would preclude shutting down a pollutant monitor with a strongly
seasonal pattern, where on average data completeness is high enough only
because of being extremely high in the nonpeak season while low in the
peak season.
22
-------
Data completeness criteria should be stricter (i.e., more data should be
required) for short-term standards than for long-term standards. This is
because the probability of attainment misclassification is higher for
short-term standards than for long-term standards for a given level of
data completeness. This relationship has been demonstrated for the two
S02 standards by Pollack and Hunt (1984).
Historical Trends
The fourth criterion for shutdown is the desired trend in the annual NAAQS
summary statistics during the test period. Such a criterion must be
established so that sites with increasing pollutant concentrations are not
eligible for shutdown. An example of a statistical criterion that
addresses the trend in pollutant concentrations in the test period is
provided in the example application at the end of this section.
To summarize, four criteria are used to judge whether a monitor is
eligible for a reduced sampling frequency:
The design value concentration (DVC) must be below some percent of
the level specified in the NAAQS.
The DVC criterion must persist for some number of years, the test
period.
EPA data completeness criteria must be met for each year in the test
period. .
Upward trends in the DVCs should not be evident in the DVC over the
test period.
ROTATING MONITORS AS AN ALTERNATIVE TO PERMANENT SHUTDOWN
The four criteria above must be established in such a way that when
monitors are completely shut down there is little probability of attain-
ment misclassification. Some States may wish to consider less restrictive
criteria for temporary shutdowns, i.e., rotating the operation of monitors
on one-year or two-year cycles. For example, the EPA-approved DVC cutoff
might be 50 percent for mothball ing but only 75 percent for rotation. It
may be the case that only 10 percent of the State's monitors meet the
criteria for mothballing but 30 percent meet the less restrictive criteria
for rotation. In this case the rotation option may be more cost-effective
than the mothball ing option.
23
-------
Under the rotation option pollutant monitors with similar characteristics
would be paired; one monitor would be in operation for a full year and
then shut down while the second was put in operation for a full year. The
cycle would be repeated unless there was some indication that pollutant
concentrations were increasing (this is discussed further below).
Alternatively, three monitors could operate on a rotation schedule of one
year on and two years off. The longer the period a specific monitor is
shut down, though, the more restrictive the four criteria need to be. One
could allow, for example, every-other-year rotation for monitors that have
DVC's below the required cutoff concentration and meet the validity
criterion for the most recent year only. If these criteria were met for
the most recent two years, the monitor might be considered for rotation on
a once-every-three-years schedule.
Pollutant monitors that are rotated in pairs or triplets should have
similar characteristics, because the monitor that is in operation in any
given year serves as an indicator of the pollutant concentrations at the
monitoring sites that are shut down. Monitors rotated together should,
for example, be located relatively close to each other, should have the
same EPA project classification (i.e., population-oriented versus source-
oriented), and should have similar patterns of pollutant concentrations.
The exact nature of the characteristics used to group monitors eligible
for rotation would be approved by the EPA on a case-by-case basis.
RESTARTING POLLUTANT MONITORING: THE USE OF INDICATORS
Pollutant monitors that have been shut down after meeting established
eligibility requirements' may need to be reinstated if nonattainment
becomes a possibility because of changes in emission patterns .in the
vicinity of the monitoring site. For example, if a large S02 source is
built near a shutdown S02 monitor, then certainly that monitor should be
restarted. There may be cases, however, in which a potential cause of
nonattainment is not so obvious. We therefore suggest the use of indica-
tor variables to determine whether or not monitoring should be reinstated.
The simplest example of an indicator for a shut-down monitor is the DVC of
a "nearby" monitor that was not shut down. Under the rotation option, the
operational monitor serves as an indicator for the nonoperational moni-
tor(s). At the outset of the design of a mothballing strategy, considera-
tion could be given to identifying and designating indicator monitoring
sites. Thus, if nearby monitors show DVCs approaching the established
pollutant cutoff concentration, then the mothballed monitors should be
turned back on. The definition of "nearby" and "approaching cutoff
concentration" are EPA choices and may be made on a case-by-case basis.
24
-------
Other data routinely collected by State and local agencies may also be
used as indicator variables. Examples of indicator variables are airport
visibility or increased agricultural activities for fine particle and PM-
10 concentrations; population trends (as evidence of growth, shifts in
emissions density, etc.) emissions trends (paying particular attention to
new point sources); and gasoline sales and'total vehicle registrations for
ozone and NC^.
For each specific indicator a threshold level needs to be established
above which pollutant monitoring will be reinstated. These threshold
levels require careful definition, for there is often a substantial time
delay between data collection and examination, and yet another time lag
before monitors are restarted.
It is possible that a monitor could be restarted because an indicator
exceeded a specified threshold only to reveal that pollutant concentra-
tions are still well below the NAAQS. In the restart year the indicator
may have reversed direction and move back toward the direction of reduced
nonattainment probability (i.e., lower gasoline sales, reduction in
population).* Again a set of decision rules must be applied to determine
when a pollutant monitor can again be mothballed. One might argue for a
test period shorter than the original mothballing test period, though the
DVC and data completeness requirements should probably be the same.
EXAMPLE DECISION RULES
The Quick Look Report data base described in Section 2 can be used to
determine how many pollutant monitors in each State would be eligible for
shutdown under example criteria. Two example cutoff concentrations for
each pollutant were chosen, so two complete sets of eligible sites can be
tallied. The exemplary set of decision rules is as follows.
1. Cutoff concentrations. The two cutoff values (CV) we chose are
50 percent and 75 percent of the NAAQS levels. For TSP and S02»
where there are two primary standards, pollutant monitors must have
DVCs below the CVs for both the short- and long-term standards. For
CO there is an 8-hour and a 1-hour standard, but the 8-hour standard
is the limiting standard without exception, and so we do not consider
the 1-hour standard in this analysis.
It is also possible that even if the indicator in the restart year is
still rising, pollutant concentrations are still low because atmospheric
dispersion conditions are above average. While such meteorological
fluctuations should be considered in developing a revised monitoring
strategy, this topic is outside the scope of this report.
25
-------
2. Test period and statistical rule. The average pollutant level
for the most recent three calendar years must be below the CV. In
addition, the most recent year must be below the CV.
3. Data completeness. For continuously monitored pollutants, data
completeness in the test period must be at least 75 percent. For
S02» N02» and CO this means that there must be 7alid data for at
least 6,570 hours in each of the three years. For ozone at least 75
percent of the days in the ozone season in each of the three years
must have valid daily maxima. TSP data for each of the three years
must satisfy the NADB data completeness criteria. (In some cases,
these criteria for TSP may be too lenient, and the EPA may opt for
stricter criteria.)
These criteria were applied to the three-year period 1981 to 1983, the
most recent years for which complete annual summary statistics were
available at the beginning of this study. The results for all pollutant
monitors in the nation are summarized in Table 3-1. For all pollutants,
the number of sites eligible for mothball ing is .far less than what might
be inferred from the cumulative distribution functions of the annual
summary statistics (Figures 2-1 to 2-8). There are two reasons for the
apparent discrepancy. First, the figures in Section 2 are for 1983 only;
the example criteria we are using specify that concentrations must be
below the cutoff for 1983 and for the average of the three years 1981 to
1983. More important is the fact that the example data completeness
requirements would disallow many sites. The last two columns in Table 3-1
show just how restrictive our data completeness requirements are. On
average, only about half of all pollutant monitors meet'the NADB annual
validity criteria for three years in a row.
Using the cutoff DVC of 75 percent of the NAAQS for TSP, almost half of
the monitors are eligible for shutdown; this represents a substantial
potential cost savings. Very few of the TSP monitors are below the 50
percent CV; however, the annual geometric mean cutoff concentration of
37.5 yg/nr is not much higher than continental-background concentrations
in many parts of the country.
The 50 percent CV concentration of 0.06 ppm for ozone is also in the
vicinity of the continental background concentration, and only one monitor
in the entire country is eligible for shutdown under such restrictions.
Even at a CV of 0.08 ppm (75 percent of NAAQS) only 20 of 581 monitors are
eligible for mothballing. Although national trends in Oj levels are
decreasing (EPA, 1986), ozone continues to be a pollution problem of major
concern. Carbon monoxide levels are also still too high in many parts of
26
-------
TABLE 3-1. Total number of sites nationally eligible for mothball ing under two example
cutoff values. Pollutant monitors must be below the CV in the most recent year, and
below the CV on average in the most recent three years; data completeness requirements
listed in text must also be satisfied.
Total Number of
Monitors in
Sites Eligible for Mothhalling
CV = 50% of NAAOS
CV = 75% of NAAOS
Sites Meeting
1981-1983
Data Completeness
Requi rements
Pollutant
TSP
°3
S02
CO
N02
Operation in 1983
2511
581
531
439
222
Count
185
1
215
15
63
Percent
7.4
0.2
40.5
3.4
28.4
Count
1110
20
255
62
80
Percent
44
3.4
48.0
14.1
36.0
Count
1670
327
270
223
93
Percent
66.5
56.3
50.8
50.8
41.9
27
-------
the country, despite a decreasing trend since monitoring began (EPA,
1986), and very few CO monitors qualify for shutdown under the example
criteria.
