R’Qecommended Rir Pollution Index

Prepaxed by the Federal lnteragen‘cy Task Force on Air Quality Indicators: CEQ. EPA. DOC

Issued by the
Council on . . _ _ ‘_ , . .
l Environmental ' _ _ _. - , . ' . . . _ .-
. Quality ‘ I. _ . ‘_ ' .’ . - I

September 1976 .‘ .. '   ,. ~ - 3 - . '

A‘ ‘l ‘1;‘ x.
'5 i‘C3; 1
lot":

:‘é

.

 

:

   

..
I
“6",.”
‘ :121.»
-an
emu-.-”
{.L t
.5 .3
{1
‘7
lull

3 l

r r. 5
.. E 3
J V; 3' s
‘ \
5

 

5*?)
‘ a:-

.
I;
.8“
‘3.
In

 

 

 

R
ecommended

ir
Pollution
Index .4

Prepared by the

Federal Interagency Task Force on
Air Quality Indicators:

Council on Environmental Quality
Environmental Protection Agency
Department of Commerce

September 1976

 

For sale by the Superintendent of Documents. US. Government Printing Office
Washington. DC. 20402 - Price $130

Stock No. 041-001-00138—7

FOREWORD

Millions of Americans are regularly exposed to
various levels of air pollution. How and when we are
affected is a matter of concern to many citizens of
urban areas, especially those who suffer from respira-
tory and cardiovascular illnesses. It is especially
important that the air quality information provided
daily to the public —— on television, on the radio,
and in print -— be as scientifically accurate and
understandable as possible.

This report describes a recommended method for
providing such information, the Pollutant Standards
Index. Use of this index by metropolitan and state
agencies will be on a voluntary basis.

We would like to express appreciation to the many
professional staff from state, local, and Canadian air
pollution control agencies for their valuable coopera-
tion in providing information for the 1975 CEQ/EPA study
of currently used indices and for providing careful
reviews and meaningful critiques of the Pollutant
Standards Index documentation.

Russell W. Peterson
Chairman, Council on
Environmental Quality

9/
Ru sell E. Train

Administrator, Environmental
Protection Agency

EE:::;fl_I(’71LLvafiLka-_—/

Elliot L. Richardson
Secretary of Commerce
is}. "5

m

CONTENTS
Foreword ................................................ iii
I. Summary ........ ... .................................. 1
II. Background ................................... . ..... 3
III. Developing a Uniform Index ........................ 7
IV. The Recommended Index .............................. 8
Number of Pollutants .......................... 8
Calculation Method ............................ 8
Example Reports ................. . ............. 12
V. Testing the Index ....................... . ........... 14
VI. Further Work ....................................... 17
Adoption and Implementation ................... 17
Further Testing ............................... 18
A More Sophisticated Indicator ................ 18
VII. References ........................................ 19

Appendix A. "A Critical Review of Air Pollution Index
Systems in the United States and
Canada" ................................... 20

Appendix B. Guideline for Public Reporting of Daily Air
Quality -— Pollutant Standards Index
(PSI) ...................................... 31

Members of the Task Force ............................... 65

T D
883
.2
U545
1976
PUBL

 

I. SUMMARY

The increasing public awareness of air pollution during the last
decade has resulted in many state and local air pollution control
agencies' use of daily air pollution indices to inform the public about
the magnitude and changes in air pollution levels. Awareness of these
levels is most important to those who suffer from illnesses aggravated
by air pollution. In addition, most people may well cut down on
pollution—generating activities if they are better informed about pollution
levels and their significance. The variety of index types in use is confusing
and makes air pollution control more difficult.

Studies by the National Academy of Sciences and the Congressional
Research Service in 1973 and 1974 recommended that the Council on
Environmental Quality, the Environmental Protection Agency, and other
federal agencies develop effective air quality indicators. In response,
CEQ and EPA conducted a comprehensive survey of air pollution indices

in use. Air Pollution Indices: A Compendium and Assessment of Indices

 

in the United States and Canada (December 1975) describes a confusing
and scientifically inconsistent array of reporting methods in use
today. At least 14 basically different indices are used, and few seem
to provide truly meaningful information to the public.

In the summer of 1975, CEQ organized the Federal Interagency Task
Force on Air Quality Indicators. Participating with the Council were
EPA and several offices in the Department of Commerce: the National
Oceanic and Atmospheric Administration, the National Bureau of Standards,
and the Office of Environmental Affairs. The primary mission of the
Task Force was to develop an index which the participating agencies

could recommend for nationwide daily reporting of air quality.

The Pollutant Standards Index was developed by an EPA Technical
Working Group in collaboration with CEQ and NOAA. Many of its features
are based on the 1975 CEQ/EPA survey. The PSI has been reviewed by
Federal, state, and local air pollution control agency officials. In
the judgment of the Task Force, it incorporates the best available
scientific information and technology in the areas of air pollution
health effects, air quality monitoring, and forecast meteorology.

In many respects, the index closely resembles some of the better
indices currently in use. The government agencies which have partici-
pated in the Task Force effort are recommending the PSI for nationwide
adoption.

The PSI is designed for daily reporting of local air pollution levels
to the public through the news media. Its main function is to inform
the public of the potential health implications of observed air pollu-
tion levels. The PSI is based on the National Ambient Air Quality Standards
and the federal episode control criteria. The air pollutants covered
by the index are carbon monoxide, sulfur dioxide, nitrogen dioxide,
ozone, and total suspended particulates. The index has a range of
0—500 and uses five descriptor words to identify progressively higher
levels of air pollution: "good," ”moderate," "unhealthful," "very
unhealthful," and "hazardous.” In addition, the PSI can provide information
to advise people about possible effects of various pollution levels

on their well—being and about what preventive action they might take.

