-------
•sss*i
O H "S t* 2
•_B « oa C «j
« S N a. S
a S*g
o
&
cn -O
n
es
£
5 "*
DS
en
Effect
ealt
0 •
I/I t— 1
-a
1
I
Al
3>g
it
13 t9D
0
o ^
3
55
a-
w ^.
11
p
1«
** ^ 4C
I"8!?
c ^*
•a | g
S S®
* Is
** S: 9*
4 &-S
fN (S
(N O
cK
*2\
o
Ci
"*•
00
30)
$*
fa
£ M
^ «
C
§5
"3 S°
SI
,2
^B
S.-S
D O
OQ G
'« x
§3
•o u
fl) •,
M O
13
TH flj
•c|
S d
u o
S g
"S *
!1
OT 13
"C CU
M
•c
.•^sr
•3 *
I!
r*i c
8.8
£>!
O °3
•S M
5 ~
11
a §
•II
o
>,
?l
8
a
t
t
-------
1 in total incidence, which represent from about 11 'to 45% reductions in the PM-related incidence.
2 Figure 4-10 shows the estimated percent reductions in total incidence for mortality associated
3 with long-term exposure to PM2 5 concentrations for this same air quality change in all five of
4 the risk assessment study areas that do not meet the current standards. The point estimates are in
5 the range 0.5 to nearly 4.0 percent reduction in total incidence, which represents from about 18
6 to 59% reductions in the PM-related incidence. Table 4-11 shows the estimated short- and long-
7 exposure mortality incidence to facilitate a comparison both within and across the five study
8 areas. For short-term exposure mortality, single-pollutant, non-accidental mortality estimates are
9 selected since they are available for four of the study areas, and cardiovascular mortality is
10 shown for the fifth area, Philadelphia. For long-term exposure mortality, the ACS-extended
11 estimates for total (all cause) mortality are selected for comparison. In Table 4-11 risk
12 ' reductions are expressed both as a percentage reduction in the PM25-associated mortality and as
13 a percentage of the total mortality due to PM2.5 and other causes. As expected, the reductions in
14 both short- and long-term exposure mortality.associated with PM2S are ranked in the same order
15 as the percent rollback required to bring as is concentrations down to just attaining the current
16 standards, with Los Angeles having the biggest percentage reduction in risk and Philadelphia the
17 least. Also, both the risk remaining upon just meeting the current PM2 5 standards and the size of
18 the reduction in risk in moving from as is concentrations to just meeting the current standards are
19 larger associated with long-term exposure mortality estimates.
20 .
21 4.4.2 Sensitivity Analyses
22 The base case risk assessment used a proportional rollback approach to adjust air quality
23 distributions to simulate the pattern that would occur in an area improving its air quality so that it
24 just meets the current annual average PM2 5 standard. The support for this approach is briefly
25 discussed in section 4.2.3 and in more detail in Appendix B of the TSD(Abt Associates, 2005). •
26 While the available data suggest that this is a reasonable approach, other patterns of change are
27 possible. In a sensitivity analysis an alternative air quality adjustment approach was used which
28 reduced the top 10 percent of the distribution of PM2S concentrations by 1.6 times as much as the
January 2005 4-65 Draft - Do Not Quote or Cite
-------
2
3
4
5
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
lower 90 percent of concentrations. The result of this alternative hypothetical adjustment which
reduces the highest days more than the rest of the distribution showed only a small difference
(less than 1%) in the percent change in PM-associated incidence (see Exhibit 8.2 and Appendix
E, exhibits E5-E8 in Abt Associates, 2005).
4.4.3 Key Observations
Sections 4.4.1 and 4.4.2 have presented the PM health risk estimates and sensitivity
analyses associated with just meeting the current PM25 standards. Presented below are key
observations resulting from this part of the risk assessment:
There is a wide range of reductions in PM25-related incidence across the five urban areas
analyzed which is largely due to the varying amount of reduction in ambient PM25
concentrations required in these urban areas to just meet the current PM2 5 standard. For
example, using single-pollutant models the percent of PM25-related incidence reduced for
short-term, non-accidental mortality ranges from about 45% in Los Angeles to about 18%
in St. Louis. Similarly, using the ACS-extended study the percent of PM25-related
incidence reduced for long-term exposure mortality ranges from roughly 60% in Los
Angeles to about 18% in Philadelphia.
The risk estimates associated with just meeting the current PM25 standards incorporate
several additional sources of uncertainty, including: (1) uncertainty in the pattern of air
quality concentration reductions that would be observed across the distribution of PM
concentrations in areas attaining the>standards ("rollback uncertainty") and (2)
uncertainty concerning the degree to which current PM risk coefficients may reflect
contributions from other pollutants, or the particular contribution of certain constituents
of PM2 5, and whether such constituents would be reduced in similar proportion to the
reduction in PM2 5 as a whole.
At least one alternative approach to rolling back the distribution of daily PM2 5
concentrations, in which the upper end of the distribution of concentrations was reduced
by a greater amount than the rest of the distribution, had little impact on the risk
estimates.
January 2005
4-66
Draft - Do Not Quote or Cite
-------
1 4.5 RISK ESTIMATES ASSOCIATED WITH JUST MEETING ALTERNATIVE
2 PMj s AND PM1M3 STANDARDS
3 4.5.1 Base Case Risk Estimates for Alternative PM2 5 Standards
4 The third part of the PM2.5 risk assessment estimates the risk reductions associated with
5 just meeting alternative suites of annual and-daily PM25 standards. For the five urban areas that
6 exceeded the current PM2 5 suite of standards (i.e., Detroit, Los Angeles, Philadelphia,
7 Pittsburgh, and St. Louis), the estimated risk reductions were those associated with a further
8 reduction in PM25 concentrations from just meeting the current standards to just meeting various
9 suites of alternative PM2 5 standards. For the four urban areas that met the current PM2 5
10 standards based on 2001-2003 levels (i,e., Boston, Phoenix, San Jose, and Seattle), Ihe estimated
11 risk reductions were those associated with a reduction from as is air.quality levels to just meeting
12 various suites of alternative PM2 5 standards.
13 • The selection of the suites of alternative annual and daily standards included in the risk
14 assessment was based on the preliminary staff recommendations described in Chapter 6 of the
15 draft 2003 Staff Paper (EPA, 2003) and consideration of public and CAS AC comments. Annual
16 standards of 15,14,13, and 12 |xg/m3 were each combined with 98th percentile daily standards of
17 40, 35,30, and 25 ng/m3, and 99th percentile daily standards at the same levels.21 In addition, an
18 annual standard of 15 jig/m3 was combined with a ninety -ninth percentile daily standard of 65
19 jig/m3. The combinations of annual and daily alternative standards used in the PM25 risk
20 assessment are summarized in Table 4-12. The same proportional adjustment approach used to
. +
21 simulate air quality just meeting the current standards, described previously in section 4.2.3.2
22 and in section 2.3 of Abt Associates (2005), was used to simulate air quality just meeting the
23 various alternative suites of standards. Table 4-13 provides the design values for the annual and
21In four of the five urban areas that do not meet the current suite of PlV^.s standards, annual standards
within the range of 12 to 15 pg/m3 combined with the current daily standard of 65 ng/m3, using a 98th percentile
form, require the same reduction as when these annual standards are combined with a daily standard of 40 ng/m1,
using the same daily form. Therefore, the risk assessment only included the 14 pg/m5 annual standard combined
with the current daily standard for the one location (i.e., Philadelphia) and annual standard scenario where there was
a difference in the reduction required between daily standards of 40 and 65
J
January 2005
4-67
Draft - Do Not Quote or Cite
-------
1 98* and 99th percentile daily standards for all of the PM2 5 risk assessment study areas based on
2 air quality data from 2001-2003 for the base case risk estimates.
3 The estimated risk reduction in total non-accidental mortality, presented both in terms of
4 percent reduction in total incidence and in number of cases avoided, associated with short-term
5 PM2s exposures for alternative annual standards combined with ninety-eighth and ninety-ninth
6 percentile daily standards, respectively, are given in Figures 4-11 and 4-12 for Detroit.
7 Similarly, the estimated risk reduction in total mortality associated with long-term PM25
8 exposures for these same alternative standards are given in Figures 4-13 and 4-14 for Detroit.
9 Similar figures for the other risk assessment locations and additional risk estimates for cause-
10 specific mortality, hospital admissions, and respiratory symptoms (depending on location)
11 associated with alternative standards are presented in Chapter 8 and Appendix F of Abt
12 Associates (2005). As with the estimated risk reductions presented earlier for just meeting the
13 current PM2S standards, when the percent reduction is expressed in terms of the estimated
14 reduction in PM-related incidence rather than total incidence, the changes are much larger. The
15 complete set of risk estimates is presented in Exhibits 8.5a through 8.5h for Detroit and the
16 exhibits in Appendix F for the other 4 locations in the TSD (Abt Associates, 2005).
17 Some interesting patterns can be observed in the estimated risk reductions displayed in
18 Figures 4-11 through 4-14. For example, in Figures 4-11 and 4-13 one observes there are no
19 estimated reductions in risk in going fromjust meeting the current 15 |ig/m3 annual standard/65
20 |ig/m3 98th percentile daily standard to either a 40 or 35 jig/m3 98th percentile daily standard with
21 the same 15 |ig/m3 annual standard. The reason for this is that the 28.1% reduction, required
22 based on the 3-year estimated design value, when applied to the 2003 PM2 5 distribution for the
23 composite monitor to meet the current 15 [ig/m3 annual standard, brings down the 98th percentile
24 daily value to below 35 ng/m3. Thus, there is no additional reduction in air quality or risk when
25 either a 40 or 35 ^g/m3 98th percentile daily standard is considered in combination with a 15
26 |ig/m3 annual standard. Meeting lower daily 98th percentile standards of 30 or 25 jig/m3 when
27 combined with the current annual standard do require additional air quality reductions and, thus,
28 result in additional estimated risk reductions compared to just meeting the current suite of
January 2005
4-68
Draft - Do Not Quote or Cite
-------
t
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
Table 4-12. Alternative Sets of PM,, Standards Considered in the PM,« Risk
•is •
Assessment"1
Annual
Standard
15
14
13
12
98th Percentile Daily Standard
65
X**
40
X
X
X
X
35
X
X
X
X
, 30
X
X
X
X
25
X
X
X
X
99th Percentile Daily Standard
65
X
40
X
X
X
X
35
X
X
X
X
• 30
X
X
X
X
25
X
X
X
X
*A11 standards are in
**Only in Philadelphia.
Table 4-13. Estimated Design Values for Annual and 98th and 99th Percentile Daily PM2.5
Standards Based on 2001-2003 Air Quality Data*
Location
Boston
Detroit
Los Angeles
Philadelphia
Phoenix
Pittsburgh
St. Louis
San Jose
Seattle
Annual
14.4
19.5
23.6
16.4 . .
11.5
21.2
17.5
14.6
11.1
98"1 Percentile Daily
44
44
62
51
35
63
42
47
41
99th Percentile Daily
60
48
96
89
41
70
46
53
48
*The calculation of design values is explained in Schmidt (2005). All design values are in |ag/m3.
January 2005
4-69
Draft - Do Not Quote or Cite
-------
V)
VA
fe
I
I
i ;;;;;As«8.
::::if.:
;;!!l7;
•a.
I
III
III
ill
w
'5F5
4
£
-------
0
t
^wx;:
%xsii$i\l
::::::*£"::::::::«
;<*.;
S
v/
w
I
oo
o "^
oo o\ o i— i
<— — (S CS
en
tN
-------
V)
\o
•s
fe
; *;=
...
:".:: a :.;;.:::
B S fi
= = -
£<3 §
fill
>•§ s
•8
I j 2
£ J 1
ill
|ls
a« «
.s r
s n
IIP
s.«» *3 -.
I5 §8
^ eg S ^
VJ
. -d I
C"3 AI *^ O
. M JT u o
5 O ^ Q aj «
|| | If!
J I 1 I 1 |
r« ® "gD S (S §
W W S 00 » co
s
a
<$
rp
•4
t
4
£
b
3
s
-------
t
•
p^
; :: Reduction in Incroence (Num&sr of c.
m^^^-^^^^
:;;;;;;::;:::::: i^|iflp-0f ; w*^yw*wwwww.
' -:--:••• .frjft^ ^ • i;-yy ^;.- VQ-J^A'I!
iiiiiMi
a
ft
*
*
»F
«0
»
**
kfM
ira
":«:
•4::^-
;;:::*»
:::::::&:
:::; :F*
V»B»rt«y>.
JS0
..„•„
J&
:"Aj!
**
••v
ipr
;-!*
PT ^
Sfr
l> P
»pa
cie
- • -i
:"
jsy'stei ;;:
!:: ::;::::
::; ; ::.»•;
::: : ::.«T
J::::: :::;::
|^;;|i
iiiii; ;;:in
Rolling Back PM2S
Current Daily Standard o
and Daily 99th Percentile
ng'Term Exposure Mortality Associated with 1
Current Annual Standard of 15 ug/m3 and the
t Just Meet Alternative Suites of PM2 5 Annual
d
. Estimated Annual Reduction in Lo
Concentrations that Just Meet the *
ug/m3 to PMj s Concentrations that
Standards: Detroit, MI, 2003.*
*Based on Pope et al. (2002) - ACS extende
Source : Abt Associates (2005)
Q
fe
8
S
'CS
-------
1 standards. The maximum incremental risk reduction from the current standards, with respect to
2 both short- and long-term exposure PM2 5-associated mortality, is estimated to occur for meeting
3 the daily 98th and 99th percentile daily standards set at 25 [ig/m3. For daily standards set at this
4 level the estimated risk reduction does not depend on the level of the annual standard within the
5 range of standards considered. Within four of the five study areas, just meeting 98th or 99th
6 percentile daily standards set at 30 p.g/m3 results in the same short- or long-term exposure
7 mortality risk reductions no matter which annual standards (from 12 to 15 ng/m3) they are paired
8 with. Similar, although not identical, patterns are observed in the other four risk assessment
9 locations that do not meet the current PM25 standards (see Figures Fl through F14 in the TSD
10 (Abt Associates (2005)).
11
12 4.5.2 Base Case Estimates for Alternative PMto_2S Standards
13 The second part of the PM]0.2 5 risk assessment estimates the risk reductions associated
14 with just meeting alternative daily PM10_2 5 standards for the three locations examined earlier
15 (Detroit, St. Louis, and Seattle). Estimated reductions in risk were developed for going from as is
16 levels (based on 2003 air quality) to just meeting alternative PM10.2 5 standards. Staff selected
17 the alternative daily standards to be included in the risk assessment based on the preliminary staff
18 recommendations described in Chapter 6 of the draft 2003 Staff Paper (EPA, 2003) and
19 consideration of public and CASAC comments. Table 4-14 summarizes the sets of 98th and 99th
20 percentile daily standards that were included in the PM10_2 5 risk assessment. The estimated design
21 values which were used to determine the air quality adjustment to be used in simulating just
22 meeting alternative PM10_25 standards are shown in Table 4-15.
23 The estimated annual reduction in hospital admissions for ischemic heart disease,
24 presented both in terms of percent reduction in total incidence and in number of cases avoided,
25 associated with short-term PM,0.2 5 exposures for alternative 98th and 99th percentile daily
26 standards, respectively, are given in Figure 4-15 for Detroit. Daily PM10_25 standards set at 80
27 (for 98th percentile form) and 100 or 80 (for 99th percentile form) result in no reduction in risk in
28 Detroit. The reason why no estimated risk reductions are observed with these alternative
29 standards is that the percent reduction of PM10.2 5 concentrations at the composite monitor to just
30 meet a standard is determined by comparing the alternative standard level with the design value
January 2005 4-74 Draft - Do Not Quote or Cite
S
-------
••
t
1 for that location based on 2001-2003 air quality data. In Detroit, the design value for the 98th
2 percentile daily PM10.25 standards is 70 jig/m3 whereas the 98th percentile daily value in 2003 is
3 105.9 ug/m3. Because the design value is lower than 80 ng/m3, the highest 98th percentile daily
4 PM10_2 5 standard, zero risk reductions were estimated to result from this standard, even though the
5 98th percentile daily value at the composite monitor in 2003,105.9 jig/m3, is well above the
6 standard level. Similarly, the design value for the 99th percentile daily PM10.2 5 standards is 77
7 jig/rn3 for Detroit, whereas the 99th percentile daily value at the composite monitor in Detroit in
8 2003 is substantially greater than 100 (ig/m3, the highest 99th percentile daily PM10.2 s standard. So
9 zero risk reductions were similarly estimated to result from both a 100 and 80 ng/m3 standards. In
10 general, estimated risk reductions increase and the confidence intervals around the estimates
11 widen as lower daily standards are considered.
12 As expected, the maximum reduction in risk is achieved with the 98* percentile 25 ^ig/m3
13 standard and 99th percentile 30 jig/m3 standard. The point estimate is that about a 4% reduction in
14 hospital admissions for ischemic heart disease, equating to roughly 450 fewer cases, would result
15 from meeting either of these daily standards. Similar patterns in risk reduction are observed for
16 the other hospital admission endpoints in Detroit which are included in Chapter 9 of Abt
17 Associates (2005). Additional risk estimates for hospital admissions for asthma in Seattle and
t
18 cough and lower respiratory symptoms in St. Louis can be found in Appendix G of Abt
19 Associates (2005). Based on the point estimates, there are no risk reductions associated with just
20 meeting daily 98th percentile PM10.2 5 standards of 80 |ig/m3 in Detroit, and 80,65, and 50 ng/m3 in
21 St. Louis or Seattle. Similarly, there are no risk reductions associated with just meeting daily 99th
22 percentile PM10.25 standards of 100 or 80 [ig/m3 in Detroit, and 100, 80, or 60 [ig/m3 in St. Louis..
23 or Seattle.
24
25 4.5.3 Sensitivity Analyses for Alternative PM1S and PM10.j 5 Standards
26 4.5.3.1 Hypothetical Thresholds
27 An important observation from the sensitivity analyses on estimated health risks
28 associated with "as is" PM2 5 concentrations was that the impact of hypothetical thresholds was
29 the greatest on the estimated risks.. In order to gain insight into the impact of this important
30
January 2005 • . 4-75 Draft - Do Not Quote or Cite
-------
1
2
3
4
5
6
7
$
9
10
11
12
13
14
15
16
17
18
19
20
21
22
Table 4-14. Alternative PM,n,. Standards Considered in the PM.n •, < Risk Assessment*
M8-2.5
Daily
Standards Based on the 98th Percentile
Value
80
65
50
30
25
Daily Standards Based on the 99th Percentile
Value
100
80
60
35
30
*A11 standards are in |ig/m3.
Table 4-15. Estimated Design Values for 98th and 99th Percentile Daily PM10,2S Standards
Based on 2001-2003 Air Quality Data9"
Location
Detroit
St. Louis
Seattle
. 98th Percentile Daily
70
33
31
99th Percentile Daily
77
47
39
"The calculation of design values is explained in Schmidt (2005). All design values are in
January 2005
4-76
Draft - Do Not Quote or Cite
f
-------
Reduction i
imoiiP
«
A|BP
i i
II
II
"o to
ft >,
•£ 1
•a
e
£
09
on £
5«
a
cn
e
"5
fi
1
e
S
e
• m
73 |
*S.§
S*
ss
0 B
if
I
in
O
8
•*—/
CO
I
3
S §^«
« y g
= ^1:
5 2 i
< *?« '
« R rf §
S « g-g
fe
a
t
s
R
I
ts tn
vo
oo
O ^^
en
vo
-------
1 uncertainty on the risk estimates, an additional set of sensitivity analyses was developed to
2 examine the impact of different hypothetical threshold assumptions on estimated risks associated
3 with just meeting the current and alternative PM25 standards and alternative PM10_25 standards.
4 For those locations and cases where either the current PM25 standards or any of the alternative
5 suites of standards were already met under as is air quality, the estimated risks associated with
6 "as is" PM2 5 (or PM]0_2 5) concentrations in excess of either background or the LML for the health
7 endpoint, whichever is greater, were calculated.
8 For PM25 this sensitivity analysis included estimates of risk for all cause mortality,
9 cardiopulmonary mortality, and lung cancer mortality associated with long-term exposure to
10 PM15 based on Pope et al. (2002) - ACS extended. Since the patterns observed were identical,
11 only the all cause mortality results are presented in Appendix 4B (See Abt Associates, 2005 for
12 the cause-specific mortality estimates). In addition, this sensitivity analysis also included non-
13 accidental mortality (or cause-specific if there was no suitable function for non-accidental
14 mortality available) associated with short-term exposure to PM25. As in the earlier sensitivity
15 analysis for as is air quality, hypothetical thresholds of 10,15, and 20 [ig/m3 were considered for
16 health endpoints associated with short-term exposures, and hypothetical thresholds of 10 and 12
17 M-§/m3 were considered for the mortality endpoints associated with long-term exposure.
18 The sensitivity analysis results for all-cause mortality associated with long-term exposure
19 and mortality associated with short-term exposure for Detroit, Los Angeles, Philadelphia,
20 Pittsburgh, and St. Louis are shown in Appendix 4B to this Chapter (Tables 4B-1 through 4B-10)
21 The results for cardiopulmonary and lung cancer mortality associated with long-term exposure to
22 PM25 based on Pope et al. (2002) - ACS extended are shown in Appendix H of Abt Associates
23 (2005). Not surprisingly, estimated PM-related incidences varied substantially with both
24 hypothetical threshold assumptions and alternative standards. In Detroit, for example, the
25 estimated number of cases of non-accidental mortality associated with short-term exposure to
26 PM2 5 when the current standards are just met decreases from 115, under the assumption of no
27 threshold, to 54, 26, and 12 under hypothetical threshold assumptions of 10,15, and 20 jig/m3,
28 respectively. Because meeting increasingly lower level standards removes estimated cases at the
29 higher concentrations and considering higher hypothetical thresholds increasingly removes
30 estimated cases at concentrations between background (or the LML) and the threshold, one would
31 expect to see an increase in the percent reduction associated with just meeting alternative
January 2005 4-78 Draft - Do Not Quote or Cite
-------
1 standards for higher hypothetical thresholds. This is exactly what is found. For example, as seen •
2 in Table 4B-1, going from just meeting the current standards (15 [ig/rri3 annual and 65 |ig/m3 daily •
3 98th percentile value)'to just meeting the lowest set of standards considered (12 jig/m3 annual and
4 25 ng/m3 'daily 99th percentile value) results in a reduction in short-term exposure mortality
5 incidence of (115 -75)7115 = 34.8 percent under the assumption of no threshold, but under the
6 assumption of a threshold of 10 |ig/m3 it results in a reduction of (54 - 22)/54 = 59 percent." Under
7 hypothetical short-term exposure thresholds of 15 and 20 jig/m3, the reductions are 73 percent and
8 83 percent, respectively. 'As shown in Table 4B-2 for all-cause mortality associated with long-
9 term exposure in Detroit, the reduction in mortality incidence is even more dramatic when'
r . ,
10 alternative hypothetical thresholds are considered. Going from just meeting the current standards
11 to j ust meeting the lowest set of standards considered (12 ng/m3 annual and 25 jig/m3 daily 99th
12 percentile value) results in a reduction in long:term exposure mortality incidence of (522-
13 207X522= 60% under the assumption of no threshold, but under the assumptions of a long-term
14 exposure threshold of 10 ng/m3 it results in a reduction of (282 - 0)/282 =100 percent. With a
15 hypothetical long-term exposure threshold of 12 jig/m3 estimated incidence is reduced to 41 upon
16 just meeting the current suite of standards and a 100% reduction is achieved upon meeting either
17 a. 15 jig/m3 annual standard with a 30 [ig/m3 daily 98th percentile standard or a 1'4 (ig/m3 annual
18 with a 40 (ig/m3 daily 98th percentile value. The same general patterns can be seen in all
19 locations and for all health endpoints considered:
20 The sensitivity analysis results examining alternative PM^ 5 standards with hypothetical
21 thresholds associated with short-term exposure morbidity endpoints for Detroit, Seattle, and St.
22 Louis also are shown in Appendix B to this Chapter (Tables 4B-11 through 4B-13). The health
23 endpoints included hospital admissions for ischemic heart disease in Detroit; hospital admissions
24 for asthma (age < 65) in Seattle; and days of cough among children in St. Louis, all associated
25 with short-term exposures to PM10.2 5 exposures Hypothetical short-term exposure thresholds 'of
26 10,15, and 20 jig/m3 were considered.
27 4.5.3.2 Spatial Averaging Versus Maximum Community Monitor
28 As -discussed previously in section 4.2.3.2, under the current annual PM2 5 standard urban
29 areas may, under certain circumstances,' use the average of the annual averages of several
30 monitors within an urban area to determine compliance with the annual standard, commonly
31 referred to as the "spatial averaging approach." Four of the five urban areas included in the PM25
January 2005 ' 4-79 Draft - Do Not Quote or Cite
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
J7
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
• . =• In four of the five risk assessment locations that do not meet the current PM2 s standards,
daily standards of 40 )ig/m3, 98th percentile or 65 (ig/m3, 99th percentile when combined
with the current 15 [ig/m3 annual standard provide no additional risk reduction in terms of
short-term exposure mortality. • . t
• In all five of the risk assessment locations that do not meet the current PM2.5 standards,
the maximum risk reduction with respect to both short- and long-term PM2 5-associated
mortality is estimated to occur upon meeting the 98th and 99th percentile daily standards
set at 25 jig/m3. For these standards the estimated risk reduction does not depend on the
level of the annual standard within the range of standards examined.
• For four of the five risk assessment locations the estimated risk reduction within each
area associated with meeting either a 98th or 99th percentile daily PM25 standard set at 30
|ig/m3 is the same no matter which annual standard is included within the range of
standards examined. -
• For the PM10.2 5 risk estimates, the maximum reduction in risk is achieved with the 98th
percentile 25 u.g/m3 standard or 99th percentile 30 jig/m3 standard. The point estimate is '
that about a 4% reduction in hospital admissions for ischemic heart disease, equating to .
roughly 450 fewer cases, would result from meeting either of these daily standards. The
confidence intervals get significantly larger as lower PM,0.2 5 standards are considered.
Similar patterns in risk reduction are observed for the other hospital admission endpoints
in Detroit
4
• Based on the point estimates, there are no risk reductions associated with just meeting
daily 98th percentile PM10.2 5 standards of 80 (ig/m3 in Detroit, and 80,65, and 50 [ig/m3 in
St. Louis or Seattle. Similarly, there are no risk reductions associated with just meeting
daily 99th percentile PM10.2 5 standards of 100 or 80 [ig/m3 in Detroit, and 100, 80, or 60
|ig/m3 in St. Louis or Seattle.
Section 4.5.3 presented the results of the following two sets of sensitivity analyses: (1)
considering the impact on risk estimates associated with just meeting the current and alternative
PM2 j standards and alternative PM10.2 5 standards when hypothetical threshold models are included
and (2) considering the impact on risk estimates associated with just meeting the current and
alternative PM25 standards when the average of the annual averages of several monitors within an '
urban area are used to determine compliance with the annual standard, commonly referred to as
the "spatial averaging approach." Presented below are key observations resulting from this part of
the risk assessment: • •
• For short-term exposure mortality associated with PM2 5 there is a significant decrease in •
the incidence avoided as one considers higher hypothetical thresholds. There also is a
January 2005
4-82
Draft - Do Not Quote or Cite
-------
1 significant increase observed in the percent reduction in PM-associated incidence upon
2 ' just meeting alternative standards with higher hypothetical thresholds. The reduction in
3 incidence and increase in percent reduction in PM-associated incidence is even more
4 dramatic for long-term exposure mortality as higher alternative hypothetical thresholds
5 are considered.
6
7 • For short-term exposure morbidity associated with PM10_25, there is a significant decrease
8 in the incidence avoided as one considers higher hypothetical thresholds.
9 '
10 • There is an increase in estimated short-term exposure mortality incidence associated with
11 PM15 when a spatial averaging approach is used to determine compliance with the current
12 annual standard or alternative suites of standards where the daily standard is not the
13 controlling standard.
t
January 2005 4-83 Draft - Do Not Quote or Cite
-------
I REFERENCES
2
3 Most Chapter 4 references are available at the end of Chapter 3. References not listed at the end
4 of Chapter 3 are listed here.
6 Abt Associates Inc. (1996). "A Particulate Matter Risk Assessment for Philadelphia and Los Angeles." Bethesda,
7 MD. Prepared for the Office of Air Quality Planning and Standards, U.S. Environmental Protection
8 Agency, Contract No. 68-W4-0029. July 3 (revised November). Available:
9 http:/.'Vww.epa.gov/ita/naags/standards/pm/s_pm pr td.html.
10
11 Abt Associates Inc. (1997a). Abt Associates Memorandum to U.S. EPA. Subject: Revision of Mortality Incidence
12 Estimates Based on Pope et al. (1995) in the Abt Particulate Matter Risk Assessment Report. June 5,1997.
13
14 Abt Associates Inc. (1997b). Abt Associates Memorandum to U.S. EPA. Subject: Revision of Mortality Incidence
15 Estimates Based on Pope et al. (1995) in the December 1996 Supplement to the Abt Particulate Matter Risk
16 Assessment Report. June 6,1997.
17
18 Abt Associates Inc. (2002). Proposed Methodology for Particulate Matter Risk Analyses for Selected Urban Areas:
19 Draft Report. Bethesda, MD. Prepared for the Office of Air Quality Planning and Standards, U.S.
20 Environmental Protection Agency, Contract No. 68-D-03-002. Available:
21 http://wvvw.epa.gOv/ttn/naaqs/standards/pm/s pm cr td.html.
22
23 Abt Associates Inc. (2003a). Abt Associates Memorandum to U.S. EPA. Subject: Preliminary Recommended
24 Methodology for PMi0 and PM10.2.s Risk Analyses in Light of Reanalyzed Study Results. April 8,2003.
25 Available: http://w^w.epa.aov/ttn/iiaaqs/sta!idards/piii/s pm cr td.html.
26
27 Abt Associates Inc. (2003b). Particulate Matter Health Risk Assessment for Selected Urban Areas: Draft Report.
28 Bethesda, MD: Prepared for the Office of Air Quality Planning and Standards, U.S. Environmental
29 Protection Agency, Contract No. 68-D-03-002. Available:
30 http://mvw.epa,eov/ttrj/naaqs/standards/CTn/s pm cr td.html.
31
32 Abt Associates Inc. (2005). Particulate Matter Health Risk Assessment for Selected Urban Areas. Draft Report.
33 Bethesda, MD. Prepared for the Office of Air Quality Planning and Standards, U.S. Environmental
34 Protection Agency, Contract No. 68-D-03-002. Available:
35 hUo://www epa.aov/ttn/naaQ3/stendards/pra/s_pm cr ld.html.
36
37 Center for Disease Control (2001). CDC Wonder. Available: hitp://wonder.cdc.aov/.
38
39 Deck, L. B.; Post, E.S.; Smith, E.; Wiener, M.; Cunningham, K.; Richmond, H. (2001). Estimates of the health risk
40 reductions associated with attainment of alternative particulate matter standards in two U.S. cities. Risk
41 Anal. 21(5): 821-835.
42
43 Environmental Protection Agency (2001). Particulate Matter NAAQS Risk Analysis Scoping Plan, Draft. Research
44 Triangle Park, NC: Office of Air Quality Planning and Standards. Available:
45 httr>://www.eKa.eov/ttn,''!iaaqs/standai'ds/pfii/sj3m cr tdJitml.
46
47 Hopke, P. (2002). Letter from Dr. Phil Hopke, Chair, Clean Air Scientific Advisory Committee (CASAC) to
48 Honorable Christine Todd Whitman, Administrator, U.S. EPA. Final advisory review report by the CASAC
49 Particulate Matter Review Panel on the proposed particulate matter risk assessment May 23,2002.
50 Available: http:/AA^vTv.epa.gpy/sab/pdt7casacady02002.pdi.
51
January 2005 4-84 Draft - Do Not Quote or Cite
t
-------
1 Langstaff, J. (2004). OAQPS Staff Memorandum to PMNAAQS Review Docket (OAR-2001-0017). Subject A
2 Methodology for Incorporating Short-term Variable Background Concentrations in Risk Assessments.
3 December 17,2004. Available: http://www.epa.gOv/ttii/riaaqs/standards/Pin/s otn.cr td.html.
*4
5 Langstaff, J. (2005). OAQPS Staff Memorandum to PMNAAQS Review Docket (OAR-2001-0017). Subject:
6 Estimation of Policy-Relevant Background Concentrations of Particulate Matter. January 27,2005.
7 Available: htlp://w\vw.er>a gov/tfo'naacis/standards/pm/s pm crtd.html.
8
9 National Academy of Sciences (2002). Estimating the Public Health Benefits of Proposed Air Pollution Regulations.
10 Washington, D.C.: The National Academy Press. Available:
11 httn://www.rnD.edu/books/0309086094/hurd/. !
12 . .
13 Post, E.; Deck, L.; Laratz, K.; Hoaglin. D. (2001). An application of an empirical Bayes estimation technique to the
14 estimation of mortality related to short-term exposure to particulate matter. Risk Anal. 21(5): 837-842.
15
16 Schmidt, M.: Mintz, D.; Rao, V.; McCluney, L. (2005). U.S. EPA Memorandum to File. Subject: Draft Analyses of
17 2001-2003 PM Data for the PMNAAQS Review. January 31,2005. Available:
18 btto:/Avww.ena.eov/oar/oaaiiS/mn25/docs.httul.
19
20 Science Advisory Board (2004). Advisory on Plans for Health Effects Analysis in the Analytical Plan for EPA's
21 Second Prospective Analysis - Benefits and Costs of the Clean Air Act, 1990-2000. Advisory by the Health
22 Effects Subcommittee of the Advisory Council for Clean Air Compliance Analysis. EPA SAB Council -
23 ADV-04-002. March. Available: http://ww\^gpa.gov/sciencel/{xif/coujicil adv 04QQ2.pdf.
24
t
January 2005 4-85 Draft - Do Not Quote or Cite
-------
1 5. STAFF CONCLUSIONS AND RECOMMENDATIONS ON PRIMARY PM NAAQS
2 5.1 INTRODUCTION
3 This chapter presents staff conclusions and recommendations for the Administrator to
4 consider in deciding whether the existing primary PM standards should be revised and, if so,
5 what revised standards are appropriate.1 The existing suite of primary PM standards includes
6 annual and 24-hour PM2 s standards, to protect public health from exposure to fine particles, and
7 annual and 24-hour PM!0 standards, to protect public health from exposure to thoracic coarse
8 particles. Each of these standards is defined in terms of four basic elements: indicator,
9 averaging time, level and form. Staff conclusions and recommendations on these standards are
10 based on the assessment and integrative synthesis of information presented in the CD and on
11 staff analyses and evaluations presented in Chapters 2 through 4 herein.
*
12 N In recommending a range of primary standard options for the Administrator to consider,
13 staff notes mat the final decision is largely a public health policy judgment. A final decision
14 must draw upon scientific information and analyses about health effects and risks, as well as
/
15 judgments about how to deal with the range of uncertainties that are inherent in the scientific
16 evidence and analyses. The staffs approach to informing these judgments, discussed more fully
17 below, is based on a recognition that the available health effects evidence generally reflects a
18 continuum consisting of ambient levels at which scientists generally agree that health effects are
19 likely to occur through lower levels at which the likelihood and magnitude of the response
20 become increasingly uncertain. This approach is consistent with the requirements of the
21 NAAQS provisions of the Act and with how EPA and the courts have historically interpreted the
22 Act. These provisions require the Administrator to establish primary standards that are requisite
23 to protect public health with an adequate margin of safety. In so doing, the Administrator seeks
24 to establish standards that are neither more nor less stringent than necessary for this purpose.
25 The provisions do not require that primary standards be set at a zero-risk level, but rather at a
26 level that avoids unacceptable risks to public health.
1 As noted in Chapter 1, staff conclusions and recommendations presented herein are provisional; final staff
conclusions and recommendations, to be included in the final version of this document, will be informed by
comments received from CASAC and the public in their reviews of this draft document
January 2005 5-1 • Draft - Do Not Quote or Cite
-------
1 " 5.2 APPROACH
2 As an initial matter, PM25 standards for fine particles and PM]0 standards for thoracic
3 coarse particles are addressed separately, consistent with the decision made by EPA in the last
4 review and with the conclusion in the CD that fine and thoracic coarse particles should continue-
5' to be considered as separate subclasses of PM pollution. As discussed in Chapter 3, section •
6 3.2.3, this conclusion is based in part on long-established information on the differences in
7 sources, properties; and atmospheric behavior between fine and coarse particles, and is
8 reinforced by new information that advances our understanding of differences in human
9 exposure relationships and dosimetric patterns, and the apparent independence of health effects
10 that have been associated with these two subclasses of PM pollution in epidemiologic studies.
11 "In general, in evaluating whether the current primary standards are adequate or whether
12 revisions are appropriate, and in developing recommendations on the elements of possible
13 alternative standards for consideration, staffs approach in this review builds upon and broadens
14 the general approach used by EPA in the last review. In setting PM25 standards in 1997, the
15 Agency mainly used an evidence-based approach that placed primary emphasis on epidemiologic
16 evidence from short-term exposure studies of fine particles, judged to be the strongest evidence
17 at that time, in reaching decisions to set a generally controlling annual PM25 standard and a 24-
18 hour PM2 5 standard to provide supplemental protection. The risk assessment conducted in the
19 last review provided qualitative insights, but was judged to be too limited to serve as a
20 quantitative basis for decisions on the standards. In this review, the more extensive and stronger
21 body of evidence now available on health effects related to both short- and long-term exposure
22 to PM25, together with the availability of much more extensive PM2 5 air quality data, have .
23 facilitated a more comprehensive risk assessment for PM25. As a result, staff has used'a broader
24 approach in this review of the PM2 5 standards that takes into account both evidence-based and
25 quantitative risk-based considerations, placing greater emphasis on evidence from long-term
26 exposure studies and quantitative risk assessment results for fine particles than was done in the
27 last review. Staff has applied this approach to a more limited degree in reviewing the PM10
28 standards, reflecting the far more limited nature of the health effects evidence and air quality
29 data available for thoracic coarse particles.
January 2005
5-2
Draft - Do Not Quote or Cite
-------
1 Staff has taken into account evidence-based considerations primarily by assessing the
2 epidemiologic evidence of associations with health endpoints that the CD has judged to be likely
3 causal based on an integrative synthesis of the entire body of evidence. Less weight is given to
4 evidence of associations that are judged to be only suggestive of possible causal relationships,
5 taking this information into account as part of margin of safety considerations. In so doing, staff
6 has placed greater weight on U.S. and Canadian studies reporting statistically significant
7 associations, providing relatively more precise effects estimates, using relatively more reliable
8 air quality data, and reporting associations that are generally robust to alternative model
9 specifications and the inclusion of potentially confounding co-pollutants. By considering the
10 ambient particle levels present during specific studies, staff has reached conclusions as to the
11 degree to which alternative standards could be expected to protect against the observed health
12 effects, while being mindful of the inherent limitations and uncertainties in such evidence.
13 Staff has also taken into account quantitative risk-based considerations, drawn from.the
14 results of the risk assessment conducted in several example urban areas (discussed in Chapter 4).
15 More specifically, staff has considered estimates of the magnitude of PM-related risks associated
16 with current air quality levels, as well as the risk reductions likely to be associated with attaining
17 the current or alternative standards. In so doing, staff recognizes the considerable uncertainties
18 inherent in such risk estimates, and has taken such uncertainties into account by considering the
19 sensitivity of the risk estimates to alternative assumptions likely to have substantial impact on
20 the estimates.
21 More specifically, in this review a series of questions frames staffs approach to reaching
22 conclusions and recommendations, based on the available evidence and information, as to
23 whether consideration should be given to retaining or revising the current primary PM standards.
24 Staffs review of the adequacy of the current standards begins by considering whether the
25 currently available body of evidence assessed in the CD suggests that revision of any of the basic
26 elements of the standards would be appropriate. This evaluation of the adequacy of the current
27 standards involves addressing questions such as the following:
28 • To what extent does newly available information reinforce or call into question evidence
29 of associations with effects identified in the last review?
January 2005
5-3
Draft - Do Not Quote or Cite
-------
1 • To what extent does newly available information reinforce or call into question any of the
2 basic elements of the current standards?
3 • To what extent have important uncertainties identified in the last review been reduced
4 and have new uncertainties emerged?
5 To the extent that the evidence suggests that revision of the current standards would be
6 appropriate, staff then considers whether the currently available body of evidence supports
7 consideration of standards that are either more or less protective by addressing the following •
8 questions:
9 • Is there evidence that associations, especially likely causal associations, extend to air
10 quality levels that are as low as or lower than had previously been observed, and what are
11 • the important uncertainties associated with that evidence?
12 • Are health risks estimated to occur in areas that meet the current standards; are they
13 important from a public health perspective; and what are the important uncertainties
14 associated with the estimated risks?
15 To the extent that there is support for consideration of revised standards, staff then identifies
16 ranges of standards (in terms of indicators, averaging times, levels and forms) that would reflect
17 a range of alternative public health policy judgments, based on the currently available evidence,
18 as to the degree of protection that is requisite to .protect public health with an adequate margin of
19 safety. In so doing, staff addresses the following questions:
20 • Does the evidence provide support for considering different PM indicators?
21 • Does the evidence provide support for considering different averaging times?
22 • What range of levels and forms of alternative standards is supported by the evidence, and
23 what are the uncertainties and limitations in that evidence?
24 • To what extent do specific levels and forms of alternative standards reduce the estimated
25 risks attributable to PM, and what are the uncertainties in the estimated risk reductions?
26s Based on the evidence, estimated risk reductions, and related uncertainties, staff makes'
27 recommendations as to ranges of alternative standards for the Administrator's consideration in
28 reaching decisions as to whether to retain or revise the primary PM NAAQS. •'
January 2005 5-4 Draft - Do Not Quote or Cite
-------
1 Standards for fine particles are addressed in section 5.3 below, beginning with staffs
2 consideration of the adequacy of the current primary PM25 standards. Subsequent subsections
3 address each of the major elements that define specific PM standards: pollutant indicator,
4 averaging time, level and form. Staff has evaluated separately the protection that a suite of PM25
5 standards would likely provide against effects associated with long-term exposures (section
6 5.3.4) and^those associated with short-term exposures (section 5.3.5). These separate evaluations
7 provide the basis for integrated recommendations on alternative suites of standards that protect
8 against effects associated with both long- and short-term exposures, based on considering how a
9 suite of standards operate together to protect public health. In a similar manner, standards for
10 thoracic coarse particles are addressed in section 5.4 below. This chapter concludes with a
11 summary of key uncertainties associated with establishing primary PM standards and related
12 staff research recommendations in section 5.5.
13 5.3 FINE PARTICLE STANDARDS
14 5.3.1 Adequacy of Current PMZ s Standards
15 In considering the adequacy of the current PM25 standards, staff has first considered the
16 extent to which newly available information reinforces or calls into question evidence of
17 associations with effects identified in the last review, as well as considering the extent to which
18 important uncertainties have been reduced or have resurfaced as being more important than
19 previously understood. In looking across the extensive epidemiologic evidence available in this
20 review, the CD addresses these questions by concluding that "the available findings demonstrate
21 well that human health outcomes are associated with ambient PM" (CD, p. 9-24) and, more
22 specifically, that there is now "strong epidemiological evidence" for PM2 5 linking short-term
23 exposures with cardiovascular and respiratory mortality and morbidity, and long-term exposures
24 with cardiovascular and lung cancer mortality and respiratory morbidity (CD, p. 9-46). This
25 latter conclusion reflects greater strength in the epidemiologic evidence specifically linking
26 PM2 5 and various health endpoints than was observed in the last review, when the CD concluded
27 that the epidemiologic evidence for PM-related effects was "fairly strong," noting that the
January 2005
5-5
Draft - Do Not Quote or Cite
-------
t
1 studies "nonetheless provide ample reason to be concerned" about health effects attributable to
2 PM at levels below the then-current PM NAAQS (EPA, 1996, p. 13-92).
3 As discussed in Chapter 3 (section 3.5) and the CD (section 9.2.2), the CD concludes that
4 the extensive body of epidemiologic evidence now available continues to support likely causal
5 associations between PM25 and the above health outcomes based on an assessment of strength,
6 robustness, and consistency in results. The CD finds "substantial strength" in the evidence of
7 PM23 associations, especially for total and cardiovascular mortality (CD, p. 9-28). The CD
8 recognizes that while the relative risk estimates are generally small in magnitude, a number of
9 new studies provide more precise estimates that are generally positive and often statistically
10 significant.' Overall, the CD firids the new evidence substantiates that the associations are
11 generally robust to confounding by co-pollutants, noting that much progress has been made in
12 sorting out contributions to observed health effects of PM and its components relative to other
13 co-pollutants. On the other hand, the CD notes that effect estimates are generally more sensitive
14 than previously recognized to different modeling strategies to adjust for temporal trends and
15 weather variables. While some studies showed little sensitivity, different modeling strategies
16 . altered conclusions in other studies.
17 Although greater variability in effects estimates across study locations is seen in the
18 much larger set of studies now available, in particular in the new multi-city studies, the CD finds
19 i much consistency in the epidemiologic evidence particularly in studies with the most precision.
20 There also are persuasive reasons why variation in associations in different locations could be
21 expected. Further, the CD concludes that new source apportionment studies and "found
22 experiments," showing improvements in community health resulting from reductions in PM and
23 other air pollutants, lend additional support to the results of other studies that focused
24 specifically on'PM2 5.
25 Looking more broadly to integrate epidemiologic evidence with that from exposure-
26 related, dosimetric and toxicologic studies, the CD (section 9.2.3) considered the coherence of
27 the evidence and-the extent to which the new evidence provides insights into mechanisms by
28 which PM, especially fine particles, may be affecting human health. Progress made in gaining
29 insights into mechanisms lends support to the biological plausibility of results observed in
January 2005 5 -6 Draft - Do Not Quote or Cite
-------
1 epidemiologic studies. For cardiovascular effects, the CD finds that the convergence of
2 important new epidemiologic and toxicologic evidence builds support for the plausibility of
3 associations especially between fine particles and physiological endpoints indicative of increased
4 risk of cardiovascular disease and changes in cardiac rhythm. This finding is supported by new
5 cardiovascular effects research focused on fine particles that has notably advanced our
6 understanding of potential mechanisms by which PM exposure, especially in susceptible
7 individuals, could result in changes in cardiac function or blood characteristics that are risk
8 factors for cardiovascular disease. For respiratory effects, the CD finds that toxicologic studies
9 have provided evidence that supports plausible biological pathways for fine particles, including
10 inflammatory responses, increased airway responsiveness, or altered responses to infectious
11 , agents. Further, the CD finds coherence across a broad range of cardiovascular and respiratory
12 , health outcomes from epidemiologic and toxicologic studies done in the same location,
13 particularly noting, for example, the series of studies conducted in or evaluating ambient PM
14 from Boston and the Utah Valley. The CD also finds that toxicologic evidence examining
15 combustion-related particles supports the plausibility of the observed relationship between fine
16 particles and lung cancer mortality. With regard to PM-related infant mortality and
17 developmental effects, the CD finds this to be an emerging area of concern, but notes that current
18 information is still very limited in support of the plausibility of potential ambient PM
19 relationships.
20 Based on the above considerations and findings from the CD, staff concludes that the
21 newly available information generally reinforces the associations between PM2 5 and mortality
22 and morbidity effects observed in the last review. Staff recognizes that important uncertainties
23 and research questions remain, notably including questions regarding modeling strategies to
24 adjust for temporal trends and weather variables in time-series epidemiologic studies.
25 Nonetheless, staff notes mat progress has been made in reducing some key uncertainties since
26 the last review, including important progress in advancing our understanding of potential
27 mechanisms by which ambient PM2 5, alone and in combination with other pollutants, is causally
28 linked with cardiovascular, respiratory, and lung cancer associations observed in epidemiologic
January 2005 5-7 . Draft - Do Not Quote or Cite
t
-------
1 studies. Thus, staff finds clear support in the available evidence, as assessed in the CD, for fine
2 particle standards that are at least as protective as the current PM2 5 standards.
3 Having reached this initial conclusion, staff also has addressed the question of whether
4 the available evidence supports consideration of standards that are more protective than the
5 current PM25 standards. In so doing, staff has considered first whether there is evidence that
6 health effects associations with short- and long-term exposures to fine particles extend to lower
7 air quality levels than had previously been observed, or to levels below the current standards. In
8 addressing this question, staff first notes that the available evidence does not either support or
9 refute the existence of thresholds for the effects of PM on mortality across the range of
10 concentrations in the studies, as discussed in Chapter 3 (section 3.4.6) and the CD (section
11 9.2.2.5). More specifically, while there are likely threshold levels for individuals and specific
12 health responses, existing studies show little evidence for thresholds for PM-mortality
13 relationships in populations, for either long-term or short-term PM exposures (CD, p. 9-44).
14 Further, the CD notes that in the multi-city and most single-city studies, statistical tests
15 comparing linear and various nonlinear or threshold models have not shown statistically
16 significant distinctions between them (CD, p. 9-44). Even in those few studies with suggestive
17 evidence for thresholds, the potential thresholds are at fairly low concentrations (CD, p. 9-45).
18 While acknowledging that for some health endpoints, such as total nonaccidental mortality, it is
19 likely to be extremely difficult to detect thresholds, the CD concludes that "epidemiblogic
20 studies suggest no evidence for clear thresholds in PM-mortality relationships within the range
21 of ambient PM concentrations observed in these studies." (CD, p. 9-48).
22 In considering the available epidemiologic evidence (summarized in Chapter 3, section
23 3.3 and Appendices 3 A and 3B), staff has focused on specific epidemiologic studies that show
24 statistically significant associations between PMZ5 and health effects for which the CD judges
25 associations with PM25 to be likely causal. Many more U.S. and Canadian studies are now
26 available in the current review that provide evidence of associations between PM?5 and serious
27 health effects in areas with air quality at and above the level of the current annual PM25 standard
28 (15 ug/m3), which was set to provide protection against health effects related to both short- and
29 long-term exposures to fine particles. Notably, a few of the newly available short-term exposure
January 2005 5-8 . Draft - Do Not Quote or Cite
-------
1 mortality studies provide evidence of statistically significant associations with PM25 in areas
2 with long-term average air quality below that ambient level (summarized in Appendix 3A). In
3 considering this group of studies, staff has focused on those studies that include adequate
4 gravimetric PM2 5.mass measurements, and where the associations are generally robust to
5 alternative model specification and to the inclusion of potentially confounding co-pollutants.
6 Three such studies conducted in Phoenix (Mar et al., 1999,2003), Santa Clara County, CA
7 (Fairley, 1999,2003) and eight Canadian cities (Burnett et al., 2000 and Burnett and Goldberg,
8 2003) report statistically significant associations between short-term PM2 5 exposure and total
9 and cardiovascular mortality in areas in which long-term average PM2 5 concentrations ranged
10 between 13 and 14 ug/m3. These studies were reanalyzed to address questions about the use of
11 GAM with default convergence criteria, and the study results from Phoenix and Santa Clara
12 County were little changed in alternative models (Mar et al., 2003; Fairley, 2003), although
13 Burnett and Goldberg (2003) reported that their results were sensitive to using different temporal
14 smoothing methods.
15 Beyond .these mortality studies, other studies provide evidence of statistically significant
16 associations with morbidity. Three studies of emergency department visits were conducted in
17 areas where the mean PM25 concentrations were approximately 12 ng/m3 or below, although
18 these studies either had not been reanalyzed to address the default convergence criteria problem
19 with GAM, did not assess the potential for confounding by co-pollutants or were not robust to
20 the inclusion of co-pollutants, or were done only during a single season. Another new study
21 reported statistically significant associations with incidence of myocardial infarction where the
22 mean PM2.5 concentration was just above 12 ug/m3; however, the CD urges caution in
23 interpreting the results of the new body of evidence related to such cardiovascular effects (CD, p.
24 8-166). Thus, these studies provide no clear evidence of statistically significant associations
25 with PM2 5 at such low concentrations.
26 New evidence is also available from U.S. and Canadian studies of long-term exposure to
27 fine particles (summarized in Appendix 3B). In evaluating this evidence (CD, section 9.2.3), the
28 CD notes that new studies have built upon studies available in the last review and these studies
29 have confirmed and strengthened the evidence of associations for both mortality and respiratory
January 2005 5 -9 Draft - Do Not Quote or Cite
-------
1 morbidity. For mortality, the CD places greatest weight on the reanalyses and extensions of the
2 Six Cities and the ACS studies, finding that these studies provide "strong evidence" for
3 associations with fine particles (CD, p. 9-34), notwithstanding the lack of consistent results in
4 other long-term exposure studies. For morbidity, the CD finds that new studies of a cohort of
5 children in Southern California have built upon earlier limited evidence to provide "fairly
6 strong" evidence that long-term exposure to fine particles is associated with development of
7 chronic respiratory disease and reduced lung function growth (CD, p. 9-34). ;
8 As discussed in the CD and in Chapter 3 above, mortality studies of the Six Cities and
9 ACS cohorts available in the last review had aggregate long-term mean PM2 5 concentrations of
10 18 ug/m3 (ranging from approximately 11 to 30 ug/m3 across cities) and 21 ug/m3 (ranging from
11 approximately 9 to 34 ug/m3 across cities), respectively. Reanalyses of data from these cohorts
12 continued to report significant associations with PM25, using essentially the same air quality
13 distributions. The extended analyses using the ACS cohort also continued to report statistically
14 significant associations with PM2 5 with the inclusion of more recent PM2_s air quality data, with
15 an average range across the old and new time periods from about 7.5 to 30 ug/m3 (from figure 1,
16 Pope et al., 2002) with a long-term mean of approximately 17.7 ug/m3 (Pope et al., 2002). As
17 with the earlier cohort studies, no evidence of a threshold was observed in the relationships with
18 total, cardiovascular, and lung cancer mortality reported in this extended study. In the morbidity
19 studies of the Southern California children's cohort, the means of 2-week average PM25
20 concentrations ranged from approximately 7 to 32 ug/m3, with an across-city average of
21 approximately 15 ug/m3 (Peters et al., 1999). Staff notes that in figures depicting relationships
22 between lung function growth and average PM concentration, there is no evidence of a threshold
23 in this study (Gauderman et al., 2000,2002).
24 Beyond the epidemiologic studies using PM2 5 as an indicator of fine particles, a large
25 body of newly available evidence from studies that used PM10, as well as other indicators or
26 components of fine particles (e.g.,' sulfates, combustion-related components), provides additional
27 support for the conclusions reached in the last review as to the likely causal role of ambient PM,
28 and the likely importance of fine particles in contributing to observed health effects. Such
29 studies notably include new multi-city studies, intervention studies (that relate reductions in
January 2005 5-10 Draft - Do Not Quote or Cite
-------
1 ambient PM to observed improvements in respiratory or cardiovascular health), and source-
2 oriented studies (e.g., suggesting associations with combustion- and vehicle-related sources of
3 fine particles). Further, the CD concludes that new epidemiologic studies of ambient PM
4 associations with potential PM-related infant mortality and/or developmental effects are very
5 limited, although if further substantiated by future research, would significantly increase
6 estimates of the extent of life shortening due to PM-related premature mortality (CD, p, 9-94).
7 The CD also notes that new epidemiologic studies of asthma-related increased physicians visits
8 and symptoms, as well as new studies of cardiac-related risk factors, suggest likely much larger
9 public health impacts due to ambient fine particles than just those indexed by the
10 mortality/morbidity effects considered in the last review (CD, p. 9-94).
11 Staff recognizes, however, mat important limitations and uncertainties associated with
12 this expanded body of evidence for PM2 5 and other indicators or components of fine particles, as
13 discussed in Chapter 3 herein and section 9.2.2 of the CD, need to be carefully considered in
14 determining the weight to be placed on the studies available in this review. For example, the CD
15 notes that while PM-effects associations continue to be observed across most new studies, the
16 newer findings do not fully resolve the extent to which the associations are properly attributed to
17 PM acting alone or in combination with other gaseous co-pollutants, or to the gaseous co-
18 pollutants themselves. The CD notes that available statistical methods for assessing potential
19 confounding by gaseous co-pollutants may not yet be fully adequate, although the various
20 approaches that have now been used to evaluate this issue tend to substantiate that associations
21 for various PM indicators with mortality and morbidity are robust to confounding by co-
22 pollutants (CD, p. 9-37).
23 Another issue of particular importance is the sensitivity of various statistical models to
24 the approach used to address potential confounding by weather- and time-related variables in
25 time-series epidemiological studies. As discussed in section 3.5.3 herein and in section 9.2.2 of
26 the CD, this issue resurfaced in the course of reanalyses of a number of the newer studies that
27 were being conducted to address a more narrow issue related to problems associated with the use
28 of commonly used statistical software. These reanalyses suggest that weather continues to be a
29 potential confounder of concern and highlight that no one model is likely to be most appropriate
January 2005 .. 5-11 Draft - Do Not Quote or Cite
-------
1 in all cases. The HEI Review Panel, in reviewing these reanalyses, concluded that this
2 awareness introduces a degree of uncertainty in evaluating the findings from time-series
3 epidemiological studies that had heretofore not been widely appreciated.
4 ' In looking beyond PM mass indicators, a number of newly available studies highlight the
5 issue of the extent to which observed health effects may be associated with various specific
6 chemical components within the mix of fine particles. The potential for various fine particle
7 components to have differing relative toxicities with regard to the various health endpoints being
8 considered adds complexity to the interpretation of the study results. The CD recognizes that
9 more research is needed to address uncertainties about the extent to which various components
10 may be relatively more or less toxic than'others, or than undifferentiated PM2 5 mass across the
11 range of health endpoints studied.
12 While the limitations and uncertainties in the available evidence suggest caution in
13 interpreting the epidemiologic studies at the lower levels'of air quality observed in the studies,
\
14 staff concludes that the evidence now available provides strong support for considering fine
15 particle standards that would provide increased protection from that afforded by. the current
* " . .
16 PM2 5 standards. More protective standards would reflect the generally stronger and broader .
17 body of evidence of associations with mortality and morbidity now available in this review, at
18 lower levels of air quality and at levels below the current standards, and with more
19 understanding of possible underlying mechanisms.