In 1983 and 1984 virtually all S02 and N02 monitors had concentrations
below the NAAQS. Nevertheless, at the 75 percent CV fewer than half of
the monitors are eligible for shutdown, not because of high pollutant
concentrations but because of the three-year data completeness require-
ment. For example, Table 3-1 shows that whereas only 51 percent (270) of
the 862 monitors meet the data completeness requirements, nearly all of
those (255) meet the 75 percent CV requirements.
Table 3-2 lists the number and percent of pollutant monitors in each State
that could be eligible for shutdown under the example criteria. Since
pollutant standards are commonly met in some States while being exceeded
or approached in others, there is much variation across States in the
percent of pollutant monitors eligible for shutdown. For example, in the
arid western States, where TSP concentrations can be high, especially in
the summer, fevy of the monitoring sites meet the example criteria.
MISCLASSIFICATION PROBABILITIES AFTER MONITOR SHUTDOWN
The criteria for permanent or temporary shutdown, particularly the cutoff
concentration, should be restrictive enough so that there is little chance
that air pollution standards will be exceeded in an area in a given year
where the pollutant monitor has been shut down for the year. The likeli-
hood of attainment misclassification in the absence of monitoring (because
monitors have been shut down after meeting eligibility criteria) can be
examined by using historical data. We demonstrate these important
calculations by applying our example shutdown criteria to the 1978-1983
data in the Quick Look Report.
We applied the example criteria to historical pollutant concentrations for
1978-1983 and calculated the probability of exceeding a NAAQS one, two,
and three years after a monitor meets the criteria and is turned off. The
results are presented in Table 3-3. As an example of the calculations,
consider the 24-hour standard for TSP, under a cutoff concentration of 75
percent of NAAQS (i.e., 195 yg/m^) and the three-year data completeness
requirement. Using the 1980 to 1982 data, 1,253 TSP monitors satisfy the
criteria; of these, only two monitors exceeded the 24-hour standard in the
following year, 1983. Using the 1979 to 1981 data, 1,121 TSP monitors
satisfy the criteria; of these only one exceeded the standard in 1982.
Finally, using the 1978 to 1980 data, 953 TSP monitors were found to meet
the criteria, and only three of these exceeded the standard in 1981. In
total, then, of 3,327 cases (953 + 1,121 + 1,253) where, the example
28
-------
TABLE 3-2a. Number and percent of sites satisfying example snutaown
criteria, 1981-1983 data, by state, total suspended participates.
State/Territory
Alabama
Alaska
Arizona
Arkansas
Cal tfornia
Colorado
Connecticut
Delaware
District of Columb
Florida
Georg i a
Hawaii
Idaho
11 1 inois
Indiana
Iowa
Kansas
Kentucky
Louisiana
Maine
Maryland
Massachusetts
Michigan
Minnesota
Mississippi
Missouri
Montana
Nebraska
Nevada
New Hampsh i re
New Jersey
New Mexico
New York
North Carol ina
North Dakota
Ohio
Okl ahoma
Oregon
Pennsy Ivania
Puerto Rico
Rhode Island
South Carol Ina
South Dakota
Tennessee
Texas
Utah
Vermont
V i rg i n i a
Washington
West Virginia
Wisconsin
Wyoming
American Samoa
Gaum
Virgin Islands
No. of
Sites
68
13
40
26
105
61
40
9
ia 9
101
48
11
18
91
79
50
20
61
26
14
36
26
76
44
20
41
21
34
33
20
22
51
198
76
20
240
34
35
133
14
12
16
20
62
73
11
6
78
37
31
79
14
0
4
4
CV = 50*
No. Of
Sites
5
2
7
0
4
1
8
0
0
26
0
4
3
0
0
2
0
0
1
1
3
3
6
4
1
2
0
2
3
1
2
0
39
6
6
3
0
6
2
0
0
0
6
1
0
0
1
3
5
0
10
5
.
1
0
of NAAQS
% of
Sites
7
15
18
0
4
2
20
0
0
26
0
36
17
0
0
4
0
0
4
7
8
12
8
9
5
5
0
6
9
5
9 '
0
20
8
30
1
0
17
2
0
0
0
30
2
" 0
0
17
4
14
0
13
36
_
25
0
CV = 15%
No. of
Sites
41
2
14
12
39
9
33
9
7
71
9
6
8
28
30
26
5
19
15
7
15
17
39
20
17
16
0
12
11
8
7
0
117
62
15
117
9
24
61
6
8
9
16
7
15
0
2
20
21
8
33
7
»
1
0
of NAAQS
% Of
Sites
60
15
35
46
37
15
83
100
78
70
19
55
44
31
38
52
25
31
58
50
42
65
51
45
85
39
0
35
33
40
32
0
59
82
75
49
26
69
46
43
67
56
80
11
21
0
33
26
57
26
42
50
25
0
1 1
-------
State/Territory
Al abama
Alaska
Arizona
Arkansas
Cal Ifornfa
Colorado
Connecticut
Delaware
District of Columb
Florida
Georg 1 a
Hawal I
Idaho
II 1 Inols
Indiana
Iowa
Kansas
Kentucky
Louisiana
Maine
Mary 1 and
Massachusetts
Michigan
Minnesota
Mississippi
Missouri
Montana
Nebraska
Nevada
New Hampshire
New Jersey
New Mexico
New York
North Carol Ina
North Dakota
Ohio
Okl ahoma
Oregon
Pennsy 1 vania
Puerto Rico
Rhode Island
South Carol Ina
South Dakota
Tennessee
Texas
Utah
Vermont
V Irglnla
Washington
West Virginia
Wisconsin
Wyoming
American Samoa
Gaum
Virgin Islands
Total
No. of
Sites
3
0
12
2
52
2
14
7
ta 2
20
9
0
4
26
15
6
2
12
6
6
9
17
17
12
2
11
1
1
0
7
14
9
39
6
4
42
8
2
41
2
4
2
0
4
11
6
3
14
11
11
19
0
0
0
2
531
cv « 5m
No. of
Sites
0
1
1
36
2
6
3
2
10
0
-
1
16
5
1
2
3
1
1
1
7
3
5
1
6
0
0
-
0
5
'5
18
2
4
13
1
2
19
0
3
0
-
1
2
1
1
12
8
0
4
-
-
0
215
5 of NAAQS
% of
Sites
0
8
50
69
100
43
43
100
50
0
-
25
62
33
17
100
25
17
17
11
41
18
42
50
55
0
0
-
0
36
56
46
33
100
31
13
100
46
0
75
0
-
25
18
17
33
.86
73
0
21
-
-
-
0
40
CV = 75<
No. Of
Sites
0
-
3
1
36
2
7
5
2
11
0
-
2
17
8
2
2
4
1
1
1
8
4
5
1
7
0
0
-
0
8
6
23
2
4
18
1
2
26
0
3
0
-
1
2
2
1
12
8
1
5
-
.
_
0
255
of NAAQS
% of
Sites
0
-
25
50
69
100
50
71
100
55
0
-
50
65
53
33
100
33
17
17
11
47
24
42
50
64
0
0
-
0
57
67
59
33
100
43
13
100
63
0
75
0
_
25
18
33
33
86
73
9
26
-
_
_
0
48
-------
State/Territory
Alabama
Alaska
Arizona
Arkansas
Cal Ifornla
Colorado
Connecticut
Delaware
District of Columb
Florida
Georg i a
Hawaii
Idaho
II 1 inois
Indiana
Iowa
Kansas
Kentucky
Louisiana
Maine
Maryland
Massachusetts
Michigan
Minnesota
Mississippi
Missouri
Montana
Nebraska
Nevada
New Hampshire
New Jersey
New Mexico
New York
North Carol ina
North Dakota
Ohio
Okl ahoma
Oregon
Pennsy 1 van! a
Puerto Rico
Rhode Island
South Carol Ina
South Dakota
Tennessee
Texas
Utah
Vermont
Virginia
Washington
West Virginia
Wisconsin
Wyoming
American Samoa
Gaum
Virgin Islands
No. of
Sites
6
0
12
2
109
11
9
4
ia 2
18
4
1
0
34
13
7
3
12
14
3
17
15
14
6
0
13
0
3
7
6
11
8
23
10
3
29
7
5
33
1
2
7
0
11
28
7
2
15
9
4
21
0
0
0
0
CV - 50*
NO. Of
Sites
0
»
0
0
0
0
0
0
0
0
0
1
-
0
0
0
0
0
0
0
0
0
0
0
-
0
-
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
-
0
0
0
0
0
0
0
0
-
-
-
-
of NAAOS
% of
Sites
0
-
0
0
0
0
0
0
0
0
0
100
-
0
0
0
0
0
0
0
0
0
0
0
-
0
-
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
-
0
0
0
0
0
0
0
0
-
_
-
CV = 75*
No. of
Sites
0
-
0
0
8
0
0
0
0
2
0
1
-
0
0
0
0
0
0
0
0
0
0
1
-
0
-
2
0
0
0
3
0
0
1
0
0
0
0
0
0
0
-
0
0
0
0
0
1
0
1
_
_
_
_
of NAAQS
% of
Sites
0
-
0
0
7
0
0
0
0
11
0
100
-
0
0
0
0
0
0
0
0
0
0
17
-
0
-
67
0
0
0
38
0
0
33
0
0
0
0
0
0
0
0
0
0
0
0
11
0
5
»
-B
_
_
Total
581
20
-------
State/Territory
At abama
Alaska
Arizona
Arkansas
Cal Ifornla
Colorado
Connect i cut
Del aware
District of ColumbI
Florida
Georg I a
Hawai i
Idaho
II 1 Inois
Indiana
Iowa
Kansas
Kentucky
Louisiana
Maine
Maryland
Massachusetts
Michigan
Minnesota
Mississippi
Missouri
Montana
Nebraska
Nevada
New Hampshire
New Jersey
New Mexico
New York
North Carol Ina
North Dakota
Ohio
Oklahoma
Oregon
Pennsy 1 vania
Puerto Rico
Rhode Island.