II. BACKGROUND
During late 1974 and early 1975, a comprehensive study of air
pollution indices currently in use was conducted by the U.S. Environ-
mental Protection Agency and the President's Council on Environmental
Quality. This survey was in partial response to recommendations made
by the National Academy of Sciences1 and the Congressional Research

Service.2 The CEQ/EPA findings were reported in Air Pollution Indices:

 

A Compendium and Assessment of Indices Used in the United States and

 

Canada.3 A condensation and update of this report are provided in "A
Critical Review of Air Pollution Indices in the United States and Canada"4
(see Appendix A).

The study examined the nature, characteristics, and reporting
formats of air pollution indices used by air pollution control agencies
and reported in the literature. From these data were developed the
criteria for the design of a uniform air pollution index as well as the
structure of two candidate indices. Information was gathered through:

(1) an extensive review of the literature on air pollution indices, (2)

a telephone survey soliciting information from 55 air pollution control
agencies in the United States and Canadian provinces, and (3) a case

study of a three-state metropolitan area (Steubenville—Wheeling—Pittsburgh)
where an attempt was being made locally to adopt a uniform tri—state air
quality index. After an index classification system was developed, it

was possible to analyze and compare the various indices reported in the

scientific literature with those used daily by the air pollution control

agencies.

0f the 55 U.S. metropolitan air pollution control agencies surveyed,
35 (64 percent) routinely use some form of air pollution index. Six
states and the two Canadian Provinces surveyed have uniform indices.
Fourteen different types of indices are in use by the 43 agencies
surveyed. With two minor exceptions, when descriptor words are taken

into account, no two indices are the same.

 

The survey found 41 air pollution descriptor words in currently
used indices. These descriptor words and their frequency of use are
shown in Table 1. Other evidence of the great diversity and lack of
consistency among air pollution indices is the meaning given index
numbers in 14 cities. Table 2 shows that a reported index value of
"25" is accompanied by 13 different descriptor words in these cities.
An individual who does not differentiate between index types and does
not understand how index values are calculated would be very confused
by descriptor words ranging from "extremely light" to "unhealthy” ——
all describing the same index number seen or heard on the news. There
is little uniformity in the indices used to inform the public about air
quality.

Air monitoring networks require substantial public investments
in both personnel and equipment. One of the primary uses of the
data obtained from these networks is informing the public. Pollution
indices provide a link between very technical information and information
for the lay public. The overall lack of uniformity among indices and
often of a firm scientific rationale is viewed by the Task Force as a

serious national communications problem.

FREQUENCY DISTRIBUTION OF THE 41 WORDS USED
FOR DESCRIPTOR CATEGORIES BY 43
AIR POLLUTION CONTROL AGENCIES

Heavy

Good

Light
Unsatisfactory
Unhealthy(ful)
Moderate
Emergency

Poor

Extremely Heavy
Fair

Medium
Satisfactory
Severe

Alert

Clean
Extremely Light
Very Poor
Acceptable
Dangerous
Excellent

Hazardous

10

H
O

NNNNwwwwbbbbbkflwO‘O‘VO

Normal

Stage 1

Stage 2

Stage 3

Very Heavy
Warning

Above Average
Acute

Average

Below Average
Endangerment
Extremely Poor
Harmful

High

Low
Significant
Slight

Very Dangerous
Very Good

Very Light

Source: Gary C. Thom and Wayne R. Ott, Air Pollution Indices:
A Compendium and Assessment of Indices Used in the United States
and Canada, December 1975.

H H H H H H H H H H H H H H m N N N m N

DESCRIPTOR WORDS ASSOCIATED WITH
AN INDEX VALUE OF "25" IN 14 U.S. CITIES*

City

Detroit, Michigan
Tampa, Florida
Ontario, Canada
Alberta, Canada
Washington, D.C.
Denver, Colorado
Nashville, Tennessee
Louisville, Kentucky
Miami, Florida
Albany, New York

Los Angeles, California
Portland, Oregon
New York, New York

San Francisco, California

Air Pollution Descrigtion

Extremely Light
Moderate
Acceptable
Clean

Fair

Fair
Slight
Good
Normal
High
Stage 1
Light
Unhealthy

Severe

*An index value of 25 is calculated from individual indexes or is the
actual concentration which is the basis of some indices.

Source: National Academy of Sciences, Planning Committee on Environmental
Indicee, "Planning for Environmental Indices," February 1975.

The Task Force solution was to develop a uniform air pollution
index for local use throughout the United States. Based on the findings
and recommendations of the 1975 CEQ/EPA study, the EPA Technical Working
Group set forth the criteria for a uniform index. The criteria for
design of the Pollutant Standards Index were:

It must be easily understood by the public.

It must be based on the health effects of the criteria
pollutants for which standards have been set; it must

be capable of including new pollutants and new standards
in the future.

It must be easily calculated from data obtained through
existing monitoring networks.

It must reflect day—to—day variations commensurate with
perceived changes in air pollution levels.

It must include forecast capability when appropriate.

III. DEVELOPING A UNIFORM INDEX

In July 1975 CEQ organized the Federal Interagency Task Force
on Air Quality Indicators to bring together the federal agencies
involved in the national air quality management effort and those
with special expertise. The Task Force included representatives
from CEQ; EPA's Offices of Research and Development, Air and Waste
Management, and Planning and Management; and the Department of
Commerce's Office of Environmental Affairs; the National Oceanic
and Atmospheric Administration, and the National Bureau of Standards.
All were selected for their competence in areas related to the develop—
ment and implementation of a uniform air pollution index: EPA in
index design, health effects, monitoring, data processing, and data
management; the Commerce Department's Office of Environmental Affairs
in disseminating information and liaison with the industrial sector;
NOAA in meteorology and forecasting; and NBS in measurement standardi—
zation.

As an integral part of the Task Force activity, EPA established
a Technical Working Group to develop a recommended index. It put
together all the elements of a viable index and prepared the "Guide—
line for Public Reporting of Daily Air Quality —— Pollutant Standards
Index" (PSI) (Appendix B). NOAA prepared an appendix to the Guideline
which addresses the forecasting aspects of index usage. The earlier
CEQ/EPA study was used in developing this Guideline, which EPA will

officially issue to state and local agencies.

