Neal Fann/RTP/USEPA/US

03/10/2008 12:18 PM

	

To

"Schwab, Margo" <Margo_Schwab@omb.eop.gov>

cc

Bryan Hubbell/RTP/USEPA/US@EPA, "King, Heidi R."
<Heidi_R._King@omb.eop.gov>, Lydia Wegman/RTP/USEPA/US@EPA, Ron
Evans/RTP/USEPA/US@EPA, Tricia Crabtree/RTP/USEPA/US@EPA, Nathalie
Simon/DC/USEPA/US@EPA

Subject

RE: Ron, can you share this with Bryan Hubbell also









Margo,

Attached is the revised version of section 7.5.1, as we discussed.

Thanks,

Neal

_________________________________

Neal Fann

USEPA

Office of Air Quality Planning & Standards

Air Benefit and Cost Group

C439-02

RTP, NC 27711

voice: (919) 541-0209

fax:     (919) 541-0839

"Schwab, Margo" <Margo_Schwab@omb.eop.gov> 

03/10/2008 10:59 AM

	

To

Ron Evans/RTP/USEPA/US@EPA

cc

Bryan Hubbell/RTP/USEPA/US@EPA, "King, Heidi R."
<Heidi_R._King@omb.eop.gov>, Lydia Wegman/RTP/USEPA/US@EPA, Neal
Fann/RTP/USEPA/US@EPA, Tricia Crabtree/RTP/USEPA/US@EPA

Subject

RE: Ron, can you share this with Bryan Hubbell also









Ok. I'm at 202 395 5647 

-----Original Message-----

From: Evans.Ron@epamail.epa.gov [mailto:Evans.Ron@epamail.epa.gov] 

Sent: Monday, March 10, 2008 10:57 AM

To: Schwab, Margo

Cc: Hubbell.Bryan@epamail.epa.gov; King, Heidi R.;

Wegman.Lydia@epamail.epa.gov; fann.neal@epa.gov;

Crabtree.Tricia@epamail.epa.gov

Subject: RE: Ron, can you share this with Bryan Hubbell also

We will try to arrange something in the next 10 minutes.

Ron Evans

Leader, Air Benefit & Cost Group

HEID/OAQPS/OAR/EPA

Mail Drop C-439-02

919-541-5488

919-541-0839 fax

We do one thing: overwhelm people's impossible expectations.

                                                                        

             "Schwab, Margo"                                            

             <Margo_Schwab@om                                           

             b.eop.gov>                                              To 

                                      Ron Evans/RTP/USEPA/US@EPA,       

             03/10/2008 10:43         "King, Heidi R."                  

             AM                       <Heidi_R._King@omb.eop.gov>       

                                                                     cc 

                                      Neal Fann/RTP/USEPA/US@EPA, Bryan 

                                      Hubbell/RTP/USEPA/US@EPA, Lydia   

                                      Wegman/RTP/USEPA/US@EPA           

                                                                Subject 

                                      RE: Ron, can you share this with  

                                      Bryan Hubbell also                

                                                                        

                                                                        

                                                                        

                                                                        

                                                                        

                                                                        

 Ron: Heidi is busy, but could talk briefly with you and Neal about some

technical questions?  I'm open all morning.  Margo

-----Original Message-----

From: Evans.Ron@epamail.epa.gov [mailto:Evans.Ron@epamail.epa.gov]

Sent: Monday, March 10, 2008 10:02 AM

To: King, Heidi R.

Cc: Schwab, Margo; fann.neal@epa.gov; Johansson, Robert;

Hubbell.Bryan@epamail.epa.gov; Cakir.Janet@epamail.epa.gov;

Wegman.Lydia@epamail.epa.gov; Crabtree.Tricia@epamail.epa.gov;

simon.nathalie@epa.gov; Davidson.Ken@epamail.epa.gov

Subject: Re: Ron, can you share this with Bryan Hubbell also

Heidi, we have reviewed Margo's suggested changes to discussion on 8 vs

24 hour metrics.  The blue shows our minor edits, otherwise we are

comfortable with making the changes as requested.  Let me know if you

have any questions.