South Carol Ina
South Dakota
Tennessee
Texas
Utah
Vermont
Virginia
Washington
West Virginia
Wisconsin
Wyom i ng
American Samoa
Gaum
Virgin Islands
No. of
Sites
4
6
14
0
73
13
5
3
a 3
27
7
2
2
12
7
5
3
8
3
1
7
9
10
9
2
9
4
4
7
2
10
9
14
10
0
17
6
7
24
3
2
2
0
11
13
10
1
11
17
3
8
0
0
0
0
CV = 50<
No. of
Sites
0
0
0
-
11
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
-
1
0
0
0
0
0
0
-
0
1
0
1
1
0
0
0
-
-
-
-
of NAAQS
% of
Sites
0
0
0
-
15
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
-
6
0
0
0
0
0
0
-
0
8
0
100
9
0
0
0
-
-
-
-
CV = 75*
No. of
Sites
1
0
1
-
23
0
0
0
0
3
0
1
0
1
0
1
1
2
1
1
0
0
3
0
0
1
0
1
0
0
0
2
3
0
-
3
0
0
0
0
1
0
-
1
3
0
1
4
1
0
2
-
-
-
-
of NAAQS
% Of
Sites
25
0
7
-
32
0
0
0
0
11
0
50
0
8
0
20
33
25
33
100
0
0
30
0
0
11
0
25
0
0
0
22
21
0
-
18
0
0
0
0
50
0
-
9
23
0
100
36
6
0
25
-
-
_
-
Total
439
15
62
14
-------
TABLE 3-2e. Nitrogen Dioxide.
State/Territory
Alabama
Al aska
Arizona
Arkansas
Cal I torn I a
Colorado
Connecticut
Del aware
District of Col
Florida
Georg i a
Hawai i
Idaho
II 1 I no is
Indiana
Iowa
Kansas
Kentucky
Louisiana
Maine
Maryland
Massachusetts
Michigan
Minnesota
Mississippi
Missouri
Montana
Nebraska
Nevada
New Hampshire
New Jersey
New Mexico
New York
North Carol Ina
North Dakota
Ohio
Ok 1 ahoma
Oregon
Pennsy 1 van la
Puerto Rico
Rhode Island
South Carol Ina
South Dakota
Tennessee
Texas
Utah
Vermont
V I rg I n 1 a
Washington
West Virginia
Wisconsin
Wyoming
American Samoa
Gaum
Virgin Islands
No. Of
Sites
0
0
4
1
64
4
3
2
umb i a 2
10
2
1
0
7
3
0
1
6
4
0
8
6
3
2
0
9
0
0
2
0
5
1
6
0
2
7
4
1
22
0
1
0
0
0
9
2
0
10
2
4
2
0
0
0
0
CV = 50*
No. Of
Sites
-
-
1
1
25
1
1
0
1
1
0
0
-
0
0
-
0
3
0
-
1
1
0
1
-
2
-
-
0
-
0
1
0
-
0
2
1
0
14
-
0
-
_
-
1
0
4
0
1
0
-
-
-
-
of NAAQS
% of
Sites
-
-
25
100
39
25
33
0
50
10
0
0
-
0
0
-
0
50
0
-
13
17
0
50
-
22
-
-
0
-
0
100
0
-
0
29
25
0
64
-
0
-
-
11
0
-
40
0
25
0
-
_
_
-
CV = 75*
No. Of
Sites
-
-
1
1
30
2
1
0
1
1
0
0
-
2
1
-
0
3
0
-
1
1
0
1
-
2
-
-
1
-
3
1
0
-
0
3
1
0
16
-
0
_
-
1
0
-
5
0
1
0
-
_
_
-
of NAAQS
% of
Sites
-
-
25
100
47
50
33
0
50
10
0
0
-
29
33
-
0
50
0
-
13
17
0
50
-
22
-
-
50
-
60
100
0
-
0
43
25
0
73
-
0
_
_
11
0
_
50
0
25
0
_
_
_
-
Total
222
63
28
80
36
-------
o»
c
*^»
o<
^
£
L.
OJ
*J
*C
trt
(.
fQ
jc
Ol 4->
c
o x
^ l/l
t- L.
fO *O
^3 OJ
c >,
4J 0)
V) 4J
^
JT c/1
U
fl3 C
Ol O
O* ^h
c
5 o
o>
OJ "C
(j a;
X in
OI IQ
£>
I/I ^^
Ol
4J ^5
^" *^-
in «_
O>
Vw 4.J
O --
(_
«- 0
Ol
-C c
E X
3 0
Z -0
4J
3
J^
OO ^
1
ro ^
F
UJ Q.
i E
CO «0
f X
1 0)
U
«e
19
Oi
JC
in
^
o<
-C
H-
u
^J
i
3
Z
u
Oi
3
Z
C
o
US
O>
f
S
t^
^
*o
HI
>-
O
H-
L-
W
jp
e
£
t_
ai
JS
E
3
z
to
c
fQ
O)
^
tfC
L.
(Q
O)
^^
o*
c
o
L.
O)
jfp
S
3
2T
u
o>
f
£
J
^J
O
JC
lo
ji»*.
^*
._-
^*
^J
TD
L_
*O
^D
C
fQ
+*
t^
^j
C
t3
^j
2
^B
^H»
0
o.
CO C C
*u c^ vo in
t*» cc m CM
*r cc *r
-« «a- co
oc o o- co
vc "w « vo
co* « « en
^^ ^*^
CO 10 CO
CC ^^ CO ^^
vo fvj ro ro
c^ m CM oc
^M CO ^^
c^ ^^ c^ ^^
t/^ r"*» i/> f"^
c
*o
o>
L.
3
O ^
JC 3
' C
« e
CM <
Q.
co
t
CO OC 00 00
coi mcM in~< cv/
r^ CO VO CTi CM CM
CO CO C-* CO
ec* voco vvo coo
f-» « VOCM »^^ CMUD
-i CV I CM
oo oo c i co
vc*r r«»CM KOO ^cn
inco «« CMen voen
co « co «
Oun oun our> om
mr>» mr*» inf^ mr>.
c c
0) Ol
a: at
t_
3 i I. t
O C JC C
« c . i e
CM «C CO «S
CM CM
o o o
CO O Z
«*
Oi
u
c
^3
"S
Ol
u
X
OJ
^^
oc
en
^M
*
«c
'C
o
o
CO
en
^*
i
CO
r
en
^^
o
4-1
^y
Q)
^B
^*
Q.
ex
fQ
(Q
f*
L.
O)
^
c.
0
TD
O)
c
^B
JO
E
o
u
Ifl
4J
pi»
3
in
O)
L.
<*-
O
in
+)
Ol
in
o>
V
L.
j^
H"
c
3
O
U
in
O)
U
c
fO
^
o<
01
u
X
O)
cv
oc
o-
*-^
k
tJ* O
r^ c
^f\ ^3
« TS
Ol
o o>
4-> 0
X
^3 Oi
Ol
ZI oc
Q. 0-
ta.
*
re ^
^ cc
4-1 CT>
C < '
3
0 0
O 4J
in
O>
O
-g
O)
u
X
Ol
CM
ec
cr
*
i£
"Jo
c
c
oc
cr
i
oc
f-
ff
o
"C
Oi
p»w
Q.
CL
o
c
IW
c
O)
0)
u
X
O)
CO
cc
^^
k
re
4.)
<0
c
c
cc
cr
i
or
cr
o
4-1
-c
Ol
^»-
a.
a.
fO
«^-
t_
(U
4-}
p>»
^
CJ
O
34
-------
criteria were met, only 6 (3 + 1 + 2), or 0.2 percent, exceeded the
standard in the year following shutdown. Based on historical data, then,
there is a very, very small chance that shutting down a site after the
three-year criteria were met would result in erroneously classifying the
site as not exceeding the standard.
Similarly, we applied the three-year criteria to 1978-1980 and 1979-1981
data for all pollutants and counted how many monitors exceeded each
standard two years after meeting the criteria, i.e., in 1981 and 1982,
respectively. Finally, we counted how many sites exceeded the standard in
1983, three years after meeting the criteria applied to 1978-1980 data.