IV. THE RECOMMENDED INDEX

The Pollutant Standards Index is a health—related air pollution
index based generally on the primary short—term National Ambient Air
Quality Standards (NAAQS) and the health-related Significant Harm
Levels.5 The development of the PSI followed the 1975 CEQ/EPA survey.
The structure of the PSI is described below.
Number of Pollutants

PSI includes the five pollutants (commonly called "criteria
pollutants"): carbon monoxide (CO), sulfur dioxide (502), total
suspended particulates (TSP), ozone (or photochemical oxidants)
(03), and nitrogen dioxide (N02). An additional number, the product
of the TSP concentration times the $02 concentration, is included
because it has been designated a Federal episode control criterion.
The calculation method for the PSI permits the addition or subtraction
of pollutants used in the index without changing the index equations
or index values for the pollutants covered.
Calculation Method

The index relates air pollution concentrations to numbers from
0 to 500. The relationships for each pollutant

are shown in Figure 1.

The PSI breakpoints,* their equivalent concentrations, health
effects descriptors, and advisory information are shown in Table 3.
An arbitrary breakpoint of 50 for all pollutants, thus creating the

"moderate" descriptor, is provided to demonstrate gradation from

*A breakpoint on a graph is a point where a shape or direction
distinctly changes.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

       
 

 

 

 

     
 

m - m -
nu [ I I I
,— — _ _.
:mnusucv uncncv
an ._ uvn u ._ m - mm. m —
_ -—I ._ —.
E unnmn ; mun-a
E m .. mm :4 — E an — mu. m —
__ — _. ._
‘LEIT ALEIY
m ._. um 11 -— 19' — uvu m —
_. __1 _ _
Ill —— I' mmnv nuns — m — mm" _
“All
— mmunv «no: — - 7‘ “nun mm" —
IAA”
. I I I I I _ I I | I I
I II n u u u u I m m m m mo ml
CAIIDI unuoxm In... nuns-us AVERAGE). «(-1 sum-nu: numuun nun" nu... name avmu, “1.1)
”' I I I I I '"
an autumn
uu _ IAMIIIHI uvn m. ._ _ “V“ II. —
% _- _
.. “ mum
3 Hanan 3 a! _ “van ' _
E 3.. ._ mu. m. — E
m _ um um _ 1n — m _
I“ -— m "mum _
III I—- 35 mun um — “A“,
"I __ —|' sex mun: —
Mill-III. 'IIMAIIV IAAB I I l
I l l I . l I
' I no mo mu no. no. mu m m m m “9' '3"
atom: (um "mum. uni-.1
:uuua nmxm! Imu- nuuuma AvmnatL ”M
III
.. _
mun-u
u- - uvn _
m -— .—
3 mum“ _‘
nu —' _ a Human
2 ma
‘E' at — uvn ._..
Altnv
2" — uvn ma -— mm
at — ._
mu u
la — —
In _— __
. l l | l I I I l J
I ma 1m an: an new un- .

Irrnocu mm: m— nvunzud-J

 

 

 

I" 1' JD M I.

mm swmntn umcuun It so, Inn-Mun! unsunuun. v01 lad-HI

Figure l — Relationship Between Pollutant Concentration
and Index Values of PSI.

Source:

Environmental Protection Agency, "Guideline for

Public Reporting of Daily Air Quality -- Pollutant
Standards Index"

OI

TABLE 3 .

COMPARISON OF PSI VALUES WITH POLLUTANT CONCENTRATIONS, DESCRIPTOR WORDS ,

GENERALIZED HEALTH EFFECTS, AND CAUTIONARY STATEMENTS

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

POLLUTANT LEVELS
TSP 802 CO 03 N02 HEALTH
lNl'i X AIR OUALITV (24-hour), (24-hour), (8-hour), (1-hour). (1-hour), EFFECT
VALUE LEVEL pwm3 pym3 mym pym3 nym3 oescnwron GENERALHEALTHEFFECTS CAUWONARYSTATEMENTS
— 500 —-—5'G:fl§/IANT 1000 2620 57.5 1200 3750
Premature death of ill and elderly. All persons should remain indoors,
Healthy people will experience ad- keeping windows and doors closed.
verse symptoms that affect their All persons should minimize PhYSI‘
normal activity. cal exertion and avoid traffic.
'—-‘OO —— EMERGENCY 875 2100 46.0 1000 3000-— HAZARDOUS
Premature onset of certain diseases Elderly and persons with existing
in addition to significant aggrava- diseases should stay indoors and
tion of symptoms and decreased avoid physical exertion. General
exercise tolerance in healthy persons. population should avoid outdoor
activity.
--— 300 -—--—- WARNING 625 1600 34.0 800 2260
Significant aggravation of symptoms Elderly and persons with existing
VERY and decreased exercise tolerance in heart or lung disease should stay
UNHEALTHFUL persons with heart or lung disease, indoors and reduce physical
with widespread symptoms in the activity.
healthy population.
-— 200 ALERT 375 800 17.0 400° 1130
Mild aggravation of symptoms in Persons with existing heart or
susceptible persons, with irritation respiratory ailments should reduce
UNHEALTHFUL symptoms in the healthy popula physical exertion and outdoor
tion. activity.
— 100 NAAOS 260 365 10.0 160 I
MODERATE
-— so ———50% or NAAOS-——— 75b sob 5.0 so a
GOOD
L——0 n 0 o n a

 

 

‘No index values reported at concentration levels below those specified by "Alert Level" criteria.
bAnnual primary NAAOS.
c400 ug/m3 was used instead of the O3 Alert Level of 200 pg/m3

Source: Environmental Protection Agency, "Guideline for Public Reporting of Daily Air Quality “ Pollutant

Standards Index"

”moderate" region

"good" to "unhealthful." The designation of a
recognizes the fact that although daily levels of a pollutant may
never exceed its short—term standard, over the course of a year its
average may violate its respective annual standard.