Both the NMMAPS analysis and the individual time series studies upon

which the meta analyses were based use the 24-hour average or 1-hour

maximum ozone levels as exposure metrics.  The 24-hour average is not

the most relevant ozone exposure metric to characterize population-level

exposure. Given that the majority of the people tend to be outdoors

during the daylight hours and concentrations are highest during the

daylight hours, the 24-hour average metric is not appropriate.

Moreover, the 1-hour maximum metric uses an exposure window different

than that that used for the current ozone NAAQS.  Together, this means

that the most biologically relevant metric, and the one used in the

ozone NAAQS since 1997 is the 8-hour maximum standard.  Thus, although

our analysis at proposal calculated impact functions based on either the

24 hour average or 1-hour maximum ozone levels originally reported in

the epidemiogical studies, for the final rule analysis, we have

converted ozone mortality health impact functions that use a 24-hour

average or 1-hour maximum ozone metric to maximum 8-hour average ozone

concentration using standard conversion functions.

   This practice is consistent both with the available exposure modeling

   and with the form of the current ozone standard. This conversion also

   does not affect the relative magnitude of the health impact function.

   An equivalent change in the 24-hour average, 1-hour maximum and

   8-hour maximum will provide the same overall change in incidence of a

   health effect. The conversion ratios are based on observed

   relationships between the 24-hour average and 8-hour maximum ozone

   values. For example, in the Bell et al., 2004 analysis of

   ozone-related premature mortality, the authors found that the

   relationship between the 24-hour average, the 8-hour maximum, and the

   1-hour maximum was 2:1.5:1, so that the derived health impact effect

   estimate based on the 1-hour maximum should be half that of the

   effect estimate based on the 24-hour values (and the 8-hour maximum

   three-quarters of the 24-hour effect estimate).

   Ron Evans

   Leader, Air Benefit & Cost Group

   HEID/OAQPS/OAR/EPA

   Mail Drop C-439-02

   919-541-5488

   919-541-0839 fax

   We do one thing: overwhelm people's impossible expectations.

             "King, Heidi R."

             <Heidi_R._King@o

             mb.eop.gov>                                             To

                                      Ron Evans/RTP/USEPA/US@EPA, Neal

             03/07/2008 01:59         Fann/RTP/USEPA/US@EPA

             PM                                                      cc

                                      "Schwab, Margo"

                                      <Margo_Schwab@omb.eop.gov>,

                                      "Johansson, Robert"

                                      <Robert_C._Johansson@omb.eop.gov>

                                                                Subject

                                      Ron, can you share this with

                                      Bryan Hubbell also

Ron, a note from Margo, can you pls fwd to Bryan?

Bryan - per our discussion yesterday, below is language to replace the

entire 'OZONE EXPOSURE METRIC' subsection of the mortality section

(PAGES 7-13 TO 7-14).  The first paragraph is a slightly edited version

of something that I sent over during L and M.  The second paragraph is

the only piece that I suggest you keep from the original write up (the

original write up is copied below).

Both the NMMAPS analysis and the individual time series studies upon

which the meta analyses were based use the 24-hour average ozone level

as an exposure metric.  The 24-hour average is not the most relevant

ozone exposure metric to characterize population-level exposure given

that the majority of the people tend to be outdoors during the daylight

hours.  Furthermore, concentrations are highest during the daylight

hours.  Together, this means that the most biologically relevant metric,

and the one used in the ozone NAAQS since 1997 is the 8-hour maximum

standard.  Thus, although our analysis at proposal calculated impact

functions based on the 24 hour average ozone levels originally reported

in the epidemiogical studies,  for the final rule analysis, we have

converted health impact functions that use a 24-hour average ozone

metric to maximum 8-hour average ozone concentration using standard

conversion functions.

This practice is consistent both with the available exposure modeling

and with the form

of the current ozone standard. This conversion also does not affect the

relative magnitude of the

health impact function. An equivalent change in the 24-hour average and

8-hour maximum will

provide the same overall change in incidence of a health effect. The

conversion ratios are based

on observed relationships between the 24-hour average and 8-hour maximum

ozone values. For

example, in the Bell et al., 2004 analysis of ozone-related premature

mortality, the authors found

that the relationship between the 24-hour average, the 8-hour maximum,

and the 1-hour

maximum was 2:1.5:1, so that the derived health impact effect estimate

based on the 1-hour

maximum should be half that of the effect estimate based on the 24-hour

values (and the 8-hour

maximum three-quarters of the 24-hour effect estimate).