The complete set of misclassification results for TSP, S02» CO, and N02
are contained in Table 3-3. (Ozone is not included in this analysis
because so few sites meet the example criteria.) When these shutdown
criteria were applied with a DVC cutoff of either 50 or 75 percent of the
NAAQS, only a few monitors for the 24-hour TSP standard and one for the 8-
hour CO standard were counted.
We have not looked closely at those few sites that exceeded the standard
after meeting the criteria, but it is possible that there would have been
prior indications that the standard would be exceeded (e.g., new source
permits for nearby emissions) so that pollutant monitoring at these sites
would have been reinstated. Thus, the example criteria are restrictive
enough that the chances of misclassification of a site are extremely
small. Such small misclassification probabilities seem to suggest that
the example criteria are unnecessarily restrictive, and that some relaxa-
tion resulting in acceptably low misclassification probabilities is
possible.
35
-------
4 SEASONAL MONITORING
Regulations for seasonal monitoring of ozone were recently proposed.*
Ozone concentrations reach their peak in the hot summer months when
meteorological conditions are most conducive to ozone formation and
precursor emissions increase. The promulgated seasonal monitoring
schedule was determined on a state-by-state basis; monitoring is not
required in those months when exceedances are unlikely to occur. April
through October monitoring is required for over half of the States, but
the season can be as short as June to September (for Montana and South
Dakota). In States where ozone pollution is a year-round problem, such as
California, Florida, and Texas, full-year monitoring is still required.
In most areas of the country, seasonal monitoring for ozone is a cost-
effective measure and entails virtually no chance of site misclassifica-
tion. The ozone standard is a short-term standard (maximum daily concen-
tration not to be exceeded on the average more than once per year) and is
ideally suited for seasonal monitoring. We have considered the possibi-
lity of seasonal monitoring for other criteria pollutants, and recommend
seasonal monitoring only for carbon monoxide. In this section we discuss
first why seasonal monitoring is not recommended for TSP, S02, and N02»
and then when and how seasonal monitoring of CO is a feasible option. As
noted in Section 1, seasonal monitoring is also not applicable to the NAMS
network.
TSP, SULFUR DIOXIDE, AND NITROGEN DIOXIDE
Controlling Standards
TSP and S02 have both short-term and long-term standards. One or the
other is the controlling standard at a site. The controlling standard is
determined by the ratio of the percent DVC for the short-term standard
divided by the percent DVC for the long-term standard.
second maximum 24-hour average annual average
!... | I I. . - - . , . I , . . i -.
24-hour NAAQS ' annual NAAQS
* 51 Fed. Reg. 9597 (1980) (to be codified at 40 CFR. part 58) (proposed
March 19, 1986).
36
-------
If this ratio is greater than one, the 24-hour standard is the controlling
standard; conversely, if the ratio is less than one, the annual standard
is the controlling one.
There may be strong seasonal variation in short-term averaging periods
depending on the location of the monitor and the types of emission sources
that influence the air quality at the site. For example, a monitor
located near a large facility with high emissions may have strong seasonal
peaks. If the GEP stack height of the facility is low, the monitor might
record winter/spring peaks; if it is high, the monitor might record summer
peaks, provided the facility is in flat terrain. Alternatively, an S02
monitor that is influenced primarily by emissions from area sources may
record winter peaks.
If there is seasonal variation in the 24-hour averages at a site, in the
sense that the short-term standard is exceeded only during some months of
the year, then one might consider seasonal monitoring as an option. But
seasonal monitoring should be considered only if the short-term standard
is clearly the controlling standard. The controlling standard, however,
often changes from year to year at a site. Scatterplots of the ratios
determining the controlling standards for TSP and S02 at all monitors
satisfying NADB completeness requirements in both years for 1982 versus
1983 are shown in Figures 4-1 and 4-2 (based on data from the Quick Look
Reports). The horizontal line corresponds to a ratio of one for 1982, and
the vertical line corresponds to a ratio of one for 1983.
For TSP (Figure 4-1) the long-term standard, the annual geometric mean, is
the controlling standard at the majority of sites in both years, as most
of the sites are in the lower left quadrant of the plot. For a small
percentage of sites the short-term (24-hour) standard controls in both
1982 and 1983 (upper right quadrant). But there are many sites at which
the 24-hour standard controls in one year and the annual standard in the
other. If 1981 data are included, there are relatively few sites where
the short-term standard is the controlling standard for three consecutive
years.
For 502 (Figure 4-2) a larger percentage of the sites are controlled by
the short-term standard in any one year than is the case for TSP, but
there are still many sites in the lower right and upper left quadrants of
the plot. That is, at many monitors, the controlling S02 standard changes
from one year to the next. As in the case of TSP, if the 1981 data are
included, there are even fewer sites where the short-term standard is the
controlling one for three consecutive years.
37
-------
ft
£
41
4
ft
«
*
« 41
41
*
41
W
« 4
* *
{I g
41 « 4
41 4
41 41 4
41
* 41 CM
CM * « 4
41
41 * « « 4
* 41 * C
41
41 « 41 4
4
« 41 41 41 n 4
41 « C
41 < 41
CM C
« 41 * 41 41
CM 41 * C
« 41 CM
CM 41 41 C
4
> 41 en r* CM CM o 0
< ^ ro 41 10 to 0t u) ^
1 CM CM 41 10 f^ 10 + *0 If)
< CM4i cMcotncoeoio
1 I*) « 41 * r»» 10 40 CK C*
i CM rn D) S £ + £ + a.
« * CMp>rsi^r-r^ +
41 41 CM ^ CM 1C CD
41 CJ M
CM
*
« * 41
41
41
41
41
^Illl^llll + llll + llll + llll+llll + llll
^ ". * * ~ » »
oi4BH pjepuns 5u LI [0.^1103 dSJ
+ CR
*
«
« 41
41 ' 41
41
<
41 41
«
« 41
n *
41 41 «
CM CM
C
^ «
fM CM CJ 41
^ 41 CM 41
10 41 *ff *
m£cM« «
+ + + O » 41 41
n ^ «
« CM * m * *
41 41 41
CM * <
*
«
H
41
CM
tQ
CD
_^
b
U)
^^
is.
.
~*
c^
CM
-
«!
^
< «
CQ
00
«
6)
1C
ts
B>
*W
.
=5
^9
(M
S
Kl
+ 5
+ IIII + IIII + I1II+ is
VO ^ CM &)
. Z86I
O
l\
to
f-y
__
i-
«3
^3
C
t/5
c
rm
f
|MV
O
i.
4 »
C
o
tJ
o.
to
ro
CO
CTl
to
o
5;
o
CD
lt
0
K
o
t-
z
0
a.
(M
O)
to
eu
C I/I
o '
0,
OJ S..
CO CU
C^ d
4-1
c
a_ -
t/1)
h^ ^
M- 0
to s_
o o
0 C
i- 0
-a o
L. CO
ro
C C
ro ro
to '
00
C7i QJ
C 4-^
f *^*-
_ to
^^
0 0
s- S
40 4->
C
O to
t-J QJ
ro
0
1 "5
«3- C
«^-
UJ
a:
=3 <\J
0 -
IZ ro
38
-------
. «
ft ft
ft ft
ft ft
« ft
ft ft ft ft
« « ft ft
ft
«
*
ft
ft ft
ft
c
ft * ft ft ft ft ft i
ft ft ft
« * *
tM
ft ft CM ft
ft ft ft ft ft
ft ft ft ft ft CM CM
CM «
ft ft CM
« ft CSI ft
« * * *
« < ft CM«
ft ft ft
« ft ft ft CM « <
* ft ft 4
ft ft ft ft .
ft ft CM* (
* n < '
« * « « (
« ft ft « CM
« CMM ft ft
ft ft C ft « C
« « CM «
ft <
CM «
ft «
ft CM « (
ft ft ft
ft ft ft
<
ft
*
ft
*
*
§ g 2 2 3 I
oi^ey pjepueis
.
«
*
i
«
« _
i
s
« «
«
CM ft ft
ft
ft
» « « csi
« «
«
« <
CM *
ft « ft CM
CM
ft P> ft CM « ft
1 ft CM ft CM
CMCMnCMftft «« «
) « ft CM ft ft
n « CM «
ft ft CM ft CM « «
CM N ft ft ft ft ft
«» ft ft ft ft ft
« « « cn « in
CM ft CM ft T n ft
n «
CM ft C"> CM »» ft CM
i « n < « CM «
« » CSI ft
ft ft ft fO ft ft ft CM «
ft ft ft ft ft
« ft ft
ft ft
CM ft <» « ft ft
ft CM CM ft
ft ft
ft ft
ft (M
ft
ft
ft
ft
* *
5iu i_[0j^iio3 ^QS Z86I
.§
CM
CD
*£»
\o
«
tS?
^f
(SI
>*
§
**
CO
ej
tit
VD
^
(O
cc
o
L.
«3
o
C
to
CO
01
c
f^
P"»
c
o
CSJ
o
oo
CO
00
Wi
f~^
CA
21
o
03
u.
o
o
z
o
a.