In order to maintain continuity between the health—related PSI
and presently established local episode prevention protocols, the
Technical Working Group set the index breakpoints between the primary
standards and the Significant Harm Levels at the recommended Federal
Episode Alert, Warning, and Emergency Levels, except for photochemical
oxidants. For oxidants, 400 micrograms/cubic meter was selected at the

breakpoint between 'unhealthful" and "very unhealthful" because this
level appears more consistent with health effects information than the
currently suggested administrative alert level of 200 micrograms/

cubic meter. For N02 and 802 times TSP, no index value is reported

below the Alert Level because no short-term standards have been established
for this range.

Each pollutant is examined individually by comparing its measured
concentration with the index plot shown in the figure. The calculation
may be performed manually, graphically, or by minicomputer. The maximum
index value for each pollutant may be reported if the user wishes. At
a minimum, however, the highest single index value computed for all the
pollutants should be reported to advise the public of the worst air
pollution in the community or region. On days when two or more pollu—
tants violate their respective standards (that is, have PSI values above
100), each pollutant may be mentioned. When the air pollution index
indicates that the air quality is "unhealthful," "very unhealthful,"

or "hazardous," both the precautionary statements and the generalized

health effects should be reported by the media.

11

Example Reports

The index was designed to be as flexible as possible for reporting
air quality to the public. Either short or detailed reports may be
used. For television, the report could read: ”Today the air pollution
index is 40; the air quality is good." However, when the air pollution
level becomes unhealthful (PSI over 100), there are several possible
reports for television, the news media, or telephone recordings. For
example, when oxidant pollution reaches a concentration of 280 micro-
grams/cubic meter (0.14ppm), the report may take one of several forms:

(1) "Today, the air pollution index is 150. The air is

unhealthful. The pollutant ozone is responsible.”

(2) Add the following precautionary statements: "Persons

with existing heart or respiratory ailments should reduce

physical exertion and outdoor activity."

(3) Include the language of (l) and (2), adding that

"'unhealthful' air can cause mild aggravation of symptoms in

the healthy population."

(4) "According to our meteorology forecast, we can expect that

the level of ozone will decrease [increase or remain at about
the same level] tomorrow."

Table 3 should be consulted in preparing the daily air
pollution report for the public. Figure 2 uses the above ozone
example in a Visual for television and the newspapers. The figure
provides essential information, including the PSI value, the critical

pollutant, and health implications for the public.

12

“RHEALTHFUL

    

U1
O
a

PSI = 150
POLLUTANT: Oxidants
TODAY’S HEALTH IMPLICATIONS: Respiratory ailment

and heart disease patients should reduce exertion
and outdoor activity.

FORECAST: No change.

Figure 2 - A Possible PSI Report

Source: Environmental Protection Agency, "Guideline for Public Reporting
of Daily Air Quality -- Pollutant Standards Index"

13

V. TESTING THE INDEX

Following development of the PSI structure, EPA examined the
performance of the index with actual air quality data.5 With a
specially designed computer program, EPA evaluated air quality data from
eight monitoring stations in eight cities. Five of the cities with the
most reliable data were examined in detail: Lennox, Calif.; Los Angeles,
Calif.; Anaheim, Calif.; Camden, N.J.; and Newark, N.J. For each station
in these five cities, the index value was calculated once per day.

An example of daily PSI values from the evaluation is shown in
Figure 3. In this plot for the downtown Los Angeles station, each
vertical line represents the index value for 1 day of the year, and
the small letter at the top of each line denotes the "critical pollu-
tant" whose concentration was responsible for the maximum index value.
This figure shows that oxidants and C0 predominate as the critical
pollutants in Los Angeles. Carbon monoxide dominates the index from
mid—October to March, and oxidants are usually highest in the remaining
months. The figure also displays the variation of index values from
day to day as well as the violations of the NAAQS (values of PSI over
100). Similar diagrams for the other seven stations indicate that
the PSI values are consistent with the pollutant—by—pollutant data.

The evaluation also showed that, for cities where two pollutants
are frequently above the NAAQS, reporting only the maximum pollutant
tends to underemphasize the second pollutant. Although it is important
to alert the public to the pollutant with the highest concentration,
the problem of underemphasizing a second pollutant may be corrected by

reporting all the pollutants for which the PSI exceeds 100.