20 In addition to this evidence-based evaluation, staff has also considered the extent to
21 which health risks estimated to occur upon attainment of the current PM25 standards may be
22 judged to be important from a public health perspective, taking into account key uncertainties
23 associated with the estimated risks. Based on the risk assessment presented in Chapter 4, staff
24 considered as a base case the estimated risks attributable to PM2 5 concentrations above
25 background levels, or above the lowest measured levels in a given study if that was higher than
26 background, so as to avoid extrapolating risk estimates beyond the range of air quality upon
27 which the concentration-response functions were based. In the case of estimated risk associated
28 with long-term exposure, based on the extended ACS study, risk was estimated down to an
29 annual level of 7.5 ug/m3, the lowest measured level in that study; for estimated risk associated
January 2005 5-12 Draft - Do Not Quote or Cite
-------
1 with short-term exposure, risk was estimated down to daily levels ranging from 2.5 to 4 ug/m3,
2 based on estimated background or the lowest measured level in a particular study.
3 In the absence of evidence for clear thresholds in any of the studies used in this risk
4 assessment, the base case estimates in this analysis reflect the linear or near-linear concentration-
5 response functions reported in the studies. To reflect the uncertainty as to whether thresholds
6 may exist within the range of air quality observed in the studies, but may not be discernable with
7 currently applied statistical methods, staff also has considered estimates of risk based on
8 concentration-response functions modified to incorporate various assumed hypothetical
9 threshold levels, as discussed in Chapter 4. Based on the sensitivity analyses conducted as part
10 of the risk assessment, the uncertainly associated with alternative hypothetical thresholds had by
11 • far the greatest impact on estimated risks. Other uncertainties have a more moderate and often
12 variable impact on the risk estimates in some or all of the cities, including the use of single-
13 versus multi-pollutant models, single- versus multi-city models, use of a distributed lag model,
14 alternative assumptions about the relevant air quality for long-term exposure mortality, and
15 alternative constant or varying background levels. v
16 - Table 5-1 summarizes the estimated PM25-related annual incidence of total mortality
17 associated with long- and short-term exposure for the base case and for alternative hypothetical
18 thresholds in the nine example urban areas included in the risk assessment. In looking
19 particularly at the annual incidence of PM2 5-related mortality estimated to occur upon attainment
20 of the current PM25 standards in the five study areas that do not meet the current standards based
21 on 2001 -2003 air quality data, staff notes that there is a fairly wide range of estimated incidence
22 across the areas. Such variation would be expected considering, for example, differences in total
23 population, demographics, exposure considerations (e.g., degree of air conditioning use),
24 presence of co-pollutants and other environmental stressors, and exposure measurement error
25 across urban areas; as well as differences in concentration-response relationships across studies
26 that might be due in part to variation in these factors across locations. Staff also recognizes that
27 mere are uncertainties associated with the procedure used to simulate air quality that would just
28 attain the current standards and in the degree to which various components of the fine particle
29 mix would likely be reduced in similar proportion to the simulated reduction in PM25 as a whole.
January 2005 5-13 Draft - Do Not Quote or Cite
-------
X
1
2
3
4
5
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
Table 5-1 Estimated PM2 s-related Annual Incidence of Total Mortality when Current
PM2 5 Standards are Met (Base Case and Assumed Alternative Hypothetical
Thresholds)*
Short4erm Exposure:
Annual Non-Accidental Mortality
(except as noted)
Base case
Estimate,
95% Cl
Assumed Hypothetical Short-term
Exposure Thresholds
10(jg/m3
15pg/m*
20 jjg/m3
Long-term Exposure:
Annual All-Cause Mortality
Base case
Estimate,
95% Cl
Assumed Hypothetical
Long-term Exposure
Thresholds
10 ug/m3
12 |jg/m3
Risks associated with just meeting current PMts standards
Detroit
Los Angeles
Philadelphia
{short-term: cardiovascular
mortality)
Pittsburgh .
(short-term: over age 74)
St. Louis
115
-116 to 338
248
-31 to 51 9
367
175 to 560
50
-108 to 200
191
70 to 311
54
-55 to 159
115
"-14 to 240
189
90 to 288
. 22
-48 to 87
75
28 to 122
26
-27 to 77
58
-7 to 121
106 -
51 to 162
10
-23 to 41
29
11 to46
12
-12 to 35
29
-4 to 61
57
27 to 87
5
-11 to 18
9
3 to 14
522
181 to 910
1,507 .
531 to 2,587
536
185 to 943
403
141 to 699
596
206 to 1,047
282
98 to 494
823
290 to 1415
' 338
116 to 597
215
75 to 373
311
107 to 548
41
14 to 72
138
48 to 237
137
47 to 244
25
9 to 43
23
8 to 40
Risks associated with "as is "air quality (in areas that moot current PM2S standards)
Boston
Phoenix
(short-term: cardiovascular
mortality over age 64)
San.Jose
Seattle"
390
265 to 514
323
97 to 536
218
45 to 387
-
173
118 to 228
115
35 to 190
80
17 to 141
82
56 to 109
67
21 to 109
44
9 to 77
41
28 to 53
43
13 to 69
28
6 to 50
594
204 to 1053
349
119 to 620
172
59 to 306
50
17 to 89
309
106 to 551
76
26 to 136
58
20 to 104
0
OtoO
20
7to36
0
OtoO
0
OtoO •
0
OtoO
* These estimates of annual incidence of PM2.5-related mortality are based on using the maximum monitor in an area
to calculate the percent rollback needed to just attain the current PM^s annual standard, and applying that percent
rollback to the composite monitor in the area, as described in Chapter 4, section 4.2.3. Estimates of annual mortality
incidence based on using a spatially averaged concentration to calculate the percent rollback needed to just attain the
current standard, where this is allowed, would be higher than the estimates shown here.
** No short-term exposure concentration-response function is available for mortality in Seattle.
26 Staff observes that base case point estimates of annual incidence of total PMZ5-related
27 mortality associated with just meeting the current PM25 standards in the five areas shown range
January 2005
5-14
Draft - Do Not Quote or Cite
-------
1 from approximately 400 to 600 in four areas (or from roughly 25 to 35 deaths per 100,000
2 population in these areas) to over 1500 annual deaths in Los Angeles (i.e., roughly 16 deaths per
3 100,000 population) associated with long-term exposure. These estimated incidences associated
4 with long-term exposure represent 2.6 to 3.2 percent of total mortality incidence due to all
5 causes. Expressing the risk estimates in terms of percentage of total incidence takes into
6 account city-to-city differences in population size and baseline mortality incidence rate. In some
7 areas, the 95% confidence ranges associated with the estimates of total annual mortality
8 incidence related to short-term exposure (but not long-term exposure) extend to below zero,
9 reflecting appreciably more uncertainty in estimates based on positive but not statistically
10 significant associations. In the other four areas mat meet the current standards based on recent
11 air quality data, base case point estimates of annual incidence of total PM2 5-related mortality
12 associated with long-term exposure range from a lower end of about 50 deaths in Seattle (which
13 represents a rate of about 3 per 100,000 population) to an upper end of almost 600 deaths in
14 Boston (a rate of 21 per 100,000 population). It is much more difficult to make comparisons
15 among the urban areas with regard to short-term exposure mortality incidence or incidence rates
16 because of the different population groups and mortality types examined in the epidemiology
17 studies'for the different locations. There also is greater variability in the estimates for mortality
18 associated with short-term exposure due to the use of different city-specific concentration-
19 response relationships.
20 In looking beyond the base case estimates, staff also considered the extent to which the
21 assumption of the presence of hypothetical thresholds in the concentration-response relationships
22 would influence the risk estimates. As expected, risk estimates are substantially smaller when
23 hypothetical threshold concentration-response functions are considered. Point estimates of
24 annual incidence of total PM25-related mortality associated with long-term exposure are roughly
25 50% of base case estimates when a hypothetical threshold of 10 ug/m3 is assumed, whereas when
26 a hypothetical threshold of 12 ug/m3 is assumed, point estimates are roughly 5 to 20% of base
27 case estimates in nonattainment areas (and even smaller in attainment areas). A similar pattern is
28 seen when considering the impact of alternative hypothetical thresholds in the range of 10 to 20
29 ng/m3 on risks associated with short-term exposure.
January 2005 5-15 , Draft-DoNot Quote or Cite
-------
1 In considering these estimates of PM2 5-related incidence of annual total mortality upon
2 . meeting the current standards in a number of example urban areas, together with the
3 uncertainties in these estimates, staff concludes that they are indicative of risks that can
4 reasonably be judged to be important from a public health perspective and provide support for
5 consideration of standards mat would provide increased protection from that afforded by the
6 current PM2 5 standards. In the absence of evidence of clear thresholds, staff believes that it is
7 appropriate to give most weight to the base case risk estimates. These estimates indicate the
8 likelihood of thousands of premature deaths per year in urban areas across the U.S. Beyond the
9 estimated incidences of mortality discussed above, staff also recognizes that similarly substantial
10 numbers of incidences of hospital admissions, emergency room visits, aggravation of asthma and
11 other respiratory symptoms, and increased cardiac-related risk are also likely in many urban
12 areas, based on risk assessment results presented in Chapter 4 and on the discussion related to •
13 the pyramid of effects drawn from section 9.2.5 of the CD. Staff also believes that it is important
14 to recognize how highly dependent these risk estimates are on the shape of the underlying
15 concentration-response functions. In so doing, staff nonetheless notes that in considering even
16 the largest assumed hypothetical thresholds, estimated mortality risks are not completely
17 eliminated when current PM2S standards are met in a number of example urban areas, including
18 all such areas that do not meet the standards based on recent air quality.
19 Staff well recognizes that as the body of available evidence has expanded, it has added
20 greatly both to our knowledge of PM-related effects, as well as to the complexity inherent in
21 interpreting the evidence in a policy-relevant context as a basis for setting appropriate standards.
22 In considering available evidence, risk estimates, and related limitations and uncertainties, staff
23 concludes that the available information clearly calls into question the adequacy of the current
24 suite of PM2 5 standards', and provides strong support for giving consideration to revising the
25 current PM2 5 standards to provide increased public health protection. Staff conclusions and
26 recommendations for indicators, averaging times, and levels and forms of alternative, more
27 protective primary standards for fine particles are discussed in the following sections.
January 2005' 5-16 Draft - Do Not Quote or Cite
-------
t
1 5.3.2 Indicators
2 " In 1997, EPA established PM25 as the indicator for fine particles. In reaching this
3 decision, the Agency first considered whether the indicator should be based on the mass of a
4 size-differentiated sample of fine particles or on one or more components within the mix of fine
5 particles. Secondly, in establishing a size-based indicator, a size cut point needed to be selected
6 that would appropriately distinguish fine particles from particles in the coarse mode.
7 In addressing the first question in the last review, EPA determined that it was more
8 appropriate to control fine particles as a group, as opposed to singling out any particular
9 component or class of fine particles based on the following considerations. Community health
10 studies had found significant associations between various indicators of fine particles (including
11 PM25 or PM10 in areas dominated by fine particles) and health effects in areas with significant
12 mass contributions of differing components or sources of fine particles, including sulfates, wood
13 smoke, nitrates, secondary organic compounds and acid sulfate aerosols. In addition, a number
14 of animal toxicologic and controlled human exposure studies had reported health effects
15 associations with high concentrations of numerous fine particle components (e.g., sulfates,
16 nitrates, transition metals, organic compounds), although such associations were not consistently
17 observed. It also was not possible to rule out any component within the mix of fine particles as
18 not contributing to the fine particle effects found in epidemiologic studies. Thus, it was
19 determined that total mass of fine particles was the most appropriate indicator for fine particle
20 standards rather than an indicator based on PM composition (62 FR 38667, July 18,1997).
21 Having selected a size-based indicator for fine particles, the Agency then based its
22 selection of a specific cut point on a number of considerations. In focusing on a cut point within
23 the size range of 1 to 3 \im (i.e., the intermodal range between fine and coarse mode particles),
24 EPA recognized that the choice of any specific sampling cut point within this range was largely a
25 policy judgment. In making this judgment, the Agency noted that the available epidemiologic
26 studies of fine particles were based largely on PM25; only very limited use of PM] monitors had
27 been made. While it was recognized that using PMj as an indicator of fine particles would
28 exclude the tail of the coarse mode in some locations, in other locations it would miss a portion
29 of the fine PM, especially under high humidity conditions, which would result in falsely low fine
January 2005
5-17
Draft - Do Not Quote or Cite
-------
1 PM measurements on days with some of the highest fine PM concentrations. The selection of a
2 2.5 um cut point reflected the regulatory importance that was placed on defining an indicator for
3 fine particle standards that would more completely capture fine particles under all conditions
4 likely to be encountered across the U.S., especially when fine particle concentrations are likely
5 to be high, while recognizing that some small coarse particles would also be captured by PM2 5
6 monitoring.2 Thus, EPA's selection of 2.5 um as the cut point for the fine particle indicator was
7 based on considerations of consistency with the epidemiologic studies, the regulatory importance
8 of more completely capturing fine particles under all conditions, and the limited potential for
9 intrusion of coarse particles in some areas; it also took into account the general availability of
10 monitoring technology (62 FR 3 8668).
11 In this current review, staff observes that the same considerations apply for selection of
12 an appropriate indicator for fine particles. As an initial matter, staff notes that the available
13 epidemiologic studies linking mortality and morbidity effects with short- and long-term
14 exposures to fine particles continue to be largely indexed by PM2 5. Some epidemiologic studies
15 also have continued to implicate various PM components (e.g., sulfates, nitrates, carbon, organic
16 compounds, and metals) as being associated with adverse effects; effects have been reported
17 with a broad range of PM components, as summarized in Table 9-13 of the CD (p. 9-31).
18 Animal toxicologic and controlled human exposure studies, evaluated in Chapter 7 of the CD,
19 have continued to link a variety of PM components or particle types (e.g., sulfates or acid
20 aerosols, metals, organic constituents, bioaerosols, diesei particles) with health effects, though
21 often at high concentrations (CD section 7.10.2). In addition, some recent studies have
22 suggested that the ultrafine subset of fine particles may also be associated with adverse effects
23 (CD, pp. 8-66,8-199). . '
24 . Staff recognizes that, for a given health response, some PM components are likely to be
25 more closely linked with that response than others (CD, p. 9-30). That different PM constituents
26 may have differing biological responses is an important source of uncertainty in interpreting
27 epidemiologic evidence. For specific effects there may be stronger correlation with individual
2 In reaching this decision, EPA indicated that it might be appropriate to address undue intrusion of coarse
mode particles resulting in violations of PM2 5 standards in the context of policies established to implement such
standards (62 FR 38668).
January 2005 5-18 Draft - Do Not Quote or Cite
-------
1 PM components than with particle mass. For example, in some toxicologic studies of
2 cardiovascular effects, such as changes in heart rate, electrocardiogram measures, or increases in
3 arrhythmia, PM exposures of equal mass did not produce the same effects, indicating that PM
4 composition was important (CD, p. 7-30). In addition, section 9.2.3.1.3 of the CD indicates that
5 particles, or particle-bound water, can act as carriers to deliver other toxic agents into the
6 respiratory tract, highlighting the fact that exposure to particles may elicit effects that are linked
7 with a mixture of components more than with any individual PM component
8 Thus, epidemiologic and toxicologic studies summarized above and discussed in the CD
9 have provided evidence for effects associated with various fine particle components or size-
10 differentiated subsets of fine particles. The CD concludes: "These studies suggest that many
11 different chemical components of fine particles and a variety of different types of source
12 categories are all associated with, and probably contribute to, mortality, either independently or
13 in combinations" (CD, p. 9-31). Conversely, the CD provides no basis to conclude that any
14 individual fine particle component cannot be associated with adverse health effects. There is no
15 evidence that would lead toward the selection of one or more PM components as being primarily
16 responsible for effects associated with fine particles, nor is there any component that can be
17 eliminated from consideration. Staff continues to recognize the importance of an indicator that
18 not only captures all of the most harmful components of fine PM (i.e., an effective indicator), but
19 also places greater emphasis for control on those constituents or fractions, including most
20 sulfates, acids, transition metals, organics, and ultrafine particles, that are most likely to result in
21 the largest risk reduction (i.e., an efficient indicator). Taking into account the above
22 considerations, staff concludes that it remains appropriate to control fine particles as a group;
23 i.e., that total mass of fine particles is the most appropriate indicator for fine particle standards.
24 With regard to an appropriate cut point for a size-based indicator of total fine particle
25 mass, the CD most generally concludes that advances in our understanding of the characteristics
26 of fine particles continue to support the use of particle size as an appropriate basis for
27 distinguishing between these subclasses, and that a nominal cut point of 2.5 um remains
28 appropriate (CD, p. 9-22). This conclusion follows from a recognition that within the intermodal
29 range of 1 to. 3 jim there is no unambiguous definition of an appropriate cut point for the
January 2005 5-19 Draft - Do Not Quote or Cite
-------
1 separation of the overlapping fine and coarse particle modes (CD, p. 9-8). Within this range,
2 staff considered cut points of both 1 jirri and 2.5 jim. Consideration of these two cut points took
3 into account that there is generally very little mass in this intermodal range, although in some
4 circumstances (e.g., windy, dusty areas) the coarse mode can extend down to and below 1 um,
5 whereas in other circumstances (e.g., high humidity conditions, usually associated with very high
6 fine particle concentrations) the fine mode can extend up to and above 2.5 um The same
7 considerations that led to the selection of a 2.5 \im cut point in the last review - that the
8 epidemiologic evidence was largely based on PM2 5 and that it was more important from a
9 regulatory perspective to more completely capture fine particles under all conditions likely to be
10 encountered across the U.S. (especially when fine particle concentrations are likely to behigh)
11 than to avoid some coarse-mode intrusion into the fine fraction in some areas - also lead to the
12 same conclusion in this review. In addition, section 9.2.1.2.3. of the CD discusses the potential
13 health significance of particles as carriers of water, oxidative compounds, and other components
14 into the respiratory system. This consideration adds to the importance of ensuring that larger
15 accumulation-mode particles are included in the fine particle size cut. Therefore, as observed
i\ 16 previously in section 3.1.2, the scientific evidence leads the CD to conclude that 2.5 uni remains
17 an appropriate upper cut point for a fine particle mass indicator.
18 . Thus, consistent with the CD's conclusion that 2.5 um remains an appropriate cut point
i
19 for including the larger accumulation-mode fine particles while limiting intrusion of coarse
20 particles, staff recommends that PM2 5 be retained as the indicator for fine particles. Staff further
21 concludes that currently available studies do not provide a sufficient basis for supplementing
22 mass-based fine particle standards with standards for any specific fine particle component or
23 subset of fine particles, or for eliminating any individual component or subset of components
24 from fine particle mass standards.
* i
25 Further, staff notes that since the' last review an extensive PM2 5 monitoring network has
26 been deployed and operated in cooperative efforts with State, local and Tribal agencies and with
27 instrument manufacturers. At the same time, EPA has been working on the development of
28 strategies and programs to implement the 1997 PM2.5 standards, based on the Federal Reference
29 Monitor (FRM).for PM2 5. The new monitoring network has provided substantial new air quality
January 2005 5-20 Draft - Do Not Quote or Cite
-------
1 potentially be used to provide information to the public based on episodic very short-term peak
2 fine particle levels that may be of public health concern.
3 In the last review, having decided to set both annual and 24-hour PM2 5 standards, EPA
4 also made judgments as to the most effective and efficient approach to establishing a suite of
5 standards that, taken together, would appropriately protect against effects associated with both
6 long- and short-term exposures. At that time, EPA selected an approach that was based on
7 treating the annual standard as the generally controlling standard for lowering the entire
8 distribution of PM2 5 concentrations, with the 24-hour standard providing additional protection
9 against the occurrence of peak 24-hour concentrations. The 24-hour standard was intended to
10 address in particular those peaks that result in localized or seasonal exposures of concern in areas
11 where the highest 24-hour-to-annual mean PM25 ratios are appreciably above the national
12 average. This approach was supported by results of the PM risk assessment from the last review
13 which indicated that peak 24-hour PM2 5 concentrations contribute a relatively small amount to
14 total health risk, such that much if not most of the aggregated annual risk associated with short-
15 term exposures results from the large number of days during which the 24-hour average •
16 ' concentrations are in the low- to mid-range. Further, no evidence suggested that risks associated
17 with long-term exposures are likely to be disproportionately driven by peak 24-hour
18 concentrations. Thus, a generally controlling annual standard was judged to reduce risks
19 associated with both short- and long-term exposures effectively and with more certainty than a
20 24-hour standard. Further, an annual standard was seen to be more stable over time, likely
21 resulting in the development of more consistent risk reduction strategies, since an area's
22 attainment status would be less likely to change due solely to year-to-year variations in
23 meteorological conditions that affect the atmospheric formation of fine particles.
24 In this review, staff recognizes that some key considerations that led to establishing a
25 generally controlling annual standard in the last review are still valid. In particular, staff
26 observes that:
27 • EPA's updated risk assessment supports the conclusion that peak 24-hour PM2 5
28 concentrations contribute a relatively small amount to the total health risk associated with
29 short-term exposures on an annual basis, such that much if not most of the aggregated
30 annual risk results from the large number of days during which the 24-hour average
31 concentrations are in the low- to mid-range, as discussed in Chapter 4 (section 4.3.3).
January 2005 , 5-23 Draft - Do Not Quote or Cite
-------
1 Support for this conclusion is also found in studies in which health effect associations
2 remain when high-concentration days are removed from the analysis (Schwartz et al,
3 " 1996; Ostroetal., 1999,2000).
4 • It continues to be the case, as discussed in section 4.2.6.1, that available short-term
5 exposure studies do not provide evidence of clear population thresholds, but rather reflect
6 relationships between health effects and ambient PM across a wide distribution of PM
7 concentrations. Thus, as in the last review, staff recognizes that these studies do not
8 provide a basis for identifying a lowest-observed-effect level that would clearly translate
9 into a 24-hour standard that would protect against all effects related to short-term
10 exposures.
11 Nonetheless, staff believes that the greatly expanded body of epidemiologic evidence and
12 air quality data provide the basis for considering alternative approaches to establishing a suite of
13 PM25 standards. Thus, staff has not focused a priori on an annual standard as the generally
14 controlling standard for protection against effects associated with both long- and short-term
15 exposures. Rather, staff has broadened its view to consider botii evidence-based and risk-based
16 approaches to evaluating the protection that a suite of PM2 5 standards can provide against effects
17 associated with long-term exposures arid against short-term exposures. These evaluations,
18 discussed in the next two sections, provide the basis for integrated recommendations on ranges
19 of alternative suites of standards that, when considered together, protect against effects
20 associated with both long- and short-term exposures.
21 5.3.4 Alternative PM2.5 Standards to Address Health Effects Related to Long-term
22 Exposure
23 In considering alternative PMj 5 standards that would provide protection against health
24 effects related to long-term exposures, staff has taken into account both evidence-based and risk-
25 based considerations. As discussed below in this section, staff has first evaluated the available
26 . evidence from long-term exposure studies, as well as the uncertainties and limitations in that
27 evidence, to assess the degree to which alternative annual PM2 5 standards can be expected to
28 provide protection against effects related to long-term exposures. Secondly, staff has considered
29 the quantitative risk estimates for long-term exposure effects, discussed in Chapter 4, to assess
30 the extent to which alternative annual and/or 24-hour standards can be expected to reduce the
January 2005
5-24
Draft - Do Not Quote or Cite
-------
1 estimated risks attributable to long-term-exposure to PM25. Staff conclusions as to ranges of
2 alternative annual and/or 24-hour standards that would provide protection against health effects
3 related to long-term exposures are summarized at the end of this section. The integrated staff
4 recommendations presented in section 5.3.7 are based in part on the conclusions from this
5 section and in part on staff conclusions from the next section, in which alternative PM2 5
6 standards to address health effects related to short-term exposures are assessed.
7 5.3.4.1 Evidence-based Considerations
8 In taking into account evidence-based considerations, staff has focused on long-term
9 exposure studies of fine particles in the U.S. As discussed above, staff notes that the reanalyses
10 and extensions of earlier studies have confirmed and strengthened the evidence of long-term
11 associations for both mortality and morbidity effects. The assessment in the CD of these
12 mortality studies, taking into account study design, the strength of the study (in terms of
13 statistical significance and precision of result), and the consistency and robustness of results,
14 concluded that it was appropriate to give the greatest weight to the reanalyses of the Six Cities
15 study and the ACS study, and in particular to the results of the extended ACS study (CD, p.
16 9-33). The assessment in the CD of the relevant morbidity studies noted in particular the results
17 of the new studies of the children's cohort in Southern California as providing evidence of
18 respiratory morbidity with long-term PM exposures.
19 Staff believes it is appropriate to consider a level for an annual PM2} standard that is
20 somewhat below the averages of the long-term concentrations across the cities in each of these
21 studies, recognizing that the evidence of an association in any such study is strongest at and
22 around the long-term average where the data in the study are most concentrated. For example,
23 the interquartile range of long-term average concentrations within a study, or a range within one
24 standard deviation around the study mean, might be used to characterize the range over which
25 the evidence of association is strongest. Staff also believes it is appropriate to consider the long-
26 term average concentration at the point where the confidence interval becomes notably wider,
27 suggestive of a concentration below which the association becomes appreciably more uncertain
28 and the possibility that an effects threshold may exist becomes more likely. Staff further notes
29 that in considering a level for a standard that is to provide protection with an adequate margin of
January 2005
5-25
Draft ~ Do Not Quote or Cite
-------
1 safety, it is appropriate to take into account evidence of effects' for which'the reported ^ !-
2 associations provide only suggestive evidence of a potentially causal association. " -"•'
3 In looking first at the long:term exposure mortality studies, staff notes that the long-term
4 mean PM2 5 concentration in the Six Cities study was 18 ug/m3, within an overall range of 11 to
5 . 30 |ig/m3. In'the studies using the ACS cohort, the long-term mean PM2 5 concentration across
6 the cities was 21 u.g/m3 in the initial study and in the reanalysis of that study, within an overall
7 range of 9 to 34 ug/m3. In the extended ACS study, the mean for the more recent time period
8 used in the analysis (from 1999 to 2000) was 14 u.g/m3; in looking at the association based on the
9 air quality averaged over both time periods (which was the basis for the concentration-response
10 functions from this study used in the risk assessment), the long-term mean PM25 concentration
11 was 17.7 ug/m3, with a standard deviation of ± 4, ranging down to 7.5 fig/m3. The CD notes that
12 the confidence intervals around the relative risk functions in this extended study, as in the initial
»
13 ACS study, start to become appreciably wider below approximately 12 to 13 u.g/m3. In
14 considering the Southern California children's cohort study showing evidence of decreased lung'
15 function growth, staff notes that the long-term mean PM2 5 concentration was 15 ug/m3, ranging
16 from 7 to 32 ng/m3 across the cities. This is approximately equal to the long-term mean PM2 j
17 concentration in the earlier 24 City study, showing effects on children's lung function; in which '
18 the long-term mean concentration was 14.5 ug/m3, ranging from 9 to 17 ug/m3 across the cities:
19 In considering this evidence, staff concludes that these studies provide a basis for
20 considering an annual PM25 standard somewhat below 15 ug/m3,' down to about -12 ug/m3. A
21 standard of 14'u.g/m3 would reflect some consideration of the more recent long-term exposure
22 studies that show associations over a somewhat lower range of air quality than had been
23 observed in the studies available in the last review. A standard of 13 jig/m3 would be consistent
24 with a judgment that appreciable weight should be accorded these long-term exposure studies,
25 particularly taking into account the most recent extended ACS mortality study' 'and the Southern
26 California children's cohort morbidity study. A standard level of 13-u.g/m3 would be well below
27 the long-term mean in the Six Cities mortality study 'and approximately 'one standard deviation
28 below the extended ACS mortality study mean, while being somewhat closer to the long-term
29 means in the morbidity studies discussed above. A standard of 12 ug/m3 would be consistent
January 2005 5-26 Draft - Do Not Quote or Cite
-------
r
I
w
o
€:
|!
3 •
10
*,
o
S3
-8
*3
1-^
^ •<
li
3
S
9-,
R
8
8 /
s
5-
3
S
8
S
S
S
* I
fe
S
a
g
S
S
fe
ft
a
*
9
i
O)
I
OS
oo Os p i— i
>r> vo
o\ o
-------
1 The alternative annual PM2 5 standards considered here include a range of levels from 15
2 to 12 ug/m3, and simulating attainment of the standards is based on a percent rollback calculated
3 using the highest monitor in an area, as noted in Table 5-1 and discussed in Chapter 4, section
4 4.2.3, The alternative 24-hour PM2 5 standards considered here include a range of levels from 65
5 to 25 ug/m3 in conjunction with two different forms, including the 98th percentile form of the
6 current 24-hour PM2 5 standard and an alternative 99th percentile form. Further discussion of
7 alternative forms of the annual and 24-hour standards is presented below in section 5.3.6.
8 In looking at the base case estimates, staff has first considered the estimated reductions
9 associated with lower levels of the annual PM2 5 standard, without changing the 24-hour
10 standard. From Table 5-2, staff observes that alternative annual standard levels of 14,13, and 12
11 ug/m3 result in generally consistent estimated risk reductions from long-term exposure to PM2 5
12 of roughly 20, 30, and 50 percent, respectively, across all five'example cities. Thus, for the base
13 case assessment in which mortality risks are estimated down to the lowest measured level in the
14 extended ACS study, estimated reductions in mortality associated with long-term exposure to
15 PM2S are no greater than 50 percent in any of the five example cities with changes in the annual
16 standard down to a level of 12 ug/m3.
17 Staff also examined the effect on mortality reduction if the 24-hour standard were to
18 change. Staff first notes that the estimated reductions in long-term mortality risk associated with
19 changes to the 24-hour standard are much more variable across cities than with changes in just
20 the annual standard. Further, no combination of standards within the ranges that staff has
21 considered result in the elimination of all estimated long-term mortality risk in all example cities.
22 This assessment indicates that estimated reductions in long-term mortality risk of approximately
23 50 percent or greater in the five example cities generally result from 24-hour standards set at 30
24 to 25 ug/m3, based on either the 98th or 99th percentile form of such a standard, depending on the
25 city. ' ''
26 Staff further considered the effects of various combinations of the annual and 24-hour
27 standard. Staff notes in particular that the base case estimates of long-term mortality risk
28 reduction associated with a 24-hour standard set at 25 ug/m3 provides the same degree of risk
29 reduction regardless of the level of the annual standard within the range of 15 to 12 ug/m3; a 24-
January2005 5-30 Draft - Do Not Quote or Cite
-------
1 hour standard set at 30 |ig/m3 provides the same degree of risk reduction in most but not all
2 cases. That is, in the range of 30 to 25 ug/m3, the 24-hour standard would be the generally
3 controlling standard in most cases relative to an annual standard in the range of 15 to 12 ug/m3;
4 and, in those cases, lowering the annual standard to as low as 12 ug/m3 would result in no
5 additional estimated reductions in long-term mortality risks.
6 Beyond this base case assessment, staff also has considered the extent to which the
7 assumption of the presence of hypothetical thresholds in the concentration-response relationships
8 would influence the estimated risk reductions. As noted above (section 5.3.1), the estimated
9 incidence of PM25-related mortality associated with long-term exposure when the current
10 standards are met are appreciably smaller, although still present, under these assumed
11 hypothetical thresholds, hi considering an assumed threshold of 10 ug/m3, staff observes that
12 lowering the annual standard to alternative levels of 14,13, and 12 ng/m3 (without changing the
13 24-hour standard) results in estimated risk reductions of roughly 30 to 40 percent, 50 to 70
14 percent, and 80 to 100 percent, respectively, across the five example cities. In considering
15 changes to the annual and/or 24-hour PM2 5 standards in this case, staff first notes that mortality
16 risk associated with long-term exposure is estimated to be reduced by 100 percent in all five
17 cities with a 24-hour standard set at 25 ug/m3, in combination with the current annual standard.
18 For a 24-hour standard set at 35 ug/m3, with a 99th percentile form, estimated risk reductions
19 remained at 100 percent in three of the cities, but were only 40 and 6 percent in the other two
20 cities. Under this assumed threshold of 10 ug/m3, similar to the base case, there is little if any
21 additional reduction obtained in lowering the annual standard below 15 ug/m3 in conjunction
22 with 24-hour standards in this range. Thus, in this case, as in the base case, changes in the 24-
23 hour standard, while retaining the current annual standard, can result in larger but much more
24 variable estimated reductions in risks associated with long-term exposures across the five cities.
25 Further, in considering an assumed hypothetical threshold of 12 ug/m3, staff observes
26 that lowering the annual standard to a level of 14 ug/m3 (without changing the 24-hour standard)
27 results in estimated risk reductions of 100 percent in all five cities. In considering changes to the
28 24-hour PM2 5 standard alone in this case, staff notes that long-term mortality risk is estimated to
January 2005 5-31 Draft - Do Not Quote or Cite
-------
1 be reduced by 100 percent in all five cities with a 24-hour standard set at 30 ug/m3, 98*
2 percentile form.
3 5.3.4.3 Summary s
4 In summary, in considering the epidemiologic evidence, estimates of risk reductions
5 associated with alternative annual and/or 24-hour standards, and the related limitations and
6 uncertainties, staff concludes that there is clear support for considering revisions to the suite of
7 current PM2 5 standards to provide additional protection against health effects associated with
8 long-term exposures.. In looking specifically at the evidence of associations between long-term
9 exposure to PM25 and serious health effects, including total, cardiovascular, and lung cancer
10 mortality, as well as respiratory-related effects on children, staff concludes that it is appropriate
11 to consider an annual PM25 standard in the range of 15 down to 12 ug/m3. In considering the
12 results of the quantitative risk assessment, in the absence of evidence of clear thresholds, staff
13 believes that it is appropriate to give significant weight to base case risk estimates, while also
14 considering the implications of potential thresholds within the range of the air quality data from
15 the relevant studies. In so doing, staff finds further support for considering an annual PM2 5
16 standard in the range of 14 to 12 ug/m3. Alternatively, staff also finds support for a revised 24-
17 hour standard, in conjunction with retaining the current annual standard, in the range of 35 to 25
18 ug/m3, with an emphasis on a 99th percentile form especially with a standard level in the middle
19 or upper end of this range. Staff notes that a 24-hour standard at a level of 40 ug/m3 is estimated
20 to provide no additional protection against the serious health effects associated with long-term
21 PM25 exposures in two or three of the five example cities (for a 99th or 98th percentile form,
22 respectively) relative to that afforded by the current annual PM25 standard, regardless of the
23 weight that is given to the potential for a threshold within the range considered by staff. Staff
24 believes that a suite of PM2 5 standards selected from the alternatives identified above could
25 provide an appropriate degree of protection against the mortality and morbidity effects
26 associated with long-term exposure to PM25 in studies in urban areas across the U.S..
January 2005
5-32
Draft - Do Not Quote or Cite
-------
1 53.5 Alternative PMW Standards to Address Health Effects Related to Short-term
1 Exposure
3 In considering alternative PM2 5 standards that would provide protection against health
4 effects related to short-term exposures, staff has similarly taken into account both evidence-
5 based and risk-based considerations. As discussed below in this section, staff has first evaluated
6 the available evidence from short-term exposure studies, as well as the uncertainties and
7 limitations in that evidence, to assess the degree to which alternative 24-hour and/or annual
8 PM2 5 standards can be expected to provide protection against effects related to short-term
9 exposures. Secondly, staff has considered the quantitative risk estimates for short-term exposure
10 effects, discussed in,Chapter 4, to assess the extent to which alternative annual and/or 24-hour
11 standards can be expected to reduce the estimated risks attributable to short-term exposure to
12 PM2 5. Staff conclusions as to ranges of alternative annual and/or 24-hour standards that would
13 provide protection against health effects related to short-term exposures are summarized at .the
14 end of this section. As noted above, the integrated staff recommendations presented in section
15 5.3.7 are based in part on the conclusions from this section and in part on staff conclusions from
16 the previous section, in which alternative PM2 5 standards to address health effects related to
17 long-term exposures are assessed.
18 53.5.1 Evidence-based Considerations
19 In taking into account evidence-based considerations, staff has evaluated the available
20 evidence from short-term exposure studies, as well as the uncertainties and limitations in that
21 evidence. In so doing, staff has focused on U.S. and Canadian short-term exposure studies of
22 fine particles (Appendix 3A). We took into account reanalyses that addressed GAM-related
23 statistical issues and considered the extent to which the studies report statistically significant and
24 relatively precise relative risk estimates; the reported associations are robust to co-pollutant
25 confounding and alternative modeling approaches; and the studies used relatively reliable air
26 quality data. In particular, staff has focused on those specific studies, identified above in section
27 5.3.1, that provide evidence of associations in areas that would have met the current annual and
28 24-hour PMZ5 standards during the time of the study. Staff believes that this body of evidence
January 2005
5-33
Draft - Do Not Quote or Cite
-------
1 can serve as a Basis for 24-hour and/or annual PM2 j standards that would provide increased
1 protection against effects related to short-term exposures; ' ; ' • • •
3 As an initial matter, staff recognizes, as discussed above, that these short-term exposure
4 studies provide' no evidence of clear thresholds, or lowest-observed-effects levels, in terms of 24-
5 hour average concentrations. Staff notes that of the two PM25 studies that explored potential
6 thresholds, one study in Phoenix provided some suggestive evidence of a threshold possibly as
7 high as 20 to 25 ug/m3, whereas the other study provided evidence suggesting that if a threshold
8 existed, it would likely be appreciably below 25 ug/m3. While'there is no evidence for clear
9 thresholds within the range of air quality observed hi the epidemiologic studies, for some health
10 endpoints (such as total nonaccidental mortality) it is likely to be extremely difficult to detect
11 threshold levels (CD, p.9-45). As a consequence, this body of evidence is difficult to'translate
12 directly into a specific 24-hour standard that would independently protect against all effects
13 associated with short-term exposures. Staff notes that the distributions of daily PM2S
14 concentrations in these studies often extend down to or below background levels, such that
15 consideration of the likely range of background concentrations across the U.S., as discussed in
16 Chapter 2, section 2,6, becomes important in identifying a lower bound of a range of 24-hour
17 standards appropriate for consideration. ' .
18 Being mindful of the difficulties posed by issues relating to threshold and background
19 levels, staff has first considered this short-term exposure epidemiologic evidence as a basis for
20 alternative 24-hour PM25 standards. In so doing, staff has focused on the upper end of the
21 distributions of daily PM25 concentrations, particularly in terms of the 98th and 99th percentile
22 values, reflecting the form of the current 24-hour standard and an alternative form considered in
23 the risk assessment, respectively. In looking at the specific studies identified in section 5.3.1 that
24 report statistically significant association in areas that would have met the current PM2 5
25 standards, including studies in Phoenix (Mar et al., 1999, 2003), Santa Clara County, CA
26 (Fairley, 1999/2003) and eight Canadian cities (Burnett et al., 2000 and Burnett and Goldberg,
27 2003), staff notes that the 98th percentile values range from approximately 32 to 39 ug/m3 in
28 Phoenix and the eight Canadian cities, up to 59 ug/m3 in Santa Clara Country; 99th percentile
29 values range from 34 to 45 ug/m3 in Phoenix and the eight Canadian cities, up to 69 'ug/m3 in •
January 2005 " 5-34 Draft - Do Not Quote or die
-------
r
1 evidence discussed above, giving little weight to the remaining uncertainties in the broader body
2 of short-term exposure evidence, including the possibility of a threshold within the range of air
3 quality in the studies and the recognition that results may be sensitive to selection of models
4 bey ond the range of models examined in these particular studies.
5 Consistent with the conclusions reached in the last review (62 FR 3 8674-7), however,
6 staff continues to believe that an annual standard cannot be expected to offer an adequate margin
7 of safety against the effects of all short-term exposures, especially in areas with unusually high
8 peak-to-mean ratios of PM25 levels, possibly associated with strong local or seasonal sources, or
9 for potential PM2 5-related effects that may be associated with shorter-than-daily exposure
10 periods (noted above in section 5.3.3). As a result, if an alternative annual standard were
11 adopted to provide primary protection against effects associated with short-term exposures, staff
12 believes it is appropriate also to consider an alternative 24-hour PM2 5 standard to provide such
13 supplemental protection. Such a supplemental 24-hour standard could reasonably be based on
14 air quality information (from 2001 to 2003) in Chapter 2, Figure 2-23, that shows the distribution
15 of 98th percentile values as a function of annual means values in urban areas across the U.S.
16 Based on this information, staff concludes that a supplemental standard in the range of
17 approximately 40 to 35 ng/m3 would limit peak concentrations in areas with relatively high
18 peak-to-mean ratios (i.e., generally in the upper quartile to the upper 5th percentile, respectively)
19 and with annual mean concentrations in the range of 12 to 15 ug/m3.
20 To assist in understanding the public health implications of various combinations of
21 alternative annual and 24-hour standards, staff assessed (based on the same air quality database)
22 the percentage of counties, and the population in those counties, that would not likely attain
23 various PM25 annual standards alone in comparison to the percentage of counties that would not
24 likely attain alternative combinations of annual and 24-hour PM2 5 standards. This assessment is
25 intended to provide some rough indication of the breadth of supplemental protection potentially
' 26 afforded by various combinations of alternative standards. The results of such an assessment,
27 based on air quality data from 562 counties, are shown in Tables 5-3(a) and (b).
January 2005 5-37 Draft - Do Not Quote or Cite
,
-------
i
^>
s.
S
1
E
€
'I
s
1
— w
s
II
I!
ja e
•I
1
1
S
s
o
i
.
S
S
8?
a
3
ffl
s
• O
O)
I
a
I
oo
^^ o^O
CNCNcNtStNcNCS
-------
!
standard anc
meet stated
not likely to
c
o
'•a
i
o
Q.
Q.
*•»
1
fe
t
S"
•a
e
cs
5
__
o
o
^
W
1
•n
c
(0
0)
1
1
£
CO
i
I
it
**
°cs
I5
CO
1
Z
Southwest
&!
= =
Industrial
Midwest
1
c/>
1
"c"
o
1
al
°f
3
8
^— .
t
a>
3.
m
*~
v>
OO
CM
g
S
V-
S
S
«
l!
6 S
-c o
-|€
« .=
8 r«
5& C
ii
P
o
fe
r-.
m
^-
R
§
£
in
SO.
o
S
CO
0
r~.
CO
o
in
^
f5
§
t-w
CO
in.
O
o
in
CO
o
t—
CO
CM
o
^g-
S
£*-*
°
f*.
E
•r-
in
•q-
CO
O
r~-
CD
CD
O
^J-
S
°
£
OO
in
CM
§
CO
r»-
i»-
co
8
o
•*_••
^-
F:
5
CO
CM
CO
m
«
C*3
CO
§
in
o>
OO
S
9
f°
£
r~
o
CO
CO
CO
o
OO
in
CO
CO
^r
3
i»-
«
£
OO
in
CM
co
o
&
CM
in
CM
3
co
p
CO
CO
S
in
CO
CM
o
fe
CM
in
CM
S
CO
0
1
S
o
in
CM
o
N-
CD
«JT
O
CM
S
CO
o
S
S
in
•*
CM
O
P—
CO
S
o
CM
S
CO
p
OO
CO
in
§
CM
r*~
&
%
o
•*
S
CO
S
e
S
o
CO
CM
CO
S
£
C*J
*r
3
r—
OO
CO
a
o>
in
CM
CM
1
£•
1*3
& fc
.Sa *
I 1
U
33
§|
•8 1
1 1 '
•s >.
« "a
3J O
M g
m °
Ii «
ra j>
•5 ^
J ,+5 .
"
1 S^ a
& £; TS **
S '^3 1
3, i -S • 8
|i a a
3 ^3 co o
IM 5
^ i 1 ^
r- TO .£ 1^-1
73 i-c 00 ^
!|| I
i|; i
Hi i
S "° >>
8 s-s §
<^ T3 HS 2
fll ^^ *.
SW 4J Jf?*
,3 H *J
s>« 1
§2*3 -1
Tfl D M -g
p^ cd s M.
» -3 S S
o
I
i— • (S m
so ox o i— '
en
>n vo
-------
i
o
B
/^^
i
o
U in
.S'S
-I I
« i
3ts
_ >n
Is
(U
&5
19 =
S.
It
.
p
•0 C
§1
S §
"I
ll
11
a. S
3
JS
H
ja
«
J.
•o
C
CO
in
*
•o
S
«
"S
I
E
S
o
1
i
fc
s>
•o
§
1
I
c
3
o
8
_«
•S
fc
C
3
s
§
**
"*
ii
of
C
u
f S
w
1
1
o
5
w
ll
°-s
Industrial
Midwest
I
o
w
1
3>
1
z
*£*
o
"So
3
-S ft
"~ z
C
J
E
in
in
CO
CN
.
§
co
CO
s
S"
S"
in
"g C
i i
p
* C
s- t*
^ .1
13 g.
^5
•
.
f
3
c
o
o
CO
"*
o
o
OJ
CM
Is-
CT>
O
CO
^r
in
o
to
in
in
o
in
T—
CO
CM
5
in
T~
O
cS
r-.
in
-
CO
r-
o
••a-
-
in
0
co
o
cS
CN
in
CN
.
CO
o
»
in
CO
o
CO
in
ro
o
r—
^
S
§
5
in
CO
in
CO
o
CO
s
«
»
&
s
£
00
CO
CO
8
in
CO
CO
e
s
.
.
g
S
en
CO
in
CO
S
in
o
fe
CO
in
o
U)
CM
CO
CM
5
in
m
CD
•*
O
55
in
o
o
kO
CM
CO
CM
5
CM
O
in
>^-
O
co
CM
O
o
CM
in
T""
CM
S
sg
CO
CM
in
•«
CO
*
o
o
c—
in
CM
CO
co
in
in
CO
o
•q-
CO
O
CO
in
en
o
CO
EM
CM
r-
^
Is-
S
••a-
CO
CO
s
CO
CO
%
&
s
CO
en
CM
CO
CO
CO
S
•*f
CO
r—
CO
s
CO
.
§1
CO
CO
s
CO
eg
in
CN
•*
0
a
i
•—< cs
-------
.
i
a
U)
CQ
?
CD
m
"S
13
"55
1
E
£
"5
"S
o
-.=
2
S
•*•*
I
1
e
o
§?
•o
a
i!
1
|
1
<£
§
•p
8
1
i
TB ^
**
c
^ 5
o
co
ts
1
•e
8
§
CO
«3 1
a- *
Q.-5
11
11
1
n
o
CO
1
z
1
ll
"8
€
8
-;?•
OJ
••w*
10
!P
00
iN
en
1
fO
S3
&
r>-
S
to
1?
1 §
•'g ^
* c
•S ^>
^ o
1!
£--
o
CO
en
in
fs
°
5?
S
o
in
CO
CO
O
to
in
?
CO
0
£
r—
5
,—,
in
CO
O
CO
CM
o
CD
°
^
CO
CO
r-
CD
CO
o
r-.
°
tn
ur>
S.
^
e^
••J-
CO
CO
o
oo
in
en
in
CO
§
in
S
S
in
CO
CO
CO
o
oo
in
oo
CO
in
en
S
8
s
f=
,—,
CO
CO
CO
f—
oo
CO
CO
en
en
co
O
S
08
in
CM
CO
O
CO
CM
in
CM
en
CO
CO
g
1
S
in
CD
CM
O
E5
CD
?
CM
OO
CO
0
£
8
C-J
in
CM
O
fe
CM
CM
O
CM
cn
oo
CO
r—
£
<0
in
CM
r—
f2
co
co
o
CM
en
OO
S
K
E.
05
O
CM
CO
S
in
en
CM
cS
CO
§
b
S
in
-------
1 For example, from Table 5-3 (a) it can be seen that for an annual standard set at 15 ug/m3,
2 24-hour standard levels ranging from 40 to 35 ug/m3, with a 98* percentile form, would add
3 approximately 3 to 13 percent to the percentage of counties nationwide that would not likely
4 attain both standards relative to the number of counties that would not likely attain the annual
5 standard alone; with a 99th percentile form, as seen in Table 5-3(b), these percentages increase to
6 13 to 30 percent. For an annual standard set at 12 ug/m3, 24-hour standard levels in this range
7 would add approximately 1 to 4 percent, or 5 to 9 percent, to the percentage of counties for
8 standards with a 98th or 99th percentile form, respectively. As seen in Tables 5-3(a) and (b), the
9 percentage of the population that would be afforded greater public health protection from these
10 alternative standards would increase somewhat more than would the percentage of counties not
11 likely to attain the standards.
12 5.3.5.2 Risk-based Considerations . .."
13 Beyond looking directly at the relevant epidemiologic evidence, staff has also considered
V14 the extent to which specific levels and forms of alternative 24-hour and annual PM25-standards
15 are likely to reduce the estimated risks attributable to short-term exposure to PM25, and the
16 uncertainties in the estimated risk reductions. As discussed above (section 5.3.1), staff has based
17 this evaluation on the risk assessment results presented in Chapter 4, in which short-term ,
18 exposure risks were estimated down to background or the lowest measured level (LML) in a
•19 particular study, whichever is higher. Staff also has considered the sensitivity of these results to
20 the uncertainty related to potential thresholds by using concentration-response functions
21 modified to incorporate assumed hypothetical threshold levels.
<~
22 Table 5-4 summarizes estimated percentage reductions in mortality attributable to short-
23 term exposure to PM2S in going from meeting the current PM25 standards to meeting alternative
24 annual and 24-hour PM2 5 standards in the five example cities that do not meet the current
25 standards based on 2001-2003 air quality data. Base case estimated percentage risk reductions
26 are given in the table, along with reductions associated with assumed alternative hypothetical
27 thresholds. The percentage reductions presented in Table 5-4 represent approximate reductions
28 relative to the total'estimated short-term mortality incidence presented above in Table 5-1.
January 2005 5-42 Draft - Do Not Quote or Cite
-------
2
0>
"3
_l
z
"I
in
*3
b-
n
f
EL
1
<
§
.±2
rS
•
2J
0
•
8
m
^
o
v"
1
.a
3
in
o
1
s
m
o
0)
I
«N
m
o
8
£
8
in
o
IMSsumea inresnoia ujgiim)
(base = backqround or LML1
o
v»
8
§
§
^~
>o
g
a
§
§
o
§
s
CO
s
s '
§
8
Cg
O
8
s
**^
Incidence Associated with
Meetina Current Standards
«
\-
\
\
(-
'\
'»
;>
S
}
,» "
\
^
,}
• »
.1.. -
r
jl
T"
;%
TO*
"H ^
i&
•p> '
$
0
o
o
0
o
CM
o
•*!•
co
o
^-
8
a
TT
O
O
o
o
o
o
o
§
^^
CO
o
-
s
8
CD
o
o
o
in
CO
h-
co
OJ
in
u>
CO
co
§
g
s
CM
m
CO
oo
co
r-
CJ
CD
Eo
&
\s
CO
M-
«
CO
CO
«
CM
r—
o
o
CO
s
S3
CO
CO
CM
CO
§
8
fc
5
s
3
e
en
XT
co
CO
s
s
CO
•«J-
R
S
OO
xr
?l
cQ
o
o
o
o
o
o
o
o
3
to
•sT
8
en
o
o
o
o
o
o
o
.£
•-P
1
0>
a.
%
f
rt
1
LO
CO
O
o
o
o
o
*r
§
S
CO
oo
CO
CM
m
CO
r-.
K
CO
CO
CO
o
o
o
C3
•*
5
3
CM
o
§
8
s
a
o
o
fe
CO
£
g
CO
CO
s
o>
•*
CO
^
CM
in
CO
R
s
9
a
o
CO
§
CO
CO
oo
CO
s
3
s
s
£
OJ
s
o
in
s
51
CO
CO
R
§
s
c
8
§
§
CM
CO
o
vo
§
1
r—
en
R
B
fe
SI
S
S
CO
r~-
%
8
?
T-,
<*> :
§2
«>
§
^
to :
«& '
9
f?
« i
^
* ..
8
fe"" ",:
» ;
13
S'
1^ •
V-
» ,
1
i^
•6
la
*
^
CO
en
CO
o
•^-
CM
CD
CO
§
CO
CO
CM
T
s
Si
t~-
O
8
CO
CM
O
o
rr
5
*—
CO
CM
CO
CO
CO
R
S
h-.
CO
S
in
m
CO
•<»-
S
S
CO
CM
O
S
S
S3
s
CM
CO
§
S
r--
r^
5
£
§
co
f-.
CT)
Tf
CO
CO
1".
CD
CO
9
R
CM
CO
•*
CO
in
CM
5
CO
CO
en
•«r
a
o
oo
s
s
s
§
s
s
in
CO
fe
T™"
o>
oS
en
to
o
in
3
CO
CO
8
§
s
-------
s»
Li
1
£
in
S
2'
o _
!!
E H
381
^1
S
s
00
8
o
W)
Associated with
tandards
Inci
M
"
w
O
•»
B
to
CO
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
The same alternative standards are considered here as were considered above in section
5.2.4. That is, the alternative annual PM2 5 standards considered here include a range of levels
from 15 to 12 ug/m3, and simulating meeting these standards is based on a percent rollback
calculated using the highest monitor in an area, as noted in Table 5-1 and discussed in Chapter 4,
section 4.2.3. The alternative 24-hour PM25 standards considered here again include a range of
levels from 65 to 25 ug/m3 in conjunction with two different forms, including the 98th percentile
form of the current 24-hour PM25 standard and an alternative 99* percentile form. Further
discussion of these alternative forms for annual and 24-hour standards is presented below in
section 5.3.6.
In looking at the base case estimates, staff first considered the estimated reductions
associated with lower levels of the annual PM25 standard, without changing the 24-hour
standard. .From Table 5-4, staff observes that lowering the annual standard to alternative levels
of 14,13, and 12 ug/m3 results in small but generally consistent estimated risk reductions of
roughly 10 to 15 percent, 15 to 20 percent, and 25 to 30 percent, respectively, across all five
example cities, Thus, for the base case assessment in which mortality risks are estimated down
to background or the lowest measured level in the relevant study, estimated reductions in
mortality associated with short-term exposure to PM2 5 are no greater than 30 percent in any of
the five example cities with changes in the annual PMZ5 down to a level of 12 ug/m3.
In considering changes to the 24-hour and/or annual PM2 5 standards for base case
estimates, staff first notes that the estimated reductions in short-term mortality risk associated
with changes to the 24-hour standard are generally larger and much more variable across cities
than with changes in just the annual standard. Further, no combination of standards within the
ranges that staff has considered results in the elimination of all estimated mortality risk
associated with short-term exposure in all example cities. More specifically, a 24-hour standard
of 25 ug/m3 results in base case estimates of reductions in short-term mortality ranging from
approximately 30 to 50 percent (98th percentile form) and 35 to 70 percent (99th percentile form)
across the five cities in conjunction with any annual standard in the range of 15 to 12 ug/m3. A
24-hour standard of 30 ug/m3 results in base case estimates of reductions in short-term mortality
ranging from approximately 25 to 35 percent (98th percentile form) and 25 to 65 percent (99th
January 2005
5-45
Draft - Do Not Quote or Cite
-------
1 percentile form) across the five cities in conjunction with an annual standard of 12 ug/m3; the
2 lower end, but not the upper end, of these ranges decreases somewhat in conjunction with annual
3 standards from 13 to 15 ug/m3. As in the assessment of risk related to long-term exposures
4 discussed in section 5.3.4.2, this assessment indicates that 24-hour standards of 30 to 25 ug/m3
5 become generally controlling standards in most cases within this range of annual standards.
6 Beyond this base case assessment, staff also has considered the extent to which the
7 assumption of the presence of hypothetical thresholds in the concentration-response relationships
8 would influence the estimated risk reductions. As noted above (section 5.3.1), the estimated
9 incidence of PM2 5-related mortality associated with short-term exposure when the current
10 standards are met are appreciably smaller under these assumed hypothetical thresholds. In
11 considering an assumed threshold of 10 ug/m3, staff observes that lowering the annual standard
12 to alternative levels of 14,13, and 12 ug/m3 (without changing the 24-hour standard) results in
\.
13 estimated risk reductions of roughly 15 to 25 percent, 30 to 35 percent, and 45 to 55 percent,
14 respectively, across all five example cities. In considering changes to the 24-hour and/or annual
15 PM25 standards in this case, staff notes that a 24-hour standard of 25 ug/m3 results in estimates
16 of reductions in short-term mortality ranging from approximately 45 to 80 percent (98*
17 percentile form) and 60 to 95 percent (99th percentile form) across the five cities in conjunction
18 with any annual standard in the range of 15 to 12 ug/m3. A 24-hour standard of 30 ug/m3 results
19 in estimates of reductions in short-term mortality ranging from approximately 45 to 60 percent
20 (98th percentile form) and 50 to 95 percent (99th percentile form) across the five cities in
21 conjunction with an annual standard of 12 ug/m3; as with the base case, the lower end, but not
22 the upper end, of these ranges decreases appreciably in conjunction with annual standards from
23 13 to 15 ug/m3. Thus, in this case, as in the base case, changes in the 24-hour standard, while
24 retaining the current annual standard, can result in generally larger but much more variable
25 estimated reductions in risks associated with short-term exposures across the five cities than with
26 changes in just the annual standard.
27 Further, in considering assumed hypothetical thresholds of 15 or 20 ug/m3, staff observes
28 that lowering the annual standard to alternative levels of 14,13, and 12 ug/m3 (without changing
29 the 24-hour standard) results in estimated risk reductions of roughly 20 to 45 percent, 40 to 65
January 2005 5-46 Draft - Do Not Quote or Cite
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
percent, and 60 to 90 percent, respectively, across all five example cities. In considering
changes to the 24-hour and/or annual PM25 standards in this case, staff notes that a 24-hour
standard of 25 ug/m3 results in estimates of reductions in short-term mortality ranging from
approximately 60 to 100 percent (98th percentile form) and 70 to 100 percent (99th percentile
form) across the five cities in conjunction with any annual standard in the range of 15 to 12
ug/m3. A 24-hour standard of 30 ug/m3 results in estimates of reductions in short-term mortality
ranging from approximately 60 to 90 percent (98th percentile form) and 60 to 100 percent (99th
percentile form) across the five cities in conjunction with an annual standard of 12 ug/m3;
similarly, the lower end, but not the upper end, of these ranges decreases appreciably in
conjunction with annual standards from 13 to 15 ug/m3. Thus, in this case as .well, changes in
«
the 24-hour standard, while retaining the current annual standard, can result in generally larger
but much more variable estimated reductions in risks associated with short-term exposures
across the five cities than with changes in just the annual standard.