-------
Estimation of the Annual Mean From Seasonal Monitoring Data
If TSP and SOj are monitored seasonally, then an annual average must be
estimated to check for exceedance of the long-term standard. The annual
mean will be some multiple, less than one, of the seasonal mean, but
selecting the appropriate multiplier is difficult. It would not be
appropriate to use the historical ratio of the seasonal mean to the annual
mean at a site as the multiplier, because changes in emissions would
require an updated multiplier. A reasonable alternative would be to use a
multiplier determined from other sites, i.e., nearby sites with similar
emissions sources and seasonal patterns, but presumably these other sites
would also qualify for seasonal monitoring. We are left with the possi-
bility of having one or two sites continuing to monitor for the entire
year and acting as indicator sites from which we would determine the ratio
of seasonal mean to annual mean. However, from a statistical point of
view, data from more than one site would be necessary to determine an
appropriate multiplier with any degree of confidence.
If seasonal monitoring were allowed for those TSP and S02 sites where the
short-term standard controls consistently from year to year and the short-
term averages exhibit seasonal behavior, we would not be able to develop a
reliable estimate of the appropriate annual mean from the seasonal mean,
and so would not be able to reliably state whether the long-term standard
was exceeded or not. (This is important for sites where both standards
are exceeded but the short-term standard is the controlling standard.)
More important, the only standard for NC^ is a long-term standard, so if
seasonal monitoring of N0£ were allowed, the annual mean would also have
to be estimated from seasonal data. In some areas, e.g., Los Angeles, N02
does exhibit seasonal patterns; N02 levels reach their peak in fall and
early winter when inversion layers are low and there is consequently less
atmospheric mixing. Because of the difficulties in reliably-estimating
annual means from seasonal data, though, we conclude that seasonal
monitoring does not appear viable for the three pollutants with annual
standards.
CARBON MONOXIDE
Ambient CO concentrations exhibit seasonal patterns in most areas of the
country. Carbon monoxide levels reach their peak in the fall and early
winter months when strong radiation cooling results in strong surface
inversion, which in turn reduces mixing and confines the spread of
pollutants from ground-level sources. An example of strong seasonality in
the maximum monthly 8-hour CO average is seen at the Denver CAMP site in
Figure 4-3. In the three years of monthly maxima plotted for the site,
40
-------
41
-------
the highest 8-hour concentrations occur in December or January of each
year. At this site the maximum monthly 8-hour average does not exceed the
NAAQS of 10 mg/m^ during the months of March through August in each of the
three years shown.
The Denver CAMP site exhibits obvious cyclical behavior, but a similar
time series of monthly maximum 8-hour averages for other sites may not
show such strong seasonal patterns. Should strong seasonality of monthly
maximum 8-hour average CO be the criterion for seasonal monitoring? If
not, how should we determine whether or not a site is eligible for
seasonal monitoring? One possibility is to perform a statistical test for
seasonal patterns. But we recommend against first deciding if there are
seasonal patterns and then determining what the monitoring season should
be.
To see why such a testing procedure is impractical, consider two hypothet-
ical sites. Site A has a strong seasonal pattern with peaks (and exceed-
ances) in the winter and valleys in the summer. Site B has relatively
constant monthly 8-hour maxima with an infrequent exceedance of the
standard. The time series of monthly 8-hour maxima for these two sites is
shown in Figure 4-4. The standard is never exceeded at site B except when
site. A has exceedances, therefore site B should not have to monitor more
than site A. Yet site A is strongly seasonal and site B is not, so if we
performed a statistical test for seasonality first, site A would be
allowed to monitor seasonally but site B would not.
Determination of Monitoring Season
We therefore recommend the following procedure for determining the CO
monitoring season. The required monitoring season, a period of fall and
winter months, is determined by examining some previous years of data with
some data completeness requirements (these decision rules are discussed
futher below). The starting month of the required monitoring season is
the earliest fall or winter month in which an exceedance of some cutoff
concentration (also discussed below) occurred. Similarly, the last month
of required monitoring is the latest winter month in which an exceedance
of the cutoff concentration occurred in the years examined.
Consider the Denver CAMP site in Figure 4-3 as an example. The earliest
fall/winter month in which there is an exceedance of the NAAQS i.n the -
three-year period is September, in both 1982 and 1983. The latest month
where an exceedance occurred is March, in 1982. The required monitoring
season for this site would therefore be September through March.
42
-------
o
o
o
o
o
(SI
o
o
o
o
z
o
ro
co c
^^ O
J3
t.
3
O
C.
I
CO
X W
c o
o
£ 4->
V
c c
*J
o
C CL
o >»
- JC.
o
tr c
«- o
a. e
o
o
o
CM
o
*
o
o
o
(£»*m/&uj)
jnoq-g
43
-------
Figure 4-5 shows a three-year record of another hypothetical site (actu-
ally the Denver CAMP data of Figure 4-3 with the May 1983 maximum
increased to a concentration just above the NAAQS). At this site the
earliest exceedance month is still September. But here the latest month
with an exceedance is May. Even though there were no exceedances in the
month of April in the three-year record, the required monitoring period
would be the continuous nine-month period of September through May.
As control programs for CO are implemented, and CO concentrations are
reduced, there may be some sites where exceedances in the first and last
months of the required monitoring period no longer occur. We therefore
recommend that if three years (or however many years are used to determine
eligibility for seasonal monitoring) pass with no exceedances in the
beginning or ending month of the required monitoring period, and if in all
years those months had acceptably complete data, then the monitoring
period can be shortened accordingly.
Decision Rules for Seasonal Monitoring of Carbon Monoxide
Three decision rules can be established in order to determine the required
monitoring season for CO. These are similar to the decision rules for the
mothballing and rotation options: (1) the cutoff concentration above
which "exceedances" are counted, (2) the length of the test period, and
(3) data completeness during the test period.
The cutoff concentration and the length of the test period must be the
same as for the mothballing option, if both shutdown and seasonal monitor-
ing are available options. Otherwise, if the requirements for seasonal
monitoring are less restrictive, we are faced with a potential paradox: a
site might not meet the requirement for mothballing but could meet the
requirements for seasonal monitoring with a 12-month off-season. Con-
sider, for example, a site where concentrations are at about 60 percent of
the standard in each and every month. If the cutoff concentrations were
50 percent of the standard for shutdown and 75 percent for seasonal
monitoring, and the remainder of the criteria are the same, then the
example site could not be mothballed but is below the seasonal cutoff
concentration in each and every month) and so qualifies for seasonal
monitoring with a 12-month off-season.
Likewise, data completeness requirements need to be as strict or stricter
for seasonal monitoring than for the mothballing and rotation options.
Rather than annual or seasonal data completeness requirements, however,
there should be monthly requirements for the seasonal monitoring option.
If this is not the case, then a site with ho exceedances of^the cutoff
concentration before October 1 in the test period would be eligible for
44
-------
o
o
o
m
o
CVJ
O
o
o
o
o
V)
rr>
eo
O
z
o
o
CO
V
cr
o
u
v
t_
3
O
I
00
X
*} OJ
C V
O £
6 «>
O
C 0.
>.
C.
C
o -
- C
** O
o -
O CJ
C C
o o
v> a
vt rj
X
- o
in c
i o
UJ C
CD oe o
o
o
o
(£*»IU/6(U)
o
o
o
O
o
*
o
jnoq-g
45
-------
shutdown in September. But one would not want to consider September part
of the off-season if in the first test year September had no valid data
and in the third test year only half a month's data. Of course, if a
given month does not meet the data completeness requirements but there is
an exceedance of the cutoff concentration in the month, then obviously
that month must be included in the required monitoring season.
46
-------
MONITORING COSTS AND AN EXAMPLE OF NETWORK COST SAVINGS
The purpose of this section is to demonstrate the potential for cost
savings if one of the options suggested in this report is followed. The
annual costs for pollutant monitoring are based on tables in the recent
report "Cost of Ambient Air Monitoring for Criteria Pollutants and Selec-
ted Toxic Pollutants" by PEI Associates, Inc. (1985) and discussions with
the PEI project manager.
The monitoring costs and cost structure in the PEI report are based on PEI
field experience and discussions with manufacturers and State and local
air pollution control agencies. The costs are estimated averages; clearly
there is much variabilty from state to state, and even from one agency to
another within the same State. We use the PEI estimates to demonstrate
how one would calculate estimates of the cost savings if one of the sug-
gested options is implemented; State and local agencies can perform simi-
lar calculations with their own known costs to determine their actual cost
savings.
COST COMPONENTS
The biggest cost component for pollutant monitoring is technical labor.
In the PEI report four classes of technical staff are identified:
"Labor requirements are based on PEI field experience in implementing
and operating monitoring networks. PEI assumed four standard labor
categories that are typical for a State or local agency. The labor
categories and associated responsibilities are as follows:
Technician IOperates the monitoring site, maintains the site
and instrument log(s), and reduces raw data from the analy-
zer(s).
Technician IIPerforms instrument precision, span, and audit
checks; does routine and remedial instrument maintenance; makes
data computations; maintains site records; and trains site
operators.
47
-------
Technical SupervisorCoordinates staff work assignments,
reviews data, develops status reports for submittal to manage-
ment, assists in equipment procurement and site selection,
reviews program problems, ensures adequate training, maintains
and reports quality assurance activities, and coordinates model-
ing efforts.