14

POLLUTANI’ SVAMDARDINDEX POLLUIANV SYANDAHD INDEX VULLUHUNSIANUAMU muu

FDLLIHANT SYANDARD INDEX

300

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

rllvtr¥llllllllllllI’l'lllYlllYlll'll'llllf'llllll"WYIIIIIIInna-.onlII.A- vI|V|ll
c——cnnaouMon-nxms
o—oxmm Very
250 t-- ~—rARncuun{ -uquAL1MuL
c
zoo—-———— _~. °—- - — , —-—— ~E-- , —-~.«
1
r l
r . . .
150- “r Cc c c r ‘uunEALturul
c : _ : r c _ r o
,c .
mo c c cc 0
C C c u v
c C CC: CC C c C c 0 no t o c 0 Moderate
51» c c c »5 [[ccc cCCOCo o
c o o c 0 c
H: I Ill 3’ “I lllhlm
fl , , , ,, ,
" JANUARV FEBRUARY l MARCH I
300 I'llIl'VlTVTfirllllllllTVTlTl'ferll‘rllilllfll'lIl'lllIll'T'KrrlIllX'l'IYTlllllT‘fl'lllYYll’r
Very
250 ~ —UNMEAL1HFUL
I 1
no ————————————— —————— ~<-—~——-—-~»——————~—¢—<
J 1 o t
150— o e n _
o o n n" UNHEALTHFUL
1 1 1 ~" 0 11 7 "A n
0 oo oo o no 6
c o o 00 u 0 D Moderate
n o o
1 o L on u o 00 c
501.4 “A. y—jpl. D 1.0 r J 0 0 m . . . '“0 . r ~ .uH 4 a 0a
0 c o t
i I l I II I l l l
' APRIL ' AY [ JUNE [M
an IlllllllllllllllY'lIIIIYIIlllvrllllII'l'llI'lrl'rlllIIIVIIIVII‘I'I'II'IIII[ITYI’T'IIII’I'YTT
Very
250 ,_ - UNHEALTNFUL
In
200————-——- -- r—‘-—-'———-—‘———-—-—————‘——U-— —-< r———«——
1 1 7 n no 1
150» 1 n 1” n 1 n n n n n) 0 n UNREALTRFUL
1 r
n o 1 1 n 1 1. C 1
Wu 0 U o '10
o 0 Moderate
5u——p DfiH-- ”-1.1 -1. m 4— . H_<_ H, ._H. E'. . -- .a
I l soon
0 JULY 1 AUGUST SEPTEMBER
no Trll'lr:‘erITTlI'lT'TTDlVllll‘lllll"!VIIIV‘IY'IVIVIllI'V‘TlYl'r'll’“71‘lIIIIIVVYIIY'IIVIl
Verv
zsn~ ~ ' unuuuuruu.
I. ~ ~r
I "n ‘ ‘ ~ r ‘-
-___ _ ___________ _ .__ E“; it ~,____
200 g 1— r. H", F [
. . r
“0 r r r c F c «
” On n . r 'c r . r P c UNNEALTHFUL
n F F r m- r F C
mu u ' r r r
C c C c nL
C C c C
0C0 C 0 cc c c C cc 0 fioderate
Sn >< —- » D F — - H >\—' r—t- NH-{H—i >4 c <
l coco
u l
OCTOBER NOVEMBER DECEMBER

 

 

Figure 3 — Time series plot of PSI in the downtown Los Angeles site in 1974

Source:

Evaluation Study,

William Hunt and Wayne Ott,

15

Pollutant

Standards Index (PSI)
Environmental Protection Agency, April 1976.

The evaluation demonstrates that the PSI presents data in an
understandable form which approximates the air quality levels measured
at each station. The PSI appears to be an effective tool for presenting
and interpreting air quality data, reflecting day-to-day variations

of air pollutant concentrations.

16

VI. FURTHER WORK
The Federal Task Force on Air Quality Indicators is not planning to
continue formally after release of this report. However, the parti-
cipating agencies will implement the report's recommendations. This
section briefly describes future efforts.

Adoption and Implementation

 

The agencies which participated in the Federal Task Force on
Air Quality Indicators will encourage and endeavor to facilitate the
adoption and use of the Pollutant Standards Index by local and state
agencies. Local and state use of the index is a voluntary effort. The
Task Force hopes that the user community and the public will recognize
the potential value of the recommended index in enhancing the meaning—
fulness and often the accuracy of public information about air quality
conditions.
Specific actions planned by the agencies are:
Through existing mechanisms, the Environmental Protection
Agency will foster and facilitate local and state use of
the recommended index.
EPA will provide technical advice to local, state, and
news media personnel regarding use of the index and any problems
which may arise.
CEQ and EPA will jointly review nationwide progress toward
adoption and use of the index. It may be appropriate at some
future time to evaluate public understanding and response to

the index.

17

Further Testing

Following the preliminary testing of the Pollutant Standards
Index described earlier, more detailed studies are being conducted
by CEQ and EPA to assess its usefulness as an indicator of national
air quality conditions and trends. The index will be tested against
historic air quality data from approximately 20 major urban regions.
If it proves valuable as an indicator of trends, it will be useful
in the preparation of various federal reports discussing air quality,
including the CEQ Annual Report.
A More Sophisticated Indicator

CEQ and EPA will continue to work toward the development of
more sophisticated and meaningful indicators of air quality conditions
and trends. Although the Task Force considers the Pollutant Standards
Index the best standardized indicator for public reporting today,
there are important factors which the index does not incorporate.
They include the spatial distribution of air pollution levels in
relation to population density, human mobility, and the variations
in human response to urban air pollution levels.

A number of approaches are being considered for addressing
these factors and eventually for developing more comprehensive

indices.

18

VII. REFERENCES

National Academy of Sciences, Planning Committee on Environmental
Indices, "Planning for Environmental Indices," February 1975.

Library of Congress, Congressional Research Service, Environ-
mental Indices, Status of Development Pursuant to Sections
102(2)(b) of the National Environmental Policy Act of 1969,

a report for the Committee on Interior and Insular Affairs,
1973.

Gary C. Thom and Wayne R. Ott, Air Pollution Indices: A Compendium
and Assessment of Indices Used in the United States and Canada,
December 1975.

Wayne R, Ott and Gary C. Thom, "A Critical Review of Air Pollution
Index Systems in the United States and Canada," Journal of the
Air Pollution Control Association, Vol. 26, No. 5, May 1976.

William Hunt and Wayne Ott, Pollutant Standards Index (PSI)
Evaluation Study, Environmental Protection Agency, April 1976.

19

APPENDIX A

a critical review of
AIR POLLUTION INDEX SYSTEMS
IN THE UNITED STATES AND CANADA

Wayne R. on
U. S. Environmental Protection Agency

and

Gary C. Thom
Consultant to the Council on Environmental Quality

An extensive survey was conducted of all the air pollu-
tion indices that are presently utilized or are available.
The data were obtained from a literature review; from
telephone discussions with personnel in State, local, and
Provincial air pollution control agencies; and from
material received from these agencies. 0f the 55 met-
ropolitan air pollution control agencies surveyed in the
United States, 35 used some form of daily air pollution
index. These indices were so varied that it was necessary
to develop a system to classify indices according to four
criteria: (1) number of variables, (2) calculation method,
(3) calculation mode, and (4) descriptor categories re-
ported with the index. Using the classification system,
14 basically different index types were identified. With
two minor exceptions, it was found that no two indices
were exactly the same. The survey results and agency
comments were used to identify the general structural
characteristics and criteria for a candidate uniform air
pollution index.