53.5.3 Summary
In summary, in considering the relevant epidemiologic evidence, estimates of risk
reductions associated with alternative annual and/or 24-hour standards, and the related
limitations and uncertainties, staff concludes that there is clear support for considering revisions
to the suite of current PM25 standards to provide additional protection against health effects
associated with short-term exposures. In looking specifically at the evidence of associations
between short-term exposure to PM2 5 and serious health effects, with a particular focus on
mortality associations, staff concludes that it is appropriate to consider a revised 24-hour
standard in the range of 30 to 25 ug/m3 in conjunction with retaining the current annual standard
level of 15 ug/m3. Alternatively, staff also believes the evidence supports consideration of a
revised annual standard, in the range of 13 to 12 ug/m3., in conjunction with a revised 24-hour
standard, to provide supplemental protection, in the range of 40 to 35 ug/m3. In considering the
results of the quantitative risk assessment, in the absence of evidence of clear thresholds, staff
believes mat it is appropriate to give significant weight to base case risk estimates, while also
considering the implications of potential thresholds within the range of the air quality data from
the relevant studies. In so doing, staff also finds support for considering a revised 24-hour
January 2005
5-47
Draft - Do Not Quote or Cite
-------
1
2
3
4
5
6
7
g
9
10
standard, in conjunction with retaining an annual standard level of 15 ug/m3, in the range of 30
to 25 ug/m3. Staff notes that a 24-hour standard at a level of 35 ug/m3 is estimated to provide
less than 30 percent reduction in mortality incidence in two or three of the five example cities
(for a 99th or 98* percentile form, respectively), in either the base case or under an assumed
hypothetical threshold of 10 ug/m3, relative to that afforded by the current annual PM25 standard
alone. Further, staff finds little support based on the risk assessment for addressing short-term
exposure effects solely with a revised annual standard in a range down to 12 ug/m3. Staff
believes that a suite of PM25 standards selected from the alternatives identified above could
provide an appropriate degree of protection against the mortality and morbidity effects
associated with short-term exposure to PM25 in studies in urban areas across the U.S..
11 5.3.6 Alternative Forms for Annual and 24-Hour PM2.5 Standards
12 53.6.1 Form of Annual Standard
13 ^ In 1997 EPA established the form of the annual PM2S standard as an annual arithmetic
14 mean, averaged over 3 years, from single or multiple community-oriented monitors. This form
15 was intended to represent a relatively stable measure of air quality and to characterize area-wide
16 PM25 concentrations. The arithmetic mean serves to represent the broad distribution of daily air
17 . quality values, and a 3-year average provides a more stable risk reduction target than a single-
18 year annual average. The annual PM2 s standard level is to be compared to measurements made
19 at the community-oriented monitoring site recording the highest level, or, if specific constraints
20 are met, measurements from multiple community-oriented monitoring sites may be averaged (62
21 FR at 38,672). The constraints on allowing the use of spatially averaged measurements were
22 intended to limit averaging across poorly correlated or widely disparate air quality values. This
23 approach was judged to be consistent with the epidemiologic studies on which the PM2 5 standard
24 was primarily based, in which air quality data were generally averaged across multiple monitors
25 in an area or were taken from a single monitor that was selected to represent community-wide
26 exposures, not localized "hot spots."
27 In this review, in conjunction with recommending that consideration be given to
28 , alternative annual standard levels, staff is also reconsidering the appropriateness of continuing to
January 2005
5-48
Draft - Do Not Quote or Cite
-------
\
1 allow spatial averaging across monitors as part of the form of an annual standard. There now
2 exist much more PM2 5 air quality data than were available in the last review. Consideration of
3 the spatial variability across urban areas that is revealed by this new database (see Chapter 2,
4 section 2.4 above, and the CD Chapter 3, section 3.2.5) raises questions as to whether an annual
5 standard that allows for spatial averaging, within currently specified or alternative constraints,
6 would provide appropriate public health protection. In conducting analyses to assess these
7 questions, as discussed below, staff has taken into account both aggregate population risk across
8 an entire urban area and the potential for disproportionate impacts on potentially vulnerable
9 subpopulations within an area
10 The effect of allowing the us e of spatial averaging on aggregate population risk was
11 considered as part of the sensitivity analyses included in the health risk assessment discussed in
12 Chapter 4. In particular, a sensitivity analysis was done in several example urban areas (Detroit,
13 Pittsburgh, and St. Louis) that compared estimated mortality risks (associated with both long-
14 and short-term exposures) based on air quality values from the highest community-oriented
15 monitor in an area with estimated risks based on air quality values averaged across all such
16 monitors within the constraints allowed by the current standard. As discussed in Chapter 4,
17 section 4.2.3, the monitored air quality values were used to determine the design value for the
18 annual standard in each area, as applied to a "composite" monitor to reflect area-wide exposures.
19 Changing the basis of the annual standard design value from the concentration at the highest
20 monitor to the average concentration across all monitors reduces the air quality adjustment
21 needed to just meet the current or alternative annual standards. As expected, the estimated risks
22 remaining upon attainment of the current annual standard are greater when spatial averaging is
23 used than when the highest monitor is used. Based on the results of this analysis in the three
24 example cities, estimated mortality incidence associated with long-term exposure based on the
25 use of spatial averaging is about 10 to over 40% higher than estimated incidence based on the
26 use of the highest monitor. For estimated mortality incidence associated with short-term
27 exposure, the use of spatial averaging results in risk estimates that range from about 5 to 25%
28 higher. In considering estimated risks remaining upon attainment of alternative suites of annual
29 and 24-hour PM25 standards, spatial averaging only has an impact in those cases when the
January 2005
5-49
Draft - Do Not Quote or Cite
-------
1 annual standard is the "controlling" standard. For such cases in the three example cities, the
2 estimated mortality incidence associated with long-term exposure in most cases ranges from
3 i about 10 to 60% higher when spatial averaging is used, and estimated mortality incidence
4 associated with shortrterm exposure in most cases ranges from about 5 to 25% higher.
5 'In considering the potential for disproportionate impacts on potentially vulnerable
6 subpopulations, staff has assessed whether any such groups are more likely to live in census
7 tracts in which the monitors recording the highest air quality values in an area are located. Data
j f
^ 8 were obtained for demographic parameters measured at the census tract level, including
9 education level, income level, and percent minority. These data from the census tract in which
10 the highest air quality values were monitored were compared to area-wide average values
11 ' (Schmidt etal., 2005). Recognizing the limitations of such cross-sectional analyses, staff-
12 observes that the results suggest that the highest concentrations in an area tend to be measured at
13 monitors located in areas where the surrounding population is more likely to have lower
14 , education and income levels, and higher percentage minority levels. Staff notes that some
~v
15 epidemiologic study results, most notably the associations between mortality and long-term
16 PM2 5 exposure in the ACS cohort, have shown larger effect estimates in the cohort subgroup
17 with lower education levels (CD, p. 8-103). As discussed in Chapter 3, section 3.4, people with
18 lower socioeconomic status (e.g., lower education and income levels), or who have greater
19 exposure to sources such as roadways, may have increased vulnerability to the effects of PM
20 exposure. Combining evidence from health studies suggesting that people with lower
21 socioeconomic status may be considered a population more vulnerable to PM-related effects
22 with indications'from air quality analyses showing that higher PM25 concentrations are measured
23 in local communities with lower socioeconomic status, staff finds that this is additional evidence
24 which supports a change.from spatial averaging across PM25 monitors to provide appropriate
25 protection from public health risks associated with exposure to ambient PM2 5.
26 In considering whether alternative constraints on the use of spatial averaging may be
27 , appropriate to consider, staff has analyzed existing data on the correlations and differences
28 '' between monitor pairs in metropolitan areas (Schmidt etal., 2005). For all pairs of PM25
29 • monitors, the median correlation coefficient based on annual air quality data is approximately
January 2005 '
5-50
Draft - Do Not Quote or Cite
-------
1 0.9; i.e., substantially higher than the current criterion for correlation of at least 0.6, which was
2 met by nearly all monitor pairs. Similarly, the current criterion that differences in mean air
3 quality values between monitors not exceed 20% was met for most monitor pairs, while the
4 annual median and mean differences for all monitor pairs are 5% and 8%, respectively. This
5 analysis also showed that in some areas with highly seasonal air quality patterns (e.g., due to
6 seasonal woodsmoke emissions), substantially lower seasonal correlations and larger seasonal
7 differences can occur relative .to those observed on an annual basis. The spatial averaging
8 requirements established in 1997 were intended to represent a relatively stable measure of air
9 quality and to characterize area-wide PM2 5 concentrations, while also precluding averaging
10 across monitors that would leave a portion of a metropolitan area with substantially greater
11 exposures than other areas (62 FR 38672). Based on the PM2 5 air quality data now available,
12 staff believes.that the existing constraints on spatial averaging may not be adequate to achieve
13 this result
14 In considering the results of the analyses discussed above, staff concludes that it is
15 appropriate to consider eliminating the provision that allows for spatial averaging from the form
16 of an annual PM15 standard. Further, staff concludes that if consideration is given to retaining an
17 allowance for spatial averaging, more restrictive criteria should be considered. Staff believes
18 that it would be appropriate to consider alternative criteria such as a correlation coefficient of at
19 least 0.9, determined on a seasonal basis, with differences between monitor values not to exceed
20 about 10%.
21 5.3.6.2 Form of 24-Hour Standard
22 In 1997 EPA established the form of the 24-hour PM2 5 standard as the 98th percentile of
23 .24-hour concentrations at each population-oriented monitor within an area, averaged over three
24 years (62 FR at 38671-74). EPA selected such a concentration-based form because of its
25 advantages over the previously used expected-exceedance form.4 A concentration-based form is
26 more reflective of the health risk posed by elevated PM2 5 concentrations because it gives
27 proportionally greater weight to days when concentrations are well above the level of the
4 The form of the 1987 24-hour PM10 standard is based on the expected number of days per year (averaged
over 3 years) on which the level of the standard is exceeded; thus, attainment with the one-expected exceedance
form is determined by comparing the fourth-highest concentration in 3 years with the level of the standard.
January 2005
5-51
Draft - Do Not Quote or Cite
-------
1 standard than to days when the concentrations are just above tiie standard. Further, a
2 concentration-based form better compensates for missing data and less-than-every-day
3 monitoring; and, when averaged over 3 years, it has greater stability and, thus, facilitates the
4 development of more stable implementation programs. After considering a range of
5 concentration percentiles from the 95th to the 99th, EPA selected the 98th percentile as an
6 appropriate balance between adequately limiting the occurrence of peak concentrations and
7 providing increased stability and robustness. Further, by basing the form of the standard on
8 concentrations measured at population-oriented'monitoring sites (as specified in 40 CFR part
9 58), EPA intended to provide protection for people residing in or near localized areas of elevated
10 concentrations.
11 : In this review, in conjunction with recommending that consideration be given to
12 alternative 24-hour standard levels, staff is also considering the appropriateness of
13 recommending that the current 98* percentile form, averaged over 3 years, be retained or
14 revised. As an initial matter, staff believes that it is appropriate to retain a concentration-based
15 form that is defined in terms of a specific percentile of the distribution of 24-hour PM2 5
16 concentrations at each population-oriented monitor within an area, averaged over 3 years, Staff
17 bases this recommendation on the same reasons that were the basis for EPA's selection of this
18 type of form in the last review. As to the specific percentile value to be considered, staff has
19 narrowed the focus of this review to the 98th and 99th percentile forms. This focus is based on the
20 observation that the current 98th percentile form already allows the level of the standard to be '
21 exceeded seven days per year, on average (with every-day monitoring), while potentially
22 allowing many more exceedance days in the worst year within the 3-year averaging period
23 (Schmidt et al., 2005). As a result, in areas that just attain the standards, EPA's communication
24 to the public through the Air Quality Index will on one hand indicate that the general level of air
25 quality is satisfactory (since the standards are being met), but on the other hand it may identify
26 many days throughout the year as being unhealthy, particularly for sensitive groups. Thus, staff
27 does not believe it would be appropriate to consider specifying the form in terms of an even
28 lower percentile value.
f
January 2005
5-52
Draft - Do Not Quote or Cite
-------
1 In considering differences between 98th and 99th percentile forms, staff believes it is
2 appropriate to take into consideration the relative risk reduction afforded by these alternative
3 forms at the same standard level. Based on the risk assessment results discussed in Chapter 4,
4 and the risk reductions associated with alternative levels and forms discussed above in sections
5 5.3.4 and 5.3.5, staff notes that the 99th percentile can, in some instances, result in appreciably
6 greater risk reductions in particular areas than that associated with a standard at the same level
7 but with a 98th percentile form. More specifically, staff considered the differences in risk
8 reductions associated with attaining alternative standards with 98th and 99th percentile forms in
9 five example urban areas (Detroit, Los Angeles, Philadelphia, Pittsburgh, and St. Louis). In
10 looking at estimated risk reductions associated with meeting a 24-hour standard of 30 |ig/m3, for
11 - example, estimated risk reductions for mortality associated with long-term exposures were
12 higher with the use of a 99th percentile form in some areas by approximately 15%, ranging up to
13 over 50% higher in Los Angeles. For estimated risk reductions for mortality associated with
14 short-term exposures, the use of a 99th percentile form resulted in estimated reductions that were
15 higher by less than 10% to over 30% across the five urban areas.
16 Staff also analyzed the available air quality data from 2001 to 2003 to compare the 98th
17 and 99th percentile forms in terms of the numbers of days that would be expected to exceed the
18 level of the standard (on average over 3 years and in the worst year within a 3-year averaging
19 period) and by how much the standard would typically be exceeded on such days (Schmidt et al.,
20 2005). In so doing, as noted above, staff observes that the current 98th percentile form allows the
21 level of the standard to be exceeded seven days per year, on average (with every-day
22 monitoring), and finds that this form allows up to about 20 days in the worst year within the 3-
23 year averaging period. A 99th percentile form would allow the level of the standard to be
24 exceeded three days per year, on average (with every-day monitoring), while allowing up to
25 about 13 days in the worst year within the 3-year averaging period. Further, staff observes that
26 for either form, daily peak concentrations in the upper 1 to 2% of the annual air quality
27 distributions are within 5 |ig/m3 of the 98* or 99th percentile value somewhat more than half the
January 2005
5-53
Draft - Do Not Quote or Cite
-------
1 time and are almost always within 10 to 15 jig/m3 above the 98th or 99th percentile values, with '
2 very few excursions above this range.5
3 Based on these considerations, staff recommends either retaining the 98th percentile form
4 or revising it to be based on the 99th percentile air quality value. In selecting between these
5 alternative forms, staff believes primary consideration should be given to the estimated level of
6 risk reduction mat is associated with standards based on either form. Staff also believes it is
7 appropriate to take into account whether the 24-hour standard is set to supplement protection
8 afforded by an annual standard, or is intended to be the primary basis for providing protection
9 against effects associated with short-term exposures. In choosing between forms of alternative
10 standards that provide generally equivalent levels of public health protection, staff believes it is
11 appropriate to consider the implications from a public health communication perspective of the
12 extent to which alternative forms allow different numbers of days in a year to be above the level
13 of the standard in areas that attain Hie standard. In particular, staff notes that the use of a 99th
14 percentile form would result in a more consistent public health message to the general public in
15 the context of the wide-spread use of the Air Quality Index.
16 5.3.7 Summary of Staff Recommendations on Primary PMZ s NAAQS
17 Staff recommendations for the Administrator's consideration in making decisions on the
18 primary PM25 standards, together with supporting conclusions from sections 5.3.1 through 5.3.6,
19 are briefly summarized below.- Staff recognizes that selecting from among alternative standards
20 will necessarily reflect consideration of the qualitative and quantitative uncertainties inherent in
21 the relevant evidence and in the assumptions that underlie the quantitative risk assessment. In
22 recommending these alternative suites of primary standards and ranges of levels for
23 consideration, staff is mindful that the Act requires standards to be set that are requisite to
5 This analysis also looked at the number of days in which the reported air quality values were "flagged" as
being heavily influenced by natural events (including forest fires, dust storms) or exceptional events, for which the
Agency's natural and exceptional events policies would likely apply. While flagged days generally account for less
than 1% of all reported 24-hour average PMj 5 concentrations, they account for about 40% of the highest 100 days
across the country. In looking at the reported values that are above the 99* or 98th percentiles of the distribution of
values, approximately 3 to 6% of the highest 2% of days (above the 98th percentile) were flagged, and approximately
5 to 10% of the highest 1% of days (above the 99th percentile) were flagged.
January 2005 5-54 Draft - Do Not Quote or Cite
-------
1 protect public health with an adequate margin of safety, such that the standards are to be neither
2 more nor less stringent than necessary. Thus, the Act does not require that NAAQS be set at
3, zero-risk levels, but rather at levels that avoid unacceptable risks to public health.
4 ( 1) Consideration should be given to revising the current PM2 5 primary standards to provide
5 increased public health protection from the effects of both long- and short-term exposures
6 to fine particles in the ambient air. This recommendation is based in general on the
7 evaluation in the CD of the newly available epidemiologic, toxicologic, dosimetric, and
8 exposure-related evidence, and more specifically on the evidence of mortality and
9 morbidity effects in areas where the current standards were met, together with judgments
10 as to the public health significance of the estimated incidence of effects upon just
11 attaining the current standards.
12 (2) The indicator for fine particle standards should continue to be PM25. This
13 recommendation is based on the conclusion that the available evidence does not provide
14 a sufficient basis for replacing or supplementing a mass-based fine particle indicator with
15 an indicator for any specific fine particle component or subset of fine particles, nor does
16 it provide a basis for excluding any components; on the evaluation in the CD of air
17 quality within the interrnodal particle size range of 1 to 3 urn; and on the policy judgment
18 made in the last review to place regulatory importance on defining an indicator that
19 would more completely capture fine particles under all conditions likely to be
20 encountered across the U.S., while recognizing that some limited intrusion of small
21 coarse particles will occur in some circumstances. Consideration should be given to
22 modifying the FRM for PM2 5 based on instrumentation and operational improvements
23 that have been made since the PM25 monitoring network was deployed in 1999, and to
24 the adoption of FEMs for appropriate continuous measurement methods.
25 ( 3) Averaging times for PM2 5 standards should continue to include annual and 24-hour
26 averages to protect against health effects associated with short-term (hours to days) and
January 2005 5-55 Draft - Do Not Quote or Cite
-------
1
2
3
4
5
6
7
8
9
10
11
long-term (seasons to years) exposure periods; consideration of other averaging times,
especially on the order of one or more hours, was limited by a lack of adequate
information at this time.
( a) Consideration should be given to revising the form of the annual standard to one
based on Hie highest community-oriented monitor in an area or, alternatively, to
one with more constrained requirements for the use of spatial averaging across
community-oriented monitors.
( b) Consideration should be given to revising the form of the 24-hour standard to a
99th percentile form or, alternatively, to retaining the 98th percentile form, based in
part on considering the level of risk reduction likely to result from a standard
using either form.
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
t
( 4) Consideration should be given'to alternative suites of PM25 standards to provide
protection against effects associated with both long- and short-term exposures, taking
into account both evidence-based and risk-based considerations. Integrated
recommendations on ranges'of alternative suites of standards that, when considered
together, protect against effects associated with both long- and short-term exposures
include: ' •
( a) Staff recommends consideration of an annual PM2 5 standard at the current level
of 15 ug/m3 together with a revised 24-hour PM2 5 standard in the range of 35 to
25 ug/m3. Staff judges that such a suite of standards, particularly in conjunction
with a 99th percentile form for a 24-hour standard set at the middle to upper end of
this range, could provide an appropriate degree of protection against serious
mortality and morbidity effects associated with long- and short-term exposures to
fine particles.
( b) Alternatively, staff also recommends consideration of a revised annual PM2 5
standard, within the range of 14 to 12 ng/m3, together with a revised 24-hour
PM2 5 standard to provide supplemental protection against episodic localized or
seasonal peaks, in the range of 40 to 35 ug/m3. Staff judges that such a suite of
January 2005
5-56
Draft - Do Not Quote or Cite
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
standards, particularly with an annual standard set at the middle or low end of this
range, could provide an appropriate degree of protection against serious mortality
and morbidity effects associated with long- and short-term exposures to fine
particles.
5.4 THORACIC COARSE PARTICLE STANDARDS
5.4.1 Adequacy of Current PM10 Standards
In considering the adequacy of the current PM10 standards to control thoracic coarse
particles, in conjunction with separate standards for PM25, staff has first considered the
appropriateness of using PM,0 as an indicator for thoracic coarse particles. In 1997, in
conjunction with establishing new PM2 5 standards, EPA determined that the new function of
PM,0 standards was to protect against potential effects associated with thoracic coarse particles
in the size range of 2.5 to 10 urn (62 FR 38,677). Although staff had given some consideration
to a more narrowly defined indicator that did not include fine particles (e.g., PM10.2 5), EPA
decided to continue to use PM10 as the indicator for standards to control thoracic coarse particles.
This decision was based in part on the recognition that the only studies of clear quantitative
relevance to health effects most likely associated with thoracic coarse particles used PM,0 in
areas where the coarse fraction was the predominant component of PM10, namely two fugitive
dust studies in areas that substantially exceeded the PM10 standards (62 FR 38,679). Also, the
decision reflected the fact that there was only very limited ambient air quality data then available
specifically on thoracic coarse particles, in contrast to the extensive monitoring network already
in place for PM10. In essence, EPA concluded at that time that it was appropriate to continue to
control thoracic coarse particles, which, like fine particles, are capable of penetrating to the
thoracic region of the respiratory tract, but that the only information available upon which to
base such standards was indexed in terms of PM10.6
As discussed in Chapter 1, however, in subsequent litigation regarding the 1997 PM NAAQS revisions,
the court held in part that PM10 is a "poorly matched indicator" for thoracic coarse particles in the context of a rule
that also includes PM25 standards because PMto includes PM25. 175 F. 3d. at 1054. Although the court found
"ample support" (idj for EPA's decision to regulate thoracic coarse particles, the court nonetheless vacated the 1997
revised PM10 standards for the control of thoracic coarse particles.
January 2005
5-57
Draft - Do Not Quote or Cite
f
-------
1 taking into account key uncertainties associated1 with the estimated risks. Estimates of risks
2 attributable to short-term exposure to PM10.2.5 are presented in Chapter 4 for Detroit, Seattle and
3 St. Louis, the urban areas in which the studies discussed above were conducted. These estimated
4 risks are attributable to PM10.25 concentrations above background levels, or above the lowest
5 measured levels in a given study if higher than background, so as to avoid extrapolating risk
6 estimates beyond the range of air quality upon which the concentration-response functions were
7 based.
8 In the absence of evidence for clear thresholds in any of the studies used in the risk
9 assessment, the base case estimates in the analysis reflect the linear or near-linear concentration-
10 response functions reported in the studies. To reflect the uncertainty as to whether thresholds
11 may exist within the range of air quality observed in the studies, but may not be discemable with
12 currently applied statistical methods, staff has also considered estimates of risk based on
r
13 concentration-response functions modified to incorporate various assumed hypothetical •
14 ^threshold levels. Based on the sensitivity analyses conducted as part of the risk assessment, the
15 'uncertainty associated with alternative hypothetical thresholds had by far the greatest impact on
16 estimated risks.
17 Table 5-5 summarizes the estimated PM10.25-related annual incidence of hospital
18 admissions and respiratory symptoms (cough) in children associated with short-term exposure
19 for the base case and for alternative hypothetical thresholds in the three example urban areas
20 included in the risk assessment. Staff observes that the base case estimates of cardiac-related
21 hospital admissions in Detroit are an order of magnitude greater than asthma-related admissions
22 in Seattle. Such large differences are in part attributable to the large differences in PM,0.2S,air
23 quality levels in these two areas, in which the 2003 annual average PM10.25 concentration in
24 Detroit (21.7 (ig/m3) is much higher than in Seattle (111.4; jig/m3).- Further, staff notes that the
25 2003 annual average PM10.2S concentration in St. Louis (12:0 iig/m3) is similarly far below that
26 in Detroit. In looking beyond the base case estimates, staff observes that, as expected, the risk
27 estimates are substantially smaller when concentration-response functions adjusted to reflect
28 hypothetical thresholds are considered.' At the largest assumed hypothetical threshold, estimates
r\
January 2005
5-60
Draft - Do Not Quote or Cite
-------
1 effects of the crustal contribution in thoracic coarse particles. The CD notes that particles of.
2 crustal origin may be linked with morbidity effects, or may serve as carriers for other more toxic
3 components, such as metals or organic compounds (CD, p. 9-63). The CD discusses some
*>
4 coarse particle components (e.g., metals, biogenic constituents) or sources contributing to coarse
5 particles (e.g., wood burning) that may be linked with health effects, but little evidence is
6 available on any of the components or sources within the coarse fraction at present (CD, p. 9-32).
7 Thus, as for fine particles, there is no evidence that would lead toward the selection of one or
8 more PM components as being primarily responsible for effects associated with coarse particles,
9 nor is there any component that can be eliminated from consideration. . ...
10 Taking into account the above considerations, staff concludes that a mass-based indicator
11 continues to be the most appropriate indicator for any thoracic coarse particle standards. Staff
12 recommends that such an indicator retain 10 urn as the upper cut point, and that the lower cut
13 point of 2.5 um be used so as to most clearly differentiate between thoracic coarse (PM10_2 5) and
14 fine (PM2 5) particles. In considering the evidence that suggests that high PM concentrations
15 linked with dust storm events may be of less concern, staff notes that EPA has historically used
16 natural events policies to address such issues in the implementation of PM standards. , , -.. ~"
17 In conjunction with considering PM]0.2 5 as an indicator for standards to address thoracic . ^^^
18 coarse particles, EPA is evaluating various ambient monitoring methods. This evaluation is
19 being performed through field studies of commercially ready and prototype methods to
20 characterize the measurement of thoracic coarse particles.8 The PM10_25 methods evaluation has
21 resulted in characterizing the performance of multiple PM10.2 5 measurement technologies under a
22 variety of aerosol and meteorological conditions. This characterization has demonstrated that
23 the majority of commercially available methods for the measurement of PM10.25 have good
24 .precision and are well correlated with filter-based gravimetric methods such as the difference
25 method that has primarily been used to date (i.e., operation of collocated PM10 and PM25 low
26 volume FRMs and calculating PM10.25 by difference). EPA is working with the instrument
27 manufacturers to address design issues that should reduce biases that have been observed among
28 methods, in preparation for another field study examining the performance of the methods.
This work is being done in consultation with the CASAC AAMM Subcommittee.
January 2005 5r63 Draft - Do Not Quote or Cite \ /
-------
1 taking into account key uncertainties associated with the estimated risks. Estimates of risks
2 attributable to short-term exposure to PM10.Z5 are presented in Chapter 4 for Detroit, Seattle and
3 St. Louis, the urban areas in which the studies discussed above were conducted. These estimated
4 risks are attributable to PM10.2.5 concentrations above background levels, or above the lowest
5 measured levels in a given study if higher than background, so as to avoid extrapolating risk
6 estimates beyond the range of air quality upon which the concentration-response functions were
7 based.
"J
8 In the absence of evidence for clear thresholds in any of the studies used in 1he risk
9 assessment, the base case estimates in the analysis reflect the linear or near-linear concentration-
\ 10 response functions reported in the studies. To reflect the uncertainty as to whether thresholds
11 may exist within the range of air quality observed in the studies, but may not be discernable with
12 currently applied statistical methods, staff has also considered estimates of risk based on
13 concentration-response functions modified to incorporate various assumed hypothetical
14 threshold levels. Based on the sensitivity analyses conducted as part of the risk assessment, the
15 uncertainty associated with alternative hypothetical thresholds had by far the greatest impact on
16 estimated risks.
17 Table 5-5 summarizes the estimated PM10_2 5-related annual incidence of hospital
18 admissions and respiratory symptoms (cough) in children associated with short-term exposure
19 for the base case and for alternative hypothetical thresholds in the three example urban areas
20 included in the risk assessment. Staff observes that the base case estimates of cardiac-related
21 hospital admissions in Detroit are an order of magnitude greater than asthma-related admissions
22 in Seattle. Such large differences are in part attributable to the large differences in PM10.25,air
23 quality levels in these two areas, in which the 2003 annual average PM10.2 s concentration in
24 Detroit (21.7 |ig/m3) is much higher than in Seattle (11'.4. n.g/m3).- Further, staff notes that the
25 2003 annual average PM10.25 concentration in St. Louis (12.0 jig/m3) is similarly far below that
26 in Detroit. In looking beyond the base case estimates, staff observes that, as expected, the risk
27 estimates are substantially smaller when concentration-response functions adjusted to reflect
28 hypothetical thresholds are considered. At the largest assumed hypothetical threshold, estimates
January 2005
5-60
Draft- Do Not Quote or Cite
-------
1 in Detroit are 50 percent smaller than base case estimates, whereas in St. Louis estimates are 90
2 percent smaller.
3
4
5
6
Table 5-5 Estimated PM10 2 rrelated Annual Incidence of Hospital Admissions and
Cough in Children with 2003 Air Quality in Areas that Meet the Current
PM10 Standards (Base Case and Assumed Alternative Hypothetical
Thresholds)
Detroit: hospital admissions for
ischemic heart disease
Seattle: hospital admissions for
asthma (age <65)
St. Louis: days of cough in
children
Short-term Exposure
Base case
Estimate,
95% Cl
654
169 to 1083
27
Oto65
27,000
11, 000 to 40,900
Assumed Hypothetical Thresholds
10 pg/m3
505
131 to 836
11
Oto26
11,500
4,700 to 17,400
15 (jg/m3
386
100 to 636
4
OtolO
5,400
2,200 to 8,000
20 Mg/m3
294
77 to 483
1
Oto3
2,600
1,100 to 3,700
7
8
9
10
11
12
13 Beyond the specific health endpoints presented in Table 5-5, for which sensitivity
14 analyses have been done, staff notes that hundreds of additional hospital admissions for other
15 cardiac- and respiratory-related diseases are also estimated in Detroit, based on risk assessment
16 results presented in Chapter 4, as are thousands of additional days in which children are likely to
17 experience other symptoms of the lower respiratory tract in St. Louis. In considering these
18 limited estimates, even when hypothetical thresholds are assumed, staff concludes that they are
19 indicative of risks that can reasonably be judged to be important from a public health
20 perspective, especially in areas in which PM,0.25 concentrations approach those observed in
21 Detroit.
22 In considering the evidence and risk estimates for thoracic coarse particles discussed
23 above, and the related limitations and uncertainties, staff concludes that this information is
24 sufficient to support consideration of revised standards for thoracic coarse particles to afford
January 2005
5-61
Draft - Do Not Quote or Cite
f
-------
1 protection from effects related to short-term exposure to current ambient levels of PM10.2.5 in
2. some urban areas. Staff conclusions and recommendations on an indicator and associated
3 monitoring methods, averaging times, and alternative levels and forms for thoracic coarse
4 particle standards mat would afford an appropriate degree of protection from'such effects are
5 discussed in the following sections.
6 5.4.2 Indicators ' ' ,
7 Section 5.4.1 above discusses EPA's decision in 1997 to continue to use PM10 as the-
8 indicator for standards intended to protect against the effects most likely associated with thoracic
9 coarse particles. In considering the adequacy of such standards, staff has taken into account
10 information now available on health effects and air quality in which thoracic coarse particles are
11 indexed by PM10.2 5, concluding that such information should form the basis for consideration of
.12 standards for thoracic coarse particles using an indicator that does not include the fine fraction of
13 PM10.
14 The CD concludes that the recent scientific information supports EPA's previous
15 decision to use an indicator based on PM mass, as discussed above in section 5.3.2 for fine
16 particles. In addition, currently available information from dosimetric studies supports retaining
17 10 urn as the appropriate cut point for particles capable of penetrating to the thoracic regions of
18 the lung. In conjunction with PM2 5 standards, an appropriate mass-based indicator for thoracic
19 coarse particles thus would be PM10.2 5. As noted above, this is the indicator that has been used
20 to index thoracic coarse particles in newly available epidemiologic studies and in
21 ' characterizations of air quality.
22 There is limited evidence to support consideration of other indicators for thoracic coarse
23 particles, such as individual components within mis PM fraction. In general, less is known about
24 the composition of thoracic coarse particles than fine particles. Even less evidence is available
25 from health studies that would allow identification of specific components or groups of
26 components of coarse particles that may be more closely linked with specific health outcomes.
27 While several studies have suggested that the crustral or geological component of fine particles
28 is not significantly associated with mortality (CD, p. 8-66), no studies have focused on potential
t
January 2005
5-62
Draft - Do Not Quote or Cite
-------
1 effects of the crustal contribution in thoracic coarse particles, The CD notes that particles of
2 crustal origin may be linked with morbidity effects, or may serve as carriers for other more toxic
3 components, such as metals or organic compounds (CD, p. 9-63). The CD discusses some
£.
4 coarse particle components (e.g., metals, biogenic constituents) or sources contributing to coarse
5 particles (e.g., wood burning) that may be linked with health effects, but little evidence is
6 available on any of the components or sources within the coarse fraction at present (CD, p.9-32).
7 Thus, as for fine particles, there is no evidence that would lead toward the selection of one or
8 more PM components as being primarily responsible for effects associated with coarse particles,
9 nor is there any component that can be eliminated from consideration.
10 Taking into account the above considerations, staff concludes that a mass-based indicator
11 continues to be the most appropriate indicator for any thoracic coarse particle standards. Staff
12 recommends that such an indicator retain 10 um as the upper cut point, and that the lower cut
13 point of 2.5 urn be used so as to most clearly differentiate between thoracic coarse (PM10.25) and
14 fine (PM2 5) particles. In considering the evidence that suggests that high PM concentrations
15 linked with dust storm events may be of less concern, staff notes that EPA has historically used
16 natural events policies to address such issues in the implementation of PM standards.
17 In conjunction with considering PM10.2 5 as an indicator for standards to address thoracic
18 coarse particles, EPA is evaluating various ambient monitoring methods. This evaluation is
19 being performed through field studies of commercially ready and prototype methods to
20 characterize the measurement of thoracic coarse particles.8 The PM10.25 methods evaluation has
21 resulted in characterizing the performance of multiple PM,0.2 5 measurement technologies under a
22 variety of aerosol and meteorological conditions. This characterization has demonstrated that
23 the majority of commercially available methods for the measurement of PM10.25 have good
24 precision and are well correlated with filter-based gravimetric methods such as the difference
25 method that has primarily been used to date (i.e., operation of collocated PM10 and PM25 low
26 volume FRMs and calculating PM10_25 by difference). EPA is working with the instrument
27 manufacturers to address design issues that should reduce biases that have been observed among
28 methods, in preparation for another field study examining the performance of the methods.
This work is being done in consultation with the CASAC AAMM Subcommittee,
January 2005 5-63 Draft - Do Not Quote or Cite
t
-------
1
2
3
4
5
6
7
8
9
10
11
EPA has begun the process of examining data quality objectives for potential PM10_2 5
standards. On the basis of preliminary analyses, it is apparent that greater sampling frequency
will be important due to the high variability of PM10.2 s in Ihe atmosphere; this would be
particularly important for a short-term PM10.2 5' standard. Due to the resource intensive nature of
filter sampling on a daily basis, staff believes that it will be critical to include continuous
monitoring in any network deployment strategy for a possible PM10.2 s standard. In addition to
providing high temporal resolution to PMj0.2 5 data, continuous monitors would also support
public reporting of PM10_2 5 episodes and inclusion of PM10_2 5 in an air quality forecasting
program. As noted above and elsewhere in this document, PM10.25 is more highly variable in the
atmosphere than PM2 5, such that spatial robustness will be a particularly important consideration
in monitoring network design.
12 5.4.3 Averaging Times
13 In the last review, EPA retained both annual and 24-hour standards to provide protection
14 against the known and potential effects of short- and long-term exposures to thoracic coarse
15 particles (62 FR at'38,677-79). This decision was based in part on qualitative considerations
16 related to the expectation that deposition of thoracic coarse particles in the respiratory system
17 could aggravate effects in individuals with asthma. In addition, quantitative support came from
18 limited epidemiologic evidence suggesting that aggravation of asthma and respiratory infection
19 and symptoms may be associated with daily or episodic increases in PM10, where dominated by
20 thoracic coarse particles including fugitive dust. Further, potential build-up of insoluble thoracic
21 coarse particles in the lung after long-term exposures to high levels was also considered
22 plausible.
23 Information available in this review on thoracic coarse particles, while still limited,
24 represents a significant expansion of the evidence available in the last review. As discussed '
25 above in section 5.4.1, a number of epidemiologic studies are now available that report
26 statistically significant associations between short-term (24-hour) exposure to PMi0.2 s and
27 morbidity effects, which the CD concludes are suggestive of causal associations, and mortality,
28 which the CD concludes provide less support for possible causal associations. With regard to
January 2005
5-64
Draft - Do Not Quote or Cite
-------
1 long-term exposure studies, while one recent study reported a link between reduced lung
2 function growth and long-term exposure to PM,0.2 5 and PM2.5? the CD concludes that the
3 . evidence is not sufficient to be suggestive of a causal association. Staff notes that no evidence is
4 available to suggest associations between PM10_2 5 and very short exposure periods of one or
5 more hours.
6 Based on these considerations, staff concludes that the newly available evidence provides
7 support for considering a 24-hour standard for control of thoracic coarse particles, based
8 primarily on evidence suggestive of associations between short-term exposure and morbidity
9 effects, reflecting as well the potential for associations with mortality. Noting the absence of
10 evidence judged to be even suggestive of an association with long-term exposures, staff
11 concludes that there is little support for an annual standard, although staff recognizes that it may
12 be appropriate to consider an annual standard to provide a margin of safety against possible
13 effects related to long-term exposure to thoracic coarse particles that future research may reveal.
14 Staff observes, however, that a 24-hour standard that would reduce 24-hour exposures would
15 also likely reduce long-term average exposures, thus providing some margin of safety against the
16 possibility of health effects associations with long-term exposures.
17 5.4.4 Alternative PM10.2.S Standards to Address Health Effects Related to Short-term
18 Exposure
19 In the last review, EPA's decision to retain the level of the 24-hour PMIO standard of 150
20 ug/mf (with revision of the form of the standard) was based on two community studies of
21 exposure to fugitive dust that showed health effects only in areas experiencing large exceedances
22 of that standard, as well as on qualitative information regarding the potential for health effects
23 related to short-term exposure to thoracic coarse particles. Because of the very limited nature of
24 this evidence, staff concluded mat while it supported retention of a standard to control thoracic
25 coarse particles, it provided no basis for considering a more protective standard. However,
26 because of concerns about the expected-exceedance-based form of the 1987 PM!0 standard,
27 primarily related to the stability of the attainment status of an area overtime and complex data
28 handling conventions needed in conjunction with less-than-every-day sampling, EPA adopted a
January 2005
5-65
Draft - Do Not Quote or Cite
-------
1 concentration-based form for the 24-hour standard, as was done for the 24-hour PM2 5 standard,
2 as discussed above in section 5.3.6. In making this change, EPA selected a 99th percentile form,9
3 in contrast to the 98th percentile form adopted for the 24-hour PM2 5 standard, so as not to allow
4 any relaxation in the level of protection that had been afforded by the previous 1-expected-
5 exceedance form. •
6 Since the last review, as discussed above in section 5.4.1, new evidence specific to
7 PM10.2 5 has become available mat suggests associations between short-term PM10.2 5
8 concentrations and morbidity effects and, to a lesser degree,- mortality. In considering this
9 evidence as a basis for setting a 24-hour PM10_2 5 standard, staff has focused on U.S. and
10 Canadian short-term exposure studies of thoracic coarse particles (Appendix 3 A). In so doing,
11 staff has taken into account reanalyses that addressed GAM-related statistical issues and has
12 considered the extent to which the studies report statistically significant and relatively precise
13 relative risk estimates; the reported associations are robust to co-pollutant confounding and
14 alternative modeling approaches; and the studies used relatively reliable air quality data. In
15 particular, staff has focused first on those specific morbidity studies that provide evidence of
16 associations in areas that would have met the current PM10 standards during the time of the
17 study.
18 As an initial matter, staff recognizes, as discussed in Chapter 3 (section 3.6.6), that these
19 short-term exposure studies provide no evidence of clear thresholds, or lowest-observed-effects
20 levels, in terms of 24-hour average concentrations. Staff notes that in the one study that explored
21 a potential PM10.2 5 threshold, conducted in Phoenix, no evidence of a threshold was observed for
22 PM10.2 5, even though that study provided some suggestion of a potential threshold for PM2 5. The
23 CD concludes that while' there is no evidence of a clear threshold within the range of air quality
24 observed in the studies, for some health endpoints (such as total nonaccidental mortality) it is
25 likely to be extremely difficult to detect threshold levels (CD, p. 9-45). As a consequence, mis
26 body of evidence is difficult to translate directly into a specific 24-hour standard that would
27 protect against all effects associated with short-term exposures. Staff notes that the distributions
9 As noted above, the court vacated the 1997 24-hour PMlo standard that had been revised to incorporate a
1 99th percentile form.
January 2005
5-66
Draft - Do Not Quote or Cite
-------
1 of daily PM10.25 concentrations in these studies often extend down to or below background
2 levels, such that the likely range of background concentrations across the U.S., as discussed in
3 Chapter 2, section 2.6, could be a relevant consideration in this policy evaluation. Staff
4 recognizes, however, that there is insufficient data to estimate daily distributions of background
5 PMi0.25 levels (as was done for background PM25 levels, as discussed in Chapter 2, section 2.6).
6 Being mindful of the difficulties posed by uncertainties related to potential thresholds and
7 insufficient data to characterize daily distributions of PM10.2 5 background concentrations, as well
8 as the limited nature of the available evidence, staff has considered the short-term exposure
9 epidemiologic evidence as a basis for alternative 24-hour PM10.2 5 standards. In so doing, staff
10 has focused on the upper end of the distributions of daily PM10.2.5 concentrations, particularly in
11 terms of the 98th and 99th percentile values, consistent with the forms considered in section 5.3.6
12 above for PM25. In looking at the specific morbidity studies identified in section 5.4.1 that
13 report statistically significant associations with respiratory- and cardiac-related hospital
14 admissions in areas that had ambient air quality levels that would have met the current PM10
15 standards at the time of the study, including studies in Toronto (Burnett et al., 1997), Seattle
16 (Sheppard et al.,1999, 2003), and Detroit (Lippmann et al., 2000; Ito, 2003), staff notes that the
17 reported 98th percentile values range from approximately 30 to 36 ug/m3 in all three areas, and
18 the 99th percentile values range from 36 to 40 ug/m3 (Ross and Langstaff, 2005).
19 In looking more closely at these studies, staff recognizes that the uncertainty related to
20 exposure measurement error is potentially quite large in epidemiologic studies linking effects to
21 PM10.25 air quality measures. For example, in looking specifically at the Detroit study, staff
22 notes that the PM,0.2 5 air quality values were based on air quality monitors located in Windsor,
23 Canada The study authors determined that the air quality values from these monitors were
24 generally well correlated with air quality values monitored in Detroit, where the hospital
25 admissions data were gathered, and, thus concluded that these monitors were appropriate for use
26 in exploring the association between air quality and hospital admissions in Detroit. Staff has
27 observed, however, mat the PM10.2 5 levels reported in this study are significantly lower than the
28 PMto.25 levels measured at some of the Detroit monitors in 2003 - an annual mean level of 13.3
29 ug/m3 is reported in the study, based on 1992 to 1994 data, as compared to an average annual
January 2005
5-67
Draft - Do Not Quote or Cite
-------
1 mean level of 21.7 ug/m3 measured at two urban-center monitors in 2003 (which is used as the
2 basis for the risk assessment presented in Chapter'4). This observation prompted staff to further
3 explore the comparison between PM10.2 5 levels monitored at Detroit arid Windsor sites.' This
4 exploration has shown that in'fecent years, based on available Windsor and Detroit data from
5 1999 to 2003, the Windsor monitors used in this study typically have recorded PM1(W5 levels that
6 are generally less man half the levels recorded at urban-center Detroit monitors, though the
7 concentrations measured in Windsor are more similar to concentrations reported for suburban
8 areas well outside the city (Ross and Langstaff, 2005). These observations lead staff to conclude
9 that the statistically significant, generally robust association with hospital admissions in Detroit
10 reflects population exposures that may be appreciably higher than what would be estimated
11 using data from the Windsor monitors. Taking these observations into account, staff nonetheless
12 believes that these studies in general, and the Detroit study in particular, do provide evidence of
13 associations between short-term exposures to PM10.25 and hospital admissions. Staff does
14 conclude, however, that the association observed in the Detroit study, which' staff judges to be
15 the strongest of these studies, likely reflects exposure levels potentially much higher in the
16 central city area than those reported in that study. Based on this information, staff believes that
17 alternative 24-hour PM]0_2 5 standards appropriate for consideration in this review need not ;
18 necessarily extend to levels down to or below the ranges reported in these studies in order to'
19 provide protection from the morbidity effects related to short-term exposures to PM,0.2 5.
20 Staff has also looked at the evidence from U.S. and Canadian studies thai report
21 statistically significant and generally robust associations with mortality and short-term exposures
22 to PM10.25. As discussed in section 9.2.3 of the CD, the evidence associating mortality with
23 short-term exposures to PM10_25 is too uncertain to infer a likely causal relationship, although it
24 is suggestive of a possible causal relationship. Staff identified two such studies, conducted in
25 Phoenix (Mar et al, 2000,2003) and Coachella Valley, CA (Ostro et al., 2000,2003), that report
26 98th percentile PM10_25 values of approximately 70 and 107 ug/m3, and 99th percentile values of
27 75 and 134 ng/m3, respectively. Staff notes that these studies were conducted in areas with air
28 quality levels that would not have met the current PM10 standards. A staff analysis of PM10 and
29 estimated PM10.25 concentrations from the AQS database for 2001 to 2003 suggests that 98th
January 2005
5-68
Draft - Do Not Quote or Cite
-------
s
S «
If
ss
Jg
-1
"
I "
c •?
8 s
g g
fli PH
1!
s
S O
§3
•3"
12
•s
l.-I
o «
eo
•5
•g
CD
*&•
\M
8
R
Jj
€ ^
1l
« J
!l
a
8
a
s
s
S
s
^
s
s
ffi
s
s
88
•s &.
and
ilia
u
a
o
O)
I
I
S
cn
-------
1 Beyond looking direcHy at the relevant epidemiologic evidence, staff has also considered
2 the extent to which the PM,0.2 5 risk assessment results discussed in Chapter 4 can help inform
3 consideration of alternative 24-hour PM10_2.5 standards. While one of the goals of the PMi0.2 s
4 risk assessment was to provide estimates of the risk reductions associated with just meeting
5 alternative PM10.2 5 standards, staff has concluded that the nature and magnitude of the
6 uncertainties and concerns associated with this portion of the risk assessment weigh against use
7 of these risk estimates as a basis for recommending specific standard levels. These uncertainties
8 and concerns include, but are not limited to the following:
9 • as discussed above, concerns that the current PM10.2 5 levels measured at ambient
10 monitoring sites in recent years may be quite different from the levels used to
11 characterize exposure in the original epidemiologic studies based on monitoring sites in
12 different location, thus possibly over- or underestimating population risk levels; "
13 • greater uncertainty about the reasonableness of the use of proportional rollback to
14 simulate attainment of alternative PM10.2.s daily standards in any urban area due to the
15 limited availability of PM10.2 5 air quality data over time;
16 • concerns that the locations used in the risk assessment are not representative of urban
17 areas in the U.S. that experience the most significant 24-hour peak PM10.2.5
18 concentrations, and thus, observations about relative risk reductions associated with
19 alternative standards may not be relevant to the areas expected to have the greatest health
20 'risks associated with elevated ambient PM10L25 levels; and
21 • concerns about the much smaller health effects database that supplies the C-R
22 relationships used in the risk assessment, compared to that available for PM2 5, which
23 limits our ability to evaluate the robustness of the risk estimates for the same health
24 endpoints across different locations.
25 In summary, in considering the relevant epidemiologic evidence and the related
26 limitations and uncertainties, staff concludes that there is support for considering a 24-hour
27 PM10.2 5 standard to replace the current PM10 standards to provide protection against health
28 effects associated with short-term exposures to thoracic coarse particles. In looking primarily at
29 the evidence of associations between short-term exposure to PM10.2.5 and mortality, staff
30 concludes that it is appropriate to consider a 24-hour standard in the range of 65 to 75 ug/m3,
31 with a 98th percentile form, or in the range of 75 to 85 ug/m3, with a 99th percentile form. A
January 2005 5-72 Draft ~ Do Not Quote or Cite
-------
1 standard set within either of these ranges could be expected to provide a margin of safety to
2 protect against the potential, but uncertain, mortality effects of PM,0.2 5, while continuing to
3 provide protection against the effects of PM10.2 5 associated with high levels of PM10 that were the
4 basis for the decision made by EPA in 1997 to retain the levels of the PMj0 standards. In
5 addition, staff observes that several epidemiologic studies have reported associations with
6 morbidity effects in areas with lower PM10.2 5 that could support consideration of standard levels
7 as low as approximately 30 ug/m3,98th percehtile, or 35 ug/m3, 99th percentile.
8 Staff recognizes, however, that the epidemiologic evidence on morbidity and mortality
9 effects related to PM10.25 exposure is very limited at this time. A key area of uncertainty in this
10 evidence is the potentially quite large uncertainty related to exposure measurement error for
11 PM10.2.5, as compared with fine particles. PM10.2 5 concentrations can vary substantially across a
12 metropolitan area and thoracic coarse particles are less able to penetrate into buildings than fine
13 particles; thus, the ambient concentrations reported in epidemiologic studies may not well
14 represent area-wide population exposure levels. Other key uncertainties include the lack of
15 information on the composition of thoracic coarse particles and the effects of thoracic coarse
16 particles from various sources, and the lack of evidence on potential mechanisms for effects of
17 thoracic coarse particles. Staff believes that taking these uncertainties into account leads to
18 consideration of standard levels toward the upper end of the ranges identified above.
19 5.4.5 Summary of Staff Recommendations on Primary PMto.2 s NAAQS
20 Staff recommendations for the Administrator's consideration in making decisions on
21 standards for thoracic coarse particles, together with supporting conclusions from sections 5.4.1
22 through 5.4.4, are briefly summarized below. In making these recommendations, staff is mindful
23 that the Act requires standards to be set that are requisite to protect public health with an
24 adequate margin of safety, such that the standards are to be neither more nor less stringent than
25 necessary. Thus, the Act does not require that NAAQS be set at zero-risk levels, but rather at
26 levels that avoid unacceptable risks to public health.
January 2005
5-73
Draft - Do Not Quote or Cite
-------
1 (1) The current primary PM10 standards should be revised in part by replacing the PM10
2 indicator with an indicator of thoracic coarse particles that does not include fine particles.
3 Any such revised standards should be based on available health effects evidence and air
4 quality data generally indexed by PMJO_2 5, to provide public health protection more
5 specifically directed toward effects related to exposure to thoracic coarse particles in the
6 ambient air.
7 (2) The indicator for a thoracic coarse particle standard should be PM ]0_2 5, consistent with
8 the recommendation made in section 5.3.7 to retain PM25 as the indicator for fine particle
9 standards.
10 ( a) As noted above, this recommendation is based primarily on the evaluation in the
1 1 CD of air quality within the intermodal particle size range of 1 to 3 um and the
12 . policy judgment made in the last review to place regulatory importance on
»
13 - . defining an indicator that would more completely capture fine particles under all
14 , conditions likely to be encountered across the U.S., while.recognizing that some
15 limited intrusion of small coarse particles will occur in some circumstances.
16 ( b) In support of this recommendation, work should continue on the development of
17 an FRM for PM10.2 5 based on the ongoing fiel'd program to evaluate various types
18 of monitors, and consideration should be given to the adoption of FEMs for
19 appropriate continuous measurement methods.
20 ( 3) A 24-hour averaging time should be retained for a PM^s standard to protect against
21 health effects associated with short-term exposure periods, with consideration given to
22 the use of either a 98th or 99th percentile form. Consideration could also be given to
23 retaining an annual averaging time, in considering the appropriate margin of safety
24 against possible health effects that might be associated with long-term exposure periods.
25 ( 4) Consideration should be given to setting a 24-hour PM10.2 5 standard about as protective
26 as the current daily PM10 standard, with a level in the range of approximately 65 to 75
January 2005 5-74 Draft - Do Not Quote or Cite
-------
1 Hg/m3, 98* percentile, or approximately 75 to 85 ng/rn3, 99th percentile. Staff also
2 believes there is some support for consideration of a PM10.2 5 standard level down to
3 approximately 30 ug/m3, 98th percentile, or 35 ng/m3,99th percentile, while recognizing
4 that a standard set at such a relatively low level would place a great deal of weight on
5 very limited and uncertain epidemiologic associations. Consideration of PM10.25
6 standards within Ihe ranges recommended above, and design considerations for an
7 associated PM10.2 5 monitoring network, should take into account the especially large
8 variability seen in currently available information on ambient concentrations and
9 composition of PM10.25.
10 5.5 SUMMARY OF KEY UNCERTAINTIES AND RESEARCH
11 RECOMMENDATIONS RELATED TO SETTING PRIMARY PM STANDARDS
12 Staff believes it is important to continue to highlight the unusually large uncertainties
13 associated with establishing standards for PM relative to other single component pollutants for
14 which NAAQS have been set. Key uncertainties and staff research recommendations on health-
15 related topics are outlined below. In some cases, research in these areas can go beyond aiding in
16 standard setting to aiding in the development of more efficient and effective control strategies.
17 Staff notes, however, that a full set of research recommendations to meet standards
18 implementation and strategy development needs is beyond the scope of this discussion.
19 The 1996 PM Staff Paper included a discussion of uncertainties and research
20 recommendations (EPA, 1996b, pp. VII-41-44) that addressed the following issues related to
21 understanding health effects associated with exposure to PM:
22 • lack of demonstrated biological mechanisms for PM-related effects, •
23 • potential influence of measurement error and exposure error,
24 • potential confounding by copollutants,
25 • evaluation of the effects of components or characteristics of particles,
26 • the shape of concentration-response relationships,
27 • methodological uncertainties in epidemiological analyses,
28 • the extent of life span shortening,
January 2005 % 5-75 Draft - Do Not Quote or Cite
-------
1
2
characterization of annual'and daily background concentrations, and
understanding of the effects of coarse fraction particles.
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
As has been discussed in depth in the CD, especially in Chapters 5 through 8, an
extensive body of new studies related to understanding health effects associated with exposure to
PM is now available that provides important information on many of the topics listed above. For
example, regarding the lack of demonstrated biological mechanisms, new evidence from
toxicologic and controlled human exposure studies has provided information on an array of
potential mechanisms for effects on the cardiac and respiratory systems, as discussed in Chapters
7 and 9 of the CD. Still, the CD emphasizes that much remains to be learned to fully understand
the pathways or mechanisms by which PM is linked with different health endpoints. For each of
> *
the issues listed above, new evidence has become available that helps to reduce uncertainties,
although uncertainty has been reduced in some areas more than others. Staff has identified the
following key uncertainties and research questions that have been highlighted in this review of
the health-based primary standards
(1)
The body of evidence on effects of thoracic coarse particles has been expanded, but the
uncertainties regarding thoracic coarse particles are still much greater than those for fine
particles. As discussed in Chapter 2, the spatial variability of thoracic coarse particles is
generally greater than that for fine particles, which will increase uncertainty in the
associations between health effects and thoracic coarse particles measured at central site
monitors. Additional exposure research is needed to understand the influence of
measurement error and exposure error on thoracic coarse particle epidemiology results.
In addition, little is known about coarse particle composition, and less about the health
effects associated with individual components or sources of thoracic coarse particles, but
it is possible mat there are components of thoracic coarse particles (e.g., crustal material)
that are less likely to have adverse effects, at least at lower concentrations, than other
components.
January 2005
5-76
Draft - Do Not Quote or Cite
-------
1 (2) Identification of specific components, properties, and sources of fine particles that are
2 linked with health effects remains an important research need. Available evidence
3 provides no basis for expecting that any one component would be solely responsible for
4 PM2 5-related effects, but it is likely that some components are more closely linked with
5 specific effects than others. Continued source characterization, exposure,
6 epidemiological, and toxicological research is needed to help identify components,
7 characteristics, or sources of particles that may be more closely linked with various
8 specific effects to aid in our understanding of causal agents and in the development of
9 efficient and effective control strategies for reducing health risks. Conducting human
10 . exposure research in parallel with such health studies will help reduce the uncertainty
11 associated with interpreting health studies and provide a stronger basis for drawing
12 conclusions regarding observed effects.
13 (3) An important aspect in characterizing risk and making decisions regarding air quality
14 standard levels is the shape of concentration-response functions for PM, including
15 identification of potential threshold levels. Recent studies continue to show no evidence
16 for a clear threshold level in relationships between various PM indicators and mortality,
17 within the range of concentrations observed in the studies, though some studies have
18 suggested potential levels.
19 (4) The relationship between PM and other air pollutants in causing health effects remains an
20 important question in reducing public health risk from air pollution. Numerous new
21 analyses have indicated that associations found between PM and adverse health effects
22 are not simply reflecting actual associations with some other pollutant. However, effects
23 have been found with the gaseous co-pollutants, and it is possible that pollutants may
24 interact or modify effects of one another. Further understanding of the sources,
25 exposures, and effects of PM and other air pollutants can assist in the design of effective
26 strategies for public health protection.
27 (5) Methodological issues in epidemiology studies were discussed at length in the previous
28 review, and it appeared at the time that the epidemiology study results were not greatly
29 affected by selection of differing statistical approaches or methods of controlling for
January 2005
5-77
Draft - Do Not Quote or Cite
-------
1 other variables, such as weather. However, investigation of recently discovered
2 questions on the use of generalized additive models in time-series epidemiology studies
3 has again raised model specification issues. While reanalyses of studies using different
4 modeling approaches generally did not result in substantial differences in model results,
5 some studies showed marked sensitivity of the PM effect estimate to different methods of
6 adjusting for weather variables. There remains a need for further study on the selection
7 of appropriate modeling strategies and appropriate methods to control for time-varying
8 factors, such as temperature.
9 (6) Selection of appropriate averaging times for PM air quality standards is important for
10 public health protection, and available information suggests that some effects, including
11 cardiac-related risk factors, may be linked to exposures of very short duration (e.g., one
12 or more hours). Data on effects linked with such peak exposures, such as those related to
13 wildfires, agricultural burning, or other episodic events, would be an important aid to
14 public health response and communication programs. Investigation into the PM exposure
15 time periods that are linked with effects will provide valuable information both for the
16 standard-setting process and for risk communication and management efforts.
17 (7) There remain significant uncertainties in the characterization of annual and daily
18 background concentrations for fine particles and especially for thoracic coarse particles.
19 Further analyses of air quality monitoring and modeling that improved these background
20 characterizations would help reduce uncertainties in estimating health risks relevant for
21 . standard setting (i.e., those risks associated with exposure to PM in excess of background
22 levels) and would aid in the development and implementation of associated control
23 programs.