ManagementReviews, analyzes, and evaluates monitoring objec-
tives; regulates budgets and procurement activities; and reviews
results to ensure that program objectives are met.
The assumed labor rates for costs developed in this report are based
on an average of rates from three Ohio and Indiana air pollution con-
trol agencies. The rates represent base wages multiplied (burdened)
by a factor of two to account for benefits and agency overhead. Both
the base wage rates and the burdening factor will differ among
agencies, depending on geographic location, agency size, and training
and length of service of employees."
The burdened hourly rates are $16.00 for Technician I, $20.00 for Techni-
cian II, $22.00 for Technical Supervisor, and $26.00 for Management.
These labor rates and all other costs in the PEI report are for 1984 dol-
lars; to reflect 1986 dollars all costs are increased by eight percent,
which is approximately the increase in the U.S. Consumer Price Index for
Urban Wage Earners over the past two years (1984 and 1985).
The annual operating costs for criteria pollutant monitors (except lead)
are presented in Tables 5-1 and 5-2. TSP sampltng costs are the most
straightforward, since no shelter is required. Table 5-1 shows annual
operating costs for intermittent (once every six days) 24-hour TSP
sampling using the required hi-volume sampler. The total annual cost for
sample collection, analysis, maintenance, and quality assurance for a
single TSP monitor is $2,972.
For the other four criteria pollutants under consideration, a temperature-
controlled environment is required for continuous monitoring. This is
normally an aluminum shelter or a room in an office building. The annual
costs for operating such a shelter are listed in Table 5-2a, along with
selected operating costs; these costs are dependent on the number of pol-
lutants monitored at the site. As noted in the table, the time.spent on
routine visits and on precision and span checks varies with the number of
continuous monitors at the site. The total of these costs, summarized in
Table 5-2b, for one or two pollutants is $6,312; for three pollutants the
cost is $9,526; and for four pollutants the total is $12,740.
48
-------
TABLE 5-1. Annual operating costs for TSP sampling. (Adapted from
Table 2-5, PEI, 1985)
"Annual Cost
Cost Item (dollars)
Sampling1190
Utilities at $7/month 84
Sample recovery
Labor (1 nour per sample, Technician I*, 61 samples) 976
100 Type A glass-fiber filters 60
100 filter holders/envelopes ' 50
100 recorder charts/inks . 20
Analysis 340
Tare weigning, numbering, conditioning3 nours,
Technician II* 60
Filter weigning2 nours per quarter, Technician II, 160
Data reduction6 nours, Technician II 120
Maintenance/repair 878
Routine maintenance/calibration32 hours, Technician II 640
Remedial maintenance8 hours, Technician II 160 '
Supplies
4 brush sets at $6 each 24
4 motor cushions at $6 each ' 24
1 Neoprene gasket at $6 each 6
4 filter holder gaskets at $6 each 24
Quality assurance and supervision 344
Reweighing of samples, review of calibrations and audits
12 hours, Technical Supervisor* 264
Certification of calibration and audit units4 hours,
Technician II 80
Total, 1984 dollars . 2752
Total. 1986 dollars (1.08 x 1984 dollars) 2972
* Assumed hourly labor rates are $16 for Technician I, $20 for
Technician II, and $22 for Technical Supervisor, 1984 dollars.
49
-------
TABLE 5-2a. Annual facility costs and selected operating costs for a
continuous monitoring site. (Adapted from Table 2-9, PEI, 1985)
Annual Cost
Cost Item (dollars)
Utilities, at $75 per month
Scheduling/supervision24 hours, Technical Supervisor
(2 hours per month, $22 per hour)
Routine site visits--156 hours, Technician I*
(three visits per week, 1 hour per visit, $16 per hour)
Precision/span checks--96 hours, Technician II*
(two 4-hour visits per month, $20 per hour)
Total, 1984 dollars
Total, 1986 dollars (1.08 x 1984 dollars)-
900
528
2496
1920
5844
6312
* Cost for one or two pollutant monitors. Add 1 hour per visit for each
additional pollutant monitor.
TABLE 5-2b. Facility and operating costs from
Table 5-2a summarized by the number of continuous
pollutant monitors at the site.
Annual Cost for Speci-
fied Number of Contin-
uous Monitors (dollars)
Cost Item
Utilities
Schedul i ng/supervi si on
Routine site visits
Precision/span checks
Total, 1984 dollars
Total, 1986 dollars
1 or 2
900
528
2,496
1,920
5,844
6,312
3
900
528
4,992
2,400
8,820
9,526
4
900
528
7,488
2,880
11,796
12,740
50
-------
TABLE 5-2c. Remaining annual operating costs for continuous pollutant monitors,
(Adapted from Table 2-9, PEI, 1985).
Cost Item
Annual Cost
(1984 Dollars)
Operating Supplies
Data reduction
Automated:
Cassette
Cassette
pickup,
pickup,
1 hour/week, Technician I*
.25 hour/month, Technician
Manual :
6 hours/month, Technician I
1 hour/month, Technician II
1 hour/week to pick up strip charts, Technician I
4 hours/month data entry, Technician I
Maintenance/Repair
Routine maintenance:
16 hours, Technician II
8 hours, Technical Supervisor*
Remedial maintenance, 8 hours, Technician II
Replacement components
Calibration
Routine calibration, 4 hours/quarter, Technician II
Supervision, 1 hour/quarter, Technical supervisor
Calibration gases or permeation tubes:
S02
NO?
CO
420
832
60
1152
240
832
768
320
176
160
400
320
460
460
1560
Quality Assurance and Reporting
Data validation, 1.5 hours/week, Technician I
Data assessment and reporting, 3 hours/quarter,
Technical supervisor
Audits:
4 hours, Technician I
4 hours, Technician II
1 hour, Technical supervisor
Audit gases or permeation tubes:
SO?
NOo
CO
1248
264
64
80
22
230
230
780
* Assumed hourly labor rates are $16 for Technician I, $20 for
Technician II, and $22 for Technical Supervisor, 1984 dollars.
51
-------
TABLE 5-2d. Summary of Table 5-2c operating costs
for continuous pollutant monitors.
Data Reduction Method Average Cost
Pollutant Automated Manual 1984 1986
S02 $5144 $7244 $6194 $6690
N02 5144 7244 6194 6690
03 4454 6554 5504 5944
CO 6794 8894 7844 8472
52
-------
The remaining annual operating costs for continuous pollutant analyzers--
operating supplies, data reduction, maintenance and repair, calibration,
and quality assuranceare listed in Table 5-2c. (These costs assume that
data is picked up at the site rather than transmitted automatically
because relatively few sites currently have telemetric equipment.) The
costs for each pollutant are summarized in Table 5-2d.* Because of the
labor involved in data reduction from strip charts, operating costs for
automated samplers are far lower (about 25 to 30 percent) than for manual
samplers. In the network example below we use the average cost between
the automated and manual data reduction methods since about half of the
samplers currently use automated data loggers.
EXAMPLE SAVINGS
To demonstrate how to calculate the potential cost savings if one of the
recommended monitoring reduction options is implemented, we constructed a
network of 100 sites. The distribution of pollutants monitored at each of
the 100 sites in the network was chosen to be representative of most State
networks (see Table 2-4). Although just seven States have more than 100
criteria pollutant monitoring sites (California, Florida, Illinois, New
York, Ohio, Pennsylvania, and Wisconsin; 1983 data), we chose a large net-
work to demonstrate the cost savings at a variety of sites, each with dif-
ferent numbers of monitors. Our purpose is to indicate the percentage
reduction in costs, not the actual amount of cost savings, that may be
possible in many States.
Table 5-3 shows the pollutants monitored at each of the 100 sites in the
example network, and the annual operating costs at each of the sites. At
59 of the 100 sites, only TSP is monitored, so no shelter is required; the
annual cost of $2,972 for these sites is the total cost in Table 5-1.
There are 15 sites where only one pollutant is continuously monitored (12
for 03, 1 for S02, and 2 for CO), and eight sites where all five pollu-
tants are monitored. The site costs for those sites where at least one
pollutant is continuously monitored are derived from Table 5-2. The total
annual cost for operation of the 100-site network is $1,138,218, or about
$11,000 per site per year. We also note that approximately 15 percent of
estimated annual costs are attributable to sites with TSP monitors only
and approximately 13 percent to sites with ozone monitors only. All sites
with single monitors (74 of the 100 sites) account for approximately 32
percent of the estimated annual operating costs. Another 31 percent of
the estimated cost is attributable to the eight sites that monitor five
* The annual operating costs in Tables 5-2a and 5-2c were split this way
because it was convenient for cost savings calculations.
53
-------
00
t/>
o>
o
c
t_
o
c.
O Ol
H- a.
0)"
*
cvi
o
c.
o
o
c
tv
o
o
a.