20

In recent years, many state and local air pollution control
agencies have responded to the need to inform the public of
daily air quality levels by developing or adopting air pollution
indices. Awareness of these levels is important not only to
those who suffer from illnesses aggravated by air pollution but
also to members of the general public, who, if cognizant of
daily variations in air pollution levels, may choose to alter their
activities accordingly. However, because many different index
types have come into routine use, a citizen who travels to
different cities will receive a confusing picture of air pollution
levels in each city. Further, existing indices, because of their
diversity, cannot be used to assemble a national picture of air
pollution levels or trends.

When we initially reviewed the air pollution indices pro-
posed in the literature, we found a confusing array of different
index types. Further, although many indices had been de-
veloped, few of those proposed in the literature were being
used in practice—either by governmental decision makers or
by state and local agencies seeking to report air quality levels
to the public. Instead, the agencies tended to develop their
own indices. Two important questions were evident: (1) What
are the technical characteristics of all the air pollution indices
proposed or in use? (2) Does any “common index" emerge
from these characteristics and, if so, what does it look like?

Journal of the Air Pollution Control Association

To develop an extensive data base on all the indices in ex-
istence, whether proposed or currently in use, we surveyed the
55 largest metropolitan air pollution control agencies in the
United States. State air pollution control agencies and Ca-
nadian provincial agencies known to use air pollution indices
were also surveyed. The complete results of this survey are
reported in Air Pollution Indices: A Compendium and As—
sessment of Indices Used in the United States and Canada.1
This review is a condensation and update of the Compendium.

The need for a uniform national air pollution index has been
noted in several recent reports?‘5 Although the literature
reveals a variety of attempts to develop candidate air pollution
indicesf’“8 none of these indices has received widespread ac-
ceptance by state and local air pollution control agencies,
probably because none has received the active support or
endorsement of the Federal Government. However, in re-
sponse to our study of this problem, progress has been made
at the federal level toward the goal of recommending a na»
tional air pollution index.9 In July 1975, the Council on En-
vironmental Quality established a Federal Task Force on Air
Quality Indicators, with members from the Environmental
Protection Agency’s Offices of Research and Development,
Air and Waste Management, and Planning and Management;
the Department of Commerce's Office of Environmental Af~
fairs; and the National Oceanic and Atmospheric Adminis-
tration. The Environmental Protection Agency has proposed
a uniform index structure to this Task Force and is consid.
ering publishing it in its guideline document series10 as a
recommended national air pollution index. Hopefully a uni-
form national air pollution index would reduce the confusion
created by the diverse array of indices currently in use.

Detlnltlon ot Index

An “air pollution index" may be defined as a scheme that
transforms the (weighted) values of individual air pollution~
related parameters (for example, carbon monoxide concen-
tration or visibility) into a single number, or set of numbers
(Figure 1). The result is a set of rules (for example, an equa-
tion) that translates parameter values—by means of a nu-
merical manipulation—into a more parsimonious form. (In
set theory, this process is viewed as mapping of elements
contained in one sample space into another sample space.)

The following evaluations were made to determine whether
an agency used an “index." If an agency reported just the ac-
tual air pollutant concentration values to the public—mi-
crograms per cubic meter (ug/m3) or parts per million
(ppm)—or concentration values along with standards, this was
not considered an “index." Rather, an index must be based
on some set of rules which translates the values into a new
variable or makes interpretations of these values. At the very
least, an index is any system in which specific concentration
ranges are grouped into air quality “descriptor categories."
For example, a system which designates 0-3 ppm carbon
monoxide as “good,” 3—15 ppm as “satisfactory," and 15—40
ppm as “unsatisfactory" was considered to be an index. In its
most elaborate form, an index is an equation which combines
many pollutants in some mathematical expression to arrive
at a single number for air quality.

Some cities use episode warning systems that report de-
scriptor categories whenever concentrationsexceed specified
levels. If this happens frequently, the distinction between an
episode warning system and a daily informational index he-
comes blurred. Thus, air pollution agencies with episode
warning systems are classified as having indices.

 

Parameter X1-’

Pa —.
rameter X 2 Transformation

Parameter X3- 80“. X2. X3, ”'xn)

Parameter Xn"

<— Index

 

 

 

Flguro 1. Index calculation.

May 1976 Volume 26, No. 5

A Clasoltlcatlon System tor Indlcos

To facilitate comparisons among the various air pollution
indices, we developed an index classification system based on
four criteria:

- Number of variables included in the index.

- Calculation method used to compute the index.
- Calculation mode (combined or uncombined).
- Descriptor categories reported with the index.

Number 0! Varlablu

This number designates the number of variables incorpo~
rated into an air pollution index. These variables include the
NAAQS (National Ambient Air Quality Standards), visibility,
particle scattering, and other measures of air pollution.

Calculation Method

Our index classification system employs four calculation
methods, three of which (Types A, B, and C) involve an
equation:

A. Nonlinear. Exponential function with coefficient or other
nonlinear relationship. Coefficients may be constant or
may vary, but the relationship contains at least one vari-
able raised to a power.

The equations for the nonlinear (Type A) index may take
several forms. One is the Oak Ridge Air Quality Index
(ORAQI),7 which may include one to five pollutants, where
C denotes concentration and S denotes the air quality stan-
dard for pollutant i.

ORAQI = [5.7 i (Ci/801137
[-1

ORAQI values may also be determined with the aid of a no-
mograph. Another Type A index is the Ontario Air Pollution
Index (API):11

API = 0.2(30.5 COH + 126.0 8010135

It includes only two pollutants—coefficient of haze (COH) and
sulfur dioxide (802). A third example is the Measure of Un-
desirable Respirable Contaminants (M.U.R.C.) Index.l2 It
includes only one pollutant and is calculated using the equa-
tion:

M.U.R.C. = 70(COI-I)°-7

B. Segmented linear function. Linear function of one or more
variables with nonconstant coefficients. There are no ex-
ponents, but the slopes are different for different ranges
of the pollutant variable.