January 2005 5-78 Draft - Do Not Quote or Cite
-------
REFERENCES
2 Burnett, R. T,; Cakmak, S.; Brook, J. R; Krewski, D. (1 997) The role of particulate size and chemistry in the
3 association between summertime ambient air pollution and hospitalization for cardiorespiratory diseases.
4 Environ. Health Perspect. 105:614-620.
5 Burnett, R. T.; Brook, J.; Dann, T.; Delocla, C.; Philips, O.; Cakmak, S.; Vincent, R.; Goldberg, M. S.; Krewski, D.
6 (2000) Association between particulate- and gas-phase components of urban air pollution and daily
7 mortality in eight Canadian cities. Inhalation Toxicol. 1 2(suppl. 4): 15-39.
8 Burnett, R. T.; Goldberg, M. S. (2003) Size-fractionated particulate mass and daily mortality in eight Canadian
9 cities. In: Revised analyses of time-series studies of air pollution and health. Special report. Boston, MA:
10 . Health Effects Institute; pp. 85-90. Available: http://www.healtheffects.org/news.htm [16 May, 2003].
1 1 EPA. (1 996) Air Quality Criteria for Particulate Matter. Research Triangle Park, NC: National Center for
1 2 Environmental Assessment-RTP Office; report no. EPA/600/P-95/001 aF-cF. 3v.
1 3 EPA. (2004) Air Quality Criteria for Particulate Matter. Research Triangle Park, NC : National Center for
1 4 Environmental Assessment-RTP Office; report no. EPA/600/P-99/002aD.
1 5 Fairley, D. (1 999) Daily mortality and air pollution in Santa Clara County, California: 1 989-1 996. Environ. Health
16 Perspect. 107:637-641.
1 7 Fairley, D. (2003) Mortality and air pollution for Santa Clara County, California, 1 989-1 996. In: Revised analyses of
1 8 time-series studies of air pollution and health. Special report. Boston, MA: Health Effects Institute; pp.
19 97-106. Available: http://www.healtheffects.org/Pubs/TimeSeries.pdf [18 October, 2004]. '
20 Gauderman, W. J.; McConnell, R; Gilliland, P.; London, S.; Thomas, D.; Avol, E.; Vora, H.; Berhane, K.;
21 Rappaport, E. B.; Lurmann, F.; Margolis, H. G.; Peters, J. (2000) Association between air pollution and
22 lung function growth in southern California children. Am. J. Respir. Crit. Care Med. 162: 1383-1390.
23 Gauderman, W. J.; Gilliland, G. F.; Vora, H.; Avol, E.; Stram, D.; McConnell, R.; Thomas, D.; Lurmann, F.;
24 Margolis, H. G.; Rappaport, E. B.; Berhane, K.; Peters, J. M. (2002) Association between air pollution and
25 lung function growth in southern California children: results from a second cohort. Am. J. Respir. Crit. Care
26 Med. 166: 76-84.
27 Ito, K. (2003) Associations of particulate matter components with daily mortality and morbidity in Detroit,
28 Michigan. In: Revised analyses of time-series studies of air pollution and health. Special report. Boston,
29 MA: Health Effects Institute; pp. 143-1 56. Available: http://www.healtheffects.org/Pubs/TimeSeries.pdf
30 [12 May, 2004].
3 1 Langstaff, J. (2004). Estimation of Policy- Relevant Background Concentrations of Particulate Matter. Memorandum
32 to PM NAAQS review docket OAR-2001-0017. January 27, 2005.
33 Lippmann, M.; Ito, K.; Nadas, A.; Burnett, R. T. (2000) Association of particulate matter components with daily
34 mortality and morbidity in urban populations. Cambridge, MA: Health Effects Institute; research report 95.
35 Mar, T. F.; Norris, G. A.; Koenig, J. Q.; Larson, T. V. (2000) Associations between air pollution and mortality in
36 Phoenix, 1995-1997. Environ. Health Perspect. 108:347-353.
January 2005 5-79 Draft - Do Not Quote or Cite
-------
1 Mat, T. F.; Morris, G. A.; Larson, T. V.; Wilson, W.'E.; Koenig, J. Q. (2003) Air pollution and cardiovascular
2 mortality in Phoenix, 1995-1997. In: Revised analyses of time-series studies of air pollution and health.
3 Special report. Boston, MA: Health Effects Institute; pp. 177-182. Available:
4 http:/Avww.healtheffects.org/Pubs/rimeSeries.pdf [18 October, 2004].
5 Ostro, B, D.; Hurley, S.; Lipsett, M. J. (1999) Air pollution and daily mortality in the Coacheila Valley, California:
6 a study of PM10 dominated by coarse particles. Environ. Res. 81: 231-238.
7 Ostro, B. D.; Broadwin, R.; Lipsett, M. J. (2000) Coarse and fine particles and daily mortality in the Coacheila
8 Valley, CA: a follow-up study. J. Exposure Anal. Environ. Epidemiol. 10:412-419.
• <
9 Ostro, B. D.; Broadwin, R.; Lipsett, M. J. (2003) Coarse particles and daily mortality in Coacheila Valley,
10 California. In: Revised analyses of time-series studies of air pollution and health. Special report. Boston,
11 MA: Health Effects Institute; pp. 199-204. Available: http://www.healtheffects.org/Pubs/TimeSeries.pdf
12 [18 October, 2004].
13 Peters, J. M.; Avol, E.; Navidi, W.; London, S. J.; Qauderman, W. J.; Lurmann, F.; Linn, W. S.; Margolis, H.;
14 ' Rappaport, E.; Gong, H., Jr.; Thomas, D. C. (1999) A study of twelve southern California communities
15 with differing levels and types of air pollution. I. Prevalence of respiratory morbidity. Am. J. Respir. Crit.
16 Care Med. 159:760-767.
17 Pope, C. A., Ill; Burnett, R. T.; Thun, M. J.; Calle, E. E.; Krewski, D.; Ito, K.; Thurston, G. D. (2002) Lung cancer,
18 cardiopulmonary mortality, and long-term exposure to fine particulate air pollution. J. Am. Med. Assoc.
19 287:1132-1141.
20 Ross, M.; Langstaff, J. (2005) Updated statistical information on air quality data from epidemiologic studies.
21 Memorandum to PMNAAQS review docket OAR-2001-0017. January 31, 2005.
22 Schmidt et al., (2005) Draft analysis of PM ambient air quality data for the PM NAAQS review. Memorandum to
23 PMNAAQS review docket OAR-2001-0017. January 31, 2005.
24 Schwartz, J.; Dockery, D. W.; Neas, L. M. (1996a) Is daily mortality associated specifically with fine particles? J.
25 Air Waste Manage. Assoc. 46:927-939.
26 Schwartz, J.; Neas, L. M. (2000) Fine particles are more strongly associated than coarse particles with acute
27 respiratory health effects in schoolchildren. Epidemiology 11:6-10.
28 Sheppard, L.; Levy, D.; Norris, G.; Larson, T. V.; Koenig, J. Q. (1999) Effects'of ambient air pollution on
29 nonelderly asthma hospital admissions in Seattle, Washington, 1987-1994. Epidemiology 10: 23-30.
30 Sheppard, L. (2003) Ambient air pollution and nonelderly asthma hospital admissions in Seattle, Washington,
31 1987-1994. Iri: Revised analyses of time-series studies of air pollution and health. Special report. Boston,
32 MA: Health Effects Institute; pp. 227-230. Available: http://www.healtheffects.org/Pubs/TimeSeries.pdf
33 [18 October, 2004].
January 2005 ' 5-80 Draft - Do Not Quote or Cite
-------
1 6. POLICY-RELEVANT ASSESSMENT OF PM-RELATED WELFARE EFFECTS
2 .
3 6.1 INTRODUCTION
4 This chapter assesses key policy-relevant information on the known and potential effects
5 on public welfare associated with ambient PM, alone and in combination with other pollutants
6 commonly present in the ambient air, drawing upon the most relevant information contained in
7 the CD and other significant reports referenced therein. The welfare effects to be considered in
8 this review of the secondary PM NAAQS include effects on visibility (section 6.2), vegetation
9 and ecosystems (section 6.3), materials (section 6.4), and climate change processes1 (section
10 6.5). For each category of effects, this chapter presents a summary of the relevant scientific
11 information and a staff assessment of whether the available information is sufficient to be •
12 considered as the basis for secondary standards distinct from primary standards for PM. Staff
13 conclusions and recommendations related to secondary standards for PM are presented in
14 Chapter 7.
15 It is important to note that discussion of PM-related effects on visibility, vegetation and
16 ecosystems, and climate change processes in Chapters 4 and 9 of the CD builds upon and
17 includes by reference extensive information from several other significant scientific reviews of
18 these topics. Most notably, these reports include the Recommendations of the Grand Canyon
19 Visibility Transport Commission (1996), the National Research Council's Protecting Visibility
20 in National Parks and Wilderness Areas (1993), reports of the National Acid Precipitation
21 Assessment Program (1991,1998), previous EPA Criteria Documents, including^//- Quality
22 Criteria for Particulate Matter and Sulfur Oxides (EPA, 1982) and Air Quality Criteria for
23 Oxides of Nitrogen (EPA, 1993), recent reports of the National Academy of Sciences (NAS,
24 2001) and the Intergovernmental Panel on Climate Change (IPCC, 1998, 2001a,b), and
25 numerous other U.S. and international assessments of stratospheric ozone depletion and global
26 climate change carried out under U.S. Federal interagency programs (e.g., the U.S. Global
27 Climate Change Research Program), the World Meteorological Organization (WMO), and the
28 United Nations Environment Programme (UNEP).
In assessing information on PM-related effects on climate change processes, consideration is given to
potential indirect impacts on human health and the environment that may be a consequence of changes in climate
and solar radiation attributable to changes in ambient PM.
January 2005 6-1 Draft - Do Not Quote or Cite
-------
1
2
3
4
5
6 •,
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33 .
6.2 EFFECTS ON VISIBILITY
Visibility can be defined as the degree to which the atmosphere is transparent to visible
light (NRCi, 1993; CD, 4-153). Visibility impairment is the most noticeable effect of fine
particles present in the atmosphere. Particle pollution degrades the visual appearance and
perceived color of distant objects to an observer and reduces the range at which they can be
distinguished from the background.
This section discusses the role of ambient PM in the impairment of visibility, drawing
upon the most relevant information contained in the CD (Chapters 4 and 9), as well as significant
reports on the science of visibility referenced therein, and building upon information presented in
section 2.8 of this document. In particular, this section includes new information on the
following topics:
• Summary findings of analyses of hourly PM25 measurements and reconstructed light
extinction coefficients for urban areas, for 2003, that demonstrate a significant
correlation between PM25 and light extinction across the U.S. during daylight hours.
• An overview of visibility programs, goals, and methods for the evaluation of visibility
impairment as a basis for standard setting, in the U.S. and abroad, illustrating Ihe
significant value placed on visual air quality, as demonstrated by efforts to improve \
visibility in national parks and wilderness areas, as well as in urban areas.
This section summarizes available information as follows: (1) information on the general
types of visibility impairment; (2) trends and conditions in Class I and non-urban areas; (3)
visibility conditions in urban areas; (4) studies of the economic value of improving visual air
quality; (5) current policy approaches to addressing visibility impairment; and (6) approaches to
evaluating public perceptions of visibility impairment and judgments about the acceptability of
varying degrees of visibility impairment.
6.2.1 Overview of Visibility Impairment
Visibility effects are manifested in two principal ways: as local impairment (e.g.,
localized hazes and plumes) and as regional haze. This distinction is significant because this
difference impacts both how visibility goals may be set and how air quality management
strategies may be devised.
January 2005
6-2
Draft - Do Not Quote or Cite
-------
1 Local-scale visibility degradation commonly occurs either in the form of a plume
2 resulting from the emissions of a specific source or small group of sources, or in the form of a
3 localized haze, such as an urban "brown cloud." Visibility impairment caused by a specific
4 source or small group of sources has been generally termed "reasonably attributable"
5 impairment. Plumes are comprised of smoke, dust, or colored gas that obscure the sky or
6 horizon relatively near sources. Sources of locally visible plumes, such as the plume from an
7 industrial facility or a burning field, are often easy to identify. There have been a limited number
8 of cases in which Federal land managers have certified the existence of visibility impairment in a
9 Class I area (i.e., 156 national parks, wilderness areas, and international parks identified for
10 visibility protection in section 162(a) of the Act) as being "reasonably attributable" to a
11 particular source.2
12 The second type of impairment, regional haze, results from pollutant emissions from a
13 multitude of source's located across a broad geographic region. Regional haze impairs visibility
14 in every direction over a large area, in some cases over multi-state regions. It also masks objects
15 on the horizon and reduces the contrast of nearby objects. The formation, extent, and intensity
16 of regional haze is a function of meteorological and chemical processes, which sometimes cause
17 fine particle loadings to remain suspended in the atmosphere for several days and to be
18 transported hundreds of kilometers from their sources (NRC, 1993). It is this second type of
19 visibility degradation, regional haze, that is principally responsible for impairment in national
20 parks and wilderness areas across the country (NRC, 1993).
21 While visibility impairment in urban areas at times may be dominated by local sources, it
22 often may be significantly affected by long-range transport of haze due to the multi-day
23 residence times of fine particles in the atmosphere. Fine particles transported from urban and
24 industrialized areas, in turn, may be significant contributors to regional-scale impairment in
25 Class I and other rural areas.
. 2Two of the most notable cases leading to emissions controls involved the Navajo Generating Station in
Arizona and the Mohave power plant in Nevada. For both plants, it was found that sulfur dioxide emissions were
contributing to visibility impairment in Grand Canyon National Park.
January 2005
6-3
Draft - Do Not Quote or Cite
-------
1 6.2.2 Visibility Trends and Current Conditions in Class I and Non-Urban Areas
2 In conjunction with the National Park Service, other Federal land managers, and State
3 organizations, EPA has supported visibility monitoring in national parks and wilderness areas
4 since 1988. The monitoring network was originally established at 20 sites, but it has now been
5 expanded to 110 sites that represent all but one (Bering Sea) of the 156 mandatory Federal Class
6 I areas across the country. This long-term visibility monitoring network is known as IMPROVE
7 (Interagency Monitoring of PROtected Visual Environments).
8 IMPROVE provides direct measurement of fine particles and precursors that contribute
9 to visibility impairment The IMPROVE network employs aerosol measurements at all sites, and
10 optical and scene measurements at some of the sites. Aerosol measurements are taken for PM10
11 and PM2 5 mass, and for key constituents of PM2 *5, such as sulfate, nitrate, organic and elemental
12 carbon, soil dust, and several other elements. Measurements for specific aerosol constituents are
13 used to calculate "reconstructed" aerosol light extinction by multiplying the mass for each
14 constituent by its empirically-derived scattering and/or absorption efficiency, with adjustment
15 for the relative humidity. Knowledge of the main constituents of a site's light extinction
16 "budget" is critical for source apportionment and control strategy development Optical
17 measurements are used to directly measure light extinction or its components. Such
18 measurements are taken principally with either a transmissometer, which measures total light
19 extinction, or a nephelometer, which measures particle scattering (the largest human-caused
20 component of total extinction). Scene characteristics are typically recorded 3 times daily with 35
21 millimeter photography and are used to determine the quality of visibility conditions (such as
22 effects on color and contrast) associated with specific levels of light extinction as measured
23 under both direct and aerosol-related methods. Directly measured light extinction is used under
24 the IMPROVE protocol to cross-check that the aerosol-derived light extinction levels are
25 reasonable in establishing current visibility conditions. Aerosol-derived light extinction is used
26 to document spatial and temporal trends and to determine how proposed changes in atmospheric
27 constituents would affect future visibility conditions.
28 Annual average visibility conditions (reflecting light extinction due to both
29 anthropogenic and non-anthropogenic sources) vary regionally across the U.S. The rural East
January 2005
6-4
Draft - Do Not Quote or Cite
-------
1 generally has higher levels of impairment than remote sites in the West, with the exception of
2 urban-influenced sites such as San Gorgonio Wilderness (CA) and Point Reyes National
3 Seashore (CA), which have annual average levels comparable to certain sites in the Northeast.
4 Regional differences are illustrated by Figures 4-3 9a and 4-39b in the CD, which show that, for
5 Class I areas, visibility levels on the 20% haziest days in the West are about equal to levels on
6 the 20% best days in the East (CD, p 4-179).
7 Higher visibility impairment levels in the East are due to generally higher concentrations
8 of anthropogenic fine particles, particularly sulfates, and higher average relative humidity levels.
9 In fact, sulfates account for 60-86% of the haziness in eastern sites (CD, 4-236). Aerosol light
10 extinction due to sulfate on the 20% haziest days is significantly larger in eastern Class I areas as
11 compared to western areas (CD, p. 4-182; Figures 4-40a and 4-40b). With the exception of
12 remote sites in the northwestern U.S., visibility is typically worse in the summer months. This is
13 particularly true in the Appalachian region, where average light extinction in the summer
14 exceeds the annual average by 40% (Sisler et al., 1996).
15 Regional trends in Class I area visibility are updated and presented in the EPA's National
16 Air Quality and Emissions Trends Report (EPA, 2001). Eastern trends for the 20% haziest days
17 from 1992-1999 showed a 1.5 deciview improvement, or about a 16% improvement. However,
'18 visibility in the East remains significantly impaired, with an average visual range of
19 approximately 20 km on the 20% haziest days. In western Class I areas, aggregate trends
20 showed little change during 1990-1999 for the 20% haziest days, and modest improvements on
21 the 20% mid-range and clearest days. Average visual range on the 20% haziest days in western
22 Class I areas is approximately 100 km.
23 .
24 6.2.3 Visibility Conditions in Urban Areas
25 Urban visibility impairment often results from the combined effect of stationary, mobile,
26 and area source emissions. Complex local meteorological conditions may contribute to such
27 impairment as well. Localized or layered haze often results from emissions from many sources
28 located across an urban or metropolitan area. A common manifestation of this type of visibility
29 impairment is the "brown cloud" situation experienced in some cities particularly in the winter
January 2005 6-5 Draft - Do Not Quote or Cite
-------
t
1 ' months, when cooler temperatures limit vertical mixing of the atmosphere. The long-range
2 transport of emissions from sources outside the urban area may also contribute to urban haze
3 levels.
4 Visibility impairment has been studied in several major cities in the past decades because
5 of concerns about fine particles and their potentially significant impacts (e.g., health-related and
6 aesthetic) on the residents of large metropolitan areas (e.g., Middletori, 1993). Urban areas•
7 generally have higher loadings of PM2 5 and, thus, higher visibility impairment than monitored
8 Class I areas. As discussed in Chapter 2, sections 2.4 and 2.5, annual meanlevels of 24-hour
9 average PM2 s levels are generally higher in urban areas than those found in the IMPROVE
10 database for rural Class I areas. Urban areas have higher concentrations of organic carbon,
11 elemental carbon, and paniculate nitrate than rural areas due to a higher density of fuel
12 combustion and diesel emissions.
13 6.2.3.1 ASOS Airport Visibility Monitoring Network
14 For many years, urban visibility has been characterized using data describing airport
15 visibility conditions. Until the mid-1990's, airport visibility was typically reported on an hourly
16 basis by human observers. An extensive database of these assessments has been maintained and
17 analyzed to characterize visibility trends from the late-1940's to mid-1990's (Schichtel et al.,
18 2001).
19 In 1992, the National Weather Service (NWS), Federal Aviation Administration (FAA),
20 and Department of Defense began deployment of the Automated Surface Observing System
21 (ASOS). ASOS is now the largest instrument-based visibility monitoring network in the U.S. •
22 (CD, p. 4-174). The ASOS visibility monitoring instrument is a forward scatter meter that has
23 been found to correlate well with light extinction measurements from the Optec transmissometer
24 (NWS, 1998). It is designed to provide consistent, real-time visibility and meteorological
25 measurements to assist with air traffic control operations. A total of 569 FAA-spbnsored and
26 313 NWS-sponsored automated observing systems are installed at airports throughout the
27 country. ASOS visibility data are typically reported for aviation use in small increments up to a
28 maximum of 10 miles visibility. While these truncated data are not ideal for characterizing
29 actual visibility levels, the raw, non-truncated data from the 1-minute light extinction and
January 2005 6-6 Draft - Do Not Quote or Cite
-------
1 meteorological readings are now archived and available for analysis for a subset of the ASOS
2 sites.3
3 6.2.3.2 Correlation between Urban Visibility and PM2S Mass
4 In an effort to better characterize urban visibility, staff has analyzed the extensive new
5 data now available on PM2 5 primarily in urban areas. This rapidly expanding national database,
6 including FRM measurements of PM25 mass, continuous measurements of hourly PM2.5 mass,
7 and PM25 chemical speciation measurements, has now provided the opportunity to conduct such
8 an analysis. In this analysis, described below and documented in detail in Schmidt et al. (2005),
9 staff has sought to explore the factors that have historically complicated efforts to address
10 visibility impairment nationally, including regional differences related to levels of primarily fine
11 particles and relative humidity. Taking these factors into account, staff has compared
12 correlations between visibility, in terms of reconstructed light extinction (using the IMPROVE
13 methodology discussed in Chapter 2, section 2.8), with hourly PM25 concentrations in urban
14 areas across the U.S. and in eastern and western regions.
15 As an initial matter, staff has explored the factors contributing to the substantial
16 East/West differences that have been characterized primarily for Class I areas across the country,
17 as discussed above in section 6.2.2. In considering fine particle levels, staff notes that East/West
18 differences are substantially smaller in urban areas than in rural areas. As shown in Figure 6-1,
19 24-hour average PM2 5 concentrations in urban areas in the East and West are much more similar
t
20 than in rural areas. A significantly lower East/West ratio is observed in urban areas, based on
21 data from either the FRM or the EPA Speciation Network, than in rural areas, based on data from
22 the IMPROVE network.
23 In considering relative humidity levels, staff notes that, while the average daily relative
24 humidity levels are generally higher in eastern than western areas, in both regions relative
25 humidity levels are appreciably lower during daylight as compared to night time hours. These
3 A preliminary analysis of the archived data for 63 cities across the U.S. was presented in the first draft
Staff Paper (August 2003), but further analysis has not been conducted. While the preliminary analysis
demonstrated relatively well-characterized correlations between predicted PM, j concentrations (based on ASOS
extinction values) and measured PM25 concentrations in some urban areas, such correlations were not consistently
observed in urban areas across the country.
January 2005 6-7 Draft - Do Not Quote or Cite
f
-------
a
t
00
vi
m
I
•a
•s
^
&
I
-------
1 differences can be seen in Figure 6-2, based on data from National Weather Service (NWS) sites.
2 As discussed in Chapter 2, section 2.8, the reconstructed light extinction coefficient, for a given
3 mass and concentration, increases sharply as relative humidity rises. Thus, visibility impacts
4 related to East/West differences in average relative humidity are minimized during daylight
5 hours, when relative humidity is generally lower.
6 Taking these factors into account, staff has considered both 24-hour and shorter-term
7 daylight hour averaging periods in evaluating correlations between PM2 5 concentrations in urban,
8 areas and visibility, in terms of reconstructed light extinction (RE), in eastern and western areas,
9 as well as nationwide. Figure 6-3 shows clear and similarly strong correlations between RE and
10 24-hour average PM2 s in eastern, western, and all urban areas. Figure 6-3 is based on data from
11 161 urban continuous PM25 mass monitoring sites across the country with co-located or nearby
12 24-hour PM2 5 speciation data. RE values were calculated based on a constructed hourly
13 speciated PMZ5 data set, hourly relative humidity data (either co-located or from nearby NWS
14 sites), and a coarse PM data set (estimated either by difference method from the continuous
15 PM25 and co-located continuous PMi0 instruments, or based on regional ratios of PM fractions)
1.6 (Schmidt et al., 2005). In calculating RE, the relative humidity was capped at 95%, reflecting
17 the lack of accuracy in higher relative humidity values and their highly disproportionate impact
18 on RE.
19 For these analyses, staff has considered both 10 years of relative humidily data,
20 converted to 10-year average hourly f(RH)4 values (Figure 6-3, panel a), as well as actual hourly
21 relative humidity data for 2003, converted to f(RH) values (Figure 6-3, panel b). Staff
22 recognizes that 10-year average hourly f(RH) data are more reflective of long-term humidity
23 patterns, and may provide a more appropriate basis for relating ambient PM2 5 levels to visibility
24 impairment in the context of consideration of a potential secondary standard to protect against
25 PM-related visibility impairment. On the other hand, since there can be significant day-to-day
26 variance in relative humidity that is not reflected in long-term average f(RH) data, actual hourly
*f(RH) is the relative humidily adjustment factor; it increases significantly with higher humidity. See
section 2.8.1 and Chapter 4 of the CD (CD, pp. 4-149 to 4-170) for further information.
January 2005
6-9
Draft - Do Not Quote or Cite
-------
CM
I—I
H:
Ht
i i
^
Hi
Ht
H{
I i"
HH±t
^
I
>h
>h
^
HP
ir.|... I,,.,!.,,,-!.
u
•* «w
S IS
CM
tM
CM
si
£ *
M
CM
3
89 4>
s
> ]
•I ts
DA O
2 ffl
**
« 2
3
-------
Figure 6-3. Relationship between reconstructed light extinction (RE) and 24- hour
average PM2 s, 2003. RE in top panel (a) computed with 10-year average
f(RH)\ RE in bottom panel (b) computed using actual f(RH).
Source: Schmidt et al. (2005)
January 2005
6-11
Draft - Do Not Quote or Cite
-------
1 f(RH) data were also included in the analyses, to reflect the potential ranges of high and low
2 relative humidity levels likely to occur over the course of a year.
3 In considering shorter-term daylight hour averaging periods, staff also evaluated the
4 slope and strength of the correlations between RE and PM25 concentrations on an hourly basis
5 (Schmidt et al., 2005). Figure 6-4 shows plots of the average slope of the correlation between
6 hourly RE and corresponding PM2 5 concentrations (i.e., the increase in RE due to Ihe
7 incremental increase in PM2 s) by region, in eastern and western areas, and nationwide. The
8 slopes are all lower during daytime hours when the disproportionate effects of relative humidity
9 on the light extinction coefficients for fine particle sulfates and nitrates are diminished. Thus,
10 during daylight hours, the slope more closely represents the influence of PM25 mass on visibility
11 than the influence of relative humidity. In addition, Figure 6-4 shows that the slopes (and hence,
12 the relationships between RE and PM2 5) are more comparable across regions during daylight
13 hours. In considering the strength of these correlations, staff notes that the correlations between
14 RE and PM25, as indicated by the model R2 values, are strong for individual daylight hours,
15 similar to that for the 24-hour average (Schmidt et al., 2005).' On a national basis, daytime (9
16 a.m. to 6 p.m.) hourly model R2 values are all above 0.6 for the RE's calculated with actual f(RH)
11 values and above 0.8 for the RE's calculated with 10-year average f(RH) values (Schmidt et al.,
18 2005).
19 On the basis of lower slopes and more inter-region comparability, staff selected anumber
20 of daylight time periods to consider in evaluating additional correlations between PM2 5
21 concentrations and RE in eastern and western regions, as well as nationwide. Evaluated time'
22 periods included 7 a.m. to 7 p.m.; 9 a.m. to 5 p.m.; 10 a.m. to 6 p.m.; 10 a.m. to 4 p.m.; 12 p.m.
23 to 4 p.m.; and 8 a.m. to 4 p.m. With a focus on minimizing slope, minimizing regional and
24 East/West slope differences, maximizing R2 values, and considering other related factors, staff
25 selected the 12 p.m. to 4 p.m. time period for further analyses (Schmidt et al., 2005).
26 Using the same data as were used for Figure 6-3, Figure 6-5 shows examples of the
27 correlations between RE and PM25 concentrations averaged over a 4-hour time period, for 10-
28 year average hourly f(RH) data (panel a) and for actual hourly f(RH) data in 2003 (panel b). As
29 seen in this figure, the correlations between RE and PM2 5 concentrations during daylight hours
January 2005
6-12
Draft - Do Not Quote or Cite
-------
i U>
CO
o> CO
If
6
o
tj **
2 w
O (0
CO UJ
II
III
CO Z
«t *
di
I *
.2 o>
51
to ^
I s
1^
in (0
<8
-------
Figure 6-5. Relationship between reconstructed light extinction (RE) and 12 p.m. -
4 p.m. average PM2 5,2003. RE in top panel (a) computed with 10-year
avsragsffRH); RE in bottom panel (b) computed using actual f(RH).
Source: Schmidt et al. (2005)
January 2005
6-15
Draft - Do Not Quote or Cite
-------
1 in urban areas are comparably strong (similar R2 values), yet more reflective of PM2 5 mass rather
2 than relative humidity effects (i.e., lower slopes), in comparison to the correlations based on a
3 24-hour averaging time. Further, these correlations in urban areas are generally similar in the
4 East and West, in sharp contrast to the East/West differences observed in rural areas.
5
6 6.2.4 Economic and Societal Value of Improving Visual Air Quality
7 Visibility is an air quality-related value having direct significance to people's enjoyment
8 of daily activities in all parts of the country. Survey research on public awareness of visual air
9 quality using direct questioning typically reveals that 80% or more of the respondents are aware
10 of poor visual air quality (Cohen et al., 1986). The importance of visual air quality to public
11 welfare across the country has been demonstrated by a number of studies designed to quantify
12 the benefits (or willingness to pay) associated with potential improvements in visibility
13 (Chestnut and Rowe, 1991).
14 Individuals value good visibility for the sense of well-being it provides them directly,
15 both in the places where they live and work, and in the places where they enjoy recreational
16 opportunities. Millions of Americans appreciate the scenic vistas in national parks and
17 wilderness areas each year. Visitors consistently rate "clean, clear air" as one of the most
18 important features desired in visiting these areas (Department of Interior, 1998). A 1998 survey
19 of 590 representative households by researchers at Colorado State University found that 88% of
20 the respondents believed that "preserving America's most significant places for future
21 generations" is very important, and 87% of the respondents supported efforts to clean up air
22 pollution that impacts national parks (Hass and Wakefteld, 1998).
23 . Economists have performed many studies in an attempt to quantify the economic benefits
24 associated with improvements in current visibility conditions both in national parks and in urban
25 areas. These economic benefits are often described by economists as either use values or non-
26 use values. Use values are those aspects of environmental quality that directly affect an
27 individual's welfare. These include improved aesthetics during daily activities (e.g., driving or
28 walking, looking out windows, daily recreations), for special activities (e.g., visiting parks and
January 2005
6-15
Draft - Do Not Quote or Cite
-------
1 scenic vistasj hiking, hunting), and for viewing scenic photography. Aesthetic benefits of better
2 visibility also include improved road and air safety.
3 Non-use values are those for which an individual is willing to pay for reasons that do not
4 relate to the direct use or enjoyment of any environmental benefit The component of non-use
5 value that is related to the use of the resource by others in the future is referred to as the bequest
6 value. This value is typically thought of as altruistic in nature. Another potential component of
7 non-use value is the value that is related to preservation of the resource for its own sake, even if
8 there is no human use of the resource. This component of non-use value is sometimes referred to
9 as existence value or preservation value. Non-use values are not traded, directly or indirectly, in
10 markets. For this reason, the estimation of non-use values has proved to be significantly more
11 difficult than the estimation of use values. Non-use values may be related to the desire that a
12 clean environment be available for the use of others now and in the future, or they may be related
13 to the desire to know that the resource is being preserved for its own sake, regardless of human
14 , use. Non-use values may be a more important component of value for recreational areas,
15 particularly national parks and monuments, and for wilderness areas.
16 In addition, staff notes that the concept of option value is a key component of the
17 measured values. The option value represents the value that is tied to preserving improved
18 visibility in the event of a visit, even though a visit is not certain. This component is considered
19 by some as a use value and by others as a non-use value.
20 Tourism in the U.S. is a significant contributor to the economy. A1998 Department of
21 Interior study found that travel-related expenditures by national park visitors alone average $.14.5
22 billion annually (1996 dollars) and support 210,000 jobs (Peacock et al., i998). A similar
23 estimate of economic benefits resulting from visitation to national forests and other public lands
24 could increase this estimate significantly.
25 It is well recognized in the U.S. and abroad that there is an important relationship
26 between good air quality and economic benefits due to tourism. McNeill and Roberge (2000)
27 studied the impact of poor visibility episodes on tourism revenues in Greater Vancouver and the
28 Lower Fraser Valley in British Columbia as part of the Georgia Basin Ecosystem Initiative of
29 Environment Canada Through this analysis, a model was developed that predicts future tourist
January 2005
6-16
Draft - Do Not Quote or Cite
-------
1 revenue losses that would result from a single extreme visibility episode. They found that such
2 an episode would result in a $7.45 million loss in the Greater Vancouver area and $1.32 million
3 loss in the Fraser Valley.
4 The results of several valuation studies addressing both urban and rural visibility are
5 presented in the CD (CD, pp. 4-187 to 4-190), the 1996 Criteria Document (EPA, 1996a, p. 8-83,
6 Table 8-5; p. 8-85, Table 8-6) and in Chestnut and Rowe (1991) and Chestnut et al. (1994). Past
7 studies by Schulze et al. (1983) and Chestnut and Rowe (1990) have estimated the preservation
l
8 values associated with improving the visibility in national parks in the Southwest to be in the
9 range of approximately $2-6 billion annually. An analysis of the residential visibility benefits in
10 the eastern U.S. due to reduced sulfur dioxide emissions under the acid rain program suggests an
11 annual value of $2.3 billion (in 1994 dollars) in the year 2010 (Chestnut and Dennis, 1997). The
12 authors suggest that these results could be as much as $1-2 billion more because the above
13 estimate does not include any value placed on eastern air quality improvements by households in
14 the western U.S.
15 Estimating benefits for improvements in visibility can be difficult because visibility is not
16 directly or indirectly valued in markets. Many of the studies cited above are based on a
17 valuation method known as contingent valuation (CV). Concerns have been identified about the
18 reliability of value estimates from contingent valuation studies because research has shown that
19 bias can be introduced easily into these studies if they are not carefully conducted. Accurately
20 estimating willingness-to-pay for avoided health and welfare losses depends on the reliability
21 and validity of the data collected. However, there is an extensive scientific literature and body of
22 practice on both the theory and technique of contingent valuation. EPA believes that well-
23 designed and well-executed CV studies are useful for estimating the benefits of environmental
24 effects such as improved visibility (EPA, 2000).
25 Some of the studies cited above used an alternative valuation method known as hedonic
26 pricing. Hedonic pricing is a technique used to measure components of property value (e.g.,
27 proximity to schools). It relies on the measurement of differentials in property values under
28 various environmental quality conditions, including air pollution and environmental amenities,
29 such as aesthetic views. This method works by analyzing the way that market prices change
January 2005
6-17
Draft - Do Not Quote or Cite
-------
1
2
3
4
5
6
7
8
9
10
II
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
with changes in environmental quality or amenity. EPA believes that well-designed and well-
executed hedonic valuation studies, in combination with public perception surveys, are useful for
estimating the benefits of environmental effects such as improved visibility.
Society also values visibility because of the significant role it plays in transportation
safety. Serious episodes of visibility impairment can increase the risk of unsafe air
transportation, particularly in urban areas with high air traffic levels (EPA, 1982). In some
cases, extreme haze episodes have led to flight delays or the shutdown of major airports,
resulting in economic impacts on air carriers, related businesses, and air travelers. For example,
on May 15,1998 in St. Louis, Missouri, it was reported that a haze episode attributed to
i.
wildfires in central America resulted in a reduction in landing rates and significant flight delays
at Lambert International Airport. The 24-hour PM25 levels reached 68 ng/m3 during that
episode. In addition, the National Transportation Safety Board (NTSB) has concluded in
accident reports that high levels of pollution and haze, such as those experienced during the July
1999 air pollution episode in the northeastern U. S., have played a role in air transportation
accidents and loss of life (NTSB, 2000). During this episode, 24-hour levels of PM2 5 ranged
from 35-52 ug/m3 in the New England states.
6.2.5 Programs and Goals for Improving Visual Air Quality
Specific discussion is provided below on regional visibility programs in the U.S.; as well
as local visibility programs established by States, localities, and other countries in an effort to
protect visual air quality.
6.2.5.1 Regional Protection
Due to differences in visibility impairment levels (due to differences in chemical
composition of haze and in relative humidity levels) between the'East and West, EPA, land
managers, and States have taken a regional approach, rather man a national approach, to
protecting visibility in non-urban areas in the U.S.. Protection against visibility impairment in
special areas is provided for in sections 169A/169B, and 165 of the Act, in addition to that
provided by the secondary NAAQS. Section 169A, added by the 1977 CAA Amendments,
established a national visibility goal to "remedy existing impairment and prevent future
January 2005
6-18
Draft - Do Not Quote or Cite
-------
1 impairment" in 156 national parks and wilderness areas (Class I areas). The Amendments also
2 called for EPA to issue regulations requiring States to develop long-term strategies to make
3 "reasonable progress" toward the national goal. EPA issued initial regulations in 1980 focusing
4 on visibility problems that could be linked to a single source or small group of sources. Action
5 was deferred on regional haze until monitoring, modeling, and source apportionment methods
6 could be improved.
7 The 1990 CAA Amendments placed additional emphasis on regional haze issues through
8 the addition of section 169B. In accordance with this section, EPA established the Grand
9 Canyon Visibility Transport Commission (GCVTC) in 1991 to address adverse visibility impacts
10 on 16 Class I national parks and wilderness areas on the Colorado Plateau. The GCVTC was
11 comprised of the Governors of nine western states and leaders from a number of Tribal nations.
12 The GCVTC issued its recommendations to EPA in 1996, triggering a requirement in section
13 169B for EPA issuance of regional haze regulations.
14 EPA accordingly promulgated a final regional haze rule in 1999 (EPA, 1999; 65 FR
15 35713). Under the regional haze program, States are required to establish goals for improving
16 visibility on the 20% most impaired days in each Class I area, and for allowing no degradation
17 on the 20% least impaired days. Each state must also adopt emission reduction strategies which,
18 in combination with the strategies of contributing States, assure that Class I area visibility
19 improvement goals are met. The first State implementation plans are to be adopted in the 2003-
20 2008 time period, with the first implementation period extending until 2018. Five multistate
21 planning organizations are evaluating the sources of PM2 5 contributing to Class I area visibility
22 impairment to lay the technical foundation for developing strategies, coordinated among many
23 States, in order to make reasonable progress in Class I areas across the country.
24 6.2.5.2 Local, State, and International Goals and Programs
25 The value placed on protecting visual air quality is further demonstrated by the existence
26 of a number of programs, goals, standards, and planning efforts that have been established in the
27 U.S. and abroad to address visibility concerns in urban and non-urban areas. These regulatory
28 and planning activities are of particular interest because they are illustrative of the significant
29 value that the public places on improving visibility, and because they have made use of
January 2005
6-19
Draft - Do Not Quote or Cite
-------
f'.
1 developed methods for evaluating public perception's and judgments about the acceptability of
2 varying degrees of visibility impairment.
3 Several state and local governments have developed programs to improve visual air
4 quality in specific urban areas, including Denver, CO; Phoenix, AZ; and, Lake Tahoe, C A. At
5 least two States have established statewide standards to protect visibility. In addition, visibility
6 protection efforts have been undertaken in other countries, including Australia, New Zealand,
7 and Canada. Examples of these efforts are highlighted below.
8 In 1990, the State of Colorado adopted a visibility standard for the city of Denver. The
9 Denver standard is a short-term standard that establishes a limit of a four-hour average light
10 extinction level of 76 Mm"1 (equivalent to a visual range of approximately 50 km) during the
11 hours between 8 a.m. arid 4 p.m. .(Ely et al., 1991). In 2003, the Arizona Department of
12 Environmental Quality created the Phoenix Region Visibility Index, which focuses on an
13 averaging time of 4 hours during actual daylight hours. This visiblity index establishes visual air
14 quality categories (i.e., excellent to very poor) and establishes the goals of moving days in the
15 poor/very poor categories up to the fair category, and moving days in the fair category up to the
16 good/excellent categories (Arizona Department of Environmental Quality, 2003). This approach
17 results in a focus on improving visibility to a visual range of approximately 48-36 km. In 1989,
18 the state of California revised the visibility standard for the Lake Tahoe Air Basin and
19 established an 8-hour visibility' standard equal to a visual range of 30 miles (approximately 48
20 km) (California Code of Regulations).
21 California and Vermont each have standards to protect visibility, though they are based
22 on different measures. Since 1959, the state of California has had an air quality standard for
23 particle pollution where the "adverse" level was defined as the "level at which there will be...
24 reduction in visibility or similar effects." California's general statewide visibility standard is a .
25' visual range of 10 miles (approximately 16 km) (California Code'of Regulations). In 1985,
26 Vermont established a state visibility standard that is expressed as a summer seasonal sulfate
27 concentration of 2 ug/m3, that equates to a visual range of approximately 50 km. This standard
28 was established to represent "reasonable progress toward attaining the congressional visibility
t
January 2005
6-20
Draft - Do Not Quote or Cite
-------
1 goal for the Class 1 Lye Brook National Wilderness Area, and applies to this Class 1 area and to
2 all other areas of the state with elevations greater than 2500 ft.
3 Outside of the U.S., efforts have also been made to protect visibility. The Australian
4 state of Victoriahas established a visibility objective (State Government of Victoria, 2000a and
5 2000b), and a visibility guideline is under consideration in New Zealand (New Zealand National
6 Institute of Water & Atmospheric Research, 2000a and 2000b; New Zealand Ministry of
7 Environment, 2000). A survey was undertaken for the Lower Fraser Valley in British Columbia,
8 with responses from this pilot study being supportive of a standard in terms of a visual range of
9 approximately 40 km for the suburban township of Chilliwack and 60 km for the suburban
10 township of Abbotsford, although no visibility standard has been adopted for the Lower Fraser
11 Valley at this time.
12
13 6.2.6 Approaches to Evaluating Public Perceptions and Attitudes
14 New methods and tools have been developed to communicate and evaluate public
15 perceptions of varying visual effects associated with alternative levels of visibility impairment
16 relative to varying pollution levels and environmental conditions. New survey methods have
17 been applied and evaluated in various studies, such as those for Denver, Phoenix, and the Lower
18 Fraser Valley in British Columbia, and these studies are described below in more detail. These
19 methods are intended to assess public perceptions as to the acceptability of varying levels of
20 visual air quality, considered in these studies to be an appropriate basis for developing goals and
21 standards for visibility protection. For the Denver and British Columbia studies, actual slides
22 taken in the areas of interest, and matched with transmissometer and nephelometer readings,
23 respectively, were used to assess public perceptions about visual air quality. For the Phoenix
24 study, WinHaze, a newly available image modeling program, discussed below, was used for
25 simulating images. Staff finds that, even with variations in each study's approaches, the survey
26 methods used for the Denver, Phoenix, and British Columbia studies produced reasonably
27 consistent results from location to location, each with a majority of participants finding visual
28 ranges within about 40 to 60 km to be acceptable.
January 2005
6-21
Draft - Do Not Quote or Cite
t
-------
1 6.2.6.1 Photographic Representations of Visual Air Quality
2 In the past, the principal method for recording and describing visual air quality has been
3 through 35 millimeter photographs. Under the IMPROVE program, EPA, federal land
4 management agencies, and Air Resource Specialists, Inc. (ARS) have developed an extensive
5 archive of visual air quality photos for national parks and wilderness areas. In comparison, we
6 have only a limited archive of photos of urban areas.
7 The CD discusses some of the methods that are now available to represent different
8 levels of visual air quality (CD, p. 4-174). In particular, Molenar et al. (1994) describes a
9 sophisticated visual air quality simulation technique, incorporated into the WinHaze program
10 developed'by ARS, which combined various modeling systems under development for Ihe past
11 20 years. The technique relies on first obtaining an original base image slide of the scene of
12 interest. The slide should be of a cloudless sky under the cleanest air quality conditions possible.
13 The light extinction represented by the scene should be derived from aerosol and optical data
14 associated with the day the image was taken, or it should be estimated from contrast
15 measurements of features in the image. The image is then digitized to assign an optical density
16 to each pixel. At this point, the radiance level for each pixel is estimated. Using a detailed
17 topographic map, technicians identify the specific location from which the photo was taken, and
18 they determine the distances to various landmarks and objects in the scene. With this
19 information, a specific distance and elevation is assigned to each pixel.
20 Using the digital imaging information, the system then computes the physical and optical
21 properties of an assumed aerosol mix. These properties are input into a radiative transfer model
22 in order to simulate the optical properties of varying pollutant concentrations on the scene.
23 WinHaze, an image modeling program for personal computers that employs simplified
24 algorithms based on the sophisticated modeling technique, is now available (Air Resource
25 Specialists, 2003).
26 The simulation technique has the advantage of being readily applicable to any location
27 as long as a very clear base photo is available for that location. In addition, the lack of clouds
28 and consistent sun angle in all images, in effect, standardizes the perception of the images and
29 enables researchers to avoid potentially biased responses due to these factors. An alternative
January 2005 6-22 Draft - Do Not Quote or Cite
-------
1 technique to using simulated images is to obtain actual photographs of the site of interest at
2 different ambient pollution levels. However, long-term photo archives of this type exist for only
3 a few cities. In addition, studies have shown that observers will perceive an image with a cloud-
4 filled sky as having a higher degree of visibility impairment than one without clouds, even
5 though the PM concentration on both days is the same.
6 As part of a pilot study5 in Washington, D.C., both survey and photographic techniques
7 were applied (Abt Associates, 2001). In conjunction with this pilot project, images that illustrate
8 , visual air quality in Washington, DC under a range of visibility conditions were prepared and are
9 available at http://www,epa.goy/ttii/naaqs/sta^^ (labeled as Attachment
10 6-A: Images of Visual Air Quality in Selected Urban Areas in the U.S.). Included as part of
11 Attachment 6-A, this website also contains actual photographs of Chicago illustrating visibility
12 conditions associated with a range of PM23 concentrations, as well as simulated images for
13 Denver and Phoenix, as discussed below.
14 6.2.6.2 Survey Methods
15 Denver, Colorado: Visibility Standard
16 The process by which the Denver visibility standard was developed relied on citizen
17 judgments of acceptable and unacceptable levels of visual air quality (Ely et al., 1991).
18 Representatives from Colorado Department of Public Health and Environment (CDPHE)
19 conducted a series of meetings with 17 civic and community groups in which atotal of 214
20 individuals were asked to rate slides having varying levels of visual air quality for a well-known
21 vista in Denver. The CDPHE representatives asked the participants to base their judgments on
22 three factors: 1) the standard was for an urban area, not a pristine national park area where the
23 standards might be more strict; 2) standard violations should be at visual air quality levels
24 considered to be unreasonable, objectionable, and unacceptable visually; and 3) judgments of
25 standards violations should be based on visual air quality only, not on health effects.
26 The participants were shown slides in 3 stages. First, they were shown seven warm-up
27 slides describing the range of conditions to be presented. Second, they rated 25 randomly-
5 A small pilot study for Washington, D.C. was conducted by EPA and was briefly discussed in the
preliminary draft staff paper (2001).
January 2005
6-23
Draft - Do Not Quote or Cite
f
-------
* 1 ordered slides based on a scale of 1 (poor) to 7 (excellent), with 5 duplicates included. Third,
2 they were asked to judge whether the slide would violate what they would consider to be an
3 appropriate urban visibility standard (i.e., whether the level of impairment was "acceptable" or
4 "unacceptable"). . . .
5 The Denver visibility standard setting process produced the following findings:
6 ' . ' , * * ••'- ' '; ' ' '
7 • Individuals' judgments of a slide's visual air quality and whether the slide violated a •
8 visibility standard are highly correlated (Pearson correlation coefficient greater than
9 " 80%) with the group average. ' •''•
10 - •. '...-..
11 • When participants judged duplicate slides, group averages of the first and second ratings
12 were highly correlated.
13 . . ' •• . . . /,'..:
14 • Group averages of visual air quality ratings and "standard violations" were highly
15 correlated.' The strong relationship of standard violation judgments with the visual air ;
16 quality ratings is cited as the best evidence available from this study for the validity of
17 standard violation judgments (Ely et al., 1991).
18 ' -. ' •
19 The CDPHE researchers sorted the ratings for each slide by increasing order of light
20 extinction and calculated the percent of participants that judged each slide to violate the
21 standard. The Denver visibility standard1 was then established based on a 50% acceptability
22 criterion. Under this approach, the standard was identified as the light extinction level that
23 divides the slides into two groups: those found to be acceptable and those found to be
24 unacceptable by a majority of study participants. The CDPHE researchers found this level to be
25 reasonable because, for the slides at this level and above, a majority of the study participants
26 judged the light extinction levels to be unacceptable. In fact, when researchers evaluated all
27 citizen judgments made on all slides at this level and above as a single group, more than 85% of
28 the participants found visibility impairment at and above the level of the selected standard to be
29 unacceptable.'
30 Though images used in the Denver study were actual photographs, more recently,
31 WinHaze has been used to generate images that illustrate visual air quality in Denver under a
32 range of visibility conditions (generally corresponding to 10th, 20*, 30*, 40th, 50*, 60th 80*, and
33 90* percentile values), and these images are available in Attachment 6-A at
34 http://www.epa.gov/ttn/naaqs/staiidards/pni/s_pm_cr_sp.htiTil.
January 2005 6-24 Draft - Do Not Quote or Cite
-------
1 Data analyses using extensive new monitoring data now available on.PM2 3 primarily in
2 urban areas show a consistently high correlation between hourly PM2.S data and RE coefficients ,
3 for urban areas across regions of the U.S. during daylight hours. These correlations in urban
4 areas are generally similar in the East and West, in sharp contrast to the East/West differences
5 observed in rural areas.
6 The importance of visual air quality to public welfare across the country has been
7 demonstrated by a number of studies designed to quantify the benefits (or willingness to pay).
8 associated with potential improvements in visibility. The value placed on protecting visual air
9 quality is further demonstrated by the existence of a number of programs, goals, standards, and
10 planning efforts that have been established in the U.S. and abroad to address visibility concerns
11 in urban and non-urban areas.
12 In some urban areas, poor visibility has led to more localized efforts to better
13 characterize, as well as improve, urban visibility conditions. The public perception survey -
14 approach used in the Denver, Phoenix, and British Columbia studies yielded reasonably
15 consistent results, with each study indicating that a maj ority of citizens find value in protecting
16 local visibility to within a visual range of about 40 to 60 km. In the cases of Denver and
17 Phoenix, these studies provided the basis for the establishment of their visibility standards and
18 goals. ..,.-. ,
19 Staff believes ,that the findings of the new data analyses, in combination with recognized
20 benefits to public welfare of improved visual air quality and an established approach for
21 • determining acceptable visual range, provide a basis for considering revisions to the secondary
22 PM25 standards to protect against PM-related visibility effects in urban areas.
23 ....,--,
24 6.3 EFFECTS ON VEGETATION AND ECOSYSTEMS
25 Information and conclusions regarding what is currently known about the impacts of
26 ambient PM on ecosystems and individual components of ecosystems such as vegetation, soils,
27 water, and wildlife are discussed in Chapters 4 and 9 of the CD. This section seeks to build upon
28 and focus this body of science using EPA's ecological risk paradigm in a manner that highlights
29 the usefulness and policy relevance of the scientific information. In so doing, staff has drawn
January 2005
6-27
Draft - Do Not Quote or Cite
t
-------
1 from EPA's Guidelines for Ecological Risk Assessment (Guidelines) (EPA, 1998), which
2 expanded upon the earlier document, Framework for Ecological Risk Assessment (EPA, 1992),
3 with the goal of improving the quality of ecological risk assessments and increasing the
\ '
4 consistency of assessments across the Agency.
5 According to the Guidelines document, the three main phases of ecological risk
6 assessment are problem formulation, analysis, and risk characterization. In problem formulation,
7 the purpose for the assessment is articulated, the problem is defined, assessment endpoints are
8 selected, a conceptual model is prepared and an analysis plan is developed: Initial work in
9 problem formulation includes the integration of available information on sources, stressors,
10 effects, and ecosystem and receptor characteristics.
11 In the analysis phase data are evaluated to determine how exposure to stressors is likely
12 to occur (exposure profile) and the relationship between stressor levels and ecological effects
13 (stressor-response profile). These products provide the basis for the risk characterization phase.
14 During the third phase, risk characterization, the exposure and stressor-response profiles
15 are integrated through the risk estimation process. Risk characterization includes a summary of
16 assumptions, scientific uncertainties, and strengths and limitations of the analyses. The final
17 product is a risk description in which the results of the integration are presented, including an
18 interpretation of ecological adversity and description of uncertainty and lines of evidence.
19 Keeping these goals and guidelines in mind, this section organizes information into the
20 following seven subsections: major ecosystem stressors in PM (6.3.1); direct vegetation effects
21 of PM stressor deposition (6.3.2); ecosystem effects of PM stressor deposition (6.3.3);
22 characteristics and location of sensitive ecosystems within the U.S. (6.3.4); ecosystem exposures
23 to PM deposition (6.3.5); consideration of critical loads as an approach for effects
24 characterization and/or as a management tool (6.3.6); and summary and conclusions (6.3.7).
25 This review will also consider and reference where applicable the extent to which PM
26 affects the essential ecological attributes (EEAs) outlined in the Framework for Assessing and
21 Reporting on Ecological Condition, recommended by the Ecological Processes and Effects
28 Committee (EPEC) of EPA's Science Advisory Board (hereafter EPEC Framework; SAB,
29 2002), ad described in subsections 4.2.1 and 4.2.3 of the CD.
January 2005 * 6-28 Draft - Do Not Quote or Cite
-------
1 6.3.1 Major Ecosystem Stressors in PM
2 As previously discussed, PM is not a single pollutant, but a heterogeneous mixture of
3 particles differing in size, origin, and chemical composition. This heterogeneity of PM exists not
4 only within individual particles or samples from individual sites, but to an even greater extent,
5 between samples from different sites. Since vegetation and other ecosystem components are
6 affected more by particulate chemistry than size fraction, exposure to a given mass concentration
7 of airborne PM may lead to widely differing plant or ecosystem responses, depending on the
8 particular mix of deposited particles. Though the chemical constitution of individual particles
9 can be strongly correlated with size, the relationship between particle size and particle
10 composition can also be quite complex, making it difficult in most cases to use particle size as a
11 surrogate for chemistry. Because PM size classes do not necessarily have specific differential
12 relevance for vegetation or ecosystem effects (Whitby, 1978; EPA, 1996a), it is the opinion of
13 the staff that an ecologically relevant indicator for PM would be based on one or multiple
14 chemical stressors found in ambient PM. At this time it remains to be studied as to what extent
15 NAAQS standards focused on a given size fraction would result in reductions of the ecologically
16 relevant constituents of PM for any given area.
17 A number of different chemical species found within ambient PM and their effects on
18 vegetation and ecosystems were discussed in chapter 4 of the PM CD. In particular, the CD
19 focused on nitrates and sulfates, concluding that these PM constituents are considered to be the
20 stressors of greatest environmental significance (CD, p. 9-114): Other components of PM, such
21 as dust, trace metals, and organics, which can also be toxic to plants and other organisms at high
22 levels, were also discussed. However, because such high levels occur only near a few limited
23 point sources and/or on a very local scale, they do not appear significant at the national level.
24 Therefore, the remainder of this section will narrow its focus to consideration of the impacts of
25 particulate nitrates and sulfates, both separately and in combination with acidifying compounds,
26 on sensitive ecosystem components and essential ecological attributes, which in turn, impact
27 overall ecosystem structure and function.
28
January 2005
6-29
Draft - Do Not Quote or Cite
f
-------
1 6.3.2 Direct Vegetation Effects of PMStressor Deposition
2 Nitrogen is a critical limiting nutrient for plant growth. The process of photosynthesis
3 uses approximately 75% of the nitrogen in a plant leaf, and,.thus, to a large extent, governs the
4 utilization of other nutrients such as phosphorus, potassium (CD, p. 4-95). Plants usually absorb
5 nitrogen (as NH4+ of N03") through their roots. However, particle deposition of nitrate, together •
6 with other nitrogen-containing gaseous and precipitation-derived sources, can represent a
• 7 substantial fraction of total nitrogen reaching vegetation. In nitrogen-limited ecosystems, this
8 influx of N can act as a fertilizer. Though it is known that foliar uptake of nitrate can occur, the
9 mechanism of foliar uptake is not well established, and it is not currently possible to distinguish
f
10 sources of chemicals deposited as gases or particles using foliar extraction. Since it has proven
11 difficult to quantify the percentage of nitrogen uptake by leaves that is contributed by ambient
12 particles, direct foliar effects of nitrogen-containing particles have not been documented. (CD,
13 pp. 4-69, 4-70).
14 Similar.to nitrogen, sulfur is an essential plant nutrient that can deposit on vegetation in
15 the form of sulfate particles, or be taken up by plants in gaseous form." Greater than 90% of
16 anthropogenic sulfur emissions are as sulfur dioxide (SO2)S with most of the remaining emissions
17 in the form of sulfate. However, sulfur dioxide is rapidly transformed in the atmosphere to
18 sulfate, which is approximately 30-fold less phytotoxic than SO2. Low dosages of sulfur can
19 also serve as a fertilizer, particularly for plants growing in sulfur-deficient soils. There are only
20 a few field demonstrations of foliar sulfate uptake, however, and the relative importance of foliar
21 leachate and prior dry-deposited sulfate particles remains difficult to quantify. Though current
22 levels of sulfate deposition reportedly exceed the capacity of most vegetative canopies to
23 immobilize the sulfur, sulfate additions in excess of needs do not typically lead to plant injury
24 (CD, .pp. 4-71,4-72). , .-.-.'
25 Staff therefore conclude that at current ambient levels, risks to vegetation from short term
26 exposures to dry deposited particulate nitrate or sulfate are low. Additional studies are needed,
27 however, on the effects of sulfate particles on physiological characteristics of plants following
28 chronic exposures (CD, p. 4-72).
January 2005
6-30
Draft - Do Not Quote or Cite
-------
1 Though dry deposition of nitrate and sulfate particles does not appear to induce foliar
2 injury at current ambient exposures, when found in acidic precipitation, such particles do have
3 the potential to cause direct foliar injury. This is especially true when the acidic precipitation is
4 in the form of fog and clouds, which may contain solute concentrations many times those found
5 in rain. In experiments on seedling and sapling trees, both coniferous and deciduous species
6 showed significant effects on leaf surface structures after exposure to simulated acid rain or acid
7 mist at pH 3.5, .while some species have shown subtle effects at pH 4 and above. Epicuticular
8 waxes, which function to prevent water loss from plant leaves, can be destroyed by acid rain in a
9 few weeks, which suggests links between acidic precipitation and aging. Due to their longevity
10 and evergreen foliage, the function of epicuticular wax is more crucial in conifers. For example,
11 red spruce seedlings, which have been extensively studied, appear to be more sensitive to acid
12 precipitation (mist and fog) when compared with other species (CD, pp. 4-72, 4-73). In addition
13 to accelerated weathering of leaf cuticular surfaces, other direct responses of forest trees to
^14 acidic precipitation include increased permeability of leaf surfaces to toxic materials, water, and
15 disease agents; increased leaching of nutrients from foliage; and altered reproductive processes
16 (CD, p. 4-86). All of these effects serve to weaken trees so that they are more susceptible to
17 other stresses (e.g., extreme weather, pests, pathogens).