«3- 00
O m ID in in ID in in
oo o ooo oo
vo vo vo vo vo vo vc vo
>* >
vovo vo vovovo vovo
cvj cvi cvj cv cvi cvj cvi cvievicvicvj
r»« r^r^r^-r^r^r*. rvr~.r»r>»
« AAMAMA »*»!
cvj cvjcvievicvjcvicvj evjcvjcvjevi
<l ilr-liIJ^rf^rt^-i^-1 t-H t-4 CM CM CNJ CM <
CTiCMiiCVICO«iCVJ-^tirO^<«iCVJ-^»-ii100 O
in i O
0s1
CVJ
O cvj O O
o o z o
z
CVJ
CVJ O CO CO CO CO
O O Z O CVjO OOO
Z O O O
cvi« «>cvicvj
fjeviOOcocoOO
e\j
«O
»cvicvj
coOO
a. cvj fvi«o-a.a.Q.a-o.
vo ooOOO cotr»uo t K H *
CJ
«
o.
cooo
O t
a. o- o.
OO !/1 OO
CO
o
u
T5 C
3 -
C L.
a.
«/> o
o c
U f-
cu ^~
eo e
o
U3
CD
54
-------
criteria pollutants. Approximately 21 percent of estimated annual costs
is attributable to TSP monitoring.
We chose the every-other-year rotation option to demonstrate the calcula-
tions; it has a greater potential for cost savings than a seasonal moni-
toring strategy, but probably a smaller cost savings than the mothballing
approach. The number of monitors assumed eligible for rotation was based
on the example criteria in Section 3: concentrations less than 75 percent
of the NAAQS (less than 75 percent of the controlling NAAQS when there are
two standards) in the most recent year and on average for the most recent
three years, with 75 percent data completeness requirement for each of the
three years. The percentage of sites in the U.S. eligible under these
criteria applied to 1981-1983 data are listed in Table 3-1; these percent-
ages were applied to the sites in the example 100-site network, with the
constraint that sites must be rotated in pairs and therefore only an even
number can be rotated. For example, Table 3-1 shows that 48 percent of
SC>2 monitors nationally meet the eligibility criteria. There are 17 SOg
monitors distributed throughout the 100-site network, as shown in Table
5-3. Eight of the 17 monitors, or about 48 percent, are considered
eligible for rotation in this example. Similarly, 32 of the 79 TSP moni-
tors, none of the 34 03 monitors, 2 of the 19 CO monitors, and 4 of the 15
N0£ monitors in the example network are considered to be eligible for
every-other--year rotation.
Those monitors eligible for rotation were distributed among the monitoring
site types listed-in Table 5-3 in a somewhat arbitrary manner, since our
purpose is to demonstrate the cost savings calculations. Those sites with
at least one pollutant monitor eligible for rotation are listed in Table
5-4; the underlined pollutants at each site indicate which monitors are to
be rotated. The table shows the two-year rotation cycle for the pollutant
monitors at each of the sites. Since the purpose of this example is to
demonstrate cost savings and not how to derive a rotation schedule, we
assume that the eligible pollutant monitors were paired for every-other-
year rotation using an EPA-approved method as discussed in Section 3 of
this report.
The last two columns of Table 5-4 show the annual cost savings at each
monitoring site for each of the two rotation years. The annual savings
were computed from the costs in Tables 5-1 and 5-2. The simplest example
of the calculations is for the 22 TSP monitors. Since the total annual
operating cost for a monitoring site with only a TSP monitor is $2,792
(first line of Table 5-3), the savings is $2,972 in the shutdown year and
none in the monitoring year. As a second example of the calculated
savings, consider a site with TSP and CO monitors. Since both monitors
are eligible for rotation, the site can be completely shut down for one
year at a time. The total annual operating cost for such a monitoring
55
-------
TABLE 5-4. Annual operating costs savings for the 100-slte example network under an every-
ottier-year rotation option. Underlined pollutants are those meeting eligibility requirements
for rotation.
Pollutant Monitors Shut Down
Pollutants Monitored
TSP
TSP
TSP
TSP
TSP
TSP
TSP
TSP
TSP
TSP
TSP
TSP
TSP
TSP
TSP
TSP
JSP
TSP
TSP
TSP
TSP
JSP
!£?
SO;;. 03
TSPj CO
TSP, CO
TSP. S02
TSP. 03, CO
TSP. S02, N02
TSP. SOj, 03
TSP. S02, 03, CO
TSP, 037 CO. N02
TSP. S02, 03, CO, N02
TSP. SO^. 03, CO, "SoJ
TSP. S02, 03, CO, N02
TSP. SO?. 03. CO. NC£
Year A
TSP
None
TSP
None
TSP
None
TSP
None
TSP
None
TSP
None
TSP
None
TSP
None -
TSP
None
TSP
None
TSP
None
S02
None
TSP, CO
None
TSP, S02
None
None
S02
None
TSP
S02, N02
None
TSP, N02
TSP
Year B
None
TSP
None
TSP
None
TSP
None
TSP
None
TSP
None
TSP
None
TSP
None
TSP
None
TSP
None
TSP
None
TSP
None
S02
None
TSP. CO
None
TSP
TSP, $03. N02
TSP
S02
None
None
TSP
None
S02, NOj
Total Savings
Annual Operating
Cost Savings
(1986 Dollars)
Year A
2,972
2,972
2,972
2,972
2,972
2,972
2,972
. 2,972
2,972
2,972
2,972
13,002
17,756
15,083
6.690
2,972
19,808
12,876
2,972
$123,851
Year B
2,972
2,972
2.972
2,972
2,972
2,972
2,972
2,972
2.972
2.972
2,972
6,690
17,756
2,972
22,664
2,972
9,904
2,972
19,808
$118.430
56
-------
site is seen to be $17,756 (seventh line of Table 5-3), so the savings is
$17,756 in the year the monitoring site is completely shut down and none
in the year when both monitors are in operation.
For a final, more complex example of the calculations, consider the last
site in Table 5-4. At this site all five criteria pollutants are moni-
tored, and the TSP, SOj, and N02 monitors are considered for a rotation
operation schedule. In Year A, the TSP monitor is shut down, and the
annual cost savings is $2,972 (see Table 5-1). In Year B the continuous
S02 and NC^ monitors are shut down. The cost savings is $6,690 for shut-
ting down the SC^ monitor, plus $6,690 for shutting down the N02 monitor
(see Table 5-2d), plus $6,428 for reducing the number of continuous moni-
tors from four to two (see Table 5-2b), for a total annual cost savings of
$19,808.
In this example, the total 100-site network cost savings is $123,851 in
Year A and $118,430 in Year B, or between 10 and 11 percent of the
$1,138,218 total annual operating costs for the network in each year.
Approximately 40 percent ($50K) of the annual cost savings is attributed
to reduction in TSP monitoring, of which $33K (27 percent) is due to rota-
tion at sites that have only TSP monitors. Remember that in this hypo-
thetical network approximately 21 percent of estimated annual operating
costs are attributable to TSP monitoring.
If the same eligibility criteria used in the example qualified the moni-
tors for mothballing instead of rotation, then the total annual savings
would be $242,281, or about 21 percent of the annual network operating
cost. However, this savings would not likely be the net savings since
other data (not air quality) that can serve as indicators of air quality
would have to be gathered and analyzed, and there may be additional shut-
down costs.
We believe that somewhat greater than 11 percent annual operating costs
savings is realistic for many State monitoring networks. The example may
be conservative for three reasons. First, annual facility costs do not
include amortization of shelter capital costs (these costs were not calcu-
lated in the PEI report); if mobile shelters are used for those monitoring
locations at which complete shutdown is possible, then the total number of
shelters in the network can be reduced. Second, the example includes
rotated monitors only; seasonal monitoring for nonrotated CO monitors
would result in additional savings. Third, and most important,, the
example was based on a three-year test period for determining shutdown
eligibility. Since only about half of all pollutant monitors satisfy data
completeness requirements for three consecutive years (see Table 3-1), a
one- or two-year test period would result in substantially more monitors
eligible for rotation. We believe that these additional cost savings
57
-------
would be greater than any costs associated with monitor and site shut-
downs.
58
-------
FURTHER STRATEGIES FOR POTENTIAL COST
REDUCTIONS AND CONCLUDING REMARKS
In this section we first consider the possibility of combining the three
suggested options for maximal monitoring costs reduction, and then discuss
the adjustment of statewide trend statistics after one or more of the
options have been implemented. We then summarize the three options
considered, the eligibility criteria, and the overall potential for state
monitoring costs reduction.
COMBINATIONS OF OPTIONS: MAXIMIZING REDUCTION OF MONITORING COSTS
States are not limited to using only one of the three options to reduce
the effective number of criteria pollutant monitors. Some combination of
all three options may in fact be the most cost-effective approach. For
example, a state might opt for seasonal monitoring of CO, and mothballing
or rotation of other monitors. Permanent shutdown might be considered for
monitors meeting a cutoff of 50 percent of NAAQS concentration during a
specified test period, and rotation for those monitors between 50 and 75
percent of the NAAQS.
Consider the number of sites in the U.S. that meet our example criteria
(Table 3-1). Suppose that the EPA approved a 50 cutoff concentration as
the criterion for mothballing and a 75 percent cutoff concentration for
rotation in a particular state, and that the percent of sites meeting
these criteria for each pollutant were similar to the national percents in
the table. In this case it would likely be most cost-effective to
permanently shut down the eligible S02 monitors but rotate the eligible
TSP monitors. This is because most of the SOg sites that are below the 75
percent cutoff concentration are also below the 50 percent cutoff concen-
tration; the limiting criterion for SOg (as shown in Table 3-1) is data
completeness rather than annual summary statistics. For TSP, on the other
hand, a majority of the sites meet the data completeness criterion, and
the cutoff concentration is the limiting criterion.