Figure 2 shows a plot of the segmented linear function for

carbon monoxide (CO) used in the Washington, DC index.

Mathematically, this function can be represented by the fol-

lowing equation, where I denotes the index value, C denotes

pollutant concentration (ppm CO), and each pair of values (xj,

y,) are the coordinates of breakpoint j (represented as a dot):

1=ZJ'+_ILVJ(C_XJ.)+”
xJ'H—Ij
wherej=0,l,...,5

for Xj < C 5x141

The function shown consists of five straight line segments,

each with different slopes, K, = (yin - yj)/(Xj+1— 35,-). The

index value for any CO concentration can be determined di-

rectly from the curve. For example, at 80 ppm CO, the index

value is 200.

C. Linear. Linear function of one or more variables in which
the slope K, for each variable i remains constant:

I = KiC.‘

Coefficients may be chosen as K ,- = 1/C,,,, where C,, is the
standard for pollutant i, giving a proportionate relation-
ship; or they may be chosen as K,- = 100/C,,, giving a per-
centage relationship; or they may be arbitrary and not
related to any standard.

D. Actual concentrations. Concentrations reported in sci-
entific units (pg/m3, ppm) or standard units from some
commonly used measurement technique (coefficient of
haze, for example).

An agency reporting just actual concentration values is
classified as not having an index and is not coded; however,
when the agency reports actual concentrations and descriptor
categories, its index is Type D.

Calculation Mode

Another important aspect of the calculation method is how
the index variables are treated. Does the agency report indi-
vidual index values for each variable? Does the agency report
an index value only for the variable which has the maximum
value of all the index variables? Does the agency’s index
combine the variables in some fashion? Thus, the mode
identifies whether the index is combined or uncombined.
Uncombined indices include those in the individual or max-
imum mode category; combined indices are sometimes re-
ferred to as aggregated indices. The mode is indicated by ap-

pending a subscript to the calculation method classification:

1. Individual. An index value is reported for each variable
comprising the index.

2. Maximum. Only the index value for the maximum variable
is reported.

3. Combined. The index variables are aggregated, through
some type of mathematical manipulation, to give one index
value.

Descriptor Categories

The descriptor categories result when the index range is
subdivided. The words assigned to each category describe
qualitatively the air quality. For example, an index may list
0—25 as “good," 26—50 as “satisfactory," 51—99 as “unsatis-

 

(110.750)
(X5. Y5)
e Index breakpoints

I, index value
§ §
l I

‘s’

(90,250)
(X4. Y. )

§

(60.!00) (X3. m
(35.50)
(20.25) X2, Y2)
o 1 X1. Y1 l i 1
0 20 40 60 80 100
CO concentration, ppm
Figure 2. Example oi a segmented linear iunction ior carbon
monoxide.

 

 

 

120

22

factory," and 100—199 as “unhealthy.” If an index reports
actual pollutant concentrations, several concentration ranges
may be used for the descriptor categories. Index descriptor
categories can be based on standards, episode criteria, or an
arbitrary basis:

A. Standards. The category breakdown is based on Federal,
State, or local ambient air quality standards—for example,
index values above 100 exceed the Federal Primary
NAAQS and those below fall into several categories par-
tially based on the Federal Secondary NAAQS. If actual
concentrations are reported, then these concentrations are
related to the standards.

B. Standards and episode criteria. Type A (above) is ex-
tended to accommodate index values above 100. These
values are based on the Federal, State, or local episode
criteria—for example, 100 is the Alert Stage, 200 is the
Warning Stage, etc. For indices reporting actual pollutant
concentrations, these concentrations are related to the
episode criteria.

C. Arbitrary. Categories of this type are semiempirically
based and usually designed to fit the specific requirements
of the index values. This classification also covers indices
with no descriptor categories.

Summary and Example Application ot the index Classification
System

The result of the classification system is a four character
code which describes any index. Thus ORAQI can be coded
as “5A3A" (Figure 3). The number “5” indicates that the index
includes five pollutants or variables; “A3" denotes the calcu-
lation method and mode (i.e., it is nonlinear and the variables
are combined to give one index value); “A" refers to the basis
for the descriptor categories (i.e., the categories reported with
this index are based on the NAAQS).

Survey at Existing Indlces
lndlcee Reported In the Literature

The literature on air pollution indicess‘s" 1—17 has previously
focused on the development of long-term trend indices. Little
has been published on the short-term indices commonly used
by state and local air pollution agencies. Although many
long-term indices have appeared in the literature, our dis-
cussions with governmental personnel have revealed few cases
in which such indices have actually been used to evaluate air
pollution data to make decisions or policies.

Each of the 11 air pollution indices reported in the literature
(Table I) differs either in terms of the number of applicable
pollutants, method of index calculation, or descriptor cate-
gories. As a result, the overall meaning of each index is dif-
ferent. Several of the indices were designed for specific pur-
poses; for example, the Extreme Value Index was designed to
describe air pollution episodes. For a more complete de-
scription of each index, the reader should consult the literature
reference. Three of these indices are currently in use by
agencies: M.U.R.C. (Detroit, Memphis), Ontario Air Pollution
Index (Ontario), and ORAQI (Tampa, Minnesota).

indices Used by Agencies

The air pollution literature provides little information about
the routine use of indices by air pollution control agencies. To
learn which air pollution indices are in common use and to
gain insight into the experiences of air pollution control
agencies with these indices, an in-depth survey of these
agencies was required. In this survey, agencies throughout the
United States and Canada were telephoned and asked to send
information describing their index. The data base in this in-
vestigation was assembled from notes taken during the tele-

Journal of the Air Pollution Control Association

Calculation Descriptor
method categories
5A3 A
Number of 2.
oolluhnb Calculation
mode ’
1. 2. 3.
A: Nonlinear 1: Individual A: Standards
a; Segmented linear 2: Maximum 3: Standards and
c: Linear 3: Combined episode criteria
D: Actual concentrations C1 Nbim'y

Figure 3. index classification system.

phone conversations and from written materials received from
the agencies.