18 Acid precipitation with levels of acidity associated with the foliar effects described above
19 are currently found in some locations in the U.S.. For example, in the eastern U.S., the mean
20 ' precipitation pH ranges from 4.3 (Pennsylvania and New York) to 4.8 (Maine)(EPA, 2003). It
21 can be assumed that occult (mist or fog) deposition impacting high elevations more frequently,
22 would contain even higher concentrations of acidity. Thus, staff conclude that the risks of foliar
23 injury occurring from acid deposition is high. The contribution of particulate sulfates and
24 nitrates to the total acidity found in the acid deposition impacting eastern vegetation is not clear.
25
26 6.3.3 Ecosystem Effects of PMStressor Deposition
27 Ecosystem-level responses related to PM occur when the effects of PM deposition on the
28 biological and physical components of ecosystems become sufficiently widespread as to impact
29 essential ecological attributes such as nutrient cycling and/or shifts in biodiversity. The most
January 2005
6-31
Draft - Do Not Quote or Cite
-------
1 significant PM-related ecosystem-level effects result from long-term cumulative deposition of a
2 given chemical species (e.g., nitrate) or mix (e.g., acidic deposition) that exceeds the natural
3 buffering or storage capacity of the ecosystem and/or affects the nutrient status of the ecosystem,
4 usually by indirectly changing soil chemistry, populations of bacteria involved in nutrient
5 cycling, and/or populations of fungi involved in plant nutrient uptake (CD, pp. 4-90,4-91). To
6 understand these effects, long-term, detailed ecosystem or site-specific data usually are required.
7 The availability of this type of long-term data is limited. The following discussion is organized
, 8 according to the speciated effects of PM on ecosystems.
9 63.3.1 Environmental Effects of Reactive Nitrogen (Nr) Deposition
10 In the environment, nitrogen may be divided into two types: nonreactive, molecular
11 nitrogen (N2) and reactive nitrogen (Nr). Molecular nitrogen is the most abundant element in the
12 atmosphere. However, it only becomes available to support the growth of plants and
13 microorganisms after it is converted into a reactive form. In nature, Nr creation is accomplished
14 by certain organisms that have developed the capability of converting N2 to biologically active
15 reduced forms (Galloway and Cowling, 2002; Homung and Langan, 1999; EPA, 1993). By the
16 mid-1960's, however, Nr creation through natural terrestrial processes'had been overtaken by Nr
17 creation as a result of human processes (CD, p. 4-95). The deposition of nitrogen in the U. S.
18 from human activity doubled between 1961 and 1997, with the largest increase occurring in the
19 1960s and 1970s (CD, p. 4-98), Reactive nitrogen is now accumulating in the environment on
20 all spatial scales - local, regional and global. The three main sources of anthropogenic Nr are:.
21 (1) the Haber-Bosch process, which converts N2 to Nr to sustain food production and some
22 industrial activities; (2) widespread cultivation of legumes, rice and other crops that promote the
23 conversion of N2 to organic nitrogen through biological nitrogen fixation; and (3) combustion of
24 fossil fuels, which converts both atmospheric N2 and fossil nitrogen to reactive NOX (CD, pp. 4-
25 95,4-96; Galloway and Cowling, 2002; Galloway et al., 2003). Currently available forms of -
26 reactive nitrogen include inorganic reduced forms (e.g., ammonia [NH3] and ammonium [NH4+]),
27 inorganic oxidized forms (e.g., nitrogen oxides [NOJ, nitric acid [HNO3], nitrous oxide [N2O],
28 and nitrate [NO3'])S and organic compounds (e.g., urea, amine, proteins, and nucleic acids (CD,
29 p. 4-95).
January 2005
6-32
Draft - Do Not Quote or Cite
-------
1 Emissions of nitrogen oxides from fuel burning increased exponentially from!940 until
2 the 1970s, leveled off after the passage of the I970 amendments to the Clean Air Act, and
3 stabilized at approximately 7 Tg NOX /yr in the late 1990s. Contemporary emissions of NO* in
4 the U.S. from fossil fuel burning are nearly two-thirds the rate of Nr released from the use of
5 inorganic fertilizers and comprise 30% of the global emissions of NOX from fossil fuel
6 combustion. Despite decreases in emissions from fossil fuel burning industries, emissions from
7 automobiles have increased approximately 10% since 1970 due to greater total miles driven
8 (Howarth et at, 2002). Some NOX emissions are transformed into a portion of ambient air PM
9 (paniculate nitrate) and deposited onto sensitive ecosystems.
10 The term "nitrogen cascade" refers to the sequential transfers and transformations of Nr
11 molecules as they move from one environmental system or reservoir (atmosphere, biosphere,
12 hydrosphere) to another, and the multiple linkages that develop among the different ecological
13 components, as shown in Figure 6-6. Because of these linkages, adding anthropogenic Nr alters
14 a wide range of biogeochernical processes and exchanges as the Nr moves among the different
15 environmental reservoirs, with the consequences accumulating through time (Galloway and
16 Cowling, 2002; Galloway et at., 2003). These changes in the nitrogen cycle are contributing to
17 both beneficial and detrimental effects to the health and welfare of humans and ecosystems
18 (Rabalais, 2002; van Egmond et al., 2002; Galloway, 1998).
19 Large uncertainties, still exist, however, concerning the rates of Nr accumulation in Hie
20 various environmental reservoirs which limit our ability to determine the temporal and spatial
21 distribution of environmental effects for a given input of Nr. These uncertainties are of great
22 significance because of the sequential nature of Nr effects on environmental processes. Reactive
23 nitrogen does not cascade at the same rate through all environmental systems. The only way to
24 eliminate Nr accumulation and stop the cascade is to convert Nr back to nonreactive N2
25 (Galloway et al., 2003).
January 2005
'6-33
Draft - Do Not Quote or Cite
f
-------
I
2
3
4
5
6
7
8
9
10
11
12
13
14
t
NOx
Energy
Production
5
s
3
*
^
Oz<
EITc
j
AAAAAAAAAAA
Atmosphere
Food
Production
NHy
People
(Food; Fiber)
Human Activities
The Nitrogen
Cascade
- Indicates dtnltrlflcitlon potential
Figure 6-6 Illustration of the nitrogen cascade showing the movement of human-
produced reactive nitrogen (Nr) as it cycles through the various
environmental reservoirs in the atmosphere and in terrestrial and aquatic
ecosystems (Galloway et al., 2003; Figure 4-15, CD p. 4-97).
Some of the more significant detrimental effects resulting from chronic increased inputs
of atmospheric Nr (e.g., participate nitrates) include: (1) decreased.productivity, increased
mortality, and/or shifts in terrestrial plant community composition, often leading to decreased
biodiversity in many natural habitats wherever atmospheric Nr deposition increases significantly
and critical thresholds are exceeded (Aber et al., 1995); (2) leaching of excess nitrate and
associated base cations from terrestrial soils into streams, lakes and rivers and mobilization of
soil aluminum; (3) eutrophication, hypoxia, loss of biodiversity, and habitat degradation in
coastal ecosystems, now considered a major pollution problem in coastal waters (Rabalais,
January 2005
6-34
Draft - Do Not Quote or Cite
-------
1 2002); (4) acidification and loss of aquatic flora and fauna biodiversity in lakes and streams in
2 many regions of the world when associated with sulfur deposition (Vitousek et al., 1997); and
3 (5) alteration of ecosystem processes such as nutrient and energy cycles through changes in the
4 functioning and species composition of beneficial soil organisms (Galloway and Cowling 2002).
5 Additional, indirect detrimental effects of excess Nr on societal values include: (1)
6 increases in fine PM resulting in regional hazes that decrease visibility at scenic rural and urban
7 vistas and airports (discussed above in section 6.2); (2) depletion of stratospheric ozone by N2O
8 emissions which can in turn affect ecosystems and human health; (3) global climate change
9 induced by emissions of N2O (Galloway et al., 2003); (4) formation of O3 and ozone-induced
10 injury to crops, forests, and natural ecosystems and the resulting predisposition to attack by
11 pathogens and insects, as well as human health related impacts (EPA, 1996); (5) decrease in
12 quantity or quality of available critical habitat for threatened and endangered species (Fenn et al.,
13 2003); and (6) alteration of fire cycles in a variety of ecoystem types (Fenn et al., 2003).
14 A number of the more significant effects of chronic, long-term deposition of Nr on
15 terrestrial and aquatic ecosystems will be discussed below, specifically those effects which seem
16 to pose the greatest long-term risks to species or ecosystem health and sustainability or that
17 threaten ecosystem flows of goods and services important to human welfare.
18 Nitrogen Saturation of Terrestrial Ecosystems
19 Long-term, chronic additions of Nr (including nitrate deposition from ambient PM) to
20 terrestrial ecosystems is resulting in numerous ecosystems shifting to a detrimental ecological
21 condition known as "nitrogen saturation." Nitrogen saturation does not occur at a specific point
22 in time, but is a set of gradually developing critical changes in ecosystem processes which
23 represent the integrated response of a system to increased nitrogen availability over time (Aber,
24 1992). It occurs when nitrogen inputs exceed the capacity of plants and soil microorganisms to
25 utilize and retain the nitrogen (Aber et al., 1989, 1998; Garner, 1994; EPA, 1993). Under
26 conditions of nitrogen saturation, some other resource generally replaces nitrogen in limiting
27 biotic functions. The appearance of nitrate in soil solution (leaching) is an early symptom of
28 excess Nr accumulation.
29 Not all vegetation, organisms, or ecosystems react in the same manner to increased Nr
30 availability from atmospheric deposition. This is due in part to the variation both within and
January 2005 6-35 Draft - Do Not Quote or Cite
-------
1 across species in their inherent capacity to utilize additional Mr and the suite of other factors that
2 influence the range of community or ecosystem types possible at any given location. Such
3 factors can include the mineral composition of the underlying bedrock, the existing soil nutrient
4 pools, the local climatic.conditions including weather extremes such as drought, High/low
5 temperatures, topography, elevations, natural/land use history, and fire regimes.
6 In U.S. ecosystems, the nutrient whose supply most often sets the limit of possible
7 primary productivity at a given site is biologically available nitrogen. However, in any given
8 ecosystem, not all plants are equally capable of utilizing extra nitrogen. Those plants that are
9 predisposed to capitalize on any increases in Nr availability gain an advantage over those that are
10 not as responsive to added nutrients. Over time, this shift in the competitive advantage may lead
11 to shifts in overall plant community composition. Whether or not this shift is considered adverse
12 would depend on the management context within which that ecosystem falls and the ripple
13 effects of this shift on other ecosystem components, essential ecological attributes (EEAs), and
14 ecosystems. '
15 The effect of additions of nitrates on plant community succession patterns and
^ 16 biodiversity has been studied in several long-term nitrogen fertilization studies in both the U.S.
(
17 and Europe. These studies suggest that some forests receiving chronic inputs of nitrogen may
18 decline in productivity and experience greater mortality (Fenn et al. 1998). For example,
19 fertilization and nitrogen gradient experiments at Mount Ascutney, VT suggest that nitrogen
20 saturation may lead to the replacement of slow-growing, slow nitrogen-cycling spruce-fir forest
21 stands by fast-growing deciduous forests that cycle nitrogen rapidly (Fenn et al. 1998).
22 Similarly, experimental studies of the effects of Nr deposition overa 12-year period on
23 Minnesota grasslands dominated by native warm-season grasses observed the shift to low-
24 diversity mixtures dominated by cool-season grasses at all but the lowest rates of Nr addition
25 (Wedin and Tilman, 1996). The shift to low-diversity mixtures was associated with the decrease
26 in biomass carbon to N (C:N) ratios, increased Nr mineralization, increased soil nitrate, high
27 nitrogen losses, and low carbon storage. Grasslands with high nitrogen retention arid carbon
28 storage rates were the most vulnerable to loss of species and major shifts in nitrogen cycling.
29 (Wedin and Tilman, 1996).
t
January 2005 6-36 Draff - Do Not Quote or Cite
-------
1 The carbon-to-nitrogen (C:N) ratio of the forest floor can be changed by nitrogen
2 deposition over time. In Europe, low C:N ratios coincide with high deposition regions. A strong
3 decrease in forest floor root biomass has also been observed with increased nitrogen availability,
4 and appears to occur when the ecosystem becomes nitrogen saturated. If root growth and
5 mycorrhizal formation are impaired by excessive nitrogen deposition, the stability of the forest
6 floor vegetation may be affected. The forest floor C:N ratio has been used as a rough indicator
7 of ecosystem nitrogen status in mature coniferous forests and the risk of nitrate leaching. Nitrate
8 leaching has been significantly correlated with forest floor nitrate status, but not with nitrate
9 deposition. Therefore, to predict the rate of changes in nitrate leaching, it is necessary to be able
10 to predict the rate of changes in the forest floor C:N ratio. Understanding the variability in forest
11 ecosystem response to nitrogen input is essential in assessing pollution risks (Gundersen et al,
12 1998; CD, pp. 4-106,4-107).
13 In the U.S., forests that are now showing severe symptoms of nitrogen saturation include:
14 the northern hardwoods and mixed conifer forests in the Adirondack and Catskill Mountains of
15 New York; the red spruce forests at Whitetop Mountain, Virginia, and Great Smoky Mountains
16 National Park, North Carolina; mixed hardwood watersheds at Femow Experimental Forest in
17 West Virginia; American beech forests in Great Smoky Mountains National Park, Tennessee;
18 mixed conifer forests and chaparral watersheds in southern California and the southwestern
19 Sierra Nevada in Central California; .the alpine tundra/subalpine conifer forests of the Colorado
20 Front Range; and red alder forests in the Cascade Mountains in Washington. All these systems
21 have been exposed to elevated nitrogen deposition, and nitrogen saturated watersheds have been
22 reported in the above-mentioned areas. Annual nitrogen additions through deposition in the
23 southwestern Sierra Nevada are similar in magnitude to nitrogen storage in vegetation growth
24 increments of western forests, suggesting that current nitrogen deposition rates may be near the
25 assimilation capacity of the overstory vegetation. Ongoing urban expansion will increase the
26 potential for nitrogen saturation of forests from urban sources (e.g., Salt Lake City, Seattle,
27 Tucson, Denver, central and southern California) unless there are improved emission controls
28 (Fenn et al., 1998).
January 2005
6-37
Draft - Do Not Quote or Cite
t
-------
'1 The composition and structure of the plant community within an ecosystem in large part
2 determines the food supply and habitat types available for use by other organisms. In terrestrial
3 systems, plants serve as the integrators between above-ground and below-ground environments
4 and are influenced by and influence conditions in each. It is because of these linkages that
5 chronic excess Nr additions can lead to complex, dramatic, and severe ecosystem level/wide
6 changes/responses. Changes in soil Nr influence below ground communities as well. A
7 mutualistic relationship exists in the rhizosphere (plant root zone) between plant roots, fungi, and
8 microbes. Because the rhizosphere is an important region of nutrient dynamics, its function is
9 critical for the growth of the organisms involved. The plant roots provide shelter and carbon for
10 the symbionts, whereas the symbionts provide access to limiting nutrients such as nitrogen and
11 phosphorus for the plant. Bacteria make N, S, Ca, P, Mg, and K available for plant use while
12 fungi in association with plant roots form mycorrhizae that are essential in the uptake by plants
13 of mineral nutrients, such as N and P (Section 4.3.3; Wall and Moore, 1999; Rovira and Davy,
14 1974). Mycorrhizal fungal diversity is associated with above-ground plant biodiversity,
15 ecosystem variability, and productivity (Wall and Moore, 1999). Studies suggest that during
16 nitrogen saturation, soil microbial communities change from being predominately fungal, and
17 dominated by mycdrrhizae, to being dominated by bacteria (Aber et al., 1998; CD, pp. 4-107,4-
18 108), dramatically affecting both above- and belowiground ecosytems. These types of effects
19 have been observed in the field. For example, the coastal sage scrub (CSS) community in
20 California has been declining in land area and in drought deciduous shrub density over the past
21 60 years, and is being replaced in many areas by Mediterranean annual grasses. At the same
22 time, larger-spored below-ground fungal species (Scutellospora and Gigaspord), due to a failure
23 to sporulate, decreased in number with a concomitant proliferation of small-spored species of
24 Glomus aggregatum, G leptotichum, and G. geosporum, indicating a strong selective pressure
25 for the smaller spored species of fungi (Edgerton-Warburton and Allen, 2000). These results
26 demonstrate that nitrogen enrichment of the soil significantly alters the arbuscular mycorrhizal
27 species composition and richness, and markedly decreases the overall diversity of the arbuscular
28 mycorrhizal community. The decline in the coastal sage scrub species can be directly linked to
f
January 2005 6-3 8 Draft - Do Not Quote or Cite
-------
1 the decline of the arbuscular mycorrhizal community (Edgerton-Warburton and Allen, 2000;
2 Allen et al., 1998; Padgett et al., 1999)(CD, pp. 4-108,4-109). -
3 Impacts on threatened and endangered species. In some rare and unique U.S.
4 ecosystems, the chronic additions of atmospherically-derived nitrogen have already had some
5 dire and perhaps irreversible consequences. For example, California has many species that occur
6 in shrub, forb, and grasslands affected by N deposition, with up to 200 sensitive plant species in
7 southern California CSS alone (Skinner and Pavlik, 1994). Some 25 plant species are already
8 extinct in California, most of them annual and perennial forbs that occurred in sites now
9 experiencing conversion to annual grassland. As CSS converts more extensively to annual
10 grassland dominated by invasive species, loss of additional rare species may be inevitable.
11 Though invasive species are often identified as the main threat to rare species, it is more likely
12 that invasive species combine with other factors, such as excess N deposition, to promote
13 increased productivity of invasive species and resulting species shifts.
14 Not surprisingly, as sensitive vegetation is lost, wildlife that depend on these plants are
15 adversely affected. Included among these wildlife species are several threatened or endangered
16 species listed by the U.S. Fish and Wildlife Service, such as the desert tortoise and checkerspot
17 butterfly. A native to San Francisco Bay area, the bay checkerspot butterfly (Euphydryas editha
18 bayensis), has been declining steadily over the past decade, with local extirpations in some
19 reserves. This decline has been associated with the invasion of exotic grasses replacing the
20 native forbs on which the butterfly depends. In particular, the larval stage is dependent on
21 primarily one host plant, Plantago erecta, which is increasingly being out-competed by exotic
22 grasses.
23 Similarly, the desert tortoise has declined due to a number of co-occurring stresses,
24 including grazing, habitat destruction, drought, disease, and a declining food base. In the desert
25 shrub inter-spaces, sites where native forbs once flourished, invasive grasses now dominate,
26 reducing the nutritional quality of foods available to the tortoise (Fenn et al., 2003; Nagy et al.,
27 1998). Nitrogen deposition contributes to the productivity and density of N-fertilized grasses at
28 the expense of native forbs (Brooks, 2003). "Thus, protection of endangered species will
29 require increased exotic grass control, but local land management strategies to protect these
January 2005 • 6-39 Draft - Do Not Quote or Cite
f
-------
1 endangered species may not succeed unless they are accompanied by policy changes at the
2 regional or national level that reduce air pollution" (Fenn et al., 2003).
3 Community composition of epiphytic lichens is readily altered by small increases in
4 nitrogen deposition, an effect that seems to be widespread in the West (Fenn et al., 2003). Most
5 epiphytic lichens meet their nutritional requirements from atmospheric deposition and can store
6 N in excess of their nutritional needs (van Herk, 1999). In the San Bernardino Mountains, up to
7 50% of the lichen species that occurred in the region in the early 1900s have disappeared, with a
8 disproportionate number of the locally extinct species being (epiphytic) cyanolichens (Fenn et
9 al., 2003; Nash and Sigal, 1999). The Pacific Northwest, in contrast, still has widespread
10 populations of pollution-sensitive lichens (Fenn et al., 2003). However, in urban areas, intensive
11 agricultural zones and downwind of major urban and industrial centers, there is a sparsity of
12 sensitive lichen species and high levels of N concentrations have been measured in lichen tissue
13 (Fenn et al., 2003). Replacement of sensitive lichens by nitrophilous species has'undesirable
14 ecological consequences. In late-successional, naturally N-limited forests of the Coast Range
15 and western Cascades, epiphytic cyanolichens make important contributions to mineral cycling
16 and soil fertility (Pike 1978, Sollins et al., 1980, Antoine, 2001), and together with other large,
17 pollution-sensitive macrolichens, are an integral part of the food web for large and small
18 mammals, insects and birds (McCune and Geiser, 1997).
19 Alteration of native fire cycles. Several lines of evidence suggest that N deposition may
20 be contributing to greater fuel loads and thus altering the fire cycle in a variety of ecosystem
21 types, although further study is needed (Fenn et al., 2003). Invasive grasses promote a rapid fire
22 cycle in many locations (D'Antonio and Vitousek, 1992). The increased productivity of
23 flammable understory grasses increases the spread of fire and has been hypothesized as one
24 mechanism for the recent conversion of CSS to grassland (Minnich and Dezzani, 1998).
25 Thus, through its effect on habitat suitability, genetic diversity, community dynamics and
26 composition, nutrient status, energy and nutrient cycling, and frequency and intensity of natural
27 disturbance regimes (fire), excess Nr deposition is having profound and adverse impact on the
28 essential ecological attributes associated with terrestrial ecosystems. Strong correlation between
29 the stressor and adverse environmental response exists in many locations, and N-addition studies
January 2005 . 6-40 Draft - Do Not Quote or Cite
-------
1 have confirmed this relationship between stressor and response. Loss of species and genetic
2 diversity are clearly adverse ecological effects and adverse to the public welfare. Research
3 efforts should be made to elucidate what role paniculate deposition is playing in contributing to
4 these effects so as to facilitate the mitigation of such effects.
5 Effects of Nitrogen Addition on Aquatic Habitats
6 Aquatic ecosystems (streams, rivers, lakes, estuaries or oceans) receive increased
7 nitrogen inputs either from direct atmospheric deposition (including nitrogen-containing
8 particles), surface runoff, or leaching from nitrogen saturated soils into ground or surface waters.
9 The primary pathways of Nr loss from forest ecosystems are hydrological transport beyond the
10 rooting zone into groundwater or stream water, or surface flows of organic nitrogen as nitrate
11 and Nr loss associated with soil erosion (Fenn et al., 1998). In the east, high nitrate
12 concentrations have been observed in streams draining nitrogen saturated watersheds in the
13 southern Appalachian Mountains (Fenn et al., 1998). The Great Smoky Mountains National
14 Park in Tennessee and North Carolina receives elevated levels of total atmospheric deposition of
15 sulfur and nitrogen. A major portion of the atmospheric loading is from dry and cloud
16 deposition. Nitrogen saturation of the watershed resulted in extremely high exports of nitrate
17 and promoted both chronic and episodic stream acidification in streams draining undisturbed
18 watersheds. Significant export of base cations was also observed (CD, pp. 4-110,4-111; see also
19 section 6.3.3.2 on acidification from PM deposition).
20 In the west, the Los Angeles Air Basin exhibited the highest stream water NCy
21 concentrations in wilderness areas of North America (Bytnerowicz and Fenn, 1996; Fenn et al.,
22 1998). Chronic N deposition in southern California, in the southwestern Sierra Nevada, and in
23 the Colorado Front Range leads to increased net N mineralization and nitrification rates in soil
24 and to elevated NO3" concentrations in lakes and streams. These symptoms occur in low- and
25 mid-elevation, high-deposition areas (>15 kg N/ha/yr) and in high elevation sites with relatively
26 low N deposition (4 to 8 kg N/ha/yr) but little capacity to assimilate and retain added N.
27 Estuaries are among the most intensely fertilized systems on Earth (Fenn et al., 1998).
28 They receive far greater nutrient inputs than other systems. In the Northeast, for example,
29 nitrogen is the element most responsible for eutrophication in coastal waters of the region. Since
30 the early 1900s, there has been a 3r to 8-fold increase in nitrogen flux fromlO watersheds in the
January 2005 - 6-41 Draft - Do Not Quote or Cite
f
-------
northeast These increases are associated with nitrogen oxide emissions from combustion which
2 have increased 5-fold. Riverine nitrogen fluxes have been correlated with atmospheric
3 deposition onto their landscapes and also with nitrogen oxides emissions into their airsheds.
4 Data from 10 benchmark watersheds with good historical records indicate that about 36-80% of
5 the riverine total nitrogen export, averaging approximately 64%, was derived directly or
6 indirectly from nitrogen oxide emissions (CD, pp. 4-109,4-110).
7 The Pamlico Sound, NC estuarine complex, which serves as a key fisheries nursery
8 supporting an estimated 80% of commercial and recreational finfish and shellfish catches in the
9 southeastern U.S. Atlantic coastal region, has also been the subject of recent research (Paerl et
10 al., 2001) to characterize the effects of nitrogen deposition on the estuary. Direct atmospheric
11 nitrogen deposition onto waterways feeding into the Pamlico Sound or onto the Sound itself and
12 indirect nitrogen inputs via runoff from upstream watersheds contribute to conditions of severe
13 water oxygen depletion; formation of algae blooms in portions of the Pamlico Sound estuarine
14 complex; altered fish distributions, catches, and physiological states; and increases in the
15 incidence of disease. Especially under extreme rainfall events (e.g., hurricanes), massive -
16 influxes of nitrogen (in combination with excess loadings of metals or other nutrients) into
17 watersheds and sounds can lead to dramati c decreases of oxygen in water and the creation of
18 widespread "dead zones" and/or increases in algae blooms that can cause extensive fish kills and
• ' * ' -
19 damage to commercial fish and sea food harvesting (Paerl et al., 2001; CD, pp. 4-109,4-110).
20 6.3.3.2 Environmental Effects of PM-Related Acidic Deposition.
21 <' Acidic deposition has emerged over the past quarter century as a critical environmental
22 stress that affects diverse terrestrial and aquatic ecosystems in North America, Europe, and Asia
23 (Driscoll et al., 2001). In the eastern U.S. for example, the current acidity in precipitation is at
24 least twice as high as in pre-industrial times, with mean precipitation pH ranges from 4.3
25 (Pennsylvania and New York) to 4.8 (Maine) (EPA, 2003). Acidic deposition is highly variable
26 across space and time, can originate from transboundary air pollution, can travel hundreds of
27 miles before being deposited, thereby affecting large geographic areas. It is composed of ions,
28 gases, and particles derived from the precursor gaseous emissions of SO2, NOX, NH3 and
January 2005 6-42 Draft - Do Not Quote or Cite
-------
1 particulate emissions of other acidifying compounds. Acid deposition disturbs forest and aquatic
2 ecosystems by giving rise to harmful chemical conditions (Dricoll et al., 2001).
3 Terrestrial Effects
4 Acidic deposition has changed the chemical composition of soils by depleting the content
5 of available plant nutrient cations (e.g., Ca2+, Mg2+, K+) by increasing the mobility of Al, and by
6 increasing the S and N content (Driscoll et al., 2001). Soil leaching is often of major
7 importance in cation cycles, and many forest ecosystems show a net loss of base cations (CD, pp.
8 4-118). In acid sensitive soils, mineral weathering (the primary source of base cations in most
9 watersheds) is insufficient to keep pace with leaching rates accelerated by acid deposition
10 (Driscoll et al.3 2001).
11 In the absence of acid deposition, cation leaching in northeastern forest soils is driven
12 largely by naturally occurring organic acids derived from the decomposition of organic matter.
13 Organic acids tend to mobilize Al through formation of organic-Al complexes, most of which are
14 deposited lower in the soil profile through adsorption to mineral surfaces. This process, termed
15 podzolization, results in surface waters with low concentrations of Al. Such concentrations are
16 primarily in a nontoxic, organic form (Driscoll et al., 1998). Acid deposition, however, has
17 altered podzolization by solubilizing Al with mobile inorganic anions, facilitating the transport .
18 of inorganic Al into surface waters. In forest soils with base saturation values less than 20%,
19 acidic deposition leads to increased Al mobilization and a shift in chemical speciation of Al from
20 organic to inorganic forms that are toxic to terrestrial and aquatic biota.
21 The toxic effect of Al on forest vegetation is attributed to its interference with plant
22 uptake of essential nutrients, such as Ca and Mg. Because Ca plays a major role in cell
23 membrane integrity and cell wall structure, reductions in Ca uptake suppress cambial growth,
24 reduce the rate of wood formation, decrease the amount of functional sapwood and live crown,
25 and predispose trees to disease and injury from stress agents when the functional sapwood
26 becomes less than 25% of cross sectional stem area (Smith, 1990a). There are large variations in
27 Al sensitivity among ecotypes, between and within species, due to differences in nutritional
28 demands and physiological status, that are related to age and climate, which change over time
29 (CD, pp. 4-126).
January 2005 ' 6-43 Draft - Do Not Quote or Cite
f
-------
N1 Acidic deposition has been firmly implicated as a causal factor in the northeastern high-
2 elevation decline of red spruce (DeHayes et al., 1999). Red spruce is common in Maine, where
3 it is an important commercial species. It is also common at high elevations in mountainous
4 regions throughout the Northeast, where it is valued for recreation and aesthetics, as well as for
5 providing a habitat for unique and endangered species. Dieback has been most severe at high
6 elevations in the Adirondack and Green Mountains, where more than 50% of the canopy trees
7 died during the 1970s and 1980s. In the White Mountains, about 25% of the canopy spruce died
8 during that same period (Craig and Friedland 1991). Dieback of red spruce trees has also been
9 observed in mixed hardwood-conifer stands at relatively low elevations in the western
10 Adirondack Mountains, areas that receive high inputs of acidic deposition (Shortie et al., 1997).
11 Results of controlled exposure studies show that acidic mist or cloud water reduces the cold
12 tolerance of current-year red spruce needles by 3-10 degrees C (DeHayes et al., 1999). This
13 increased susceptibility to freezing occurs due to the loss of membrane-associated Ca2+ from
14 needles through leaching caused by the hydrogen ion. The increased frequency of winter injury
15 in the Adirondack and Green Mountains since 1955 coincides with increased exposure of red
16 spruce canopies to highly acidic cloud water (Johnson et al., 1984). Recent episodes of winter
17 injury have been observed throughout much of the range of red spruce' in the Northeast
18 (DeHayes et al., 1999). DeHayes etal. (1999) indicate that there is a significant positive
19 association between cold tolerance and foliar calcium in trees exhibiting deficiency in foliar
20 calcium, and further state that their studies raise the strong possibility that acid rain alteration of
21 foliar calcium is not unique to red spruce but has been demonstrated in many other northern
22 temperate forest tree species including yellow birch (Betula alleghaniensis), white spruce (Picea
23 r glaucus), red maple (Acer rubrum) eastern white pine (Pinus strobus), and sugar maple (Acer
24 saccharum) (CD, p. 4-120).
25 Although less well established, there is also a strong possibility mat low Ca to Al ratios
26 in soils may also be impacting northeastern red spruce. Cronan and Grigal (1995) concluded that
27 a Ca:Al ratio of less than 1.0 in soil water indicated a greater than 50% probability of impaired
28 growth in red spruce. They cite examples of studies from the northeast where soil solutions in
29 the field were found to exhibit Ca: Al ionic ratios less than 1.0.
January 2005 .
6-44
Draft - Do Not Quote or Cite
-------
1 Acidic deposition may also be contributing to episodic dieback of sugar maple in the
2 Northeast through depletion of nutrient cations from marginal soils. Horsley et al. (1999) found
3 that dieback at 19 sites in northwestern and north-central Pennsylvania and south-western New
4 York was correlated with combined stress from defoliation and deficiencies of Mg and Ca.
5 Dieback occurred predominately on ridgetops and on upper slopes, where soil base availability
6 was much lower than at mid and low slopes of the landscape (Bailey et al., 1999). Because
7 multiple factors such as soil mineralogy and landscape position affect soil base status, the extent
8 to which sugar maple dieback can be attributed to acidic deposition is not clear.
9 Less sensitive forests throughout the U.S. are experiencing gradual losses of base cation
10 nutrients, which in many cases will reduce the quality of forest nutrition over the long term
11 (National Science and Technology Council, 1998). In some cases, such effects may not even
12 take decades to occur because these forests have already been receiving S and N deposition for
13 many years.
14 In contrast to contributing to the adverse impacts of acid deposition, particles can also
15 provide a beneficial supply of base cations to sites with very low rates of supply from mineral
16 sources. In these areas, atmospheric inputs of bass cations can help ameliorate the acidifying
17 effects of acid particles. The Integrated Forest Study (IFS) (Johnson and Lindberg, 1992) has
18 characterized the complexity and variability of ecosystem responses to atmospheric inputs and
19 provided the most extensive data set available on the effects of atmospheric deposition, including
20 particle deposition, on the cycling of elements in forest ecosystems. This study showed that in
21 the IFS ecosystems, inputs of base cations have considerable significance, not only for base
22 • cation status, but also for the potential of incoming precipitation to acidify or alkalize the soils.
23 The actual rates, directions, and magnitudes of changes that may occur in soils (if any), however,
24 will depend on rates of inputs from weathering and vegetation outputs, as well as deposition and
25 leaching. In other words, these net losses or gains of base cations must be placed in the context
26 of the existing soil pool size of exchangeable base cations (CD, p. 4-132). Given the wide
27 ranges of paniculate deposition for each base cation across the IFS sites, however, the unique
28 characteristics of various sites need to be better understood before assumptions are made about
29 the role particulate pollution plays in ecosystem impacts (CD, pp. 4-127,4-128).
January 2005
6-45
Draft - Do Not Quote or Cite
-------
\m
1 In a follow up study, Johnson et al. (1999) used the nutrient cycling moded, NuCM, to
2 simulate the effects of reduced S, N, and base cation (CB) deposition on nutrient pools, fluxes,
3 soil, and soil solution chemistry in two contrasting southern Appalachian forest ecosystems. The
4 authors found that in an extremely acidic system, CB deposition can have a major effect on CB
5 leaching through time and S and N deposition had a major effect on Al leaching. At the less
6 acidic Coweeta site, CB deposition had only a minor effect on soils and soil solutions; whereas S
7 and N deposition had delayed but major effects on CB leaching (CD, pp. 4-136,4-137).
8 Aquatic Effects
9 Inputs of acidic deposition to regions with base-poor soils have resulted in the
10 acidification of soil waters, shallow ground waters, streams, and lakes in anumber of locations
11_ within the U.S. In addition, perched seepage lakes, which derive water largely from direct
12 precipitation inputs, are highly sensitive to acidic deposition (Charles, 1991). These processes
13 usually result in lower pH and, for drainage lakes, higher concentrations of inorganic monomeric
14 Al. Such changes in chemical conditions are toxic to fish and other aquatic animals. (Driscoll et
15 al., 2001).
16 A recent report, Response of Surface Water Chemistry to the Clean Air Act of 1990
17 (EPA, 2003), analyzes data from 1990 through 2000 obtained from EPA's Long Term
18 .Monitoring (LTM) and Temporally Integrated Monitoring of Ecosystems (TIME) projects, part
19 of EMAP (Environmental Monitoring and Assessment Program). The report assesses recent
20 changes in surface water chemistry in response to changes in deposition, in the northern and
21 eastern U.S., specifically in the acid sensitive regions defined as New England (Maine, New
22 Hampshire, Vermont and Massachusetts), the Adirondack Mountains of New York, the Northern
23 Appalachian Plateau (New York, Pennsylvania and West Virginia), the Ridge and Blue Ridge
24 Provinces of Virginia, and the Upper Midwest (Wisconsin and Michigan). Acidic waters are
25 defined as having acid neutralizing capacity '(ANC) less than zero (i.e., no acid buffering
26 capacity in the water), corresponding to a pH of about 5.2. Increases in surface water ANC
27 values and/or pH would indicate improved buffering capacity and signal the beginning of
28 recovery (EPA, 2003).
29 Using National Atmospheric Deposition Program (NADP) data, trends in sulfate and N
30 (nitrate + ammonium) deposition were analyzed, along with CB deposition, sulfate and nitrate
January 2005 6-46 Draft - Do Not Quote or Cite
-------
1 concentrations in surface waters, ANC and pH levels. Over this timeframe, sulfate deposition
2 declined significantly across all regions, while N declined slightly in the Northeast and increased
3 slightly in the Upper Midwest. Base cation deposition showed no significant changes in the East
4 and increased slightly in the Upper Midwest. Concurrently, all regions except the Ridge/Blue
5 Ridge province in the mid-Atlantic showed significant declines in sulfate concentrations in
6 surface waters, while nitrate concentrations decreased in two regions with the highest ambient
7 nitrate concentrations (Adirondacks, Northern Appalachian Plateau) but were relatively
8 unchanged in regions with low concentrations.
9 Given the declines in S and N deposition measured for these areas, one would expect to
10 find increasing values of ANC, pH or both in response. ANC values did increase in the
11 Adirondacks, Northern Appalachian Plateau and Upper Midwest, despite a decline in base
12 cations (Ca and Mg) in each region. The loss of base cations limited the extent of ANC and pH
13 increase. Toxic Al concentrations also declined slightly in the Adirondacks. In New England
14 and Ridge/Blue Ridge, however, regional surface water ANC did not change significantly (EPA,
15 2003).
16 Modest increases in ANC have reduced the number of acidic lakes and stream segments
17 in some regions. There are an estimated 150 Adirondack lakes with ANC less than 0, or 8.1% of
18 the population, compared to 13% (240 lakes) in the early 1990s. In the Upper Midwest, an
19 estimated 80 of 250 lakes that were acidic in mid-1980s are no longer acidic. TIME surveys of
20 streams in the Northern Appalachian Plateau region estimated that 8.5% (3,600 kilometers) of
21 streams remain acidic at the present time, compared to 12% (5,014 kilometers) of streams that
22 were acidic in 1993-94. In these three regions taken together, approximately one-fourth to one-
23 third of formerly acidic surface waters are no longer acidic, although still with very low ANC.
24 The report finds little evidence of regional change in the acidity status of New England or the
25 Ridge/Blue Ridge regions and infers that the numbers of acidic waters remain relatively
26 unchanged. Despite a general decline in base cations and a possible increase in natural organic
27 acidity, there is no evidence that the number of acidic waters have increased in any region (EPA,
28 2003).
January 2005 6-47 Draft - Do Not Quote or Cite
-------
1 Acidification has marked effects on the trophic structure of surface waters. Decreases in
2 pH and increases in Al concentrations contribute to declines in species richness and in the
3 abundance of zooplankton, macroinvertebrates, and fish (Schindler et al.,1985; Keller and Gunn
4 1995). Numerous studies have shown that fish species richness (the number offish species in a
5 water body) is positively correlated with pH and ANC values (Rago and Wiener, 1986 Kretser et
6 al.j 1989). Decreases in pH result in decreases in species richness by eliminating acid-sensitive
7 species (Schindler et al. 1985). Of the 53 fish species recorded by the Adirondack Lakes Survey
8 Corporation, about half (26 species) are absent from lakes with pH below 6.0. Those 26 species
9 include important recreational fishes, such as Atlantic salmon, tiger trout, redbreast sunfish,
10 bluegill, tiger musky, walleye, alewife, and kokanee (Kretser et al. 1989), plus ecologically
11 important minnows that serve as forage for sport fishes.
«
12 A clear link exists between acidic water, which results from atmospheric deposition of
13 strong acids, and fish mortality. The Episodic Response Project (ERP) study showed that
14 streams with moderate to severe acid episodes had significantly higher fish mortality during
15 bioassays than nonacidic streams (Van Sickle et al., 1996). The concentration of inorganic
16 monomeric Al was the chemical variable most strongly related to mortality in the four test
17 species (brook trout, mottled sculpin, slimy sculpin, and blacknose dace). The latter three
18 species are acid sensitive. In general, trout abundance was lower in ERP streams with median
19 episode pH less than 5.0 and inorganic monomeric Al concentrations greater than 3.7-7.4 mmol
20 L"1. Acid sensitive species were absent from streams with median episode pH less than 5.2 and
21 with a concentration of inorganic monomeric Al greater than 3.7 mmol L'1..
22 Given the significant reductions in sulfur emissions that have occurred in the U. S. and
23 Europe in recent decades, the findings of Driscoll et al. (1989,2001 j and Hedin et al. (1994) are
24 especially relevant. Driscoll et al. (1989, 2001) noted a decline in both SO4"2 and base cations in
25 both atmospheric deposition and stream water over the past two decades at Hubbard Brook
26 Watershed, NH. However, the reductions in S02 emissions in Europe and North America in
27 recent years have not been accompanied by equivalent declines in net acidity related to sulfate in
28 precipitation, and may have, to varying degrees, been offset by steep declines in atmospheric
29 base cation concentrations over the past 10 to 20 years (Hedin et al., 1994).
January 2005 6-48 Draft - Do Not Quote or Cite
-------
1 Driscoll et al. (2001) envision a recovery process that will involve two phases. Initially,
2 a decrease in acidic deposition following emissions controls will facilitate a phase of chemical
3 recovery in forest and aquatic ecosystems. Recovery time for this phase will vary widely across
4 ecosystems and will be a function of the following:
5 • the magnitude of decreases in atmospheric deposition
6 • the local depletion of exchangeable soil pools of base cations
7 • the local rate of mineral weathering and atmospheric inputs of base cations
8 • the extent to which soil pools of S and N are released as SO42" or as NO3" to drainage
9 waters and the rate of such releases (Galloway et al. 1983),
10
11 In most cases, it seems likely that chemical recovery will require decades, even with additional
12 controls on emissions. The addition of base cations, e.g., through liming, could enhance
13 chemical recovery at some sites.
14 The second phase in ecosystem recovery is biological recovery, which can occur only if
15 chemical recovery is sufficient to allow survival and reproduction of plants and animals. The
16 time required for biological recovery is uncertain. For terrestrial ecosystems, it is likely to be at
17 least decades after soil chemistry is restored because of the long life of tree species and the
18 complex interactions of soil, roots, microbes, and soil biota. For aquatic systems, research
19 suggests that stream macroinvertebrate populations may recover relatively rapidly
20 (approximately 3 years), whereas lake populations of zooplankton are likely to recover more
21 slowly (approximately 10 years) (Gunn and Mills 1998). Some fish populations may recover in
22 5 to 10 years after the recovery of zooplankton populations. Stocking could accelerate fish
23 population recovery (Driscoll et al., 2001)
24 Projections made using an acidification model (PnET-BGC) indicate that full
25 implementation of the 1990 CAAA will not afford substantial chemical recovery at Hubbard
26 Brook EF and at many similar acid-sensitive locations (Driscoll et al., 2001) . Model
27 calculations indicate that the magnitude and rate of recovery from acidic deposition in the
28 northeastern U.S. are directly proportional to the magnitude of emissions reductions. Model
29 evaluations of policy proposals calling for additional reductions in utility SO2 and NOX
January 2005
6-49
Draft - Do Not Quote or Cite
-------
t
1' emissions, year round emissions controls, and early implementation indicate greater success in
2 facilitating the recovery of sensitive ecosystems (Driscoll et al., 2001).
3 Indirect Vegetation and Ecosystem Effects from Atmospheric PM
4 In addition to the direct and indirect effects of deposited PM, ambient atmospheric PM
5 can effect radiation and climate conditions that influence overall plant/ecosystem productivity.
6 The degree to which these effects occur in any given location will depend on the chemical and
7 physical composition and concentration of the ambient PM. Because plants are adapted to the
8 overall light and temperature environments in which they grow, any PM-related changes to these
9 conditions potentially alter the overall competititive success these plants will have in that
10 ecosystem.
11 With respect to radiation, the characteristics and net receipts of solar and terrestrial
12 radiation determine rates of both photosynthesis and the heat-driven process of water cycling.
13 Atniospheric turbidity (the degree of scattering occurring in the atmosphere due to particulate
14 loading) influences the light environment of vegetative canopy in two ways: through conversion
15 of direct to diffuse radiation and by scattering or reflecting incoming radiation back out into
16 space. Diffuse radiation increases canopy photosynthetic productivity by distributing radiation
17 more uniformly throughout the canopy so that it also reaches the lower leaves and improves the
18 canopy radiation use efficiency (RUE). Acting in the opposite direction, non-absorbing,
19 scattering aerosols present in PM reduce the overall amount of radiation reaching vegetative
20 surfaces, by scattering or reflecting it back into space. It appears that global albedo has been
21 increasing due to an increasing abundance of atmospheric particles. Using World
22 Meteorological Organization (WMO) data, Stanhill and Cohen (2001) have estimated that
23 average solar radiation receipts have declined globally by an average of 20 W m-2 since 195 8.
24 The net effect of atmospheric particles on plant productivity is not clear, however, as the
25 enrichment in photosynthetically active radiation (PAR) present in diffuse radiation may offset a
26 portion of the effect of decreased solar radiation receipts in some instances (CD, pp.' 4-92,4-93).
1 t
27 Plant processes also are sensitive to temperature. Some atmospheric particles (most
28 notably black carbon) absorb short-wavelength solar radiation, leading to atmospheric heating
29 and reducing total radiation received at the surface. Canopy temperature and transpirational
30 water use by vegetation are particularly sensitive to long-wave, infrared radiation. Atmospheric
January 2005 6-50 Draft - Do Not Quote or Cite
-------
^ \
1 heating by particles can potentially reduce photosynthetic water uptake efficiency and vertical
2 temperature gradients, potentially reducing the intensity of atmospheric turbulent mixing,
3 Stanhill and Cohen (2001) suggested that plant productivity is more affected by changes in
4 evapotranspiration induced by changes in the amount of solar radiation plants receive than by
5 changes in the amount of PAR plants receive (CD, p. 4-93).
6
• 7 63.4 Characteristics and Location of Sensitive Ecosystems in the U.S.
8 Ecosystems sensitive to anthropogenically derived nitrogen and/or acid deposition tend to
9 have similar characteristics. Some of these ecosystems and characteristics have already been
10 mentioned in earlier sections but are repeated here to provide a more comprehensive list that can
11 help ecological risk assessors/managers identify areas of known or potential concern. For
12 example, lower nitrogen and/or resource environments, such as those with infertile soils, shaded
13 understories, deserts, or tundras, are populated with organisms specifically adapted to survive
14 under those conditions. Plants adapted to these conditions have been observed to have similar
15 characteristics, including inherently slower growth rates, lower photosynthetic rates, and lower
16 capacity for nutrient uptake, and grow in soils with lower soil microbial activity. When N
17 becomes more readily available, such plants will be replaced by nitrophilic plants which are
18 better able to use increased amounts of Nr (Fenn et al., 1998).
19 Additionally, in some instances, there seem to be important regional distinctions in
20 exposure patterns, environmental stressors, and ecosystem characteristics between the eastern
21 A and western U.S.. A seminal report describing these distinctive characteristics for the western
22 U.S. (11 contiguous states located entirely west of the 100th meridian) is Fenn et al., 2003.
23 In the western U.S., vast areas receive low levels of atmospheric deposition, interspersed
24 with hotspots of elevated N deposition downwind of large, expanding metropolitan centers or
25 large agricultural operations, hi other words, spatial patterns of urbanization largely define the
26 areas where air pollution impacts are most severe. The range of air pollution levels for western
27 wildlands is extreme, spanning from near-background to the highest exposures in all of North
28 America, with the possible exception of forests downwind of Mexico City. Over the same
29 geographic expanse, climatic conditions and ecosystem types vary widely. Some regions receive
January 2005 6-51 Draft - Do Not Quote or Cite
-------
7,
1
2
3
4
5
6
7
8
9
10
II
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
more than 1000 millimeters of precipitation, namely the Pacific coastal areas, the Sierra Nevada,
x
the Colorado Rockies, and northern Idaho, while other regions are arid or semiarid, with more
than 300 clear days per year (Riebsame et al., 1997). In these latter regions, the contribution of
atmospheric dry deposition is likely to be most important These characteristics which are
unique to the West require special consideration, and often make application of models and
ecological effects thresholds developed for other regions inappropriate.
In summary, sensitive or potentially sensitive ecosystems in the west include those that:
• are located downwind of large urban source areas; regions with a mix of emissions
sources that may include urban, mobile, agricultural, and industrial sources; and/or sites
near large point.sources of N.
• contain inherently N sensitive ecosystem components, such as lichens, diatoms, or poorly
buffered watersheds which produce high streamwater NO3- levels: These sensitive
components can be affected by N deposition rates as low as 3-8 kg/ha/yr.
• occur on top of siliclastic/crystalline bedrock with little potential for buffering acidity.
• are naturally nitrogen limited. For example, the approximately 16,000 high elevation
western mountain lakes are generally oligotrophic and especially sensitive to the effects
of atmospheric deposition.
A seminal report describing key characteristics of sensitive ecosystems for the eastern
and in particular the northeastern U.S. is Driscoll et al. (2001). In the northeastern United States,
atmospheric deposition is largely a regional problem. Because S and N most often occur
together in the eastern atmosphere and deposit to the environment as acidic deposition, acidic
deposition is seen as a critical environmental stress.
Several critical chemical thresholds appear to coincide with the onset of deleterious
effects to biotic resources resulting from acid deposition. Thus, ecosystems sensitive to
additional acid inputs include those with the following characteristics:
a molar Ca:Al ratio of soil water that is less than 1.0; ' •
• soil percentage base cation saturation less than 20%;
• surface water pH less than 6.0; *
• ANC less than 50 meqL-1; and
• concentrations of inorganic monomeric Al greater than 2 mmol L-l.
Knowledge of such indicators is necessary for restoring ecosystem structure and function.
January 2005
6-52
Draft - Do Not Quote or Cite
-------
1 6.3.5 Ecosystem Exposures to PM Stressor Deposition
2 In order for any specific chemical stressor present in ambient PM to impact ecosystems,
3 it must first be removed from the atmosphere through deposition. Deposition can occur in three
4 modes: wet (rain/frozen precipitation), dry, or occult (fog, mist or cloud). At the national scale,
5 all modes of deposition must be considered in determining potential impacts to vegetation and
6 ecosystems because each mode may dominate over specific intervals of time or space. (CD, p.
7 4-8 to 4-10). For example, in large parts of the western U.S. which are arid or semiarid, dry
8 deposition may be the source of most deposited PM (Fenn, et al., 2003). However, in coastal
9 areas or high elevation forests, wet or occult deposition may predominate. Where the latter is the
10 case, deposition levels may greatly exceed PM levels measured in the ambient air. Occult
11 deposition is particularly effective for delivery of dissolved and suspended materials to
12 vegetation because: (1) concentrations of ions are often many-fold higher in clouds or fog than in
13 precipitation or ambient air (e.g., acidic cloud water, which is typically 5-20 times more acid
14 than rainwater, can increase pollutant deposition and exposure to vegetation and soils at high
15 elevation sites by more than 50% of wet and dry deposition levels); (2) PM is delivered in a
16 hy drated and bioavailable form to foliar surfaces and remains hy drated due to conditions of high
17 relative humidity and low radiation; and (3) the mechanisms of sedimentation and impaction for
18 submicron particles that would normally be low in ambient air are increased. High-elevation
19 forests can be especially at risk from depositional impacts because they receive larger paniculate
20 deposition loadings than equivalent low-elevation sites, due to a number of orographic
21 (mountain related) effects. These orographic effects include higher wind speeds that enhance the
22 rate of aerosol impaction, enhanced rainfall intensity and composition, and increased duration of
23 occult deposition. Additionally, the needle-shaped leaves of the coniferous species often found
24 growing in these high elevation sites, enhance impaction and retention of PM delivered by all
25 three deposition modes (CD, pp. 4-29,4-44).
26 In order to establish exposure-response profiles useful in ecological risk assessments, two
27 types of monitoring networks need to be in place. First, a deposition network is needed that can
28 track changes in deposition rates of PM stressors (nitrates/sulfates) occurring in sensitive or
29 symptomatic areas/ecosystems. Secondly, a network or system of networks that measure the
January 2005
6-53
Draft - Do Not Quote or Cite
-------
1 response of key ecological indicators sensitive to changes in atmospheric deposition of PM
2 stressors is also needed.
3 ' Currently in the U.S., national deposition monitoring networks routinely measure total
4 wet or dry deposition of certain compounds. Atmospheric concentrations of dry particles began
5 to be routinely measured in 1986, with the establishment of EPA's National Dry Deposition
6 Network (NDDN). After new monitoring requirements were added in the 1990 CAAA, EPA, in
7 cooperation with the National Oceanic and Atmospheric Association, created the Clean Air
8 Status and Trends Network (CASTNet) from the NDDN. C ASTNet comprises 85 sites and is
9 considered the nation's primary source for atmospheric data to estimate concentrations for
10 ground-level ozone and the chemical species that make up the'dry deposition component of total
11 acid deposition (e.g., sulfate, nitrate, ammonium, sulfur dioxide, and nitric acid), as well as the
12 associated meteorology and site characteristics data that are needed to model dry deposition
13 velocities (CD. pg. 4-21: ftittp://.wvvw.e^p..a.g<)Tv/castneu')..
14 . To provide data on wet deposition levels in the U.S., the National Atmospheric
15 Deposition Program (NADP) was initiated in the late 1970's as a cooperative program between
16 federal, state, and other public and private groups. By the mid-1980's, it had grown to nearly
17 200 sites, and it stands today as the longest running national atmospheric deposition monitoring
18 network (http://nadp.sws.. liiuc. eduA.
19 In addition to these deposition monitoring networks, other networks collect data on
20 ambient aerosol concentrations and chemical composition. Such networks include the
21 IMPROVE network, discussed above in section 2.5, and the newly implemented PM25 chemical
22 Speciation Trends Network (STN) that consists of 54 core National Ambient Monitoring
23 Stations and approximately 250 State and Local Air Monitoring Stations.
24 Data from these deposition networks demonstrate that N and S compounds are being
25 deposited onto soils and aquatic ecosystems in sufficient amounts to impact ecosystems at local,
26 regional and national scales. Though the percentages of N and S containing compounds in PM
27 vary spatially and temporally, nitrates and sulfates make up a substantial portion of the chemical
28 composition of PM. In the future, speciated data from these networks may allow better
January 2005 6-54 Draft - Do Not Quote or Cite
-------
Y
1 understanding of the specific components of total deposition that are most strongly influencing
2 PM-related ecological effects.
3 Unfortunately, at this time there is only limited long-term ecosystem response monitoring
4 taking place at the national level. Two exceptions are the Hubbard Brook Experimental Forest
5 research site, that provides the longest continuous record of precipitation and stream chemistry in
6 the U.S. (Likens and Bormann, 1995) and EPA's LTM and TIME projects which monitor
7 changes in surface water chemistry in the acid sensitive regions of the northern and eastern U.S..
8 Because the complexities of ecosystem response make predictions of the magnitude and timing
9 of chemical and biotic recovery uncertain, it is strongly recommended that this lype of long-term
10 surface water chemistry monitoring network be continued, and that a biological monitoring
11 program be added. Data from these long-term monitoring sites will be invaluable for the
12 evaluation of the response of forested watersheds and surface waters to a host of research and
13 regulatory issues related to acidic deposition, including soil and surface water recovery, controls
14 on N retention, mechanisms of base cation depletion, forest health, sinks for S in watersheds,
15 changes in dissolved organic carbon and speciation of Al, and various factors related to climate
16 change (EPA, 2003).
17
18 63.6 Critical Loads
19 The critical load (CL) has been defined as a "quantitative estimate of an exposure to one
20 or more pollutants below which significant harmful effects on specified sensitive elements of the
21 environment do not occur according to present knowledge" (Lokke et al., 1996). The critical
22 load framework originated in Europe where the concept has generally been accepted as the basis
23 for abatement strategies to reduce or prevent injury to the functioning and vitality of forest
24 ecosystems caused by long-range transboundary chronic acidic deposition. The concept is
25 useful for estimating the amounts of pollutants that sensitive ecosystems can absorb on a
26 sustained basis without experiencing measurable degradation. The estimation of ecosystem
27 critical loads requires an understanding of how an ecosystem will respond to different loading
28 rates in die long term and is a direct function of the level of sensitivity of the ecosystem to the
29 pollutant and its capability to ameliorate pollutant stress.
January 2005 6-55 Draft - Do Not Quote or Cite
-------
1 Key to the establishment of a critical load is the selection of appropriate ecological
2 endpoints or indicators that are measurable characteristics related to the structure, composition,'
3 or functioning of ecological systems (i.e., indicators of condition). In Europe, the elements used
4 in the critical load concept are a biological indicator, a chemical criterion, and a critical value
5 (CD, p. 4-124). A number of different indicators for monitoring ecosystem status have been
6 proposed. Indicators of ecosystems at risk of N saturation could include: foliar nitrogen, nutrient
7 ratios (N:P, N:cation); foliar nitrate; foliar 515 N; arginine concentration; soil C:N ratio; NO3" in
8 soil extracts or increased and prolonged NO3~ loss below the main rooting zone and in stream
9 water or in soil solution; and flux rates of nitrogenous trace gases from soil (Fenn et al., 1998),
10 Seasonal patterns of stream water nitrate concentrations are especially good indicators of
11 watershed N status. Biological indicators that have been suggested for use in the critical load
12 calculation in forest ecosystems include mycorrhizal fungi (Lokke et al., 1996) and fine roots,
13 since they are an extremely dynamic component of below-ground ecosystems and can respond
14 rapidly to stress. The physiology of carbon allocation has also been suggested as an indicator of
15 anthropogenic stress (Andersen and Rygiewicz, 1991). Lichen community composition, in
16 terrestrial ecosystems or lichen N tissue levels are also fairly responsive to changes in N
17 deposition over time (Fenn et al., 2003). In aquatic systems, diatom species composition can be
18 a good indicator of changes in water chemistry (Fenn et al., 2003). It should be kept in mind,
19 however, that the response of a biological indicator is an integration of a number of different
20 stresses. Furthermore, there may be organisms more sensitive to the pollutant(s) than the species
21 selected (Lokke et al., 1996; National Science and Technology Council, 1998) (CD, pp. 4-124 to
22 126).