For your reference, here is LANGUAGE THAT IS CURRENTLY ON PAGES 7-13,

14, 15 (and thus being replaced)

Ozone Exposure Metric

Because several of the ozone mortality epidemiology studies report their

base results relating

mortality to 24-hour average ozone, the importance of understanding the

biological relevance of

the 24-hour metric relative to other possible metrics is critical. Our

air quality modeling projects

ozone-related disbenefits to occur in small regions within specific

urban cores such as Chicago

and Detroit. When NOx reductions increase ozone in these regions,

however, it is typically

during nighttime and early morning hours, when NOx titrates ozone.

Because human exposure

to ozone is a function of the temporal and spatial patterns of ambient

concentrations of ozone in

the atmosphere, the ozone-related health impacts analysis is especially

sensitive to which ozone

exposure metric we use in the health impact functions.19

7-14

Prior to the addition of ozone-related premature mortality functions to

the health impacts

analysis, most of our ozone health impact functions have used metrics

which are less sensitive to

ozone disbenefits (e.g., 8-hour daily average). For example, emergency

department visits for

asthma are related to 1- or 8-hour maxima. School absences are based on

8-hour mean and 1-

hour maximum ozone levels. It should be noted that ozone disbenefits

that occur during daylight

hours, when ozone is higher, are accounted for in these averages.

Epidemiology studies are retrospective in nature and focus on

identifying a statistical

relationship between some measure of ozone and a health outcome. The

specific

epidemiological studies that form the basis for the ozone mortality

impact estimates use timeseries

statistical methods that estimate the relationship between daily ozone

levels and mortality

based on day to day variations in ozone and mortality. The focus of

these studies is not as much

on a specific ozone averaging time as on the day to day variation in the

ozone metrics. In fact,

epidemiologists often analyze and report results for multiple ozone

metrics, but may report

results for only one metric in the abstract of an article.

In most cases, the day to day variation in different metrics (24-hour

average vs 8-hour maximum,

for example) is highly correlated. As such, the relationships between

mortality and different

ozone metrics will be highly correlated as well. However, when we apply

the mortality impact

functions derived from these time-series results to evaluate the impacts

of a specific control

measure, we do not focus on the day to day variation in ozone levels so

much as the shift in the

overall distribution of ozone concentrations over an entire season.

Because specific emission

control strategies might result in a different diurnal profile than was

observed in the monitored

ozone data used in the studies, it is important to choose an ozone

metric that is best suited to

capturing changes in ozone that are likely to occur during hours where

populations are likely to

be exposed to the ozone.

To address this issue in the final rule analysis, we have followed a

process outlined in the

Locomotive and Marine Diesel RIA (EPA, 2008) and used standard

conversion functions to

convert ozone-related premature mortality health impact functions that

use a 24-hour average or

1-hour maximum ozone metric to functions that use an 8-hour maximum

ozone concentration

instead. This practice is consistent both with the available exposure

modeling and with the form

of the current ozone standard. This conversion also does not affect the

relative magnitude of the

health impact function. An equivalent change in the 24-hour average and

8-hour maximum will

provide the same overall change in incidence of a health effect. The

conversion ratios are based

on observed relationships between the 24-hour average and 8-hour maximum

ozone values. For

example, in the Bell et al., 2004 analysis of ozone-related premature

mortality, the authors found

that the relationship between the 24-hour average, the 8-hour maximum,

and the 1-hour

maximum was 2:1.5:1, so that the derived health impact effect estimate

based on the 1-hour

maximum should be half that of the effect estimate based on the 24-hour

values (and the 8-hour

maximum three-quarters of the 24-hour effect estimate).

The conversion of ozone metrics does not require adjustment to the air

quality modeling. It

preserves the observed patterns of ozone-related disbenefits, and allows

for disbenefits to occur

in the health impact estimates if those disbenefits occur during hours

when populations are likely

7-15

to be exposed. In future analyses, we will also convert the ozone

exposure metrics in morbidity

studies that do not use an 8-hour exposure metric.