If only the rotation option is considered, or if the rotation option is
considered in combination with other options, then the cost savings
achieved each year depends on the rotation schedule. As shown in Section
59
-------
5, since there are fixed costs associated with operating a continuous
pollutant monitoring shelter, cost savings are greatest if all of the
monitors at a site can be rotated off in the same year. If the rotation
option is chosen for more than one pollutant, then, there is not one
unique rotation schedule, but rather many. Not only are there different
rotation schedules for each pollutant, but there may be choices for
rotating groups for specific pollutants. For example, one might have SC^
monitors A, B, C, and D all eligible for rotation on an every-other-year
schedule. If all four monitors have the same characteristics, then there
are six possible pairings of the monitors. Which of the six pairings is
optimal, and what the rotation schedule for each pairing should be, is a
type of linear optimization problem. When there are many choices to be
made, it is not practical to calculate cost savings for all possible
rotation schedules. In that case, the linear optimization techniques of
dynamic programming (Dreyfus and Law, 1977) or combinational optimization
(Lawler, 1976) can be used to derive the rotation schedule resulting in
maximal annual operating cost reduction.
ADJUSTMENT OF STATE TREND STATISTICS
Many state air pollution control agencies publish annual pollution summary
statistics each year, and show trends in recent years (e.g., the Califor-
nia Air Resource Board's California Air Quality Data). If these same.
statistics are calculated after some of the suggested options are imple-
mented, then it will appear that pollutant concentrations in the state are
increasing, because the monitors with very low pollutant concentrations
have been permanently or temporarily shut down and are not included in the
averages. Some adjustments to the usual trend statistic calculations must
therefore be made.
Adjustments to trend statistics depend on which options are implemented.
If seasonal monitoring for CO is implemented, then no adjustment of trend
statistics need be made since normally only the maximum 1-hour and 8-hour
concentrations are considered. These maximum concentrations will,
presumably, occur during the season when the CO monitors are in operation.
If monitors are rotated in pairs or triplets, then the monitor in opera-
tion in any given year provides the best estimate of the monitors not in
operation that year. Therefore a weighted average across sites can be
calculated, with the rotation monitor in operation receiving double or
triple weight as appropriate. Suppose, for example, that there are four
monitors (A, B, C, D) of a pollutant, and that in 1987 monitors C and D
begin a rotation schedule, with C in operation the first year and D in
operation the second year. Suppose also that the summary statistic of
interest is the annual mean X at monitor i. Then for 1986 and preceding
60
-------
years the average^ summary statistic for the four months is calculated as
(I. + "L +_YC +_XD)/4^ For 1987 the annual average statistic of interest
would be (^ + XB + 2X/0/4» and for 1988 the average across the sites
would be (XA + Xg + 2XQ)/4.
If mothball ing or permanent shutdown is implemented, then adjustment of
trends statistics is difficult. There are three possible alternatives for
reporting statewide trends. First, weighted averages could be calculated
if one of the monitors eligible for mothball ing was kept in operation to
act as an indicator site; in this case the weight for the indicator
monitor would be one plus the number of mothballed monitors it repre-
sents. The second possibility is to base trend statistics on only those
monitors with continuous data, i.e., those monitors not mothballed. The
percent change from year to year in the average of the operational
monitors' statistics could be considered representative of the percent
change that would be seen if all monitors continued to operate, but
obviously the absolute levels of the average summer statistics would not
be representative. The third possibility is to report the averages across
time for the operating monitors, but to also show the average across all
monitors for those years before mothballing was implemented. The two
trend lines would then show the percent changes from year to year as well
as the difference between the averages across all monitors and the
averages across the operational monitors.
SUMMARY
Current air monitoring costs nationally are about $55 to $58 million
annually. Many of the more than 4,000 criteria pollutant monitors show
concentrations below the NAAQS, and so one way to reduce ambient air
monitoring costs would be to shut down those monitors in low pollutant
concentration areas for some period of time. In this report we suggest
three options, to be used singly or in combination, to reduce criteria
pollutant monitoring costs. These options are mothba.bll ing, or permanent
shutdown, of monitors; operation of monitors on an annual rotation
schedule; and seasonal monitoring recommend for CO in addition to the
current practice for 03.
Four criteria are used to judge whether a monitor is eligible for moth-
balling or rotation:
The design value concentration (DVC) must be below some percent of
the level specified in the NAAQS.
The DVC criterion must be met for some number of years, the test
period.
61
-------
EPA data completeness criteria must be met for each year in the test
period.
Upward trends in the DVCs should not be evident in the OVC over the
test period.
"To be eligible for seasonal CO monitoring, a monitor must meet data
completeness requirements for a specified period. The monitoring season
is defined by the earliest and latest months in which exceedances of some
percent of the 8-hour CO NAAQS occurred in the test period.
The number of monitors eligible for any of the three options, and thus the
potential cost savings, depends on the eligibility criteria used. We
calculated potential cost savings for an example statewide network of 100
monitoring sites with a total of 155 monitors. The number of monitors in
the example network eligible for temporary or permanent shutdown for each
pollutant was based on the national percent of monitors eligible under
strict eligibility criteria applied to recent air quality data. For this
network, we calculated that approximately 10 percent of the total annual
operating costs would be saved if the eligible monitors were rotated; if
the eligible monitors were mothballed instead then the cost savings could
be as much as 20 percent. We believe that these savings for the example
network would be achievable in many statewide ambient air quality monitor-
ing networks.
62
-------
REFERENCES
Burton, C. S., and A. K. Pollack. 1985. "Candidate Options for Improving
Criteria Pollutant Monitoring Cost-Effectiveness: Tentative Rank-
Ordering of Options by Potential Cost Savings." Systems
Applications, Inc., San Rafael, California.
Dreyfus, S. E., and A. M. Law. 1977. The Art and Theory of Dynamic
Programming. Academic Press, New York.
EPA. 1984. AEROS User's Manual, 3rd ed. U.S. Environmental Protection
Agency, Research Triangle Park, North Carolina (EPA-450/2-76-029b).
EPA. 1985. Cost of Ambient Air Monitoring For Criteria Pollutants and
Selected Toxic Pollutants. U.S. Environmental Protection Agency,
Research Triangle Park, North Carolina (EPA-450/4-85-004).
EPA. 1986. National Air Quality and Emissions Trends Report, 1984.
U.S. Environmental Protection Agency.
EPRI. 1982. The Sulfate Regional Experiment: Data Base Inventory and
Summary of Major Index File Programs. Electric Power Research
Institute, Palo Alto, California (EPRI EA-1904).
Lawler, E. L. 1976. Combinatorial Optimization: Networks and
Matroids. Holt, Rinehart, and Winston, New York.
Pollack, A. K., and W. F. Hunt. 1984. "Analysis of Trends and
Variability in Extreme and Annual Average Sulfur Dioxide
Concentrations." APCA/ASQC Speciality Conference On: Quality
Assurance in Air Pollution Measurements, Boulder, Colorado (14-18
October 1984). d
63
-------
TECHNICAL REPORT DATA
(Please read Instructions on reverse before competing)
1. REPORT NO.
EPA-450/4-86-014
3. RECIPIENT'S ACCESSION NO.
4. TITLE AND SUBTITLE
Options for Reducing the Costs of Criteria Pollutant Monitoring
5. REPORT DATE
October 1986
6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
A.K. Pollack and C.S. Burton
8. PERFORMING ORGANIZATION REPORT NO
SYSAPP-86/106
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Systems Applications Inc.
101 Lucas Valley Road
San Rafael, California 94903
10. PROGRAM ELEMENT NO.
11. CONTRACT/GRANT NO.
68-02-3889
12. SPONSORING AGENCY NAME AND ADDRESS
13. TYPE OF REPORT AND PERIOD COVERED
Final
U.S. Environmental Protection Agency
Office of Air Quality Planning and Standards
Monitoring and Data Analysis Division
Monitoring and Reports Branch
Research Triangle Park, NC 27711
14. SPONSORING AGENCY CODE
15. SUPPLEMENTARY NOTES
16. ABSTRACT
The primary purpose of this report is to provide guidance in reducing costs associated with operating
and maintaining monitoring systems, new equipment purchases, quality assurance, laboratory analysis,
maintaining computerized data bases, and data summary and reporting. Since 1980 there have been
increasing efforts to hold or reduce monitoring costs; over the same period pressures for additional
monitoring have been developed. This document provides guidance to State and local agencies on how
the monitoring of criteria pollutants can become more cost effective.
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b. IDENTIFIERS/OPEN ENDED TERMS
c COSATI Field/Group
18. DISTRIBUTION STATEMENT
Release Unlimited
19. SECURITY CLASS (Report)
Unclassified
21. NO. OF PAGES
64
20. SECURITY CLASS (Page)
Unclassified
22. PRICE
EPA Form 2220-1 (Rev. 4-77) PREVIOUS EDITION IS OBSOLETE
------- |