23 Within North America, a number of different groups have recently begun to use or
24 develop critical loads. As discussed below, these groups include the U.S. Federal Land
25 Managers (FLMs), such as the National Park Service and the Forest Service, a binational group
26 known as New England Governors/Eastern Canadian Premiers (NEG/ECP), and several
27 Canadian Provinces.
28 Federal Land Managers have hosted a number of meetings over the last few years to
29 discuss how the CL concept might be used in helping them fulfill their mandate of providing
January 2005
6-56
Draft - Do Not Quote or Cite
-------
1 protection for the lands they manage. In trying to develop a consistent approach to using CL; a
i
2 number of issues and considerations have been identified. First, the distinction between critical
3 loads (which are based on modeled or measured dose-response data) and target loads (which can
4 be based on political, economic, spatial or temporal considerations in addition to scientific
5 information) needs to be recognized. When using the critical or target load (TL) approach, one
6 must indicate the spatial (or geographic) scope, the temporal scope (timeframe to ecological or
7 ecosystem recovery), and a description of the sensitive receptors (or resource) to be protected,
8 the sensitive receptor indicators (physical, chemical biological, or social characteristics of the
9 receptor that can be measured), and the harmful effect on the receptor that is of concern.
10 Additionally, one would need to specify what is the "desired condition" that the critical or target
11 load is meant to achieve. For any given location, there may be a range or suite of possible
12 critical or target loads based on different sensitive receptors and/or receptor indicators found at
13 that site. Alternatively, one could focus on the most sensitive receptor and select a single CL or
14 TL for that receptor. Several aspects of the CL approach make it attractive for use by the FLMs.
15 Specifically, it can provide a quantitative, objective and consistent approach for evaluating
16 resource impacts. In an effort to progress the CL approach, the Forest Service is testing the
17 applicability of the European protocol to several U.S. case study sites.
18 Under the auspices of the NEG/ECP, and other binational efforts, Canadian and U.S.
19 scientists are involved in joint forest mapping projects. A Forest Mapping Work Group has been
20 tasked with conducting a regional assessment of the sensitivity of northeastern North American
21 forests to current and projected sulfur and nitrogen emissions levels, identifying specific forested
22 areas most sensitive to continued deposition and estimating deposition rates required to maintain
23 forest health and productivity. They have completed the development of methods, models and
24 mapping techniques, and identification of data requirements. Some of these data requirements
25 include: pollution loading to forest landscapes; the interaction of pollutants with forest canopies;
26 plant nutrient requirements; and the ability of soils to buffer acid inputs and replenish nutrients
27 lost due to acidification.
28 In addition to the CL measure, they have also defined a "deposition index" as the
29 difference between the CL and current deposition levels. Positive values of the index reflect the
January 2005
6-57
Draft - Do Not Quote or Cite
-------
1 capacity of a forest ecosystem to tolerate additional acidic deposition. Negative index values
2 correspond to the reduction in S and N deposition required to eliminate or deter the development
3 of future nutrient limitations. This allows an assessor to identify areas where the deposition
4 problems are most severe, and which sites might be under the CL level currently but not far from
5 reaching or exceeding that level should deposition levels increase. Currently maps exist for
6 Vermont and Newfoundland, though the goal is to develop maps that will cover Quebec and the
7 Atlantic provinces of Canada, along with the remaining New England states. These maps show
8 that 31% of Vermont forests and 23% of Newfoundland forests are sensitive (e.g., current levels
9 of S and N deposition are causing cation depletion).
10 Though these current activities hold promise for using the CLs approach in
11 environmental assessments and in informing management decisions, widespread use of CL's in
12 the U.S. is not yet possible. Critical loads is a very data-intensive approach, and, at the present
13' time, there is a paucity of ecosystem- level data for most sites. However, for a limited number of
14 areas which already have a long-term record of ecosystem monitoring, (e.g., Rocky Mountain
i
15 National Park in Colorado and me Lye Brook Wilderness in Vermont), FLMs may be able to
16 develop site-specific CLs. Further, in areas already exceeding the CL, it may be difficult to
17 determine what the management goals are/should be for each mapped area (e.g., what is the
18 "desired condition" or level of protection) without historic baseline data. More specifically, with
19 respect to PM deposition, there are insufficient data for the vast maj ority of U. S. ecosystems that
20 differentiate the PM contribution to total N or S depostion to allow for practical application'bf
21 this approach as a basis for developing national standards to protect sensitive U.S. ecosystems
22 from adverse effects related to PM deposition. Though atmospheric sources of Nr and acidifying
23 compounds, including ambient PM, are clearly contributing to the overall excess pollutant load
24 or burden entering ecosystems annually, insufficient data are available at this time to quantify
25 the contribution of ambient PM to total Nr or acidic deposition as its role varies both temporally
26 and spatially along with a number of other factors. Thus, it is not clear whether a CL could be
27 developed just for the portion of the total N or S input that is contributed by PM.
28
January 2005
6-58
Draft - Do Not Quote or Cite
-------
\
1 6.3.7 Summary and Conclusions
2 The above discussions identify a group of ecosystems known to be sensitive to excess N
3 and S inputs and a list of characteristics that can be used to predict or locate other potentially
4 sensitive ecosystems within the U.S. Further, exposures of these sensitive ecosystems to
5 atmospherically derived pollutants (e.g., N and S) have been measured and documented, in some
6 cases for decades. Clear linkages between reduced atmospheric concentrations of these
7 pollutants and reduced deposition rates have been made. The mechanisms of environmental and
8 ecosystem responses to these inputs are increasingly understood, though very complex.
9 Fertilization and acidification studies have verified observed ecosystem responses to these
10 pollutants in the field. Ecosystem-level effects associated with excess N and S inputs are
11 profound, but in most cases potentially reversible. New assessment and management tools, such
12 as critical and target loads, are being developed to better characterize the relationship between
13 deposition loads and ecosystem response. The success of these tools will depend on the
14 availability of sufficient ecosystem response data, which is currently limited to a few long-term
15 monitoring networks/sites (e.g., TIME/LTM). The current risk to sensitive* ecosystems and
16 especially sensitive species like the checkerspot butterfly, desert tortoise, epiphytic lichens,
17 native shrub and forb species, and aquatic diatom communities is high. The loss of species and
18 whole ecosystem types is adverse and should receive increased protection.
19 A number of ecosystem-level conditions (e.g., nitrogen saturation, terrestrial and aquatic
20 acidification, coastal eutrophication) have been associated with chronic, long-term exposure of
21 ecosystems to elevated inputs of compounds containing Nr, sulfur and/or associated hydrogen
22 ions. These ecosystem level changes profoundly impact almost all of the EEAs identified in the
.23 EPEC Framework (SAB, 2002) and described in sections 4.2.1 and 4.2,3 of the CD. These
24 impacted EEAs include Landscape Condition, Biotic Condition, Chemical and Physical
25 Characteristics, Ecological Processes, and Natural Disturbance Regimes. Given that humans, as
26 well as other organisms, are dependent on the services ecosystems provide, ecosystem changes
27 of mis magnitude are of concern and can lead to adverse impacts on human health and welfare.
28 Based on the information included in the above discussions and Chapters 4 and 9 of the
29 CD, staff has reached the following conclusions:
January 2005
6-59
Draft - Do Not Quote or Cite
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
• An ecologically-relevant indicator for PM would be based on one or multiple chemical
stressors found in ambient PM (e.g. N or S containing compounds).
• Ecosystem effects can be associated with long-term high or even low levels of excess
inputs. Thus, there is no bright line or threshold for effects, but rattier a "syndrome" of
complex changes over time. Additionally, ecosystem recovery can occur but may take
decades, and may require controls beyond those already established.
• Excess N or acid deposition acts in conjunction with other co-occurring stresses (e.g.,
invasive species, reduced grazing pressure) that jointly determine ecological outcomes.
Therefore, these pollution-related stresses should not be considered in isolation.
Additionally, all forms of airborne nitrogen and acidic compounds need to be considered
and managed in harmony.
• Monitoring networks may be sufficient to measure air concentrations or deposition but
are not generally sufficient to monitor ecosystem response. For example, in the West;
more environmental monitoring is needed downwind of large urban areas.
Unfortunately, our ability to relate ambient concentrations of PM to ecosystem response
is hampered by a number of significant data gaps and uncertainties. First, U.S. monitoring
networks have only recently begun to measure speciated PM. Historically, measurements were
focused only on a particular size fraction such as PM10 and, more recently, PM15. An exception
to this is the IMPROVE network, which collects speciated measurements. Additionally, except
for the IMPROVE and some CASTNet sites, much of the PM monitoring effort has focused on
urban or near urban exposures, rather than on those in sensitive ecosystems. Thus, the lack of a
long-term, historic database of annual speciated PM deposition rates precludes establishing
relationships between PM deposition (exposure) and ecosystem response at this time. As a
result, while evidence of PM-related effects clearly exists, there is insufficient information
available at this time to serve as a basis for a secondary national air quality standard for PM,
specifically selected to protect against adverse effects on vegetation and ecosystems. .
A second source of uncertainty lies in predicting deposition velocities based on ambient
concentrations of PM. There are a multitude of factors that influence the amounts of PM that get
deposited from the air onto sensitive receptors, including the mode of deposition (wet, dry, and
occult), wind speed, surface roughness or stickiness, elevation, particle characteristics (e.g., size,
January 2005
6-60
Draft - Efo Not Quote or Cite
-------
1 shape, chemical composition), and relative humidity. Therefore, modeled deposition rates, used
2 in the absence of monitored data, can be highly uncertain.
3 Third, each ecosystem has developed within a context framed by the topography,
4 underlying bedrock,'soils, climate, meteorology, hydrologic regime, natural and land use history,
5 species associations that co-occur at that location (e.g., soil organisms, plants), and successional
6 stage, making it unique from all others. Because of this variety, and insufficient baseline data on
7 each of these features for most ecosystems, it is currently not possible to extrapolate with much
8 confidence any effect from one ecosystem to another, or to predict an appropriate "critical load"
9 for the vast majority of U.S. ecosystems.
10 As additional PM speciated air quality and deposition monitoring data become available,
11 there is much room for fruitful research into the areas of uncertainty identified above. At this
12 time, however, staff concludes that there is insufficient information available to recommend for
13 consideration an ecologically defined secondary standard that is specifically targeted for
14 protection of vegetation and ecosystems against the adverse effects potentially associated with
t
15 the levels of PM-related stressors of nitrate and sulfate found in the ambient air.
16
17 6.4 EFFECTS ON MATERIALS
18 The effects of the deposition of atmospheric pollution, including ambient PM, on
19 materials are related to both physical damage and aesthetic qualities. The deposition of PM
20 (especially sulfates and nitrates) can physically affect materials, adding to the effects of natural
21 weathering processes, by potentially promoting or accelerating the corrosion of metals, by
22 degrading paints, and by deteriorating building materials such as concrete and limestone.
23 Particles contribute to these physical effects because of their electrolytic, hygroscopic and acidic
24 properties, and their ability to sorb corrosive gases (principally SO2). As noted in the last
25 review, only chemically active fine-mode or hygroscopic coarse-mode particles contribute to
26 these physical effects (EPA 1996b, p. VIII-16).
27 In addition, the deposition of ambient PM can reduce the aesthetic appeal of buildings
28 and culturally important articles through soiling. Particles consisting primarily of carbonaceous
29 compounds cause soiling of commonly used building materials and culturally important items
January 2005
6-61
Draft - Do Not Quote or Cite
I
-------
1 such as statues and works of art (CD, p. 4-191). Soiling is the deposition of particles on surfaces
2 by impingement, and the accumulation of particles on the surface of an exposed material results
3 in degradation of its appearance. Soiling can be remedied by cleaning or washing, and
4 depending on the soiled material, repainting (EPA, 1996b, p. VIII-19),
5 Building upon the information presented in the last Staff Paper (EPA, 1996b), and
6 including the limited new information presented in Chapter 4 (section 4.4) of the CD, the
7 following sections summarize the physical damage and aesthetic soiling effects of PM on
8 materials including metals, paint finishes, and stone and concrete.
9
10 6.4.1 Materials Damage Effects '
11 Physical damage such as corrosion, degradation, and deterioration occurs in metals, paint
12 finishes, and building materials such as stone and concrete, respectively: Metals are affected by
13 natural weathering processes even in the absence of atmospheric pollutants. Atmospheric
14 pollutants, most notably S02 and participate sulfates, can have an additive effect, by promoting
15 and accelerating the corrosion of metals. The rate of metal corrosion depends on a number of
16 factors, including the deposition rate and nature of the pollutants; the influence of the protective
17 corrosion film that forms on metals, slowing corrosion; the amount of moisture present;
18 variability in electrochemical reactions; the presence and concentration of other surface
19 electrolytes; and the orientation of the metal surface. Historically, studies have shown that the
20 rate of metal corrosion decreases in the absence of moisture, since surface moisture facilitates
21 the deposition of pollutants and promotes corrosive electrochemical reactions on metals (CD, pp.
22 4-192 to 4-193).
23 The CD (p. 4-194, Table 4-18) summarizes the results of a number of studies
24 investigating the roles of particles and SO2 on the corrosion of metals. The CD concludes that
25 the role of particles in the corrosion of metals is not clear (CD, p. 4-193). While several studies
26 suggest that particles can promote the corrosion of metals, others have not demonstrated a
27 correlation between particle exposure and metal corrosion. Although the corrosive effects of
28 SO2 exposure in particular have received much study, there remains insufficient evidence to
t
January 2005 6-62 Draft - Do Not Quote or Cite
-------
1 relate corrosive effects to specific participate sulfate levels or to establish a quantitative
2 relationship between ambient paniculate sulfate and corrosion.
3 Similar to metals, paints also undergo natural weathering processes, mainly from
4 exposure to environmental factors such as sunlight, moisture, fungi, and varying temperatures.
5 Beyond these natural processes, atmospheric pollutants can affect the durability of paint finishes
6 by promoting discoloration, chalking, loss of gloss, erosion, blistering, and peeling. Historical
7 evidence indicates that particles can damage painted surfaces by serving as carriers of more
8 corrosive pollutants, most notably SO2, or by serving as concentration sites for other pollutants.
9 If sufficient damage to the paint occurs, pollutants may penetrate to the underlying surface. A
10 number of studies available in the last review showed some correlation between PM exposure
11 and damage to automobile finishes. In particular, Wolff et al. (1990) concluded that damage to
12 automobile finishes resulted from calcium sulfate forming on painted surfaces by the reaction of
13 calcium from dust particles with sulfuric acid contained in rain or dew. In addition, paint films
14 permeable to water are also susceptible to penetration by acid-forming aerosols (EPA 1996b, p.
15 VHI-18). The erosion rate of oil-based house paint has reportedly been enhanced by exposure to
16 SO2 and humidity; several studies have suggested that this effect is caused by the reaction of S02
17 with extender pigments such as calcium carbonate and zinc oxide, although Miller et al. (1992)
18 suggest mat calcium carbonate acts to protect paint substrates (CD, p. 4-196).
19 With respect to damage to building stone, numerous studies discussed in the CD (pp.
20 4-196 to 4-202; Table 4-19) suggest that air pollutants, including sulfur-containing pollutants
21 and wet or dry deposition of atmospheric particles and dry deposition of gypsum particles, can
22 enhance natural weathering processes. Exposure-related damage to building stone results from
23 the formation of salts in the stone that are subsequently washed away by rain, leaving the surface
24 more susceptible to the effects of air pollutants. Dry deposition of sulfur-containing pollutants
25 and carbonaceous particles promotes the formation of gypsum on the stone's surface. Gypsum is
26 a black crusty material that occupies a larger volume than the original stone, causing the stone's
27 surface to become cracked and pitted, leaving rough surfaces that serve as sites for further
28 deposition of airborne particles (CD, p. 4-200).
January 2005
6-63
Draft - Do Not Quote or Cite
I
-------
1 The rate of stone deterioration is determined by the pollutant mix and concentration, the
2 stone's permeability and moisture content, and the pollutant deposition velocity. Dry deposition
3 of SO2 between rain events has been reported to be a major causative factor in pollutant-related
4 erosion of calcareous stones (e.g., limestone, marble, and carbonated'cement). While it is clear
5 from the available information that gaseous air pollutants, in particular SO2, will promote the
6 decay of some types of stones under specific conditions, carbonaceous particles (non-carbonate
7 carbon) and particles containing metal oxides may help to promote the decay process (CD, p.
8 4-201 ,4-202).
9
10 6.4.2 Soiling Effects
11 Soiling affects the aesthetic appeal of painted surfaces. In addition to natural factors,
12 exposure to PM may give painted surfaces a dirty appearance. Early studies demonstrated an
13 association between particle exposure and increased frequency of cleaning painted surfaces. •
14 More recently, Haynie and Lemmons (1990) conducted a study to determine how various
15 environmental factors contribute to the rate of soiling on white painted surfaces. They reported'
16 that coarse-mode particles initially contribute more to soiling of horizontal and vertical surfaces
17 man do fine-mode particles, but are more easily removed by rain, leaving stains on the painted •
18 surface. The authors concluded that the accumulation of fine-mode particles, rather than coarse-
19 mode particles, more likely promotes the need for cleaning of the painted surfaces (EPA 1996b,
20 p. VIII-21-22; CD, pp. 4-202 to 4-204). Haynie and Lemmons (1990) and Creighton et al.
21 (1990) reported that horizontal surfaces soiled faster than vertical surfaces and that large
22 particles were primarily responsible for the soiling of horizontal surfaces not exposed to rainfall.
23 Additionally, a study was conducted to determine the potential soiling of artwork in five
24 Southern California museums (Ligocki, et al., 1993). Findings were that a significant fraction of
25 fine elemental carbon and soil dust particles in the ambient air penetrates to the indoor
26 environment and may constitute a soiling hazard to displayed artwork (EPA 1996b, p. VIII-22).
27 As for stone structures, the presence of gypsum is related to soiling of the stone surface
28 by providing sites for particles of dirt to concentrate. Lorusso et al. (1997) attributed the need
S
29 for frequent cleaning and restoration of historic monuments in Rome to exposure to total
t
January 2005-
6-64
Draft - Do Not Quote or Cite
-------
1 suspended particles (TSP). Further, Davidson etal. (2000) evaluated the effects of air pollution
2 exposure on a limestone structure on the University of Pittsburgh campus using estimated
3 average TSP levels in the 1930s and 1940s and actual values for the years 1957 to 1997.
4 Monitored levels of SO2 were also available for the years 1980 to 1998. Based on the available
5 data on pollutant levels and photographs, the authors concluded that soiling began while the
6 structure was under construction. With decreasing levels of pollution, the soiled areas have been
7 slowly washed away, the process taking several decades, leaving a white, eroded surface (CD,
8 pp. 4-203).
9
10 6.4.3 Summary and Conclusions
11 Damage to building materials results from natural weathering processes that are
12 enhanced by exposure to airborne pollution, most notably sulfur-containing pollutants. Ambient
13 PM has been associated with contributing to pollution-related damage to materials, and can
14 cause significant detrimental effects by soiling painted surfaces and other building materials.
15 Available data indicate that particle-related soiling can result in increased cleaning frequency
16 and repainting, and may reduce the useful life of the soiled materials. However, to date, no
17 quantitative relationships between particle characteristics (e.g., concentrations, particle size, and
18 chemical composition) and the frequency of cleaning or repainting have been established. Thus,
19 staff-concludes that PM effects on materials can play no quantitative role in considering whether
20 any revisions of the secondary PM NAAQS are appropriate at this time.
21
22 6.5 EFFECTS ON CLIMATE CHANGE AND SOLAR RADIATION
23 Atmospheric particles alter the amount of electromagnetic radiation transmitted through
24 the earth's atmosphere by both scattering and absorbing radiation. As discussed above in .
6
25 Chapter 2 (section 2.2.6), most components of ambient PM (especially sulfates) scatter and
26 reflect incoming solar radiation back into space, thus offsetting the "greenhouse effect" to some
27 degree by having a cooling effect on climate. In contrast, some components of ambient PM
28 (especially black carbon) absorb incoming solar radiation or outgoing terrestrial radiation, and
29 are believed to contribute to some degree to atmospheric wanning. Lesser impacts of
January 2005
6-65
Draft - Do Not Quote or Cite
I
-------
1 atmospheric particles are associated with their role in altering the amount of ultraviolet solar
2 radiation (especially UV-B) penetrating through the earth's atmosphere to ground level, where it
3 can exert a variety of effects on human health, plant and animal biota, and other environmental
4 components (CD, p. 205). The extensive research and assessment efforts into global climate
5 change and stratospheric ozone depletion provide evidence that atmospheric particles play
6 important roles in these two types of atmospheric processes, not only on a global scale, but also
7 on regional and local scales as well.
8 Information on the role of atmospheric particles in these atmospheric processes and the
9 effects on human health and the environment associated with these atmospheric processes is
10 briefly summarized below, based on the information in section 4.5 of the CD and referenced"
11 reports. These effects are discussed below in conjunction with consideration of the potential
12 indirect impacts on human health and the environment that may be a consequence of climatic
13 and radiative changes attributable to local and regional changes in ambient PM.
14
15 6.5.1 Climate Change and Potential Human Health and Environmental Impacts
16 As discussed in section 4.5.1 of the CD, particles can have both direct and indirect effects
17 on climatic processes.' The direct effects are the result of the same processes responsible for
18 visibility degradation, namely radiative scattering and absorption. However, while visibility
19 impairment is caused by particle scattering in all directions, climate effects result mainly from
20 scattering light away from the earth and into space. This reflection of solar radiation back to
21 space decreases the transmission of visible radiation to the surface and results in a decrease in
22 the heating rate of the surface and the lower atmosphere. At the same time, absorption of either
23 incoming solar radiation or outgoing terrestrial radiation by particles, primarily black carbon,
24 results in an increase in the heating rate of the lower atmosphere.
25 In addition to these direct radiative effects, particles can also have a number of indirect
26 effects on climate related to their physical properties. For example, sulfate particles can-serve as
27 condensation nuclei which alter the size distribution of cloud droplets by producing more
28 droplets with smaller sizes. Because the total surface area of the cloud droplets is increased, the
29 amount of solar radiation that clouds reflect back to space is increased. Also, smaller cloud
I
January 2005
6-66
Draft - Do Not Quote or Cite
-------
1 droplets have a lower probability of precipitating, causing them to have longer atmospheric
2 lifetimes. An important consequence of this effect on cloud properties is the suppression of rain
3 and potentially major disruption of hydrological cycles downwind of pollution sources, leading
4 to a potentially significant alteration of climate in the affected regions (CD, p. 4-218).
5 The overall radiative and physical effects of particles, both direct and indirect, are not the
6 simple sum of effects caused by individual classes of particles because of interactions between
7 particles and other atmospheric gases. As discussed in Section 4.5.1.2 of the CD, the effects of
8 sulfate particles have been the most widely considered, with globally averaged radiative effects
9 of sulfate particles generally estimated to have partially offset the warming effects caused by
10 increases in greenhouse gases. On the other hand, global-scale modeling of mineral dust
11 particles suggests that even the sign as well as the magnitude of effects depends on the vertical
12 distribution and effective particle radius.
13 The CD makes clear that atmospheric particles play an important role in climatic
14 processes, but that their role at this time remains poorly quantified. In general, on a global scale,
15 the direct effect of radiative scattering by atmospheric particles is to likely exert an overall net
16 effect of cooling the atmosphere, while particle absorption may lead to wanning. The net impact
17 of indirect effects on temperature and rainfall patterns remains difficult to generalize. However,
18 deviations from global mean values can be very large even on a regional scale, with any
19 estimation of more localized effects introducing even greater complexity (CD, p. 216). The CD
20 concludes that any effort to model the impacts of local alterations in particle concentrations on
21 projected global climate change or consequent local and regional weather patterns would be
22 subject to considerable uncertainty (CD, p. 4-240).
23 More specifically, the CD notes that while current climate models are successful in
24 simulating present annual mean climate and the seasonal cycle on continental scales, they are
25 lass successful at regional scales (CD, p. 4-207). Findings from various referenced assessments
26 illustrate well the considerable uncertainties and difficulties in projecting likely climate change
27 impacts on regional or local scales. For example, uncertainties in calculating the direct radiative
28 effects of atmospheric particles arise from a lack of knowledge of their vertical and horizontal
29 variability, their size distribution, chemical composition, and the distribution of components
January 2005 6-67 Draft - Do Not Quote or Cite
f
-------
1 within individual particles. Any complete assessment of the radiative effects of PM would
2 require computationally intensive calculations that incorporate the spatial and temporal behavior
3 of particles of varying composition that have been emitted from, or formed by precursors emitted
4 from, different sources. In addition, calculations of indirect physical effects of particles on
5 climate (e.g., related to alteration of cloud properties and disruption of hydrological cycles) are
6 subj ect to much larger uncertainties than those related to the direct radiative effects of particles
7 (CD, p. 4-219). The CD concludes that at present impacts on human health and the environment
8 due to aerosol effects on the climate system can not be calculated with confidence, and notes that
9 the uncertainties associated with such aerosol-related effects will likely remain much larger than
10 those associated with greenhouse gases (CD, p. 4-219). Nevertheless, the CD concludes that
11 substantial qualitative information available from observational and modeling studies indicates
12 that different types of atmospheric aerosols (i.e., different components of PM) have both
13 warming and cooling effects on climate, both globally and regionally. Studies also suggest that
14 global and regional climate changes could potentially have both positive and negative effects on
15 human health, human welfare, and the environment.
16 '
17 6.5.2 Alterations in Solar UV-B Radiation and Potential Human Health and
18 - Environmental Impacts
19 As discussed in section 4.5.2 of the CD, the effects of particles in the lower atmosphere.
20 " on the transmission of solar UV-B radiation have been examined both by field measurements
21 and by radiative transfer model calculations. Several studies cited in the CD reinforce the idea
22 that particles can play an important role in modulating the attenuation of solar UV-B radiation,
23 although none included measurements of ambient PM concentrations, so that direct relationships
24 between PM levels and UV-B radiation transmission could not be determined. The available
25 studies, "conducted in diverse locations around the world, demonstrate that relationships between
26 particles and solar UV-B radiation transmission can vary considerably over location, conditions,
27 and time. While ambient particles are generally expected to decrease.the flux of solar UV-B
28 radiation reaching the surface, any comprehensive assessment of the radiative effects of particles
29 would be location-specific and complicated by the role of particles in photochemical activity in
30 . the lower atmosphere. Whether the photochemical production of ozone is enhanced, remains the
'January 2005 6-68 Draft - Do Not Quote or Cite
-------
\
1 same, or reduced by the presence of ambient particles will be location-specific and dependent on
2 particle composition. Also complicating any assessment of solar UV-B radiation penetration to
3 specific areas of the earth's surface are the influences of clouds, which in turn are affected by the
4 presence of ambient particles.
5 The main types of effects associated with exposure to UV-B radiation include direct
6 effects on human health and agricultural and ecological systems, indirect effects on human
7 health and ecosystems, and effects on materials (CD, p. 4-221). The study of these effects has
8 been driven by international concern over potentially serious increases in the amount of solar
9 UV-B radiation reaching the earth's surface due to the depletion of the stratospheric ozone layer
10 by the release of various man-made ozone-depleting substances. Extensive qualitative and
11 quantitative characterizations of these global effects attributable to proj ections of stratospheric
12 ozone depletion have been periodically assessed in studies carried out under WMO and UNEP
13 auspices, with the most recent projections being published in UNEP (1998,2000) and WMO
14 (1999).
15 Direct human health effects of UV-B radiation exposure include: skin damage (sunburn)
16 leading to more rapid aging and increased incidence of skin cancer; effects on the eyes, including
17 retinal damage and increased cataract formation possibly leading to blindness; and suppression
18 of some immune system components, contributing to skin cancer induction and possibly
19 increasing susceptibility to certain infectious diseases. Direct environmental effects include
20 damage to terrestrial plants, leading to possible reduced yields of some major food crops and
21 commercially important tress, as well as to biodiversity shifts in natural terrestrial ecosystems;
22 and adverse effects on aquatic life, including reductions in important components of marine food
23 chains as well as other aquatic ecosystem shifts. Indirect health and environmental effects are
24 primarily those mediated through increased tropospheric ozone formation and consequent
25 ground-level ozone-related health and environmental impacts. Effects on materials include
26 accelerated polymer weathering and other effects on man-made materials and cultural artifacts.
\
27 In addition, there are emerging complex issues regarding interactions and feedbacks between
28 climate change and changes in terrestrial and marine biogeochemical cycles due to increased
29 UV-B radiation penetration. (CD, p. 4-221, 4-222).
January 2005
6-69
Draft - Do Not Quote or Cite
-------
1 In contrast to these types of negative impacts associated with increased UV-B penetration
2 to the Earth's surface, the CD (p. 4-222,4-223) summarizes research results that are suggestive
3 of possible beneficial effects of increased UV-B radiation penetration. For example, a number of
4 studies have focused on the protective effects of UV-B radiation with regard to non-skin cancer
5 incidence, which proved suggestive evidence that UV-B radiation, acting through the production
6 of vitamin D, may be a risk-reduction factor for mortality due to several types of cancer,
7 including cancer of the breast, colon, ovary, and prostate, as well as non-Hodgkin lymphoma,
8 Hie various assessments of these types of effects that have been conducted consistently
9 note that the modeled projections quantitatively relating changes in UV-B radiation (attributable
10 to stratospheric ozone depletion) to changes in health and environmental effects are subject to
11 considerable uncertainty, with the role of atmospheric particles being one of numerous
12 complicating factors. Taking into account the complex interactions between ambient particles
13 and UV-B radiation transmission through the lower atmosphere, the CD concludes that any
14 effort to quantify projected indirect effects of variations in atmospheric PM on human health or
15 the environment due to particle impacts on transmission of solar UV-B radiation would require
16 location-specific evaluations that take into account the composition, concentration, and internal
17 structure of the particles; temporal variations in atmospheric mixing heights and depths of layers
18 containing the particles; and the abundance of ozone and other absorbers within the planetary
19 boundary layer and the free troposphere (CD, 4-226).
20 At present, models are not available to take such complex factors into account, nor is
21 sufficient data available to characterize input variables that would be necessary for any such
22 modeling. The CD concludes, however, that the outcome of such modeling efforts would likely
23 vary from location to location, even as to the direction of changes in the levels of exposures to
24 UV-B radiation, due to location-specific changes in ambient PM concentrations and/or
25 composition (CD, p. 4-227).' Beyond considering just average levels of exposures to UV-B
26 radiation in general, the CD notes that ambient PM can affect the directional characteristics of
27 UV-B radiation scattering at ground-level, and thus its biological effectiveness. Also, ambient
28 PM can affect not only biologically damaging UV-B radiation, but can also reduce the ground-
29 level ratio of photorepairing UV-A radiation to damaging UV-B radiation. Further, the CD notes
January 2005
6-70
Draft - Do Not Quote or Cite
-------
1 that ambient PM deposition is a major source of PAH in certain water bodies, which can enhance
2 the adverse effects of solar UV-B radiation on aquatic organisms, such that the net effect of
3 ambient PM in some locations may be to increase UV-B radiation-related biological damage to
4 certain aquatic and terrestrial organisms. (CD, p, 4-227).
5 /-
6 6.5.3 Summary and Conclusions
7 A number of assessments of the factors affecting global warming and climate change as
8 well as those affecting the penetration of solar UV-B radiation to the earth's surface clearly
9 recognize ambient PM as playing various roles in these processes. These assessments, however,
10 have focused on global- and regional-scale impacts, allowing for generalized assumptions to take
11 the place of specific, but unavailable, information on local-scale atmospheric parameters and
12 characteristics of the distribution of particles present in the ambient air. As such, the available
13 information provides no basis for estimating how localized changes in the temporal, spatial, and
14 composition patterns of ambient PM, likely to occur as a result of expected future emissions of
15 particles and their precursor gases across the U.S., would affect local, regional, or global changes
16 in climate or UV-B radiation penetration - even the direction of such effects on a local scale
17 remains uncertain. Moreover, similar concentrations of different particle components can
18 produce opposite net effects. It follows, therefore, that there is insufficient information available
19 to project the extent to which, or even whether, such location-specific changes in ambient PM
20 would indirectly affect human health or the environment secondary to potential changes in
21 climate and UV-B radiation.
22 - Based on currently available information, staff concludes that the potential indirect
23 effects of ambient PM on public health and welfare, secondary to potential PM-related changes
24 in climate and UV-B radiation, can play no quantitative role in considering whether any
25 revisions of the primary or secondary PMNAAQS are appropriate at this time. Even
26 qualitatively, the available information is very limited in the extent to which it can help inform
27 an assessment of the overall weight of evidence in an assessment of the net health and
28 environmental effects of PM in the ambient air, considering both its direct effects (e.g.,
January 2005
6-1 \
Draft - Do Not Quote or Cite
-------
1 inhalation-related health effects) and indirect effects mediated by other routes of exposure and
2 environmental factors (e.g., dermal exposure to UV-B radiation).
January 2005 6-72 Draft - Do Not Quote or Cite
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
REFERENCES
Section 6.2 - Visibility Impairment
Abt Associates, Inc. (2001) Assessing Public Opinions on Visibility Impairment Due to Air Pollution: Summary
Report. Prepared for EPA Office of Air Quality Planning and Standards; funded under EPA Contract No.
68-D-98-001. Bethesda, Maryland. January 2001.
Air Resource Specialists, Inc. (2003) WinHaze Air Quality Modeler, version 2.9.0. Available from
http://www.air-resource.com/whatsnew.htm
Arizona Department of Environmental Quality. (2003) Visibility Index Oversight Committee Final Report:
Recommendation for a Phoenix Area Visibility Index. March 5,2003.
htl.t>://www.pfaoenixvis.nel/R)F/vis 031403fi.ual.ndf.
BBC Research & Consulting. (2002) Phoenix Area Visibility Survey. Draft Report. October 4, 2002.
htta://www.bbcresearch,com/lit)rarvAfisibili)v draft report.pdf
California Code of Regulations. Title 17, Section 70200, Table of Standards.
Chestnut, L. G.; Rowe, R. D. (1990) Preservation values for visibility in the national parks. Washington, DC: U.S.
Environmental Protection Agency.
Chestnut, L. G.; Rowe, R. D. (1991) Economic valuation of changes invisibility: A state of the science assessment.
Sector B5 Report 27. In Acidic Depositions: State of Science and Technology Volume IV Control
Technologies, Future Emissions and Effects Valuation. P.M. Irving (ed.). The U.S. National Acid
Precipitation Assessment Program. GPO, Washington, D.C.
Chestnut, L.G.; Dennis, R. L.; Latimer, D. A. (1994) Economic benefits of improvements invisibility: acid rain
. provisions of the 1990 clean air act amendments. Proceedings of Aerosols and Atmospheric Optics:
Radiative Balance and Visual Air Quality. Air & Waste Management Association International Specialty
Conference, pp. 791-802.
Chestnut, L. G.; Dennis, R. L. (1997) Economic benefits of improvements invisibility: acid rain. Provisions of the
1990 clean air act amendments. J. Air Waste Manage. Assoc, 47:395-402.
Cohen, S.; Evans, G.W.; Stokols, D.; Krantz, D.S. (1986) Behavior, Health, and Environmental Stress. Plenum
Press. New York, NY,
Department of Interior. (1998) Air Quality in the National Parks. Natural Resources Report 98-1. National Park
Service, Air Quality Division. Denver, Colorado.
Ely, D.W.; Leary, J.T.; Stewart, T.R.; Ross, D.M. (1991) The Establishment of the Denver Visibility Standard. For
presentation at the 84th Annual Meeting & Exhibition of the Air and Waste Management Association, June
16-21,1991.
Environmental Protection Agency. (1979) Protecting Visibility: An EPA Report to Congress. Research Triangle
Park, NC: Office of Air Quality Planning and Standards. Report no. EPA-45-/5-79-008.
January 2005
6-73
Draft - Do Not Quote or Cite
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
Environmental Protection Agency. (1982) Review of the National Ambient Air Quality Standards for Particulate
Matter, Assessment of Scientific and Technical Information, OAQPS Staff Paper. Research Triangle Park,
N.C.: Office of Air Quality Planning and Standards, Strategies and Air Standards Division. Report no.
EPA-450/5-82-001.
Environmental Protection Agency. (1993) Air Quality Criteria for Oxides of Nitrogen. Research Triangle Park, NC:
Office of Health and Environmental Assessment, Environmental Criteria and Assessment Office. Report
no. EPA-600/8-91/049F.
Environmental Protection Agency. (1996a) Air Quality Criteria for Particulate Matter. Research Triangle Park, NC:
National Center for Environmental Assessment-RTF Office; report no. EPA/600/P- 95/00 laF-cF. 3v.
Environmental Protection Agency. (1 996b) Review of the National Ambient Air Quality Standards for Particulate
Matter: Policy Assessment of Scientific and Technical Information, OAQPS Staff Paper. Research
Triangle Park, NC 27711: Office of Air Quality Planning and Standards; report no. EPA-452XR-96-013.
Environmental Protection Agency. (1999) Regional Haze Regulations. 40 CFR Part 51. 300-309. 64 Federal
3571 3.
Environmental Protection Agency. (2000) Guidelines for Preparing Economic Analyses. Washington, DC: Office of
the Administrator. EPA 240-R-00-003.
Environmental Protection Agency. (2001) National Air Quality and Emissions Trends Report, 1 999. Research
Triangle Park,NC: Office of Air 'Quality Planning and Standards. Report no. EPA/454/R-01-004.' March.
Grand Canyon Visibility Transport Commission (1996) Report of the Grand Canyon Visibility Transport
Commission to the United States Environmental Protection Agency.
Hass, G. E.; Wakefield, T. J. (1 998) National Parks and the American Public: A National Public Opinion Survey of
the National Park System. Colorado State University, Department of Natural Resource Recreation and
Tourism, College of Natural Resources, Fort Collins, CO. Report prepared for the National Parks and
Conservation Association. June 1998.
McNe ill, R. and Roberge, A. (2000) The Impact of Visual Air Quality on Tourism Revenues in Greater Vancouver
and the Lower Fraser Valley. Environment Canada, Georgia Basin Ecosystem Initiative. GBEI report no. .
' EC/GB-00-028.
Middleton, P. (1993) Brown Cloud II: The Denver Air Quality Modeling Study, Final Summary Report. Metro
Denver Brown Cloud Study, Inc. Denver, CO.
Molenar, J.V.; Malm, W.C.; Johnson, C.E. (1994) Visual Air Quality Simulation Techniques. Atmospheric
Environment. Volume 28, Issue 5, 1055-1063.
Molenar, J.V. (2000) Visibility Science and Trends in the Lake Tahoe Basin: 1989-1 998. Report by Air Resource
Specialists, Inc., to Tahoe Regional Planning Agency. February 15, 2000. '
National Acid Precipitation Assessment Program (NAPAP) (1991) Acid Deposition: State of Science and
Technology. Report 24. Visibility: Existing and Historical Conditions - Causes and Effects. Washington,
DC.
National Acid Precipitation Assessment Program (NAPAP). (1998) Biennial Report to' Congress: an
Integrated Assessment.
http://dwb.unl.edu/Teacher/NSF/C14/C14Links/www.nnic.noaa.gov/CENR/NAPAP/NAPAP_96.htm
January 2005
6-74
Draft - Do Not Quote or Cite
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
National Research Council. (1993) Protecting Visibility in National Parks and Wilderness Areas. National
Academy of Sciences Committee on Haze in National Parks and Wilderness Areas. National Academy
Press: Washington, DC.
National Transportation Safety Board (NTSB). (2000) NTSB Report NYC99MAI78, July 6,2000. Report on July
16, 1999 fatal accidental Vineyard Haven, MA.
National Weather Service. (1998) Automated Surface Observing System (ASOS) User's Guide. ASOS Program
Office. Silver Spring, MD.
New Zealand Ministry for the Environment. (2000) Proposals for Revised and New
Ambient Air Quality Guidelines: Discussion Document. Air Quality Report No. 16. December.
New Zealand National Institute of Water & Atmospheric Research (NIWAR). (2000a) Visibility in New Zealand:
Amenity Value, Monitoring, Management and Potential Indicators. Air Quality Technical Report 17.
Prepared for New Zealand Ministry for the Environment. Draft report.
New Zealand National Institute of Water & Atmospheric Research (NIWAR). (2000b) Visibility in New Zealand:
National Risk Assessment. Air Quality Technical Report 18. Prepared for New Zealand Ministry for the
Environment. Draft report.
Peacock, B.; Killingsworth, C.; Simon, B. (1998) State and National Economic Impacts Associated with Travel
Related Expenditures by Recreational Visitors to Lands Managed by the U.S. Department of Interior. U.S.
Department of the Interior. January.
Pryor, S.C. (1996) Assessing Public Perception of Visibility for Standard Setting Exercises. Atmospheric
Environment, vol. 30, no. 15, pp. 2705-2716.
Schichtel, B.A., Husar, R.B., Falke, S.R., and Wilson, W.E. (2001) "Haze Trends over the United States,
1980-1995," Atmospheric Environment, vol. 35, no. 30, pp. 5205-5210.
Schmidt, S.M., Mintz, D., Rao, T., and McCluney, L. (2005) Draft analysis of PM ambient air quality data for the
PM NAAQS review. Memorandum to PM NAAQS review docket OAR-2001 -0017. January 31,2005.
Schulze, W. D.; Brookshire, D. S.; Walther, E. G.; MacFarland, K. K.; Thayer, M. A.; Whitworth, R. L.; Ben-Davis,
S.; Malm, W.; Molenar, Jr. (1983) The Economic Benefits of Preserving Visibility in the National
Parklands of the Southwest. Nat. Resour. J. 23: 149-173.
Sisler, J., Malm, W.; Molenar, J.; Gebhardt, K. (1996) Spatial and Seasonal Patterns and Long Term Variability of
the Chemical Composition of Haze in the U.S.: An Analysis of Data from the IMPROVE Network. Fort
Collins, CO: Cooperative Institute for Research in the Atmosphere, Colorado State University.
State Government of Victoria, Australia. (2000a) Draft Variation to State Environment Protection Policy (Air
Quality Management) and State Environment Protection Policy (Ambient Air Quality) and Draft Policy
Impact Assessment. Environment Protection Authority. Publication 728. Southbank, Victoria.
State Government of Victoria, Australia. (2000b) Year in Review. Environment Protection Authority. Southbank,
Victoria.
January 2005
6-75
'Draft - Do Not Quote or Cite
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
Section 6.3 - Vegetation and Ecosystems
Aber, J. D.; Nadelhoffer, K. J.; Steudler, P.; Melillo.J. M. (1989) Nitrogen saturation in northern forest ecosystems:
excess nitrogen from fossil fuel combustion may stress the biosphere. Bioscience 39: 378-386.
Aber, J. D.; Magill, A; McNulty, S. G.; Boone, R. D.; Nadelhoffer, K. J.; Downs, M.; Hallett, R. (1995) Forest
biogeochemistry and primary production altered by nitrogen saturation. Water Air Soil Pollut. 85:
1665-1670.
Aber, J.; McDowell, W.; Nadelhoffer, K.; Magill, A.; Bemtson, G.; Kamakea, M.; McNulty, S.; Currie, W.;
Rustad, L.; Fernandez, I. (1998) Nitrogen saturation in temperate forest ecosystems. BioScience 48:
921-934.
Allen, E. B.; Padgett, P. E.; Bytnerowicz, A.; Minich, R. (1998) Nitrogen deposition effects on coastal sage
vegetation of southern California. USDA Forest Service Gen. Tech. Rep. PSW-GTR-166, pp. 131-139.
Andersen, C. P.; Rygiewicz, P. T. (1991) Stress interactions and mycorrhizal plant response: understanding carbon
allocation priorities. Environ. Pollut. 73: 217-244.
Antoine, M.E. (2001) Ecophysiology of the cyanolichen.Lo&wia oregana. Master's thesis. Oregon State University,
Corvallis.
Bailey, S.W., Horsley, S.B., Long, R.P., Hallet, R.A. (1999) Influence of geologic and pedologic factors on health of
sugar maple on the Allegheny Plateau, U.S. In Horsley, S.B. and Long, R.P., eds. Sugar Maple Ecology
and Health: Proceedings of an International Symposium. Radnor, PA: U.S. Department of Agriculture,
Forest Service. General Technical Report NE-261. PP. 63-65.
Brooks, ML. (2003) Effects of increased soil nitrogen on the dominance of alien annual plants in the Mojave Desert.
Journal of Applied Ecology. 40:344-353.
Bytnerowicz, A.; Fenn, M. E. (1996) Nitrogen deposition in California forests: a review. Environ. Pollut.
92: 127-146.
Charles, D.F., ed. (1991) Acidic Deposition and Aquatic Ecosystems. Regional Case Studies. New York: Springer-
Verlag.
Craig, B.W. and Friedland, A.J. (1991) Spatial patterns in forest composition and standing dead red spruce in
montane forests of the Adirondacks and northern Appalachians. Environmental Monitoring and
Assessment. 18:129-140.
Cronan, C. S.; Grigal, D. F. (1995) Use of calcium/aluminum ratios as indicators of stress in forest ecosystems.
J. Environ. Qual. 24: 209-226.
D'Antonio, C.M. and Vitousek, P.M. (1992) Biological invasions by exotic grasses: The grass-fire cycle and global
change. Annual Review of Ecology and Systematics. 23: 63-87.
DeHayes, D. H.; Schaberg, P. G.; Hawley, G. J.; Strimbeck, G. R. (1999) Acid rain impacts on calcium nutrition and
forest health. Bioscience 49: 789-800.
Driscoll, C. T.; Wyskowski, B. J.; DeStaffan, P.; Newton, R. M. (1989) Chemistry and transfer of aluminum in a
forested watershed in the Adirondack region of New York, USA. In: Lewis, T. E., ed. Environmental /
chemistry and toxicology of aluminum. Chelsea, MI: Lewis Publishers, Inc.; pp.'83-105.
January 2005
6-76
Draft - Do Not Quote or Cite
-------
1 Driscoll, C.T., Likens, G.E., Church, MR. (1998) Recovery of surface waters in the northeastern U.S. from
2 decreases in atmospheric deposition of sulfur. Water, Airand Soil Pollution. 105:319-329.
3
4 Driscoll, C. T.; Lawrence, G. B.; Bulger, A. J.; Butler, T. J.; Cronan, C. S.; Eagar, C.; Lambert, K. F.; Likens, G. E.,
5 Stoddard, J. L.; Weathers, K. C. (2001) Acidic deposition in the northeastern United States: sources and
. 6 inputs, ecosystem effects, and management strategies. BioScience 51:180-198.
7
8 Edgerton-Warburton, L. M.; Allen, E. B. (2000) Shifts in arbuscular mycorrhizal communities along an
9 anthropogenic gradient nitrogen deposition gradient. Ecol. Appl. 10: 484-496.
10
11 Environmental Protection Agency. (1982) Air quality criteria for particulate matter and sulfur oxides. Research
12 Triangle Park, NC: Office of Health and Environmental Assessment, Environmental Criteria and
13 Assessment Office; EPA report no. EPA-600/8-82-029aF-cF. 3v. Available from: NTIS, Springfield, VA;
14 PB84-156777.
15
16 Environmental Protection Agency. (1992) Framework for Ecological Risk Assessment Washington, D.C.: Risk
17 Assessment Forum, U.S. Environmental Protection Agency. EPA/630/R-92/001.
18
19 Environmental Protection Agency. (1993) Air quality criteria for oxides of nitrogen. Research Triangle Park, NC:
20 Office of Health and Environmental Assessment, Environmental Criteria and Assessment Office; report
21 nos. EPA/600/8-91/049aF-cF. 3v. Available from: NTIS, Springfield, VA; PB95-124533, PB95-124525,
22 and PB95-124517.
23
24 Environmental Protection Agency. (1996) Air quality criteria for particulate matter. Research Triangle Park, NC:
25 National Center for Environmental Assessment-RTP Office; report nos. EPA/600/P-95/001 aF-cF. 3v.
26
27 Environmental Protection Agency. (1997) Nitrogen oxides: impacts on public health and the environment.
28 Washington, DC: Office of Air and Radiation; August. Available:
29 www.epa.gov/ttncaaal/tl/reports/noxrept.pdf [1999, November 24].
30
31 Environmental Protection Agency. (1998) Guidelines for Ecological Risk Assessment Washington, D.C: Risk
32 Assessment Forum, U.S. Environmental Protection Agency. EPA/630/R-95/002F.
33
34 Environmental Protection Agency. (2000) Deposition of air pollutants to the great waters. Third report to Congress.
35 [Executive Summary]. Research Triangle Park, NC: U.S. Environmental Protection Agency, Office of Air
36 Quality Planning and Standards; report no. EPA-453/R-00-005.
37
38 Environmental Protection Agency. (2001) Air Quality Criteria for Particulate Matter. Research Triangle Park, NC:
39 Office of Research and Development; report no. EPA/600/P-99/002.'March.
40
41 Environmental Protection Agency. (2002) A Framework for Assessing and Reporting on Ecological Condition: An
42 SAB Report. Washington, D.C.: Ecological Processes and Effects Committee, Science Advisory Board,
43 U.S. Environmental Protection Agency. EPA-SAB-EPEC-02-009.
44
45 Environmental Protection Agency. (2003) Response Of Surface Water Chemistry to the Clean Air Act Amendments
46 of 1990. National Health and Environmental Effects Research Laboratory, Office of Research and
47 Development, U.S. Environmental Protection Agency. Research Triangle Park, NC. EPA 620/R-03/001.
48
49 Fenn, M. E.; Poth, M. A.; Aber, J. D.; Baron, J. S.; Bormann, B. T.; Johnson, D. W.; Lemly, A. D.; McNulty, S. G.;
50 Ryan, D. F.; Stottlemyer, R. (1998) Nitrogen excess in North American ecosystems: predisposing factors,
51 ecosystem responses, and management strategies. Ecol. Appl. 8: 706-733.
52 •)
January 2005 6-77 Draft - Do Not Quote or Cite
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
Fenn, M. E.; Baron, J. S.; Allen, E. B.; Rueth, H. M.; Nydick, K. R.; Geiser, L.; Bowman, W. D.; Sickman, J. O.;
Meixner, T.; Johnson, D. W.; Neitlich, P. (2003) Ecological effects of nitrogen deposition in the western
United States. BioScience 53: 404-420.
Galloway, J.N, Norton, S.N., Church, MR. (1983) Freshwater acidification from atmospheric deposition of sulfuric
acid: A conceptual model. Environmental Science and Technology. 17:541 A-545A
Galloway, J. N. (1998) The global nitrogen cycle: changes and consequences. Environ. Pollut 102(suppl. 1): 15-24.
Galloway, J. N.; Cowling, E. B. (2002) Reactive nitrogen and the world: 200 years of change. Ambio 31: 64-71.
Galloway, J. R; Aber, J. D.; Erisman, J. W.; Seitzinger, S. P.; Howarth, R. W.; Cowling, E. B.; Cosby, B. J. (2003)
The nitrogen cascade. BioScience 53: 341-356.
Gamer, J. H. B. (1994) Nitrogen oxides, plant metabolism, and forest ecosystem response. In: Alscher, R. G.;
Wellburn, A. R., eds. Plant responses to the gaseous environment: molecular, metabolic and physiological
aspects, [3rd international symposium on air pollutants and plant metabolism]; June 1992; Blacksburg, VA.
London, United Kingdom: Chapman & Hall; pp. 301-314.
Gunn, J.M. and Mills, K.H. (1998) The potential for restoration of acid-damaged lake trout lakes. Restoration
Ecology. 6:390-397.
Gundersen, P.; Callesen, I.; De Vries, W. (1998) Nitrate leaching in forest ecosystems is related to forest floor C/N
ratios. Environ. Pollut 102(suppl. 1): 403-407.
Hedin, L. O.; Granat, L.; Likens, G. E.; Buishand, T. A.; Galloway, J. N.; Butler, T. J.; Rodhe, H. (1994)
Steep declines in atmospheric base cations in regions of Europe and North America. Nature (London)
367:351-354.
Hornung, M; Langan, S. J. (1999) Nitrogen deposition, sources, impacts and responses in natural and semi-natural
ecosystems. In: Bangan, S. J., ed. Impact of nitrogen deposition on natural ecosystems and semi-natural
ecosystems. Dordrect, Netherlands: Kluwer Academic Publishers; pp. 1-14. [Environmental Pollution,
no.3].
Horsley, S.B., Long, R.P., Bailey, S.W., Hallet, R.A., Hall, T.J. (1999) Factors contributing to sugar maple decline
along topographic gradients on the glaciated and unglaciated Allegheny Plateau. In Horsley, S.B. and
fLong, R.P., eds. Sugar Maple Ecology and Health: Proceedings of an International Symposium. Radnor,
'PA: U.S. Department of Agriculture, Forest Service. General Technical Report NE-261. PP. 60-62.
Howarth, R. W.; Boyer, E. W.; Pabich, W. J.; Galloway, J. N. (2002) Nitrogen use in the United States from
1961-2000 and potential future trends. Ambio 31: 88-96.
Jaworski, N. A.; Howarth, R. W.; Hetling, L. J. (1997) Atmospheric deposition of nitrogen oxides onto the landscape
contributes to coastal eutrophication in the northeast United States. Environ. Sci. Technol. 31:1995-2004.
Johnson, A.H., Friedland, A.J., Dushoff, J.G. (1984) Recent and historic red spruce mortality: Evidence of climatic
influence. Water, Air and Soil Pollutioa 30:319-330.
Johnson, D. W.; Van Miegroet, H.; Lindberg, S. E.; Todd, D. E.; Harrison, R. B. (1991) Nutrient cycling in red
spruce forests of the Great Smoky Mountains. Can. J. For. Res, 21:769-787.
January 2005 '
6-78
Draft - Do Not Quote or Cite
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
'21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
Johnson, D. W.; Lindberg, S. E., eds. (1992) Atmospheric deposition and forest nutrient cycling: a synthesis of the
integrated forest study. New York, NY: Springer-Verlag, Inc. (Billings, W. D.; Golley, F.; Lange, O. L.;
Olson, J. S.; Remmert, H., eds. Ecological studies: analysis and synthesis: v. 91).
Johnson, D. W.; Swank, W. T.; Vose, J. M. (1993) Simulated effects of atmospheric sulfur deposition on nutrient
cycling in a mixed deciduous forest. Biogeochemistry 23: 169-196.
Johnson, D.W.; Susfalk, R.B.; Brewer, P.,R; Swank, W.T. (1999) Simulated effects of reduced sulfur, nitrogen, and
base cation deposition on soils and solutions in southern Appalachian forests. J. Environ. Qual 28:1336-1346.
-Keller, W. and Gunn, J.M. (1995) Lake water quality improvements and recovering aquatic communities. In Gunn,
J.M. ed. Restoration and Recovery of an Industrial Region: Progress in Restoring the Smelter-damaged
Landscape Near Sudbury, Canada. New York: Springer-Verlag, PP. 67-80.
Kretser, W., Gallagher, J., Nicolette, J. (1989) Adirondack Lake Study. 1984-1987. An Evaluation of Fish
Communities and Water Chemistry. Ray Brook, New York: Adirondacks Lakes Survey Corporation.
Likens, G.E. and Bormann, F.H. (1995) Biogeochemistry of a Forested Ecosystem. 2nd ed., New York: Springer-
Verlag.
L0kke, H.; Bak, J.; Falkengren-Grerup, U.; Finlay, R. D.; Ilvesniemi, H.; Nygaard, P. H.; Starr, M. (1996)
Critical loads of acidic deposition for forest soils: is the current approach adequate. Ambio 25: 510-516.
(
Lovett, G. M; Traynor, M. M; Pouyat, R. V.; Carreiro, M. M.; Zhu, W.'-X.; Baxter, J. W. (2000) Atmospheric
deposition to oak forest along an urban-rural gradient. Environ. Sci. Technol. 34: 4294-4300.
Magill, A. H.; Aber, J. D.; Bemtson, G. M.; McDowell, W. H.; Nadelhoffer, K. J.; Melillo, J. M.'; Steudler, P. (2000)
Long-term nitrogen additions and nitrogen saturation in two temperate forests. Ecosystems 3: 238-253.
McCune, B. and Geiser, L. (1997) Macrolichens of the Pacific Northwest. Corvallis: Oregon State University Press.
McDonnell, M. J.; Pickett, S. T. A.; Groffman, P.; Bohlen, P.; Pouyat, R. V.; Zipperer, W. C.; Parmelee, R. W.;
Carreiro, M. M.; Medley, K. (1997) Ecosystem processes along an urban-to-rural gradient. Urban Ecosyst.
1:21-36.
Minnich, R. A. and Dezzani, R.J. (1998) Historical decline of coastal sage scrub in the Riverside-Ferris Plain,
California. Western Birds. 29:366-391.
*
Naeem, S.; Thompson, L. J.; Lawler, S. P.; Lawton, J. H.; Woodfin, R. M. (1994) Declining biodiversity can alter
the performance of ecosystems. Nature 368: 734-737.
i
Nagy, K.A., Henen, B.T., Vyas, D.B. (1998) Nutritional quality of native and introduced food plants of wild desert
tortoises. Journal of Herpetology 32: 260-267.
Nash, T.H. and Sigal, L.L. (1999) Ephiphytic lichens in the San Bernardino mountains in relation to oxidant
gradients. In Miller, P.R. McBride, J.R., eds. Oxidant Air Pollution Impacts on the Montane Forests of
Southern California: A Case Study of the San Bernardino Mountains. Ecological Studies 134. New York:
Springer-Verlag. PP. 223-234.
National Science and Technology Council. (1998) National acid precipitation assessment program biennial report to
Congress: an integrated assessment; executive summary. Silver Spring, MD: U.S. Department of
Commerce, National Oceanic and Atmospheric Administration. Available:
www.nnic.noaa.gov/CENR/NAPAP/NAPAP_96.htm [24 November 1999].
January 2005
6-79
Draft - Do Not Quote or Cite
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
Padgett, P. E.; Allen, E. B.; Bytnerowicz, A.; Minich, R. A. (1999) Changes in soil inorganic nitrogen as related to
atmospheric nitrogenous pollutants in southern California. Atmos. Environ. 33:769-781.
Paerl, H. W.; Bales, J. D.; Ausley, L. W.; Buzzelli, C. P.; Crowder, L. B.; Eby, L. A.; Go, M.; Peierls, B. L.;
Richardson, T. L.; Ramus, J. S. (2001) Ecosystem impacts of three sequential hurricanes (Dennis, Floyd,
and Irene) on the United States' largest lagoonal estuary, Pamlico Sound, NC. Proc. Nat. Acad. Sci. U. S. A.
98:5655-5611.
Pike, L.H. (1978) The importance of epiphytic lichens in mineral cycling. Bryologist 81: 247-257.
Pouyat, R. V.; McDonnell, M. J. (1991) Heavy metal accumulations in forest soils along an urban-rural gradient in
southeastern New York, USA. Water Air Soil Pollut. 57/58:797-808.
Rabalais,N.N. (2002) Nitrogen in aquatic ecosystems. Ambio 31:102-112.
Rago, PJ. and Wiener, J.G. (1986) Does pH affect fish species richness when lake area is considered? Transactions
of the American Fisheries Society. llb:438-447.
Riebsame, W.E., Robb, J.J., Gosnell, H., Theobald, D., Breding, P., Hanson, C., Rokoske, K. (1997) Atlas of the
New West: Portrait of a Changing Region. New York: W. W. Norton.
Rovira, A. D.; Davey, C. B. (1974) Biology of the rhizosphere. In: Carson, E. W., ed. The plant root and its
environment: proceedings of an institute; My, 1971; Blacksburg, VA. Charlottesville, VA: University
Press of Virginia; pp. 153-204.
Schindler, D.W., Mills, K.H., Malley, D.F., Findlay, S., Schearer, J.A., Davies, I.J., Turner, M.A., Lindsey, G.A.,
Cruikshank, D.R. (1985) Long-term ecosystem stress: Effects of years of experimental acidification.
Canadian Journal of Fisheries and Aqautic Science. 37:342-354. '
Science Advisory Board. (2002) A framework for assessing and reporting on ecological condition: an SAB report.
Young, T. F.; Sanzone, S., eds. Washington, DC: U.S. Environmental Protection Agency, Ecological
Processes and Effects Committee; report no. EPA-SAB-EPEC-02-009. Available at:
http://www.epa.gov/sciencel/rndex.html (27 January 2003).
Shortle, W. C.; Smith, K. T.; Minocha, R; Lawrence, G. B.; David, M B. (1997) Acidic deposition, cation
mobilization, and biochemical indicators of stress in healthy red spruce. J. Environ. Qual. 26: 871-876.
Skinner, M.W.; Pavlik, B.M. (1994) CNPS Inventory of Rare and Endangered Vascular Plants of California.
Sacramento: California Native Plant Society.
Smith, W. H. (1990) Forests as sinks for air contaminants: soil compartment. In: Air pollution and forests:
interactions between air contaminants and forest ecosystems. 2nd ed.' New York, -NY: Springer-Verlag;
pp. 113-146. (Springer series on environmental management).
Sollins, P., Grier, C.C., McCorison, F.M., Cromack, K., Jr.Fogel, R., Fredriksen, R.L. (1980) The internal element
cycles of an old-growth Douglas-fir ecosystem in western Oregon. Ecological Monographs 50:261-285.
Stanhill, G.; Cohen, S. (2001) Global dimming: a review of the evidence for a widespread and significant reduction
in global radiation with discussion of its probable causes and possible agricultural consequences. Agric.
For. Meteorol. 107: 255-278.
January 2005
6-80
Draft - Do Not Quote or Cite
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
Strickland, T. C.; Holdren, G. R., Jr.; Ringold, P. L.; Bernard, D.; Smythe, K.; Fallen, W. (1993) A national critical
loads framework for atmospheric deposition effects assessment: I. method summary..Environ. Manage.
17: 329-334.
Van Egmond, K.; Bresser, T.; Bouwman, L. (2002) The European nitrogen case. Ambio 31: 72-78.
van Herk, C.M. (1999) Mapping of ammonia pollution with epiphytic lichens in the Netherlands. Lichenologist
31:9-20.
Van Sickle, J., Baker, J.P., Simonin, H.A., Baldigo, B.P., Kretser, W.A., Sharpe, W.F. (1996) Episodic acidification
of small streamsin the northeastern U.S.: Fish mortality in field bioassays. Ecological Applications. 6:408-
421.
Vitousek, P. M.; Mooney, H. A.; Lubchenco, J.; Melillo, J. M. (1997) Human domination of Earth's ecosystems.
Science (Washington, DC) 277:494-499.
Wall, D. H.; Moore, J. C. (1999) Interactions underground: soil biodiversity, mutualism, and ecosystem processes.
Bioscience 49: 109-117.
Wedin, D. A.; Tilman, D. (1996) Influence of nitrogen loading and species composition on the carbon balance of
grasslands. Science 274: 1720-1723.
Wesely, M. L.; Hicks, B. B. (2000) A review of the current status of knowledge on dry deposition. Atmos. Environ.
34: 2261-2282.
Whitby, K. T. (1978) The physical characteristics of sulfur aerosols. Atmos. Environ. 12:135-159.
World Health Organization. (1997) Nitrogen oxides. 2nd ed. Geneva, Switzerland: World Health Organization.
(Environmental health criteria 188).
Section 6.4 - Man-made Materials
Creighton, P. J.; Lioy, P. J.; Haynie, F. H.; Lemmons, T. J.; Miller, J. L.; Gerhart, J. (1990) Soiling by atmospheric
aerosols in an urban industrial area, J. Air Waste Manage. Assoc. 40:1285-1289.
Davidson, C. I.; Tang, W.; Finger, S.; Etyemezian, V.; Striegel, M. F.; Sherwood, S. I. (2000) Soiling patterns on a
tall limestone building: changes over 60 years. Environ. Sci. Technol. 34: 560-565.
Environmental Protection Agency. (1996b) Review of the National Ambient Air Quality Standards for Particulate
Matter: Policy Assessment of Scientific and Technical Information, OAQPS Staff Paper. Research
Triangle Park, NC 27711: Office of Air Quality Planning and Standards; report no. EPA-452\R-96-013.
Haynie, F.H.; Lemmons, T. J. (1990) Particulate matter soiling of exterior paints at a rural site. Aerosol Sci. Technol.
13:356-367.
Ligocki, M. P.: Salmon, L. G.; Fall, T.; Jones, M. C.; Nazaroff, W. W.; Cass, G. R. (1993) Characteristics of
airborne particles inside southern California museums. Atmos. Environ. Part A 27: 697-711.
Lorusso, S.; Marabelli, M.; Troili, M. (1997) Air pollution and the deterioration of historic monuments. J. Environ.
Pathol. Toxicol. Oncol. 16:171-173.
January 2005
6-81
Draft - Do Not Quote or Cite
-------
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
Miller, W. C,; Fomes, R. E.; Gilbert, R. D.; Speer, A.; Spence, J. (1992) Removal of CaCOj extender in residential
coatings by atmospheric acidic deposition. In: Measurement of toxic and related air pollutants: proceedings
of the 1992 U. S. EPA/A&WMA international symposium. Pittsburgh, PA: Air & Waste Management
Association; pp. 129-134. (A&WMA publication VIP-25)
Wolff, G. T.; Collins, D. C.; Rodgers, W. R.; Verma, M. H.; Wong, C. A. (1990) Spotting of automotive finishes
from the interactions between dry deposition of crustal material and wet deposition of sulfate. J. Air Waste
Manage. Assoc. 40:1638-1648.
1
2
3
4
5
6
7
8
9
10
11 Section 6.5 - Climate Change and Solar Radiation
Intergovernmental Panel on Climate Change (IPCC). (1998) The regional impacts of climate change: an assessment
of vulnerability. Cambridge, United Kingdom: Cambridge University Press.
Intergovernmental Panel on Climate Change (IPCC). (2001 a) Climate change 2001: the scientific basis.
Contribution of working group I to the third assessment report of the Intergovernmental Panel on Climate
Change. Cambridge, United Kingdom: Cambridge University Press.
Intergovernmental Panel on Climate Change (IPCC). (2001 b) Climate change 2001: impacts, adaptation, and
vulnerability. Contribution of working group II to the third assessment report of the Intergovernmental
Panel on Climate Change. Cambridge, United Kingdom: Cambridge University Press.
r
National Academy of Sciences (NAS). (2001) Committee on the Science of Climate Change, National Research
Council. Climate Change Science: An Analysis of Some Key Questions, National Academy .Press,
Washington, DC.
United Nations Environment Programme (UNEP). (1998) Environmental effects of ozone depletion: 1998
assessment. J. Photochem. Photobiol. B 46:1-4.
United Nations Environment Programme (UNEP). (2000) Environmental effects of ozone depletion: interim
summary. Available at: bttp://www.gcrio.or^oyoiie/unep20(X)5ummanLhtml (9 April 2002).
World Meteorological Organization. (1999) Scientific assessment of ozone depletion: 1998. Geneva, Switzerland:
World Meteorological Organization, Global Ozone and Monitoring Project; report no. 44.
January 2005
6-82
Draft - Do Not Quote or Cite
-------
1 7. STAFF CONCLUSIONS AND RECOMMENDATIONS ON
2 SECONDARY PMNAAQS
3 7.1 INTRODUCTION
4 This chapter presents staff conclusions and recommendations for the Administrator to
5 consider in deciding whether the existing secondary PM standards should be revised and, if so,
6 what revised standards are appropriate.1 The existing suite of secondary PM standards, which is
7 identical to the suite of primary PM standards, includes annual and 24-hour PM25 standards and
8 annual and 24-hour PM,0 standards to address visibility impairment associated with fine particles
9 and materials damage and soiling related to both fine and coarse particles. Each of these
10 standards is defined in terms of four basic elements: indicator, averaging time, level and form.
11 Staff conclusions and recommendations on these standards are based on the assessment and
12 integrative synthesis of information related to welfare effects presented in the CD and on staff
13 analyses and evaluations presented in Chapters 2 and 6 herein.
14 In recommending a range of secondary standard options for the Administrator to
15 consider, staff notes that the final decision is largely apublic policy judgment. A final decision
16 must draw upon scientific evidence and analyses about effects on public welfare, as well as'
17 judgments about how to deal with the range of uncertainties that are inherent in the relevant
18 information. The NAAQS provisions of the Act require the Administrator to establish secondary
19 standards that are requisite to protect public welfare2 from any known or anticipated adverse
20 effects associated with the presence of the pollutant in the ambient air. In so doing, the
21 Administrator seeks to establish standards that are neither more nor less stringent than necessary
22 . for this purpose. The provisions do not require that secondary standards be set to eliminate all
As noted in Chapter 1, staff conclusions and recommendations presented herein are provisional; final staff
conclusions and recommendations, to be included in the final version of this document, will be informed by
comments received from CASAC and the public in their reviews of this draft document.
2 As noted in Chapter 1, welfare effects as defined in section 302(h) [42 U.S.C. 7602(h)] include, but are
not limited to, "effects on soils, water, crops, vegetation, man-made materials, animals, wildlife, weather, visibility
and climate, damage to and deterioration of property, and hazards to transportation, as well as effects on economic
values and on personal comfort and well-being."
January 2005
7-1
Draft - Do Not Quote or Cite
-------
1 welfare effects, but rather at a level requisite to protect public welfare from those effects that are
2 judged to be adverse.
3 7.2 APPROACH
4 Similar to the approach discussed in Chapter 5, section 5.2, for the review of the primary
5 NAAQS, staffs approach here can be framed by a series of questions that may be applicable for
6 , each category of PM-related welfare effects identified in the CD as being associated with the
7 presence of the pollutant in the ambient air. Staffs review of the adequacy of the current PM
8 standards for each effects category involves addressing questions such as:
9 • To what extent does the available information demonstrate or suggest that PM-related
10 effects are occurring at current ambient conditions or at levels that would meet the
11 . current standards?
12 • To what extent does the available information inform j udgments as to whether any
13 observed or anticipated effects are adverse to public welfare?
w
14 • To what extent are the current secondary standards likely to be effective in achieving
15 protection against any identified adverse effects?
16 To the extent that the available information suggests that revision of the current secondary
17 standards would be appropriate for an effects category, staff then identifies ranges of standards
18 (in terms of indicators, averaging times, levels, and forms) that would reflect a range of
19 alternative policy judgments as to the degree of protection that is requisite to protect public
20 welfare from known or anticipated adverse effects. In so doing, staff addresses questions such
21 as:
22
23
24
25
26
27
28
Does the available information provide support for considering different PM indicators?
Does the available information provide support for considering different.averaging
times?
What range of levels and forms of alternative standards is supported by the information,
and what are the uncertainties and limitations in that information?
To what extent would specific levels and forms of alternative standards reduce adverse
impacts attributable to PM, and what are the uncertainties in the estimated reductions?
January 2005
7-2
Draft - Do Not Quote or Cite
-------
1 Based on the available information, estimated reductions in adverse impacts, and related
2 uncertainties, staff makes recommendations as to ranges of alternative standards for the
3 Administrator's consideration in reaching decisions as to whether to retain or revise the
4 secondary PM NAAQS.
5 In presenting this approach, staff well recognizes that for some welfare effects the
6 currently available information falls short of what is considered sufficient to serve as a basis for
7 a distinct standard defined specifically in terms of the relationship between ambient PM and that
8 effect In the case of visibility impairment, however, the available information may well provide
9 a basis for a distinctly defined standard. In either case, staff believes it is appropriate to consider
10 the extent to which the current or recommended primary standards may afford protection against
11 the identified welfare effects.
12 Staff first considers information related to the effects of ambient PM, especially fine
13 particles, on visibility impairment in section 7.3, and makes recommendations that consideration
14 be given to a revised PM25 standard. Other PM-related welfare effects, including effects on
15 vegetation and ecosystems, materials, and global climate change processes, are addressed in
16 section 7.4. This chapter concludes with a summary of key uncertainties associated with
17 establishing secondary PM standards and related staff research recommendations in section 7.5.
18 7.3 STANDARDS TO ADDRESS VISIBILITY IMPAIRMENT
19 In 1997, EPA decided to address the effects of PM on visibility by setting secondary
20 standards identical to the suite of PM2 5 primary standards, in conjunction with the future
21 establishment of a regional haze program under sections 169A and 169B of the Act (62 FR at
22 38,679-83). In reaching this decision, EPA first concluded that PM, especially fine particles,
23 produces adverse effects on visibility in various locations across the country, including multi-
24 state regions, urban areas, and remote Class I Federal areas (e.g., national parks and wilderness
25 areas). EPA also concluded that addressing visibility impairment solely through setting more
26 stringent national secondary standards would not be an appropriate means to protect the public
27 welfare from adverse impacts of PM on visibility in all parts of the country. As a consequence,
January 2005
7-3
Draft - Do Not Quote or Cite
-------
1 EPA determined mat an approach that combined national secondary standards with a regional
2 haze program was the most appropriate and effective way to address visibility impairment.
3. In reaching these conclusions in 1997, EPA recognized, based on observations from
4 available monitoring data, primarily from rural sites in the IMPROVE monitoring network, that
5 the selection of an appropriate level for a national secondary standard to address visibility
6 protection was complicated by regional differences in visibility impairment These differences
7 were due to several factors, including background and current levels of PM, the composition of
8 PM, and average relative humidity. As a result of these regional differences, EPA noted that a
9 national standard intended to maintain or improve visibility conditions in many parts of the West
10 would have to be set at or below natural background levels in the East; conversely, a national
11 standard that would improve visibility in the East would permit further degradation in the West
12 Beyond such problems associated with regional variability, EPA also determined that there was
13 not sufficient information available to establish a standard level to protect against visibility
14 .conditions generally considered to be adverse in1 all areas.
15 These considerations led EPA to assess whether the protection afforded by the
16 combination of the selected primary PM2 5 standards and a regional haze program would provide
17 appropriate protection against the effects of PM on visibility. Based on such an assessment,
18 EPA determined that attainment of the primary PM2 5 standards through the implementation of
19 regional control strategies would be expected to result in visibility improvements in the East at
20 both urban and regional scales, but little or no change in the West, except in and near certain
21 urban areas. Further, EPA determined that a regional haze program that would make significant
22 progress toward the national visibility goal in Class I areas would also be expected to improve
23 visibility in many urban and non-Class I areas as well. EPA also noted, however, that the
24 combined effect of the PM NAAQS and regional haze programs may not address all situations in
25 which people living in certain urban areas may place a particularly high value on unique scenic
26 resources in or near these areas. EPA concluded that such situations were more appropriately
27 and effectively addressed by local visibility standards, such as those established by the city of
28 Denver, than by national standards and control programs.
29 As anticipated in the last review, EPA promulgated a regional haze program in 1999.
30 That program requires States to establish goals for improving visibility in Class I areas and to
January 2005 7-4 Draft - Do Not Quote or Cite
-------
1 adopt control strategies to achieve these goals. More specifically, States are required to establish
2 goals for improving visibility on the 20% most impaired days in each Class I area, and for
3 allowing no degradation on the 20% least impaired days. Since strategies to meet these goals are
4 to reflect a coordinated approach among States, multistate regional planning organizations have
5 been formed and are now developing strategies, to be adopted over the next few years, that will
6 make reasonable progress in meeting these goals.
7 7.3.1 Adequacy of Current PMZ 5 Standards
8 In considering the information now available in this review, as discussed in Chapters 2
9 and 6 (section 6.2), staff notes that, while new research has led to improved understanding of the
10 optical properties of particles and the effects of relative humidity on those properties, it has not
11 changed the fundamental characterization of the role of PM, especially fine particles, in visibility
12 impairment from the last review. However, extensive new information now available from
i
13 visibility and fine particle monitoring networks has allowed for updated characterizations of
14 visibility trends and current levels in urban areas, as well as Class I areas. These new data are a
15 critical component of the analysis presented in section 6.2.3 that better characterizes visibility
16 impairment in urban areas.
17 Based on this information, staff has first considered the extent to which available
18 information shows PM-related impairment of visibility at current ambient conditions in areas
19 across the U. S. Taking into account the most recent monitoring information and analyses, staff
20 makes the following observations:
21 • In Class I areas, visibility levels on the 20% haziest days in the West are about equal to
22 levels on the 20% best days in the East. Despite improvement through the 1990's,
23 visibility in the rural East remains significantly impaired, with an average visual range of
24 approximately 20 km on the 20% haziest days (compared to the naturally occurring
25 visual range of about 150 + 45 km). In the rural West, the average visual range showed
26 little change over this period, with an average visual range of approximately 100 km on
27 the 20% haziest days (compared to the naturally occurring visual range of about 230 ± 40
28 km).
29 • In urban areas, visibility levels show far less difference between eastern and western
30 regions. For example, based on reconstructed light extinction values calculated from 24-
31 hour average PM2 5 concentrations, the average visual ranges on the 20% haziest days in
32 eastern and western urban areas are approximately 21 km and 28 km, respectively. Even
January 2005
7-5
Draft - Do Not Quote or Cite
Q
-------
1 more similarity is seen in considering 4-hour (12:00 to 4:00 pm) average PM25
2 concentrations, for which the average visual ranges on the 20% haziest days in eastern
3 and western urban areas are approximately 26 km and 3 Okm, respectively. (Schmidt et
4 al.,2005)
5 Based on this information, and on the recognition that efforts are now underway to
6 address all human-caused visibility impairment in Class I areas through the regional haze
7 program implemented under sections 169A and 169B of the Act, as discussed above, staff has
8 focused in this review on visibility impairment primarily in urban areas, hi so doing, staff has
9 considered whether information now available can inform judgments as to the extent to which
10 existing levels of visibility impairment in urban areas can be considered adverse to public
11 welfare. In so doing, staff has looked at studies in the U.S. and abroad that have provided the
12 basis for the establishment of standards and programs to address specific visibility concerns in
13 v local areas, as discussed in section 6.2.5. These studies have produced new methods and tools to
14 communicate and evaluate public perceptions about varying visual effects associated with
15 alternative levels of visibility impairment relative to varying particle pollution levels and
16 environmental conditions. As discussed in section 6.2.6, methods involving the use of surveys to
17 elicit citizen judgments about the acceptability of varying levels of visual air quality in an urban
18 area have been developed by the State of Colorado, and used to develop a visibility standard for
19 Denver. These methods have now been adapted and applied in other areas, including Phoenix,
20 AZ, and the province of British Columbia, Canada, producing reasonably consistent results in
21 terms of the visual ranges found to be generally acceptable by the participants in the various
22 studies, which ranged from approximately 40 to 60 km in visual range.
23 Beyond the information available from such programs, staff believes it is appropriate to
24 make use directly of photographic representations of visibility impairment to help inform
25 judgments about the acceptability of varying levels of visual air quality in urban areas. As
26 discussed in section 6.2.6, photographic representations of varying levels of visual air quality
27 have been developed for several urban areas and are available on EPA's website
28 Outp://vvvsw.epa.gov/ttnJ'fnaaqs/statidards/pm/sjjm_crjsp.htinl') as an attachment to this •
29 document. In considering these images for Washington, D.C., Chicago, and Phoenix (for which
30 PMZ5 concentrations are reported), staff makes the following observations:
January 2005 7-6 Draft - Do Not Quote or Cite
-------
1 • At concentrations at or near the level of the current 24-hour PM25 standard, scenic views
2 (e-g-> mountains, historic monuments), as depicted in these images around and within the
3 urban areas, are significantly obscured from view.
4 • Appreciable improvement in the visual clarity of the scenic views depicted in these
5 images .occurs at PM2.5 concentrations below 35 to 40 ug/m3, or at visual ranges generally
6 above 20 km for the urban areas considered.
7 While being mindful of the limitations in using visual representations from a small
8 number of areas as a basis for considering national visibility-based secondary standards, staff
9 nonetheless concludes that the observations discussed above support consideration of revising
10 the current PM2 5 secondary standards to enhance visual air quality, particularly with a focus on
11 urban areas. Thus, in the sections that follow, staff evaluates information related to indicator,
12 averaging time, level and form to identify a range of alternative PM standards for consideration
13 that would protect visual air quality, primarily in urban areas.
14 7.3.2 Indicators
15 . As discussed in Chapter 2, section 2.8, fine particles contribute to visibility impairment
16 directly in proportion to their concentration in the ambient air. Hygroscopic components of fine
17 particles, in particular sulfates and nitrates, contribute disproportionately to visibility impairment
18 under high humidity conditions, when such components can reach particle diameters up to and
19 even above 2.5 um. Particles in the coarse mode generally contribute only marginally to
20 visibility impairment in urban areas. Thus, fine particles, as indexed by PM2i, are an appropriate
21 indicator of PM pollution to consider for the purpose of standards intended to address visibility
22 impairment.
23 In analyzing how well PM2 s concentrations correlate with visibility in urban locations
24 across the U.S., as discussed above in section 6.2.3 and in more detail in Schmidt et al. (2005),
25 staff concludes that the observed correlations are strong enough to support the use of PM2 5 as the
26 indicator for such standards. More specifically, clear correlations exist between 24-hour average
27 PM2 j concentrations and reconstructed light extinction (RE), which is directly related to visual
28 range, and these correlations are similar in eastern and western regions. These correlations are
29 less influenced by relative humidity and more consistent across regions when PM2 5
January 2005
7-7
Draft - Do Not Quote or Cite
-------
1 concentrations are averaged over shorter, daylight time periods (e.g., 4 to 8 hours). Thus, staff
2 concludes that it is appropriate to use PM2 5 as an indicator for standards to address visibility
3 impairment in urban areas, especially when the indicator is defined for a relatively short period
4 of daylight hours.
5 7.3.3 Averaging Times
6 In considering appropriate averaging times for a standard to address visibility
7 impairment, staff has considered averaging times that range from 24 to 4 hours, as discussed in
8 section 6.2.3. Within this range, as noted above, correlations between PM25 concentrations and
9 RE are generally less influenced by relative humidity and more consistent across regions as the
10 averaging time gets shorter. Based on the regional and national average statistics considered in
11 this analysis, staff observes that in the 4-hour time period between 12:00 and 4:00 p.m., the slope
12 of the correlation between PM25 concentrations and hourly RE is lowest and most consistent
13 across regions than for any other 4-hour or longer time period within a day (Chapter 6, Figure .
14 6-4). Staff also recognizes that these advantages remain in looking at a somewhat wider time
15 period, from approximately 10:00am to 6:00 pm. Staff concludes that an averaging time from 4
16 to 8 hours, generally within the time period from 10:00 am to 6:00 pm, should be considered for
17 a standard to address visibility impairment.
18 In reaching this conclusion, staff recognizes that the national PM2 5 FRM monitoring
19- network provides 24-hour average concentrations, such that implementing a standard with a less-
20 than-«24-hour averaging time would necessitate the use of continuous monitors that can provide
21 hourly time resolution. Given that the data used in the analysis discussed above are from
22 commercially available PM2} continuous monitors, such monitors clearly could provide the
23 hourly data that would be needed for comparison with a potential visibility standard with a less-
24 than-24-hour averaging time. Decisions as to which PM2.5 continuous monitors are providing
,\
25 data of sufficient quality to be used in a visibility standard would follow protocols for approval
26 of reference and equivalent methods that can provide data in at least hourly intervals.
27 Development of the criteria for approval of these reference or equivalent methods for support of
28 a visibility standard would be based upon a data quality objective process,that considers
January 2005
7-8
Draft - Do Not Quote or Cite
-------
1 uncertainties associated with the measurement system and the level of the standard under
2 consideration.
3 73.4 Alternative VM1S Standards to Address Visibility Impairment
4 In considering alternative short-term (4- to 8-hour) PM2 5 standards that would provide
5 requisite protection against PM-related impairment of visibility primarily in urban areas, staff
6 has taken into account the results of public perception and attitude surveys in the U.S. and
7 Canada, State and local visibility standards within the U.S., and visual inspection of
8 photographic representations of several urban areas across the U. S. Staff believes that these
9 sources provide a basis for bounding a range of levels appropriate for consideration in setting a
10 national visibility standard primarily for urban areas.
11 As discussed above in section 6.2, public perception and attitude surveys conducted in
12 Denver, CO and Phoenix, AZ resulted in judgments reflecting the acceptability of a visual range
13 of approximately 50 and 40 km, respectively. A similar survey approach in the Fraser Valley in
14 British Columbia, Canada reflected the acceptability of a visual range of 40 to 60 km. Visibility
15 standards established for the Lake Tahoe area in California and for areas within Vermont are
16 both targeted at a visual range of approximately 50 km. Staff notes that, in contrast to this
17 convergence of standards and goals around a visual range from 40 to 60 km, California's long-
18 standing general state-wide visibility standard is a visual range of approximately 16 km. Staff
19 believes that consideration should be given to national visibility standards for urban areas across
20 the U.S. that are somewhat less stringent than local standards and goals set to protect scenic
21 resources in and around certain urban areas that are particularly highly valued by people living in
22 those areas, suggesting an upper end of the range of consideration below 40 km.
23 Staff has also inspected the photographic representations of varying levels of visual air
24 quality that have been developed for Washington, D.C., Chicago, Phoenix, and Denver
25 (available on EPA's website, http:/viww.epa.gov/tta/naaqs/st^ as an
26 attachment to this document). Staff observes that scenic views (e.g., historic monuments, lake
27 front and mountain vistas) depicted in these images (around and within the three urban areas for
28 which PM2 5 concentrations are reported) are significantly obscured from view at PM2 5
29 concentrations of 35 to 40 ^g/m3 in Chicago, Washington, D.C., and Phoenix, corresponding to
January 2005 7-9 Draft - Do Not Quote or Cite
-------
1 reported visual ranges in Washington, D.C. and Phoenix of 12 to 20 km, respectively. Staff also
2 observes that visual air quality appears to be good in these areas at PM2 s concentrations
3 generally below 20 ug/m3, corresponding to reported visual ranges in Washington, D.C. and
4 Phoenix above approximately 25 to 35 km, respectively. In looking at the images in Denver,
5 staff observes that visual air quality appears to be generally good, specifically in terms of the .
6 ability to view nearby mountain ranges, at a visual range above 52 km. These observations are
7 interpreted by staff as suggesting consideration of a national visibility standard in the range of 30
8 to 20 jAg/m3, The upper end of this range is below the levels at which scenic views are
9 significantly obscured, and the lower end is around the level at which visual air quality generally
10 appeared to be good in these areas. Staff recognizes that the above observations about visual air
11 quality in urban areas inherently take into account the nature and location of scenic views that
12 are notable within and around a given urban area, which has implications for the appropriate
13 design of an associated monitoring network.
i
14 Building upon the analysis discussed above in section 6.2.3, staff has characterized the
15 distributions of PM2 5 concentrations, 4-hour averages in the 12:00 to 4:00 pm time frame, by
16 region, that correspond to various visual range target levels; The results are shown in Figure 7-1,
17 panels (a) through (c), for visual range levels of 25, 30, and 35 km, respectively. This figure
18 shows notable consistency across regions in the median concentrations that correspond to the
19 target visual range level, with what more variation in regional mean values as well as notable
20 variation within each region. In focusing on the median values, staff observes that 4-hour
21 average PM2.5 concentrations of approximately 30, 25, and 20 |ig/m3 correspond to the target
22 visual range levels of 25, 30, and 35 km, respectively. Thus, a standard set within the range of
23 30 to 20 ug/m3 can be expected to correspond generally to median visual range levels of
24 approximately 25 to 35 km in urban areas across the U.S.. Staff notes, however, that a standard
25 set at any specific PM2 5 concentration will necessarily result in visual ranges that vary somewhat
26 in urban areas across the country, reflecting in part the less-than-perfect correlation between
27 PM2 5 concentrations and reconstructed light extinction. Staff also notes that the range of PM2 5
28 concentrations from 30 to 20 ug/m3, suggested by staffs analysis and observations of
29 photographic representations, is generally consistent with national target visual range levels
January 2005 7-10 Draft - Do Not Quote or Cite
-------
::•:::::-:•:::::•:•::: Visual range ~ 25kltt I!!-:-:li!!•!-!! ii-lii ii-l !!ii.::1'ii--.iitiii-iii'ii-iii'iii'iii
:;;;t;;.:;it Visual ranee = 3Skm
Northeast Southeast Industrial Uppef Southwes, Northwest Southern
Midwest Midwest California
Figure 7-1. Distributions of PM25 concentrations for 12 p.m. - 4 p.m.
corresponding to visual ranges of 25km (panel a), 30km (panel b),
and 35km (panel c) — by region. Box depicts interquartile range and
median; whiskers depict 5th and 95th percentiles; star denotes mean.
Source: Schmidt et al. (2005)
January 2005 7-11 Draft - Do Not Quote or Cite
-------
I
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
below 40 km, the level suggested by the public perception surveys and the.local visibility
standards and goals discussed above.
In considering a standard down to 20 ng/m3, staff has again looked to information on i
PM2S background concentrations, as was done in considering primary PM25 standard levels in
Chapter 5, section 5.3.5. In both instances, staff recognizes that an appropriate standard level
intended to provide protection from man-made pollution should be clearly above background '
levels. In considering background levels in conjunction with a primary standard, staff focused
on the 99th percentile of the distribution of estimated background levels; consistent with
consideration of a 98th or 99th percentile form for a primary standard, concluding in that case that
25 ug/m3 was an appropriate lower end to the range of 24-hour primary PM25 standards for
consideration. For reasons discussed below, staff believes that a lower percentile form would be
appropriate to consider for a visibility standard, and thus has looked to a lower percentile in the
distribution of estimated background levels as a basis for comparison with the lower end of the
range of short-term secondary PM2 s standards for consideration. As discussed in Chapter 2,
section 2.6, staff notes that, while long-term average daily PM2 s background levels are quite low
(ranging from 1 to 5 ug/m3 across the U.S.), the estimated 90th percentile values in distributions
of daily background levels are appreciably higher, but generally well below 15 ng/m3, with
levels below 10 ug/m3 in most areas, and these levels may include some undetermined
contribution from anthropogenic emissions (Langstaff, 2005). In addition, staff again notes that
even higher daily background levels result from episodic occurrences of extreme natural events
(e.g., wildfires, global dust storms), but levels related to such events are generally excluded from
consideration under EPA's natural events policy, as noted in section 2.6. Taking these
considerations into account, staff believes that 20 ug/m3 is an appropriate lower end to the range
of short-term PM2 5 standards for visibility protection for consideration in this review.
As in the last review, staff believes that a national visibility standard should be
considered in conjunction with the regional haze program as a means of achieving appropriate
levels of protection against PM-related visibility impairment in urban, non-urban, and Class I
areas across the country. Staff recognizes that programs implemented to meet a national
standard focused primarily on urban areas can be expected to improve visual air. quality in
surrounding non-urban areas as well, as would programs now being developed to address the
>
January 2005 7-12 Draft - Do Not Quote or Cite
-------
1 requirements of the regional haze rule established for protection of visual air quality in Class I
2 areas. Staff further believes that the development of local programs continues to be an effective
3 and appropriate approach to provide additional protection for unique scenic resources in and
4 around certain urban areas that are particularly highly valued by people living in those areas.
5 Based on these considerations, and taking into account the observations and analysis discussed
6 above, staff concludes that consideration should be given to a short-term (4- to 8-hour daylight
7 average) secondary PM25 standard in the range of 30 to 20 ug/m3 for protection of visual air
8 quality primarily in urban areas.
9 7.3.5 Alternative Forms of a Short-term PM2 s Standard
10 In considering an appropriate form for a short-term PM2 5 standard for visibility, staff has
11 taken into account the same general factors that were taken into account in considering an
12 appropriate form for the primary PM25 standard, as discussed above in Chapter 5, section 5.3.6.
13 In that case, as in the last review, staff has concluded that a concentration-based form should be
14 considered because of its advantages over the previously used expected-exceedance form3. One
15 such advantage is that a concentration-based form is more reflective of the impacts posed by
16 elevated PM2 5 concentrations because it gives proportionally greater weight to days when
17 concentrations are well above the level of the standard than to days when the concentrations are
18 just above the standard. Staff notes that the same advantage would apply for a visibility standard
19 as to a health-based standard, in that it would give proportionally greater weight to days when
20 PM-related visibility impairment is substantially higher than to days just above the standard.
21 Further, staff recognizes that a concentration-based form better compensates for missing data and
22 less-than-every-day monitoring; and, when averaged over 3 years, it has greater stability and,
23 thus, facilitates the development of more stable implementation programs. Taking these factors
24 into account, staff concludes that consideration should be given to a percentile-based form for a
25 visibility standard.
The form of the 1987 24-hour PM10 standard is based on the expected number of days per year (averaged
over 3 years) on which the level of the standard is exceeded; thus, attainment with the one-expected exceedance
form is determined by comparing the fourth-highest concentration in 3 years with the level of the standard.
January 2005
7-13
Draft - Do Not Quote or Cite
-------
1 To identify a range of concentration percentiles that would be appropriate for
2 consideration, staff first concludes that constraints on the number of days in which a standard
3 can be exceeded should be appreciably tighter for a standard intended to protect against serious
4 health effects than would be appropriate for a standard intended to protect against visibility
5 impairment, as noted above. Thus, staff believes that the upper end of the range of consideration
6 should be below the 98th to 99* percentiles being considered for a 24-hour primary PM25
7 standard. Staff has also considered that the regional haze program targets the 20% most
8 impaired days for improvements in visual air quality in Class I areas. If a similar target of the
9 20% most impaired days were judged to be appropriate for protecting visual air quality in urban
10 areas, a percentile well above the 80* percentile would be appropriate to increase the likelihood
11 that days in this range would be improved by control strategies intended to attain the standard. A
12 focus on improving the 20% most impaired days suggests to staff that the 90th percentile, which
13 represents the middle of the distribution of the 20% worst days, would be an appropriate form.
14 To assist in understanding the implications of alternative percentile forms in combination
15 with alternative levels of a standard, staff assessed the.percentage of days estimated to exceed
16 various PM25 concentrations in counties across the U.S., as shown in Figure 7-2. This analysis is
17 based on 2001 to 2003 air quality data, using the 4-hour average concentration from 12:00 to
18 4:00 pm at the maximum monitor in each county. This assessment is intended to provide some
19 rough indication of the breadth of additional protection potentially afforded by alternative
20 percentile forms for a given standard level: Staff notes that a 90th percentile form, averaged over
21 3 years, that allows 10% of the days to be above the level of the standard provides additional
22 protection of visual air quality in far fewer areas at a standard level of 30 ug/m3 than at a level of
23 20 ug/m3.
24 Based on the factors discussed above, staff concludes that a percentile-based form should
25 be considered, based on a percentile at or somewhat above the 90th percentile. Staff believes that
26 a form selected from within this range could provide an appropriate balance between adequately
27 limiting the occurrence of peak concentrations and providing for a relatively stable standard.
January 2005 7-14 Draft - Do Not Quote or Cite
-------
_ o
e
41
en
,2
*
ft
u
fe
I
•d
o
CO
[3A3| Bujpsaoxa sAep 10
I
-------
1 7.3.5 Summary of Staff Recommendations
2 Staff recommendations for the Administrator's consideration in making decisions on the
3 secondary PM2 5 standards to address PM-related visibility impairment, together with supporting
4 conclusions from sections 7.3.1 through 7.3.4, are briefly summarized below. Staff recognizes
5 that selecting from among alternative standards will necessarily reflect consideration of the
6 qualitative and quantitative uncertainties inherent in the relevant information. In making the
7 following recommendations, staff is mindful that the Act requires secondary standards to be set
8 . lhat are requisite to protect public welfare from those effects that are judged to be adverse, such
v
9 that the standards are neither more nor less stringent than necessary. • The provisions do not
10 require that secondary standards be set to eliminate all welfare effects.
11 (1) Consideration should be given to revising the current suite of secondary PM25 standards
12 to provide increased and more targeted protection primarily in urban areas from visibility
13 impairment related to fine particles.
14 (2) The indicator for a fine particle visibility standard should be PM25, reflecting the strong
15 correlation between short-term average PM25 in urban areas across the U.S. and light
16 extinction, which is a direct measure of visibility impairment.
17 (3) Consideration should be given to a short-term averaging time for a PM2 5 standard, within
18 the range of 4 to 8 hours, within a daylight time period between approximately 10:00 am
19 to 6:00 pm. To facilitate implementation of such a standard, consideration should be
20 given to the adoption of FEMs for appropriate continuous methods for the measurement
21 of short-term average PM2 5 concentrations.
22 (4) Consideration should.be given to alternative PM25 standards to provide protection against
23 visibility impairment primarily in urban areas. This recommendation reflects the
24 recognition that programs implemented to meet such a standard can be expected to
25 improve visual air quality in non-urban areas as well, just as programs now being
26 developed to address the requirements of the regional haze rule, for protection of visual
January 2005 7-16 Draft - Do Not Quote or Cite
-------
1
2
3
4
5
6
7
8
9
10
11
12
air quality in Class I areas, can also be expected to improve visual air quality in some
urban areas. Recommendations on ranges of alternative levels and forms for such a
standard include:
( a) Staff recommends consideration of a 4- to 8-hour PM2 5 standard within the range
of 30 to 20 ng/m3. Staff judges that a standard within this range could provide an
appropriate degree of protection against visibility impairment, generally resulting
in a visual range of approximately 25 to 35 km, primarily in urban areas, as well
as improved visual air quality in surrounding non-urban areas.
(b) Staff also recommends consideration of a percentile-based form for such a
standard, focusing on a range at or somewhat above the 90* percentile of the
annual distribution of daily short-term PM2 s concentrations, averaged over 3
years.
13 7.4 STANDARDS TO ADDRESS OTHER PM-RELATED WELFARE EFFECTS
14 EPA's decision in 1997 to revise the suite of secondary PM standards took into account
15 not only visibility protection, but also materials damage and soiling, the other PM-related
16 welfare effect considered in the last review. Based on this broader consideration, EPA
17 established secondary standards for PM identical to the suite of primary standards, including
18 both PM2 5 and PM10 standards, to provide appropriate protection against the welfare effects
19 associated with fine and coarse particle pollution (62 FR at 38,683). This decision was based on
20 considering both visibility effects associated with fine particles, as discussed above in section
21 7.3, and materials damage and soiling effects associated with both fine and coarse particles.
22 With regard to effects on materials, EPA concluded that both fine and coarse particles can
23 contribute to materials damage and soiling effects. However, EPA also concluded that the
24 available data did not provide a sufficient basis for establishing a distinct secondary standard
25 based on materials damage or soiling alone. These considerations led EPA to consider whether
26 the reductions in fine and coarse particles likely to result from the suite of primary PM standards
27 would provide appropriate protection against the effects of PM on materials. Taking into
28 account the available information and the limitations in that information, EPA judged that setting
29 secondary standards identical to the suite of PM25 and PM10 primary standards would provide
January 2005 7-17 Draft - Do Not Quote or Cite
-------
1 increased protection against the effects of fine particles and retain an appropriate degree of
2 control on coarse particles.
3 In this review, in addition to addressing visibility impairment, the CD has broadened its
4 scope to include effects on ecosystems and vegetation, discussed in Chapter 6, section 6.3, and
5 also addresses PM-related effects on materials, discussed in section 6.4, and the role of ambient
6 PM in atmospheric processes associated with climate change and the transmission of solar
7 radiation, discussed in section 6.5. In considering the currently available evidence on each of
8 these types of PM-related welfare effects, staff notes that there is much information linking
9 ambient PM to potentially adverse effects'on materials and ecosystems and vegetation, and on
10 characterizing the role of atmospheric particles in climatic and radiative processes. However, on
11 the basis of the evaluation of the information discussed in Chapter 6, which highlighted the
12 substantial limitations in the evidence, especially with regard to the lack of evidence linking
13 various effects to specific levels of ambient PM, staff concludes that the available evidence does
14 not provide a sufficient basis for establishing distinct secondary standards based on any of these
15 ' effects alone. These considerations lead staff to address in the following sections'whether the
16 reductions in fine and coarse particles likely to result from the current secondary standards, or
17 the range of recommended revisions to the primary standards and the secondary PM25 standard
18 to address visibility impairment, would provide appropriate protection against these other PM-
19 related welfare effects.
20 7.4.1 Vegetation and Ecosystems
21 With regard to PM-related effects on ecosystems and vegetation, staff notes that the CD
22 presents evidence of such effects, particularly related to nitrate and acidic deposition, and
23 concludes that current PM levels in the U.S. "have the potential to alter ecosystem structure and
24 function in ways that may reduce their ability to meet societal needs" (CD, p, 4-153). Much of
25 the associated uncertainty surrounding the characterization of the relationships between ambient
26 PM levels and ecosystem or vegetation responses is related to the extreme complexity and
27 variability that exist in predicting particle deposition rates, which are affected by particle size
28 and composition, associated atmospheric conditions, and the properties of the surfaces being
29 impacted. Though several national deposition monitoring networks have been successfully
January 2005 ' 7-18 Draft - Do Not Quote or Cite
-------
1 measuring wet and dry deposition for several decades, they often do not distinguish the form
2 (e.g., particle, wet, and dry gaseous) in which a given chemical species is deposited, so that it is
3 difficult to know what percentage of total deposition is attributable to ambient PM. Further, data
4 from monitoring sites generally do not address all the variables affecting deposition that come
5 into play in a natural system.
6 In addition to these uncertainties, many of the documented PM-related ecosystem-level
7 effects only became evident after long-term, chronic exposures to specific chemical
8 constituents) of PM eventually exceeded the natural buffering or assimilative capacity of the
9 system. In most cases, PM .deposition is not the only source of the chemical species to the
10 affected system and the percentage of the deposition due to ambient PM is often not known.
11 Because ecosystems have different sensitivities and capacities to buffer or assimilate pollutants,
12 it is difficult to predict the rate of deposition that would be likely to lead to the observed adverse
13 effects within any particular ecosystem. Equally difficult is the prediction of recovery rates for
14 already affected areas if deposition of various chemical species were to be reduced.
15 Despite these uncertainties, a number of significant and adverse environmental effects
16 that either have already occurred or are currently occurring have been linked to chronic
17 deposition of chemical constituents found in ambient PM. Staff notes, for example, mat the
18 following effects have been linked with chronic additions of nitrate and its accumulation in
19 ecosystems:
20 • Productivity increases in forests and grasslands, followed by decreases in productivity
21 and possible decreases in biodiversity in many natural habitats wherever atmospheric
22 reactive nitrogen deposition increases significantly and critical thresholds are exceeded;
23 • Acidification and loss of biodiversity in lakes and streams in many regions, especially in
24 conjunction with sulfate deposition; and
25 • Eutrophication, hypoxia, loss of biodiversity, and habitat degradation in coastal
26 ecosystems.
27 Staff notes that effects of acidic deposition have been extensively documented, as
28 discussed in the CD and other reports referenced therein. For example, effects on some species
29 of forest trees linked to acidic deposition include increased permeability of leaf surfaces to toxic
30 materials, water, and disease agents; increased leaching of nutrients from foliage; and altered
January 2005
7-19
Draft - Do Not Quote or. Cite
-------
1 reproductive processes; all of which serve to weaken trees so that they are more susceptible to
2 other stresses (e.g., extreme weather, pests, pathogens). In particular, acidic deposition has been
3 implicated as a causal factor in the northeastern high-elevation decline of red spruce. Although
4 U. S. forest ecosystems other than the high-elevation spruce-fir forests are not currently
5 manifesting symptoms of injury directly attributable to acid deposition, less sensitive forests
6 throughout the U.S. are experiencing gradual losses of base cation nutrients, which in many
7 cases will reduce the quality of forest nutrition over the long term.
8 Taking into account the available evidence linking chemical constituents of both fine and
9 coarse PM to these types of known and potential adverse effects on ecosystems and vegetation,
10 staff believes that further reductions in ambient PM would likely contribute to long-term
11 recovery and to the prevention of further degradation of sensitive ecosystems and vegetation.
12 Staff recognizes, however, that the available evidence does not provide any quantitative basis for
13 establishing distinct national standards for ambient PM. Further, staff recognizes that due to
14 site-specific sensitivities to various components of ambient PM, differing buffering and
15 assimilative capacities, and local and regional differences in the percentage of total deposition
16 that is likely attributable to ambient PM, national standards alone may not be an appropriate
17 means to protect against adverse impacts of ambient PM on ecosystems and vegetation in all
18 parts of the country. Nonetheless, staff believes that reductions in fine and coarse particles likely
19 to result from the current suite of secondary standards or the range of recommended revisions to
20 the primary standards would contribute to increased protection against PM-related effects on
21 ecosystems and vegetation. Staff recommends that the potential for increased protection of
22 ecosystems and vegetation be taken into account in considering whether to revise the current
23 secondary PM standards. Further, staff believes that any such increased protection should be
24 considered in conjunction with protection afforded by other programs intended to address
25 various aspects of air pollution effects on ecosystems and vegetation, such as the Acid
26 Deposition.Program and other regional approaches to reducing pollutants linked to nitrate or
27 acidic deposition.
January 2005 7-20 Draft - Do Not Quote or Cite
-------
1 7.4.2 Materials Damage and Soiling
2 With regard to PM-related effects on materials, staff notes that the available evidence
3 continues to support the following observations:
4 • Materials damage and soiling that occur through natural weathering processes are
5 enhanced by exposure to atmospheric pollutants, most notably SO2 and particulate
6 sulfates.
7 * While ambient particles play a role in the corrosion of metals and in the weathering of
8 paints and building materials, no quantitative relationships between ambient particle
9 concentrations and rates of damage have been established.
10 • Similarly, while soiling associated with fine and coarse particles can result in increased
11 cleaning frequency and repainting of surfaces, no quantitative relationships between
12 particle characteristics (e.g., concentrations, particle size, and chemical composition) and
13 the frequency of cleaning or repainting have been established.
14 Staff believes that these observations and the underlying avail able evidence continue to support
15 consideration of retaining an appropriate degree of control on both fine and coarse particles.
16 Lacking any specific quantitative basis for establishing distinct standards to protect against PM-
17 related adverse effects on materials, staff recommends consideration be given to (1) retaining the
18 current secondary PM2 5 standards or revising those standards to be consistent with any revisions
19 made to the primary PM2S standards or to the secondary PM25 standards to address visibility
20 impairment, and (2) retaining secondary standards for coarse particles, using a PM,0.2 5 indicator
21 consistent with the primary standards, at a level that either retains the degree of protection
22 afforded by the current PM10 standards or that is consistent with any new PMi0.2 5 primary
23 standards.
24 7.4.3 Climate Change and Solar Radiation
25 With regard to the role of ambient PM in climate change processes and in altering the
26 penetration of solar UV-B radiation to the earth's surface, staff notes that information available
27 in this review derives primarily from broad-scale research and assessments related to the study of
28 global climate change and stratospheric ozone depletion. As such, this information is generally
29 focused on global- and regional-scale processes and impacts and provides essentially no basis for
30 characterizing how differing levels of ambient PM in areas across the U.S. would affect local,
January 2005 . 7-21 Draft - Do Not Quote or Cite
-------
1 regional, or global climatic changes or alter the penetration of UV-B radiation to the earth's
2 surface. As noted in section 6.5, even the direction of such effects on a local scale remains
3 uncertain. Moreover, similar concentrations of different particle components can produce •
4 opposite net radiative effects. Thus, staff concludes that there is insufficient information
5 available to help inform consideration of whether any revisions of the current secondary PM
6 standards are appropriate at this time based on ambient PM's role in atmospheric processes
7 related to climate or the transmission of solar radiation.
8 7.4.4 Summary of Staff Recommendations
9 Taking into account the conclusions presented in sections 7.4.1 through 7.4.3 above, staff
10 makes the following recommendations with regard to PM-related effects on vegetation and
j
11 ecosystems and materials damage and soiling:
•».
12 (1) Consideration should be given to retaining secondary standards for fine and coarse
13 particles that at a minimum retain the level of protection afforded by the current PM2 5
14 and PM10 standards so as to continue control of ambient particles, especially long-term
15 deposition of particles, especially particulate nitrates and sulfates, that contribute to
16 adverse impacts on vegetation and ecosystems and materials damage and soiling.
17 (2) For consistency with the primary standards, secondary standards for fine and coarse
18 particles should be indexed by PM25 and PMi0_25. While staff recognizes that PM-related
19 impacts on vegetation and ecosystems in particular are associated with chemical
20 components in either size fraction rather than with particle size per se, staff also
21 recognizes that sufficient information is not available at this time to recommend
22 consideration of an ecologically based indicator in terms of a specific chemical
23 ' component of PM.
24 In making these recommendations, staff has taken into account both the available
25 evidence linking fine and coarse particles with effects on vegetation and ecosystems and material
26 damage and soiling, as well as the limitations in the available evidence. In so doing, staff
January 2005 7-22 Draft - Do Not Quote or Cite
-------
1 recognizes that the available information does not provide a sufficient basis for the development
2 of distinct national secondary standards to protect against such effects beyond the protection
3 likely to be afforded by the suite of primary PM standards.
4 7.5 SUMMARY OF KEY UNCERTAINTIES AND RESEARCH
5 RECOMMENDATIONS RELATED TO STANDARD SETTING
6 Staff believes it is important to continue to highlight the unusually large uncertainties
7 associated with establishing standards for PM relative to other single component pollutants for
8 which NAAQS have been set. Key uncertainties and staff research recommendations welfare-
9 related topics are outlined below. In some cases, research in these areas can go beyond aiding in
10 standard setting to aiding in the development of more efficient and effective control strategies.
11 Staff notes, however, that a full set of research recommendations to meet standards
12 implementation and strategy development needs is beyond the scope of this discussion.
13 With regard to welfare-related effects, discussed in Chapter 4 of the CD and Chapter 6
14 herein, staff has identified the following key uncertainties and research questions that have been
15 highlighted in this review of the welfare-based secondary standards:
16 (1) Refinement and broader application of survey methods designed to elicit citizens'
17 judgments about the acceptability of varying levels of local visibility impairment could
18 help inform future reviews of a visibility-based secondary standard. Such research could
19 appropriately build upon the methodology developed by the State of Colorado and used
20 as a basis for setting a visibility standard for the city of Denver, which has been adapted
21 and applied in other areas in the U.S. and abroad.
22 (2) There remain significant uncertainties associated with the characterization and prediction
23 of particle deposition rates to natural surfaces in general, and most importantly, with
24 respect to nitrogen deposition in particular. Reduction in these uncertainties will be key
25 to developing the capability of quantitatively linking ambient PM concentrations with
26 environmental exposures and response. In order to better understand the nature of the •
27 role that PM plays in cumulative long-term environmental impacts, more research needs
January 2005 7-23 Draft - Do Not Quote or Cite
-------
3
4
5
6
7
8
9
10
11
to be conducted on the percentage of total deposition contributed by PM and where
necessary, better tools and monitoring methods should be developed.
(3) The immense variability in sensitivity to PM deposition across U.S. ecosystems has not
yet been adequately characterized, specifically the factors controlling.ecosystem
sensitivity to and recovery from chronic nitrogen and acid inputs. Data should be
collected on a long-term basis on a greater variety of ecosystems in conjunction with the
development of improved predictive models. Such research could help in future
consideration within the U.S. of the "critical loads" concept, which is generally accepted
s
in Europe as the basis for abatement strategies to reduce or prevent injury to the
functioning and vitality of forest ecosystems caused by long-range transboundary chronic
acidic deposition.4
t
4 This recommendation is consistent with the views of the National Research Council (NRC) contained in
its recent review of air quality management in the U.S. (NRC, 2004). This report recognizes that for some resources
at risk from air pollutants, including soils, groundwaters, surface waters, and coastal ecosystems, a deposition-based
standard could be appropriate, and identifies "critical loads"as one potential approach for establishing such a
deposition-based standard.
January 2005
7-24
Draft - Do Not Quote or Cite
-------
1 REFERENCES
2 Langstaff, John E. (2004). Estimation of Policy-Relevant Background Concentrations of Particulate Matter.
3 Memorandum to PM NAAQS review docket OAR-2001 -0017. January 27,2005.
4 National Research Council (NRC) (2004). Air Quality Management in the United States. Committee on Air Quality
5 ' Management in the U.S., National Research Council of the National Academy of Science. The National
6 Academies Press, Washington, D.C. ISBN 0-309-08932-8.
7 Schmidt et al. (2005) Draft analysis of PM ambient air quality data for the PM NAAQS review. Memorandum to
8 PMNAAQS review docket OAR-2001-0017. January 31,2005.
January 2005 7-25 Draft - Do Not Quote or Cite
-------
t
1 APPENDIX 2A. Source Emissions
- 2 '
3 The distribution and amount of emissions of pollutants that contribute to ambient PM can
4 provide insights into observed ambient levels. The links between source emissions and ambient
5 concentrations of PM can include complex, non-linear atmospheric processes, including gaseous
6 chemical reactions and pollution transport.
7 Source emissions can be measured using monitoring equipment or estimated using
8 emission inventory methods. For most source types, emissions inventory methods are the most
9 practical. The EPA routinely publishes national estimates of annual source emissions of
10 pollutants that contribute to ambient PM concentrations. In general, national emissions estimates
11 are uncertain, and there have been few field studies to test emission inventories against
12 observations. The draft CD concludes that uncertainties in national emissions estimates could be
13 as low as ±10 percent for the best characterized source categories (e.g., SO2 from power plants
14 measured by continuous instruments), while fugitive dust sources should be regarded as order-
15 of-magnitude (CD, p. 3-98). The EPA is working to reduce these uncertainties through advances
f 16 in the understanding of the fate and transport characteristics of fugitive dust emissions released
17 at ground level. Episodic emissions from dust storms and forest fires are difficult to quantify and
18 to allocate accurately in space and time, and discerning between natural and anthropogenic
19 "causality" for these source categories is especially challenging.
20 Table 2A-1 provides a summary of recent annual estimates of national emissions of
21 primary PM and PM precursors. While reviewing the following discussion on emissions
22 estimates, the reader should keep in mind that national estimates, while instructive, can obscure
23 important distinctions in the relative contributions of different sources across smaller geographic
24 regions, including important differences between urban and rural areas.
25
26 Primary PM Emissions
27 The majority of directly emitted anthropogenic PM is estimated to be coarse particles.
28 Though highly uncertain, recent national estimates of PMi0.25 emissions (total of all sources)
29 shown in Table 2A-1 are about 2.5 times higher than estimates of PM25 emissions -16.3 million
30 short tons compared to 6.6 million short tons. A large portion of primary PM emissions are
January 2005 2A-1 Draft - Do Not Quote or Cite
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
attributed to a variety of small area-wide sources, which are often more difficult to characterize
and are more uncertain than larger point source emissions.
National estimates of primary PM10_25 are dominated by fugitive dust and agricultural
sources. Fugitive dust sources include paved and unpaved road dust, dust from construction and
agricultural activities, and natural sources like geogenic wind erosion (not estimated in Table
2A-1). Fugitive dust is also a significant source of primary PM25. Unlike PMI0.25, where
fugitive dust emissions comprise about 75 percent of total emissions, fugitive dust emissions of
PM2S is only about one-third of total emissions. Recent research has found that about 75 percent
of these emissions are within 2 meters of the ground when measured. A significant portion of
these coarse-mode particles are removed or deposited within a few kilometers of their release
point due to turbulence associated with surface topography, and the presence of vegetation or
structures (DRI, 2000). This is consistent with the generally small amount of crustal material
found in ambient PM25 samples in most locations. As shown in Table 2A-1, direct emissions
from fuel combustion, industrial processes, fires, and motor vehicles contribute more to primary
PM2.5 than to primary PM10.2 5. Recent improvements to methodologies for estimating
emissions, reflected in the values in Table A-l, include:
• Wildfires and prescribed burning - use of state-specific fuel loading factors and improved
emission factors
• Residential wood combustion (woodstoves & fireplaces) - recalculation of emissions
using updated wood consumption data
• Condensible PM emissions - added these emissions; were not previously included
• Animal husbandry - updated NH3 emissions for this category based on recent work by
EPA's Emission Standards Division/OAQPS
• Mobile source emissions - updated estimates using the latest MOBILE and NONROAD
models
Secondary PM Precursor Emissions
Major precursors of secondarily formed fine particles include SO2, nitrogen oxides
(NOJ, which encompasses NO and NO2, and certain organic compounds. Table 2A-1 shows the
estimated contribution of various sources to nationwide emissions of SO2 NOX, VOC, and NH3.
January 2005 2A-2 Draft - Do Not Quote or Cite
t
-------
1 Fuel combustion in the power generation and industrial sectors dominates nationwide estimates
2 of SO2 emissions and contributes significantly to NOX emissions. However, emissions from
3 motor vehicles comprise the greatest portion of nationwide NOX emissions. Motor vehicle
4 emissions also make up a substantial portion of nationwide VOC emissions, with additional
5 contributions from the use of various solvents in industrial processes and commercial products.
6 The vast majority of nationwide NH3 emissions are estimated to come from livestock operations
7 • and fertilizer application, but in urban areas there is a significant contribution from light-duty
8 cars and trucks, as well as certain industrial processes.
9 The relationship between changes in precursor emissions 'and resulting changes in
10 ambient PM25 can be nonlinear. Thus, it is difficult to project the impact on ambient PM25
11 arising from expected changes in PM precursor emissions without air quality simulation models
12 that incorporate treatment of complex chemical transformation processes and meteorology.
13 Generally SO2 emissions reductions lead to reductions in sulfate aerosol, and NOX emissions
14 reductions lead to reductions in nitrate aerosol. However, the direction and extent of changes
15 will vary by location and season, depending on fluctuations in NH3 emissions and changes in
16 prevailing meteorology and photochemistry.
t
January 2005 2A-3 Draft - Do Not Quote or Cite
\
-------
p
2
w *
e «
a. s
•5 S3
A *»
I =
•g s
•- "C
2 o
"I w
il
i.i
S T3
w
Oi
vt •
3^2
fk
£3 Ofl t^
*^ S o
^ a a,
%**
6 a)
P
O jj
'u' S3 ^
^ ts
O V,
in 5
o\ (X
^^ 1H m_.
i °"^
G
g ^
^ ^"u
M f^ ^
0 ^
• S V!
.s •§
f i
<. §
^ O
Original study*
Study Location
Reanalysis study
•.
-
e
5
f
wmmmwm
'
s ^ ^ If s |
5 si xtL S-"" ^""* i
\ * * 1 3|
5 a, a, °" gs i
i i i i -
*-:
*s"'
! ! ! ! **"»"
! ! ! •
iii' ?f)
«->
^ "^ § "^
r^ ^^ ^ *-s , 1
^- sL ^L ^
s s s s
*tt tS "tt "tS
§ 3 3 § S
•*M i»j *« '*•* ?*
§o o o «
C K S !>3
!5 »5 -5 •§ *§
^ ^ ^ N ^
O C3 C) O O
O\ \j)
t |3afe,1 *
|tj ^-| ^f | j "§
1 | 11 "3r§ Jf «
•a S ^^ ^ ^ ^ s fc
a la ii il i
^ ^ 0, to <%
va
tC
T "l" ^ °0" oO \o~ OC 'O*
Ol
o
5*
t
vo
S
1^
**»
t
-------
u
g
S
32.3)
2L9)
^ Ci
aa&
vo \o 2
iX O\ —-
^ ** ^
0 »> cy
aaV
Q
7.8-45.6
. 2-29. 6
6. 1
1
means 1
s 6.6-
PM1
PU2
PMt
*f ^ _;
vi 2 ^ O 3
a 2 is if I
9 9
°° i
°! vi
P, P.
vq p
TT oo
dc5 5 3
O C3 C) O
•0
o o
'5 00
w Z
0 0
00
T ^-
rf -H"
O ^
£ !"
5
o o"
SO vi
(S -^
^i r-
•" cs
o
0
o5 &:
"S.^
"s^i
j 1
I i i
5-S -S
Scartz et al., 1
Portage, WI
Schwartz, 2003a
Schwartz et al., 1
Topeka, KS
Schwartz, 2003a
3a
2
t aL,
Over
Schwartz
6 Cities,
Schwart
«
o
•
en
1
-------
I-r
—r w
?|
a §
8*
O
S
u
u, <->
o
m
C3\
I
* It
>>-9 R
1
O 03
'
u-i \O O
P?
t-H
t~-
^ m t> -3
S3 i %
oo
ocT oC
oo m
K oo
o
O
ii
r-
(S
S I
o o
c\ ^ vq i": —i vo t
oj 2 o oo -^ in <
"*! 3 s 2 3 a
s" 2 2^^" 2 '
c-
CO
n
.a s
•30 T5
tS -S § '
CS "
(S P"
o "1
s-
S
r-
od
w "5 S
z a w
o o
r~* ^o
vo r-
•-< fS
.-. « o
q1 m t-^ c-
o •7 *7 "7
O OO O OO O OO OO
i
>-,
| o g
a s
s.-g-
s o
fi X!
_; o
S|
vb Ix
CO
*' ^"n "^w *•».*:
2j• o
5 M S
O
3
O)
-------
PM10.
Levels
"
PM10, PM2 j
Mean (Ran
Repo
u
vi
8-1
a.
o
*
>n
a!
0,
*
u
on
y
y Lo
lysis
S
R
-r^" q
i 80
p ^ V?
;«r«? 2
oo" ff
o'q^
is
o o
P
P
5SJ 5S3 §
§E g§2-
DH
*> O
00 >"
o o
q °!
o o
3
°\
f>
m
•n
I—I fl pvj t-"1
vi o1*
I
OH
o
!|I|i'|\glS
;S^f> >U> ^JJH
!J?« o3*jOS ^HnQ8>
D^M OOT500 ~^-a
o> 8 *
.$^$ a -g j» "T- a i
."3 § "3 3o"o £§S
I -Si j ° I j M a
I
I
a
I
en
-------
o
1
v>
*
^
60
"->
cs
*
in
O-l
-a -g
M O
a
"0
cs
I
r-
(N
6
o
I
s
0)
O
I
en
o\
CM
1
O
a
G
I
a
•3
ITS
00
s
o
r-;
-------
#«
&
2:
o
^ u 0
U S n
S S-
I
y
S rf
C>
vi
Vlf
» J •§
"3 >..2?
R
ts
vq
«n
O
t>
£
CO
i
a
00
00
v o ^ o
-------
Repo
t, 0*
B «
&
tS
O
— t- »n
u
Os
PM1047
a,
vo" m
IT) O,
06 od
3 o
* 'S
o o
O
ey, 1999
nta Clara Co
Fair
vo
S
vo
vo
•fi 60
=> 7; 00 __
S P " g
=>:^-Q ^.
•<. o rvi cs
00 r—
sag s « ao ^2
s
a
o
o
00
O O
vq o
fS rrj
O O
- §*
S 2*£
f-; O
fl 00
"* (N
VD
'oO
0 ~
oo a!
C-2-
r- >n
Os ON
<*>
O O
rr ~
^"«?
«> t-
ts ^
SO vo
T5
IS
3S
o o
c
o\,o o . o Q"SO
SS1O IN»J(S (S.JC-* O^S
3o -«J- »'-1-rs»>r»
M<;<^ 53"«3S S "S -oa'r
JSa),g jg o S ^3=5-3
|,3 fe S° "
etal
1 1 u 1 -
^ « l> «>-S J> « si
ooo ooo ^o
£J3 .3^,0 -go
22220
G
I
a
i
>0
-------
t
1
a
.
oo
O ^"i
i-; d
en
o
CM"-*
CO P1^
T.
TT
o
SS
O O
O
1 IB I3'
•n
I
-------
<£• *
es
33
11
£*
Ct
5P
OfJ
22?
8
CN
00
y—V 00
f: G^ ci
•n ^ in <*>
$S~$ %
o\ °; HI —i
M M « ,x
O
•^
00
tn
rf
S
cs
ON
oq
S
•° *«•
§48
. CN of es
,« JJ «
111
S < S
3> w ^P
J3J
o
T3
~O
•a d-
-5 •§
* ^
4> *J
sw
8P
I"
S
2 00
cK 9
„ \
-------
^
«i ^
_
1 ^ I
„ 31 €
«« § ^
S ^(2
S §'
g"l
G ~
x^ ^
0
vo;^
00 2
J? -
£ 2*
fa
Av
. '
1
H
^
0
0
^_^
«*1
—
P
vU
r-i
P
^
J^
C5
Z
J
O
Dysrhythmias Hospital Admi
Lippmann et al., 2000
Detroit, MI (>65 years)
Ito (2003)
v
Q
g
en
1
1
a.
1
CA
«
U
M
5
r
Heart Failure/Congestive He
"^\ vS*
00 o
?f t'
v*" ^
§ rr
5 g
•§ 1
a. g
i
j
«s
oq
"S
i
o
Schwartz and Morris, 1995
Detroit f> 65 years)
rs
2
&
«n
29 yeare
^^
in /*^v £^~ ' f^
2 £g «i ^
g vo; S ~- '
ooo 2 ^
~ 1 3 """
"> 3 2 °
oJ" ± ^
rx^fc
/— v ^
*n •
m 2
•*•' H '
vL-^
t 0\
TT
y— v / — \
O 00
12 2
•5T f| !
q oq
ob vo
c?P g
o^ ao v«
^•H ^-1 • ^
25 S?
^~i/ ^
S 5 ^
13
2 |
o § §
w O O
Lippmann et al., 2000
Detroit, MI (>65 years)
Ito, 2003
Morris and Naumova, 1 998
Chicago, IL (>65 years)
B
_o
S
a
'
3
Myocardial Infarction Hospi
fN|
2
in"
m
>n
o
2*
fc
j
i
/^^
vd
S
e
S
*"
"8
3
1
o
Linn etal., 2000
Los Angeles, CA (>29 years
X-N /~\ X-N
(N n in v-i
O O O
2" S 2
fr, E OH
i i 1
i i i
£T ST Q
O P-. c\
•*fr ts o\
ro PO ^f
CT» ^^ ^n
T *? ^
O r*^ .
• to
i— i O • >tj
1 "8 • "8
3 3 3
•s -s -s
i; S (3 C
.So go o
a -g
~ ,s
i . - 1 ^ ^
CS MX 'IT^ •« W
-S b <^ £ 2 b
"S. S "• * S
o ^ a >> 'i ^
B 3 g. oj .2 g^
.2 & I t 1 ^
1 .-3 * < ^ ^
£'o° ,2 oo •« o1^
! a.| I S4 fM
u "3S 5 "3 ?• Da "3 S
a -« < }• •«< »-«<<
1. J Mil ia
w 3 u a J oS j
1
a
<^^
i^
S
•^j
O
5
^
a
i
^
k.
Q
o
-------
1
u
^
£
'-' t, "
u u
|
PM
25
%
sr*
OS n in
OS OS
— ^ O
I"- I-"
«N 00
22
S o
b
9
I
I
(H
U
t.
o
22
OH OH
O
0
o
o
— i
«
•^ v£> ^T v:
111 I
o
•a 2
» ^
s (§
PM1035.2(S
C-
ro
ss
on
v
1 8
00
0
1
O
1
o
«" B
a o
s •
J ffl
I
0
I
a
CO
O
o
o
•a">
«a
^1
I*
-------
sa
>n
T
1—4
ON"
\o
oo
00
ts
t
O
3
1
1
o
S
8
rn
cs
8
S
\d
fi
O
o
>%
QA ^7
2^
•a'--
TtS s
tc o
gS
.a
r-
<»?
T—*
^S
00
OO
1
"S
g
*S
.2
1
•<
1
E.
M
£
i
a*
.2 5^,
g tf-
a
1 i
£ I
a
1
o*
G
I
a
i
I
cs
3
S
-------
||
f*H "1 «
g 'ST'S
•0 S a
:£t &
S *;;&
& a
P
0 ,
£ i
^ H ^
a ^"s
nj -5
Q Esfl
1 a
0
^ M
i/i 2
0^ Oi
^ S^g
g "So
g >
*
? a
1, i
1*1
g a.
1 *
» f
fi-
^ 0
Original study*
Study Location
ReanaJysis study
«jj
*»
S
i
o
s
p
'
I
oo o vo'cn'
"2 ""2
oC oT <*f in"
vi TT SM
V™> \s_X \^>
00 tr> <* m
00 00* ^ \O
5 5
O O ^Z fc
•*<-* •«-• <^F< ^r<
_65y
Zanobetti and Schwart
?ffs
I"H QQ
l-ej
s2al
*5?" n S
&, *
,_,
o »
vS
00 00
< M
OZ
S O
s
.5
•t?1 S
Lippmann et al., 2000
Detroit, MI (>65 year
Ito 2003
COPD Hospital Admi
1^1 /* "N /*^\ -^^
&=> ^ s 8j> sj !d in ,«« 1"H "o
^ Cs 2 ^" fn .: 5J
S* s 2 2^ -
S fc 0-
^v?
5J IN
II ! *o" **•
' ' 1 TT 7
^oo'
<» 2
ON1 O>"
H o
i i i ££
vo ON
1 I
O fO
-i
u ^T ^ S
|89 ofg.s
8s §13 |s2t
5^- ^s i -I, 1^^
£•* ««'§ 31 i-fs
Il«^o " " ll^
•§ Q 1 ^ s .! »3 .§! <5 ^j"
1
^v
I/I
-------
3.9
«.
1
H
1st
ji
0*
&
0,
u
£
v>
c\
|
1
O
5n
¥
C9
1
*
§
* e^
W>
v^X
!
t
ji
*~^ TJ
u
ftS
ft
g<2
<3
i
«*.£
(X
°"^
^
g
u g
01 "5b
i^
^
PL,
l\
a.
8
1
o
O
Original study*
Study Location
Reanalysis study
S§ '
r~ 4 ^
•If £j
g §
•3 «§
= *>
'ss
£ CL,
,-
o^ P S"
(^ CJ "1
r- «ri •"
8 R S
si • e 3,
S5 fc ov
VO OO (o
•* ^ cs
S? -^ '-*
<"">.?! Ov
S oo 06
-r «A ri
S "^ S
5 e d
^ C4 0
* W) p
< < 1»
O O ^ 2 ^
S o -P o jr o
O3 m 03 •— U «— i
Moolgavkar, 2000c
Los Angeles, C A (all ages)
Moolgavkar 2003
VI f*^ (S
VO VO ^
in „ -n -
t— r ^ vf
>/l ^^ tfl
rf +*•*-*
g 8 • S-
5 S S
O O C3
Asthma Hospital Admissions:
Choudbury et al, 1 997
Anchorage, AK
Medical Visits (all ages)
Jacobs et al, 1997
Butte County, CA (all ages)
Linn etal, 2000
Los Angeles, CA (>29 years)
§
oC
cs
— i
o
ex.
i
i
a\
vi
t-"
r-i
v^/
*J
5
O
fl
cs
c^-
o
Ov"
^^
s
Q\ O\ \
-*\
TV
S
3
Ts
1>
S
a o* ""i ^ vd
PK
^
1
I
Lipsett etal, 1997
Santa Clara Co, CA (all ages)
w,
00
1
1
1
Vs!
**!
fn"
o
1
5
O
Nauenberg and Basu, 1999
Los Angeles, CA (all ages)
OO
o
(X,
i
1
p
CN
H"
~
; not used
>
O
Tolbert etal, 2000
^H \D f
O "1
. PH &,
<^5 ^^
• (
SM"
vi ^
v\ v\
. o
s q.
cio
r- in
00 VO
<3~
O\ '
*-* ^2
«f ^"
2 «>
< 00
^2
i^
Atlanta, GA (<17 years)
Sheppardetal, 1999
Seattle, WA (<65 years)
-H
e*
£
Sheppard et al, 2003
O)
fe
I
8
i
<
.PI
b
-------
»- &
5 „
°<
o
£
o
.S
•i
•S
«s W*
§§
-~
K t ^
flo
ast
-------
tr,
PM10, PM25 and PM
Mean (Range) Lev
Reported**
? a
5n 3
^ H S
1 a t
N» CS
£••
o 3
a i!
^"^ JM **L-
S3 Hi "^
§ S
£5
'S tS
*
0 _ =
£h 2
\_^ . —
g "Sb
<3 "*"
* *
•§" 5
1°
Original study*
Study Location
Reanalysis study
P£> /•— \ ^"S
00 |> 00
cs vS r4 wS
oo oo oo oo
O "1 O ""!
^^ ^S^ wr* i_y
i i
: :
? ' ?
i? ^
—< £5 •' *O§H?
•3r "E . ^ ® , * « -^
p§2 #x2 pxiS.**
| j a « w a -g a a |
O -S »5 ^ w5 xS S1
rf
PM,a median 22. 1(0
159.0) (north site)
|
i
R
2
i-i
o
»— *
not used
2j
o
Vedal et al., 1998
Port Alberni, CAN
(children, cough)
O)
fe
I
-------
S
1
o
S I
¥ fc^
g
s
^ "53
8
O
etal.,1998
C
•8
§
1
o
W °^
I, J
^ a
'0 «
nose
8
o
etal.,1998
Alberni, C
(S
•-Tn
TO £*
I?
a 55
vq
s
8-
VO
O
998
i, C
^1
tJ 5
^o 1$
P
ts
s"
o o
l-
•dl"?
"S % S
o
•rfJ^
is < S
ill
r> ,o.
s
I
O
8
Ov
•3*
-s
I
>
f
C
•<] g
Jl
»:
I
*1
B§
a
J!
I
OJ
r*
i—i
<:
-------
S
nJ !
Q
£
8.
^ •< K
Ml
.3'
i
i
tudies in italics
mean (minimu
a
oo
I
b
-------
APPENDIX 3B. Mortality and Morbidity Effect Estimates and PM Concentrations
from U.S. and Canadian Studies for Long-Term Exposures to PM10, PM2 s, and PM10 2 s
Study
Indicator (Increment)
Relative Risk (95% CI)
Study
Concentrations
(us/m3)
Increased Total Mortality in Adults
Six City*
Six City8
ACS Studyc
(151 U.S. SMSA)
Six City ReanalysisD
ACS Study ReanalysisD
ACS Study Extended
Analyses5
Southern California1*
(
Southern California8
Veterans Cohort0
PM15/10(20ng/m3)
PMj.5 (10 fig/m3) .
SC-: (15 Mg/m3)
PMIMJ (10 Mg/m3)
PMj ., (10 Mg/m3)
SOJ (15 ug/m3)
PM15/10(20Mg/m3)
PMj., (10 Mg/m3)
PM15/1()(20ug/m3)(dichot)
FMu (10 Mg/m3)
PM^OOug/m3)
PM25 (10 ug/m3) (1979-83)
PMj5 (10 Mg/m3) (1999-00)
PMj5 (10 Mg/m3) (average)
PM10(20ug/m3)
PM10 (30 days/year>100 ug/m3)
PMIO (20 ug/m3)
PM10 (30 days/year>100 Mg/m3)
PH., (10 Mg/m3)
PM1M.5(10Mg/m3)
PM2.5 (10 Mg/m3) (1979-81)
1.18(1.06,1.32)
1.13(1.04,1.23)
1.54(1.15,2.07)
1.43(0.83,2.48)
1.07(1.04,1.10)
1.11(1.06,1.16)
1.19(1.06,1.34)
1.14(1.05,1.23)
1.04(1.01,1.07)
1.07(1.04, 1.10)
1.00(0.99,1.02)
1.04(1.01,1.08)
1.06(1.02,1.10)
1.06(1.02,1.11)
1.09(0.99, 1.21) (males)
1.08 (1.01, 1.16) (males)
0.95(0.87, 1.03) (females)
0.96(0.90, 1.02) (females)
1.09(0.98, 1.21) (males)
1.05(0.92, 1.21) (males)
0.90 (0.85, 0.95) (males)
NR(18,47)
MR (11, 30)
NR(5, 13)
18U(9, 34)
llu(4,24)
NR(18,47)
NR(11,30)
59(34,101)
20 (10, 38)
7.1(9,42)
21 (9, 34)
14(5,20)
18(7.5,30)
51 (0, 84)
51 (0, 84)
32 (17, 45)
27 (4, 44) '
24 (6, 42)
Increased Cardiopulmonary Mortality in Adults
Six City*
Six City ReanalysisD
ACS Studyc
ACS Study Reanalysis0
ACS Study Extended
Analyses8
PM15/10(20Mg/m3)
PMj5 (10 Mg/m3)
PM15/10(20(xg/m3)
PM2.5 (10 Mg/m3)
PMu (10 Mg/m3)
PM15/lc(20Mg/m3)(dichot)
PM2.5(10Mg/m3)
PM15.2.5(10Mg/m3)
PMJS (10 Mg/m3) (1979-83)
PM2 5 (10 Mg/m3) (1999-00)
PMj5 (10 (ig/m3) (average)
V
1.18(1.06,1.32)
1.20(1.03,1.41)
1.19(1.07,1.33)
1.12(1.07,1.17)
1.07(1.03,1.12)
1.12(1.07,1.17)
1.00(0.98,1.03)
1.06(1.02,1.10)
1.08(1.02,1.14)
1.09(1.03,1.16)
MR (18, 47)
NR(11,30)
NR(18,47)
NR(11,30)
18U(9, 34)
59 (34, 101)
20 (10, 38)
7.1(9,42)
21 (9, 34)
14(5,20)
18(7.5,30)
January 2005
3B-1
Draft - Do Not Cite or Quote
-------
Study
Southern CalifomiaF
Southern California15
Increased Lung Cancer
Six City*
Six City ReanalysisD
ACS Study0
ACS Study Reanalysis"
ACS Study Extended
Analyses6
Southern California'
Southern California11
Increased Bronchitis in
Six City1
24 CityJ
AHSMOG*
12 Southern California
communities1"
(all children)
12 Southern California
communities"
fchildren with asthma)
Indicator (Increment)
PM,o (20 ug/m3)
PMjiOOMg/m3)
PM1M.5(10Mg/m3)
Mortality in Adults
PMMO(20Mg/m3)
PM2.5(10Mg/m3)
PM1J/10(20Mg/m3)
PMj.5 (10 Mg/m3)
PM2.j (10 Mg/m3)
PHs/jo (20 ug/m3) (dichot)
PMj3 (10 Mg/m3)
PM15.2.5 (10 Mg/m3)
PM2 5 (10 Mg/m3) (1979-83)
PMj 5 (10 Mg/m3) (1999-00)
PM35 (10 Mg/m3) (average)
PM10 (20 Mg/m3)
PM-j, (10 Mg/m3)
-
Children
PM15/10(20Mg/m3)
PM2.5(10Mg/m3)
SOI (15 Mg/m3)
PMj.OOMg/m3)
PM10(20Mg/m3)
SOT (15 Mg/m3)
PM10(20Mg/m3)
(1986-1 990 data)
PM,o (20 Mg/m3)
PMj.5 (10 Mg/m3)
Relative Risk (95% CI)
1.01 (0.92, 1.10)
1.23(0.97, 1.55) (males)
1.20(0.87, 1.64) (males)
****
1.18(0.89,1.57)
1.14(0.75,1.74)
1.21(0.92,1.60)
1.01(0.91,1.12)
1.01(0.91,1.11)
1.01(0.91,1.11)
0.99(0.93,1.05)
1.08(1.01,1.16)
1.13(1.04,1.22)
1.14(1.05,1.24)
1.81(1.14, 2.86) (males)
1.39(0.79, 2.50) (males)
1.26 (0.62, 2.55) (males)
1.6(1.1,2.5)
1.3(0.9,2.0)
3.02(1.28,7.03)
1.31(0.94,1.84)
1.60(0.92,2.78)
1.39(0.99,1.92)
0.95(0.79,1.15)
1.4(1.1,1.8)
1.3(0.9,1.7),
Study
Concentrations
(Jis/m3)
51 (0, 84)
32 (17, 45)
27 (4, 44)
NR(18,47)
NR(11,30)
NR(18,47)
, NR(11,30)
18" (9, 34)
59(34,101)
20 (10, 38)
7.1(9,42)
21 (9, 34)
14(5,20)
18(7.5,30)
51 (0, 84)
32 (17, 45)
27(4,44)
NR(20,59)
NR(12,37)
4.7(0.7,7.4)
14.5(5.8,20.7)
23.8(15.4,32.7)
—
NR (28.0, 84.9)
34.8 (13.0, 70.7)
15.3(6.7,31.5)
Increased Cough in Children
12 Southern California
communitiesL
(all children)
12 Southern California
communities**
(children with asthma)
PM)0 (20 Mg/m3)
(1986- 1990 data)
PM10 (20 Mg/m3)
PM, .5 (10 Mg/m3)
1.05(0.94,1.16)
1.1(0.7,1,8)
1.2(0.8,1.8)
NR (28.0, 84.9)
13.0-.70.7
6.7-31.5
January 2005
3B-2
Draft - Do Not Cite or Quote
-------
Study
Indicator (Increment)
Relative Risk (95% CI)
Study
Concentrations
Cue/in3)
Increased Airway Obstruction in Adults
AHSMOGK
PM10(20ng/m3)
1.19(0.84,1.68)
NR
Decreased Lung Function in Children
Six City1
24 City3
12 Southern California
communities1"
(all children)
12 Southern California
communities11
(all children)
12 Southern California
communities'5
(4th grade cohort)
12 Southern California
communities'5
(4* grade cohort)
12 Southern California
communitiesR
(second 4* grade
cohort)
12 Southern California
communities11
(second 4* grade
cohort)
12 Southern California
communities11
(second 4* grade
cohort)
PM15/10(50ng/m3)
S0:(15ng/m3)
PMj L (10 ng/m3)
PM10 (20 ng/m3)
PM10(20ng/m3)
(1986-90 data)
PM10 (20 ng/m3)
(1986-1 990 data)
PM10(20ng/m3)
PM^aOug/m3)
PM1M.5(10ng/m3)
PM10(20ug/m3)
PMjsClOug/m3)
PMuwj (10 Mg/m3) -
PM10(20ug/m3)
PM^OOug/m3)
PM10(20ng/m3)
PM^s (10 ng/m3)
PM10(20|ig/m3)
PMjj (10 ng/m3)
NS Changes
-6.56% (-9.64, -3.43) FVC
-2.15% (-3.34, -0.95) FVC
-2.80% (-4.97, -0.59) FVC
-19.9 (-37.8, -2.6) FVC
-25.6 (-47.1, -5.1) MMEF
-0.23 (-0.44, -0.01) FVC %
growth
-0.1 8 (-0.36, 0.0) FVC %
growth
-0.22 (-0.47, 0.02) FVC %
growth
-0.51 (-0.94, -0.08) MMEF %
growth
•i - 0.4 (-0.75, -0.04) MMEF %
growth
-0.54 (-1 .0, -0.06) MMEF %
growth
/
-0.1 2 (-0.26, 0.24) FVC %
growth
-0.06 (-0.30, 0.18) FVC %
growth
-0.26 (-0.75, 0.23) MMEF %
growth
-0.42 (-0.84, 0.0) MMEF %
growth
-0.16 (-0.62, 0.30) PEFR %
growth
-0.20 (-0.64, 0.25) PEFR %
growth
NR(20,59)
4.7 (0.7, 7.4)
14.5(5.8,20.7)
23.8(15.4,32,7)
NR (28.0, 84.9)
NR (28.0, 84.9)
NR(15,70)X
NR(10, 35)x
NR
*
NR(15,70)X
NR(10,35)X
NR
NR(10,80)Y
NR(5,30)Y
NR(10,80)Y
NR(5,30)Y
NR(10,80)Y
NR(5,30)Y
January 2005
3B-3
Draft - Do Not Cite or Quote
-------
Study
12 Southern California
communities8
12 Southern California
communities*
12 Southern California
communities8
Indicator (Increment)
PM10 (20 ug/m3)
PM10(20ug/m3)
PM10 (20 Mg/m3)
Relative Risk (95% CI)
-3.6 (-18, 11) FVC growth
-33 (-64, -2.2) MMEF growth
-70 (-120, -20) PEFR growth
Study
Concentrations
(ue/m3)
NR (15.0, 66.2)
NR (15.0, 66.2)
NR (15.0, 66.2)
Lung Function Changes in Adults
AHSMOGT(%
predicted FEV,,
females)
AHSMOGT
(% predicted FEV,,
males)
AHSMOGT
(% predicted FEV,,
males whose parents
had asthma, bronchitis,
emphysema)
AHSMOG7
(% predicted FEV^
males)
PM,0 (cutoff of 54.2 days/year
>100 ug/m3)
PM,0 (cutoff of 54.2 days/year
>100ug/m3)
PM,0 (cutoff of 54.2 days/year
>1 00 ug/m3)
SO; (1.6 ug/m3)
+0.9% (-0.8, 2.5) FEV,
+0.3 % (-2.2, 2.8) FEV,
-7.2% (-11.5, -2.7) FEV,
-1.5%(-2.9,-0.1)FEV,
52.7(21.3,80.6)
54.1(20.0,80.6)
54.1(20.0,80.6)
7.3 (2.0, 10.1)
References:
ADockeryetal.(1993)
BEPA(1996a)
c Pope etal. (1995)
DKrewskietal. (2000)
E Pope etal. (2002)
F Abbey etal. (1999)
°Lipfertetal. (2000b)
* McDonnell et al. (2000)
'Dockery etal. (1989)
'Dockery etal. (1996)
K Abbey etal. (1995a,b,c)
L Peters etal. (1999a)
"McConnelletal. (1999)
NBerglundetal. (1999)
°Raizenne etal. (1996)
pPeters etal. (1999)
Q Gauderman et al. (2000)
R Gauderman et al. (2002)
sAvoletal. (2001)
T Abbey etal. (1998)
Note: Study concentrations are presented as mean (min, max), or mean (±SD); NS Changes = No significant changes
(no quantitative results reported); NR=not reported.
"Median
v Results only for smoking category subgroups.
x Estimated from Figure 1, Gauderman et al. (2000) ,
Y Estimated from figures available in online data supplement to Gauderman et al. (2002)
January 2005
3B-4
Draft - Do Not Cite or Quote
-------
APPENDIX 4A
Study-Specific Information on Short- and Long-term Exposure
Studies in Cities included in PM2 5 Assessment and on Short-term
Exposure Studies in Cities included in PM10.2.5 Assessment
January 2005
4A-i
Draft - Do Not Quote or Cite
-------
I
!
§r
^1
Ii
1
8
p
o
8
o
s
I
!
i o
! D>
'. <*>.
=
S.
o
I
9
IS
I
I
> CO
i -5
°
If
P •*'
8
5
i?
s
s-
•o
o
is
£
-------
II
l
«>. it
CM o
o"
-CO
E
3 O
••9
f
Ill
•fe
jfl
£ S
•o
i
Dl
ISA
fliiill
11
,
<
I
P
£
I
.e
sr
s
o'
9
N.
I
o
c
» 1
ft:
o
,
o
Q_
09
<1J
IT
•5:=:
I
2
>,
ra
•?
s-
T3
i
Q_
sr- C
CO c
5? (D
H £
•^ Q.
_£$
g
d
nj
>s
ra
73
CO
ID
00
6i< £
o n w
S
a
a.
o
o
•5=:
f ^3
'« §
>\ j\|
"
9
8
d
>,
f
&
-D
CM
i I
II
!
ra '*-'
c —
ro CD
loJ
•"*• C
Cv c
il
O
TT
-
l'
a
O
N
-s
8
6
,
CD
•o
CN
<:
ra
o
ca •-"
c
(0
E
10 "—
C _:
ra ra
i
«•!
-------
I
I
,
a.
n
i
0.
ill
I
is
o '-
•3 ff
§> ^
s= I?
<
C3 r.
8
d
s
§
9
S
O
•J
t
i
is
o *-
I S
61 £
£ &
s" s
"^ o
IPS:
S
s-
T3
O
1
S
S
d
II
"^ O
I i
8
d
ii *-
"I?
t
5 o
-------
a
Is
3 g
i
o.
f
° s
f
1'
S
1
o
0.-.
o
o
s-
•o
o
P.
8
o
3
s
s
o
g8
S
S.
j_ O
II-
tr
9
8
1
8
il
11=
8
9
S
9
S
§
9
s-
8
(3 .
i
o
•|
-o
t
o
o
S Q
< s
o ^
S
S
if!
o
b
8
6
9
5
o
?
W
f
I
8
t
18
5-
S
1
o o
I?
iti
00
"^ a
s-s
1
a
-------
I
2
:
I,
S 6
t
3
S
§
o n
s
S
s
o
t
8.
ill
s
6
s-
5
s
s
o
s
d
t
a*
is"?
s
8
o
t
!$
ll
if
t
58
-S
d
s-
XI
o
8
f!
s
6
fi
II
o
b
s-
•o
o
S
g
m o
^j*- •*—
^ o
9
K
18
o ^l
ii
if-
8
o
I
I
8
o
I
t
,
18
s
§
I
>.
I
O
I*
II
-------
I
s-
•o .
_
tj
i
o
0.
i
I
I
55
I
*
;8
d
Ife
L;
I-B
8
a
IT
*
•5
a !2
* 2.—
tff
CN
d
IS
9
R
§
f
8
8
9
8
b
•*-" "p"
•H) Q)
= £
ii
at
s
i
S
III
-------
<
a.
_2
a.
a.
cu
x
UJ
E
I
*j
o
V)
1
1
i
ig
5 u
II
c ,s
«SE
UI
ra
I
1
(0
s
S
TJ
3
(0
Q.
0)
X
UJ
o
V)
£
o
•a
I
in*
I
n
1}
i
0,
a;
X •
ui
ui
•S
o
8
9
o>
3
(0
TO
O
is i
5 2.
0)
ID
O
10
c
10
8
3
o
6
6> o
i-8
O O
01
c c
ra a
c c
a) ffl
•O -n
88
11
22
"8 "5
1
w
I
o
r
sis
8"
in
iS
o
§
R
E
fc
8
9"
1
ra
•D
O
M
-------
O
II
o
od
p
d
o>
ra
S
0)
C
o
I
S
I
82
o ~—'
CM >.
— 4,
a-s
1.2
n
in
8
9
D>
n
s-
•o
S
C
o
ra ^
u <£,
= s|
§"0 2
o
S
n
82
o ^—
n ^-
U. 0
2
w
q
d
0)
C
o
il
V
a:
ra
re
(2
ley
(O
g
8
d
8
9
8
o>
ra
*
•o
o
S
CM
O
C
o
linear,
M
ngent
ill
i
ra
O
in
'9!
.a-
a
c
(D
ey (2003)
irley(19£
3
8
d
8
I
p
d
D)
ra
ra
•o
o
in
o
L-- £
IS"
Jo
§
10
'55
m
2S
« o>
8^
CM >.
|i
£ u.
.J-s
to
S
8
d
p
d
m
o
in
o
8
in
"35
irley (2003)
Failley (19
o
p
d
p
d
O)
>>
-S-
o
in
o
(O
O
8
1
o>
•D
8
m
(0
'55
"ra
c
m
8?
s —•
CJ-ST
tl
II
U
S
u
111
£
in
csi
o
c
ls£l
5
;>.
tfi O)
111
Q.
-------
5
n.
a.
3 I
i'
!!.
O c =
s
O
i
8
d
I
I
(M
II
01
6
• s c
6.1 £
°<3J2
g
05
V
1
i
ill
o
3
d
f
ill
8
1
o
?• >— 1
£ CO ^
^ 8.-
•2
7
CM
1.1
Q
Q.
8
q
6
g
d
a-
•o
o
en
d
.1
I!
a '»
£• ™ ^™
_ s ff
3
d
a
I
TJ
O
c
'•j
O J2
I
a.
11
A M
E "5 ;
II!
o
q
6
s-
73
O
S •£
c «
.£ ra
Q
Q.
8
5T «
315
|Il
III
6
I
O
O:
8
n-5
||
Is
s-s
= •§
ll§
HI
6
s
I
I
o
I
a
I
i
n
Q
5
o>
q
d
I
I
s-
•o
i
u
I
f
1
•5.
s
6
§
9
I
r
•o
C'
o
i
i (D
3
^
o
o
s
q
o
•v
1
T
*-
tv
A
1
%
J
§
?
(0
(O
I
-------
(A
£
•o
10
cv
I
IA
O
Q.
fi
O)
O
.Q
n
in
S
•f
Q!"
t^-
^~
^
o
^J-
CD
C
E
I
o
E
re
0
t
g"
O T3
CM £
"— -o
__.* ^
(0
00
CO
S
d
CD
C-.
§
d
annual
mean
1
00
CO
o
T~
O
O
(0
(U
c
1
^>
5
CO
_
0*
$
re
u
5
o"
o
§.
"re
K
'x.
1 w
£ o
* <
s
T-
o
d
JQ
S
o
d
CM
00
o
o
d
H ^
2 co
c «
co E
JO
c
CO
CO
0
1
^_
re
CD
c
1)
G
A
CO
^
re
(U
(0
^
8
<
4
o"
CM,
"re
0)
Jx:
S w
*U
o
o
d
O)
CO
CO
S
d
CO
§
d
1 c
5 re
c £
re E
re
"c
00
co
o
T—
' CO
0
k_
(0
CO
C
^
O
******"
o
CO
—
re
CD
o>
3
re
u
_ ?
g
o
£L
"re
V
!x
is
* <
en
o
d
en
0
CM
S
°
^
CN
T-
S
d
1 c
2 co
El
$
CO
CO
o
CM
O
CO
CO
CD
c
•1
_o
o
CO
^
OJ
§
CO
o
5
o*
o
8
"re
oi
^
5 w
2 o
* <
fe
I
Itvl
o
o
5
-------
10
s
a.
£
c
o
IS
o
E
u
a>
a.
(0
<0
s
II
3 CO
s
ii
UJ
1
CD
g
O
o
o
CO
CO
o
g
9
CO
CO
t
S
E.
Q>
CO CO
§ O
S.g
CO?
Ill
IO
to
o
o
o
g
to
CN
CO
o
o
O)
t
co
s
log-linear,
GAM
(stringent
m
to
o
0.
CO
-*-l
V
CO C
III
CO
o
o
in
t*-
i
o
o
d
in
en
oo
o
o
•a
CN
O
IO
u co
•= CO
t 40
co =g
"5 ti
U) CO
—
i
I
o
o.
in
V
c
o
c
ll
.s
"S co
So
x-s (TJ
si,
S;
o i
in
o
o
CM
CO
§
9'
CN
O
o
fr.
2
8?
T3
O
O
in
co
c
o
c
~ S
A
CD
CO
CN
> £
11
c ro
O CO
0 £
(A
'S^
"co to
s-g
o s
§
Ol
a
-------
:I
m
o &
E o
£ o
8-
S-:
Ul
u>
I
co"
n
o>
(O
_c
n
.S
•«rf
to
E
§
0
i
|
CO
u
E
o
0)
a.
to
i
S
o
o
o
§
o
o
O)
I
00
03
0)
c
log-li
GAM
(stri
co
O)
Asthm
i
Q.
__^ CD
*"*"(§
i^
ISS
Q. TO -—
«" TO "CB
£ (D j_,
OT ^-a>
Sh
£.•5
p?0
§,£?
1?
S.«
SITI
2 CM
to
|
w
co"
0. B
0. 3
O. O
3 m
o °
-J CO
S
u
X
111
Ol
|
£ i
0 §
o
li
u
T3
i
55
p
o
S
in
S
p
9
in
oo
eo
CD
q
o
S
o
CN
Is
£?
S
00
8
CO
p
9
CO
§
O)
g
s-
t
CM
in
CN
CO
"o>
I
tl
1
S?
N CO
CD t
OT CN
in
CM
(0
'Si
1
T3
(D 'o
£J CD
CO CM
I
Crt
0
I
13
I
I
'&
CN
-------
APPENDIX 4B
Sensitivity Analyses: Estimated PM-Related Incidence Associated
with Short- and Long-term Exposure to PM2 5 and Short-term
Exposure to PM,0_2 5
January 2005 4B-i Draft - Do Not Quote or Cite
-------
f:l
& O
IP
|Ii
s
»
t,
I'
£
1
13
•a,?
S
il?
iss;
;? i
R
;*i
§
§ §
o o
S «i
ll
II
I-
1!
as
II
tl
il «
lit
Ul M
1*
;?8
"-IS
!
s
-S
Standards
it
i
o
a
-------
8
i
I
g
it
ai.
•6 S
I!
.f
!1! I
II I
-.3 &
I;
««
tt <
f!!
-------
QJ
Q
fe
aj (0
•S «•
.s m
S^
S 5
<•£
10
C 10
"° •§ *
111
I wo
(o a> S
ui -o
.2 c
o S;
>S
A 2
>
5 u>
gg
JlZ
«^s
I
., 10
la
^Sfi
(0
£
(0
U
ii
x °
Is
n w
E e
a) is
*^i o
t
S
2
a
£
5
1
re
ce Associ
§
2
u
c
1
jj
a
a.
00
le value ~~j
1
(U
Q.
1
o"
in
CN
^4
~—
sf
"~"
o"
to ^
00
3
3
ja>
1
&
i
o
in"
f*3
in
O
O '
o
v~ '
co.
m
9 rl
in
le value
§
1
O)
o-
co
10
0
O i
o
o"
O '
o.
So
°T
NS
le value
1
1
in"
CM
in
S
^4
• — •
CM ^
~~"
o"
in rl
00
ile value
•E
1
i
CD
in"
CD
in
£
^4
— *
0>
""""
o"
a?
in ,-
00
ile value
1
1
i
CO
o"
in
o"
CO
in"
CD ^
^ CM
o
~~"
DO
10 f2
0)
75
.S>
I
"c
S
a5
a.
1
in"
co
^r
Of
te
m
I
I
-------
0
b
1
2
Q.
H
s
l»
<
s
e
«
^«
•B
1
1
CL
to
^n
I
fj
1
(0
1
I
<
v
73
n
il
1~°
l«
o
|
•o
0
in
e!
*g
» 2
91 TT
€ "
o
Q.
X
4mt
S £
UJ % ^"1
,
S
"Bi
3
S
D
8
&
s
0)
o"
CO
•<>•
o
O i
o
o1
O i
o.
sr
^ in
fc°?
NR
le value
•E
?
o>
a.
CO
in"
CM
*a-
o
0 i
0
s
£ '.
CO
CO
in
*~
•rt- CO
CO CO
CM
S
to
0)
1
8
8.
of •
O)
o"
•^
CO
0
O i
0
M-
CD
^ *~
CO I
CO
CO
in"
a-?
"8
o>
3
§
0
•*j-
CO
0
0 i
o
S
& ',
co
CO
in
S5?
°S
le value
1
o>
a.
CO
ui"
CO
CO
o
O i
0
•*r
CD
£ '•
CO
CO
in"
TJ- CD
CO co
CM
0)
15
.a>
-s
P
CO
a.
£ •
o"
CO
CO
0
O '
0
o"
0 i
0
sr
fi»
"B
le value
•E
8
0)
Q.
.C
(S>
o>
in"
CM
CO
o
0 i
o
—
O i
0
sr
-1?
"si
e value
<=
1
8.
1
o" •
T
CM '
o
1
t
m
I
i
-------
$
U)
^
a
£
i
1
ce Asso
i
"5
c
1
i
a
a.
Standards
>
c
(Q
£
<
•o
0
Jm
vt
1 "
1- •§
13 A
O -.
v TT
£ "
Q.
X
s
o
£
II)
fi «
it
I "
gj
X
« c
g —
Q (U M
7^ S ^ ,E
111 9 D>
il~2
i" a "
CO 0)
sr
•~
3,
'ra
a
t
1
15
a
c
e
O
O i
2.
O i
2
^
O)
^- ^-
™$
•3
ra
o
•B
1
S.
1
in
CO
CM
o
0 i
0
0 i
o_
ifm^
en
T~ ^r
"i
0>
3
§
.2
a)
Q.
(-j-
00
CM
0
O i
o
0 I
.•—•.
00
O -*J-
^s
0)
"re
_0)
5
Q.
i
0)
LO'
CM
CM
O
0 i
O_
O t
^_^
O)
"*" "^
Ra
(U
75
J!J
"E
I
cl
0)
o"
CM
0
O i
Q
O 1
^ ^
o>
T™ ^
^S
O)
>
.2
^
8
d>
a.
O5
m"
CO
CM
0
O i
o.
T—
O i
0
^— ^
O)
T— ^J-
^g
o"
CO
CM
O
0 i
0
o"
0 •
/-•-.
00
t; fO
04 CM
ID
•i
1
1
.c
cr>
in"
CM
CM
•D
£
"c
i
'S
f
S.
o
•lofthesli
S
a.
™
£
jj
8!
«
1
i
x^
"5
tg
I
•g
o
u
i
M
1
j
•D O
I
o
1
in
03
-------
a
01
13
If
I- m
1 O
£
c
il
ra S!
•£
.£>:
m
o
S
*
2||
I ||l
|2 uj x 3
*l
•a a
£
•a
tfi»
Study
,f
8
8
s
r.
in
S0?
8
-------
I-
a
til
I
s
HI
I
to ^
!
e 'i
£
li
« 8
1!
-------
I
Jra
01
V II
s?
S
-------
O
fe
«J
t
|
*8
«
g
S
s
1
D.
Ill
i
£
i
I
!H
£ N
£ ^r
in
£ "i 9!
a. e
£H IT
ji'I
O
i
>
s
f
1
1
s
O i
S
O i
a
s
O i
a
intile value
A
a
si
i
m
*
o
o >
a
s
O i
a
s
O i
a
intile value
P
a
£
§
S
,
s
o >
jjT
Pm *"
N !(!
s.
§
n J:
5 <
'i
ntile value
«
S
i
§
fO
S
O i
a
.
pi. 1
N 1ft
ffi
n ^
O 1
*• «o
ntile value
8
S
8
in
n
S
o •
P,
p
18 T
£i
i
ntile value
8
8
i
n
<•>
S
o •
P.
S
O i
a
is
U>
S
o <
P.
S
O i
a
0
o •
e
mtile value
P
a
£
s
s
D
s
O i
P.
S
o •
a
o
S;
5
I
8
a
i
o
CM
S
o >
P,
S
O '
a
0
"§
ntile value
8
a
i
Ul
n
a
o
O i
3
intile value
?
a
i
o
CM
O
O •
O
O i
e
o
O i
e
mtile value
P
a
£
s
in
CM
fr£
¥ f:
-------
II
i=O
II
>
I E
'S «) «.,
.
o s .a
"3 -o -g
ilt
i!
.3
8
a
i*i
•s .
a
s
s
,ST
©
ss
Ei
§t
Ei
«s
&'
CJ
a
s
rRT
Ma
:,S
2 . !
'fiR
o
5
CVI ,
5
1
I
.2
8.
8
ff
1
S
a
fi
S
1
a>
I
1
S
§
£
s.
s
tf
-------
§
s
0
1
00 t*l <» (N
I
ical
a
£
SS
Ci
S
d
a
3?
! TI
to
Ci
S
"
°T
S
ci
Is
JS
II
s
!S«
s
I
£
I
5
I
i
s"
u
I
s
?
tf
xf
I!
•I
I E
i!
II
1
-------
§0)
III c
£. °3»
»2
§1
_l «
1 U
•a «i
S ia
•5 *
ro oj
v
«£
<£
£tf
i s
il
11
SI
!3*w
ill
g W 0
iS S s
^S 2
«! 5
p*
lit
** S«
fs'gg
S
[O (o
J
ss
1
n
>
i
a.
01
S
in
s
O i
a
5?
"S.
s
^
5
J
1
|
0)
§
in
S
O i
S
U)
N
ST
«
• ?
"I
1
i
Q.
n
!!?
in
S
O i
S
S
o >
8
S *?
^s
0)
g—
0>
&
£
§
3
u>
s
O i
s
s
o •
o
§
ST
§
2~~
i •
|
S
a
m
0
0 >
a
ftf
"•l
UJ •
5
1
s
a
c
I
01
iS
in
S
O •
S
s-
0 i
a
5
ST
§
!
i
Q.
S
§
M)
8
o •
a
s
o •
a
o
o •
9
•
o>
^
^
ffi
0>
s
in
S
O '
a
s
O i
a
s
O i
a
!
i
m
a>
S
in
S
o <
a
s
O i
S
S
0 i
a
!
i
a.
a>
S
u>
E
_
|
S
s
*
s
O '
a
s
O '
a
f
ST
O)
a
a>
J
a.
f
00
O)
in
«
^
S
o <
a
s
O I
e
5
ST
§
D
1
^
a.
i
s>
S
•»
-
-------
Of
*1
£
•o
JS
8
S
«
u
di
I
|
1
n
n
»
>
1
S3 <3
•s*l
£ M a
Ti
Hypothetic
Threshold
o10 M9'm3
ii
BASE CASE: Lov
Measured Level
Sbdy
=7.5 ug/ml
%
2?
I
|
i
1
S
O i
a
s
a >
8
S
0 <
a
1
M
C
1
S
B
T
&
O •
a.
s
0 >
a
s
o •
a
1
n
>
c
a
&
%
*
s-
o •
a
s
0 i
a
s
o •
S
|
1
8.
&
U)
C4
V
S
0 <
a
6*
^s
»&
fi|
S
5
_£
^S
<1>
&
i
§
CO
s
O '
a
8
PT
S
o> «
!
1
s
s
CO
s
o >
a
s
O i
S
5
r~
S
01
3
5
0)
:s
i
0.
i
s
CO
0
O i
a.
S1
0 <
g.
f
ST
O)
Ci
0)
i
j
c
i
a.
JZ
i
in
CM
CO
s
o >
a
8"
0 i
8
o
ST
o
a
J
or
>
s
c
k
&
s
n
S
0 i
a
o
o •
e
s
o <
a
J
5
c
Q.
S
s
CO
s
O i
a
0
O '
S
S
o •
o
d>
*
o.
§!
S
CO
S"
o •
a
c?
O i
a
s
0 •
3
1
C
1
£
in
CM
CO
S
o •
a
to
£T
CO
d
s
8*
CMg
0)
>
J
Q.
£
§
§
CM
S
(3 •
a
s
i-. »-
(O i
S
to
a^
•£
O
Q.
£
§
S
CM
S
O i
a
s
O 1
o
§
ST
§
!
m
09
A
£
S
f*
S
o <
o
0
O i
a
s
o •
a
J
CD
>
1
e
1
&
8
CM
O
0 •
O
O
O i
a
0
O i
8
0)
^
5
5
c
0.
£
tn
(N
CM
0> S 'Q
5 0.-
f*i
S a£
s*«
*f S
Is*
* *s
s i
n. *
B S
**§
li
€
8
|
*fl
-------
1
o
1
5*
Si
2
CO
1
CN
i?
Ml .£
£3
*1
CO «
SO
•n »
* 2
IS
8£
»*
< =
t!
i =
II
Sin
E
^5
in
K.
0.
I
.I
Si
S
2
CN
1
I
!i
I!
as v> g
= sl
OJ ^ 1™
wQ-P
. a *
I*
£ffl
§
CM
2
£~
S1
2
W
•i
I
I
I
1
8.
>
I
2
I
S
s
S.
£
|
S
$
6
S
f
£
I
I
§
To
-------
II
Sgtf
l!
-8
551
Si
a
1
S
1
Is
Ill
il*
a
Ill
f is
o
fc
o
I
a
-------
. fe
a
1
£
<5
•g
•
i!
o
g
• "
.2 m
8 «
ll
II
t
JC
s
I I (
s
s
o
i i
o
o
I 1
o
8
!T«
5
9
o
I I
S
,?!
S'
i
j
ii
a
i
S
a
ST
ST
Cl
> ^ «
l§li
w p i
£e!
5 2>
5
I
J
Ji
o
2
g
«
x = ±;
iu < n.
-------
II
O
a
£
S'- C
I U
s
8
T «
5
o
i C
o
i'
So
s
o
I i C
a
o
) i C
a
if!
;s •»
R
1?«
'o '
»>
f i
111
* S-ri
£
I
S3 8
ST '
°1 |
si*!
C9| ji
fill
-------
W X >. *J
h- 111 X (0
ed
2
II
- e
II
si
o> T
CO
8
S
m
0.
«
•f
in
in
8?
ID •
CM
1
0)
¥
R
S"
3.
in
<»
f
S
5"
in
m
d
O)
'S
in
$
m
O '
CO
R?
00?
in
GO
00
i i
Ci
to
I
d
in
R1?
in
in
,-?
CM
S
1C 7
J
S.
-------
8
II
CO
CM i
CO
O i
CN
CN
CM-
CM"
CM
O
|
a
&>
c4
c
w
II
O) ,
CO
to
in
s-
CO
ro
v
en
CO '
co
sa
c.
G.
"
G.
co-
in >
G
ffl T
Gi
CO i
a.
a?
t.
a?
00?
•—
s
8
'£
CM
^
HZ
§
jj
I
o>
Q.
3
S
I
a.
i
i
4
CO
R
CO
IR
8
(N
CM
CM
(N
I
-------
a
f!
o ^
j g
11
Si
11
£2
a
iy
.B
w
l!i
lil
w ? P
« So 3
j a a. J
a x >> «J
I- ui X M
ii
! £
11
£3.
0.7
s
*i
R?
s
55
3
a i
S
t 00
'is
) CO
.te.
§
s
s
-------
a
*:
Ss
Is
£3
t
87
S
§
*
,_ r-
s.
ST
S
in
a> 'V
G
S
s"?
s
.S
tf
.0
p .
i«S i
Hj 1
I 12 3
!!! !
»'!.!
s I
»«!•
!!!
it
if
' i
f1
S^
•I
111
-------
i
£
.2 «„
,>
w ^
"« S
«<
If
Q (A
•E 3
II
O V
13
oi O
ll
£ Q>
w £
!--
|i
< 3
ra
i E
1
^ w
JJ i. I
K>£ j« 7.
£?!
*• o o — -
=20
thetical
shold
Ill
I|s
o S o
O.£ S
£t- II
&'
=5 S.
*-.$£
w
o I ^
111 5 «
a
in
ID
S
ID
o> o>
in oi
8
S
s
8*?
cS
8?
s.
I
i-
'ra
•o
co
2
i
3 daily
S
8
ID
SJ
n
1
S.
I
i-
n
T3
co
percentile
N
o> °>
8? o>
pe
§
0?
10 a)
a>
a.
i
.£•
'
-------
_
|
w
JC
4>^
i
i
.2 ..
o d)
8"i
»
u 2
< E
fi?S
co 3-
*^ ^
v o>
S. c
<5
ro TJ
l§
** ni
« Si
S^
S«
W (M
c J2
•S?
o 2
.2 £
E c
35 ,
15 S 2
fi *- TS
O. in >
83 >
~> =5
x g S
7n «ff Bf
IS ^L i£
3 10 ?
il §
< «J s
•o ? &•
ill — .^
15 i
c W "»
-3 g 9
rt .£ 5
UJ IS «
*" W 3
• 1. 2
§>l^ -
« ^ S S
5 S o s
< £ S $
>, > _ 5
f ^ ^ •*
'> «q "o a
4s CM £ g
S?805
0) ^ t O «
Table4B-12. S
Exposure to Pl\
Hypothetical Tf
Seattle, WA, 20
(As Is = 11.4 ug/m:
I
&>
2
i
•n
1
8
1
•g
c
1
1
E
S n
T
o
S
CM N
O
5?
N. <°
CM i
O
g
J i
I !
f
5>
«
<
8 j
". <
N
!;
? J
•t
CM
C
^
ID ^
g
JR
CM N
O
5?
&
o
01
1
s
= :
1
1
i I
I) C
>• .
TO
D 1
n (
1
gu
(j
r ^
CM
> e
in *7
S
co
CM
a)
i
s
5
i s
n (
X
S"
i
O (
1 •
rj t
a u
r -5i
CM
> C
£
0)
1
B
S.
§ ;
n i
>.
IB
D T
•5 C
1 -
n '
r c
^~
1 C
c
•* T
g
1)
1
:
S I
>,
i 7
3 1
O C
1
R f
> «-
o •
1 C
in
CM i
9
O
ai
3
5
2 ;
E
u
2
0)
a.
i !
^
n i
? '
I ;
5 i
^
CM i
C
S
in "7
o
CO
CM "
^ i
o
r ^
CM
C
if T
o
O3
CM CM
*" O
in
Kf
o
ij
2
>
u
1 •
s
8.
1 J
J. ^
1 '
T ^
= C
1 -
— <
2»
CM1?
O
n
3
5
z*
j
:
1;
o t
1
3 c
i> T
o
oo
CM CM
O
in
CM?
0
Ol
S
>
a>
1 r
5
01
x
c j
1 i
X
s r
n
a t
o f
1
R S
CSI '
o
s
=
n
>
III
1 ;
C J
i I
X
= r
in
a i
T t
I 1
3 £
r c
O '
, S
^
CM '
0
<0 7
o
a?
o
-n 5
1
Gf
2 ^1
o n
1 *
1 *?
j t
i «
3
i
3
e
S
1
c
e
&
S.
|
.a w
J 1
JQ .3=
•g S &
J f c 8 "
I S| N
I i i 2
i ** z
tils S
fill *
"K tj « §
« | 5 S
* 3 " i
•s s o j
(- O *3 .(5"
1 g I 1
s If 1
co (S u a
a S e 1
5 § 2 f
e 8 « u
w 5 w *l
S u fl> .12
•2 w g J
U N s 8
i el %
$ s - ~8
(0 Q. r 2
— •sis
sf 2 5 -°
S S» J9
s f I I
1 if 1
i s o s
Ol
3
1
(U
•j
u
^,
|
>>
S
3
1
*
fijs I
Sni ^
S '1 I
1 f*J
fill
1 Ii
§ s! $
"In the base case, incidei
models, incidence was qu;
Note: Numbers in parenth
January 2005
-------
>*i S
w <
&•§
fl
I
£
v
§1
si
8
So
85
85
85
So
s-
i!
"8
si
§1
85
a
SI
85
8*7
CJ
•I
i
5 .
a
1
-------
(0
x
10 -8 «»
111
*1 J
ill
Hi
UJ O M
§1°
£ 'fl 6>
•I c 2
|||
S '5 E
£sa
-
!S 2 o)
^1 8
L; * °
D>
2s? 2
^
o
<
»
•
ii
o >
i^1
1|
i!
4 R
Hypothetical
Threshold
=15 ug/mS
Threshold
=10 ug/mS
BASE CASE:
Dwest Measured
Level in Study
=4
:i|
g a
Thres
=15 |j
M P
I"
s S
1 «*
c &
a CB
•i
^
*«§
s
NO!
in <7
2.
§
a
<*>
CM 1
•4
8AS
r a?
S T ^
95
ff.
O) i O
"
1*
CO
"
X*?*
0 ' (0
TS
1
I
m
"
•S
•--•
00
"
S
S
P
B
O i r-
n
«M
o
f
S
in
*«*!
O ' ID
-go.
8"
Q
ID
a
t:
i
-------
II
II,
11
&
™ m S
f*
. *
,
'88
£*
I . O
n J. IJ
:M
ID . P
2.8
II
'88
*«
°E
r!JS?
9*
s -5
'aft
i *»
1 s^
111
li |i
MIS
i ii
» -• o i
•II*
Ms!
s * a
ll|
•3 *» « g «
ililU
tlnn
-------
si
«c
i!
II
r
i
11
II
lc
i
iif'
5 3 '
£ w •
ill
:•>-• T
I c
ui —' •
< 3
..«*
CM • 0
"i- e>
a"!
FT
§:
N. V- '
S T<
l~- ~ (
sr
i o>;
i i i
'5'
?*5>
s-
1 5 j
* - •
»
i ^ "
' T- CO
10 r
e ,
8
§"7«
»*
a-
Si
,8*
™
§"*
I ^ c»
's.Si
S"^
'a?
-------
K Ot
||
Ii
< g
|?
s!
if
ing Annual
of Monitor-
K
S
fe
o
1
1
li
s j?
IT
i* w v*
lli
|lf
1
IE
1
£
3
I
8
S
ii -,
,iit
ul f
t
3
i
<.
T-8?
°S§
B-^
fill
« i ^j
8,
sr
5?
en • ^
ffi-
;s§
°°
P.8
f*
S'
is
• 0
op
°°
a
c.
og
1
2
6
*
|
! !i
! SJ
* st-
!!
s?
III
ill
'I
?
S-S
ll
\\
h
si
l
f I
H
-------
l
— 3
I
00
I
-------