Survey Population. The population surveyed in this in-
vestigation consisted of the 55 largest metropolitan (city and
county) air pollution control agencies in the United States,
along with state air pollution agencies in the United States
known to operate state-wide air pollution index systems. It
also included the Canadian Provinces with air pollution
control agency staffs of 10 or more persons and one Canadian
city which uses an index. To select the survey population, the
total number of staff members from every city and county air
pollution agency was computed using the 1973 Directory of
Governmental Air Pollution Agencies, published by the Air
Pollution Control Association.18 Only those US. air pollution
agencies having 10 or more staff members were included in
this survey population (Table II),

In the United States, the resulting survey population con-
sisted of 55 agencies. Telephone inquiries revealed that, in 14
of these cities, the index was operated as part of a general
state-wide or regional index system. Six states were operating
this type of system: Connecticut, District of Columbia, Min-
nesota, New Jersey, New York, and Ohio. These state indices
serve 67 cities (Table III), but many of these air pollution
agencies have staffs less than 10 persons. Also, the agencies
in Baltimore, MD, Boston, MA, and Portland, OR, are oper-
ated by the state but are not part of a state-wide system. In
Portland, the state not only reports the air pollution index but
it operates the entire city air pollution control agency as well.
In Baltimore, on the other hand, the state reports the index,
but the local air pollution agency is organizationally separate
from the state agency.

In Canada, only Montreal operates a city air pollution
control agency. Therefore, all provinces with staffs of 10 or
greater were included in the survey population. Alberta and
Ontario operate province-wide air quality indices. Eight cities
within these Provinces issue daily indices (Table IV).

Survey Approach. Telephone calls were made to the
agencies in the survey population from August to December
1974. For each agency, a respondent was sought who was very
familiar with the agency’s air pollution index, if any. In small
agencies, this usually turned out to be the agency’s director;
in the larger agencies, a public information specialist or a
professional in the field of monitoring and data analysis
usually was the respondent. With the respondent on the
telephone, the investigator went through an informal ques-
tion-and-answer session. Each respondent was asked, “Can
you provide any literature or description of your index?" Of
the 55 agencies on the major list (Table II), 28 supplied written

May 1976 Volume 26, No. 5

material. Typically, it covered the nature of the index, its
method of calculation, the history of its development, and the
way in which it is reported. In some cases, the telephone dis-
cussion provided sufficient information about the index and
no mailed material was necessary. Some agencies not using
indices provided material that discussed their reasons for not
adopting an index or their experience with a previously dis-
continued index.

Analysis and Evaluation oi Indices

Initially, information from the mailed responses and from
notes taken during each telephone conversation was con-
densed and compiled into tables. The tabular compilation was
found to be inadequate, however, due to the varied and
sometimes extensive information received from the agencies.
Consequently, the information was assembled into the three
appendices which appear in the Compendium.1 An index
analysis record was developed to present detailed factual in.
formation about each index in a uniform format; informal
comments from the agencies were copied into an extended
table; examples of the ways in which indices are reported by
the news media were also recorded. A summary and analysis
of the data extracted from these appendices are presented
here. For more detailed information, the reader should consult
the Compendium.

Table I. Characteristics of indices reported in the literature.

 

Variablesa

Classification CO SO, N02 0,, COH TSPOthcr

 

Green's
Combined
Index” 2A3C o ’ e
M.U.R.C.” lA,C e
Combustion
Products
Index” 2Cl C b
Ontario Air
Pollution
Index“ 2A,B I 0
Air Quality nC,C c I e e .
Index” (n = l to 5)
PINDEX‘ 7C3C e o e s o c
Oak Ridge
Air
Quality
Indcx" nAlA e o e e o
(ORAQI) (n = l to 5)
Mitre Air
Quality nA,A e e o o o
IndexI (n = l to 5)
Extreme
Value nA3A e e o 0
Index“ (7: = l to 4)
STARAQS
Air
Quality
lndcx
Model" 7B,A e o e e o e d
Canadian
Air
Quality
indices” SAJA e e o e o o c
Total 7 9 6 7 5 7

 

30x, oxidant; COH, coefficient of haze; TSP, total suspended
articulate.
Fuel burned and ventilating volume.

cHydrocarbons and SO, X TSP.

dVisibility.

cVisibility and industrial emissions.

Index Varlables

The variables included in the 11 indices reported in the
literature are given in Table I. Ten of the 11 indices include

Table II. The 55 U.S. city/county air pollution control agencies
with staffs greater than 10.

 

 

Ma-
te-
ri- In-
all dex
Agency re- in
City/County size ceived use Comments
Birmingham, AL 17 I
Phoenix, AZ 25 I Discontinued Index
Anaheim, CA 24 I Replaced Index
Los Angeles, CA 380 I I Replaced Index
Riverside, CA 26
San Bernardino, CA 53
San Diego, CA 53
San Francisco, CA 220 I I Replaced Index
Denver, C03 54 I I
New Haven, CNb I I I I Replaced Index
Washington, ch l4 . o
Bradenton, FL 1 1
Jacksonville, FL 15 I I
Miami, FL 50 I I
Sarasota, FL 21
Tampa, FL 16 I 0
Atlanta, GA 14 e 0
Chicago, IL 175 I I
Gary, IN 18
Indianapolis, IN 15
Louisville, KY 39 e 0
Baltimore, MD3 90 I I Replaced Index
Montgomery Co., MDb 10 I
Boston, MA3 87 I
Springfield, MA 12
Detroit, MI 77 I I
St. Paul, NW) 13 o 0
Kansas City, MO 15
St. Louis, MO 35
Albuquerque, NM 15
Albany, Mb 237 0
Buffalo, NW 44 e o
Mineola, NW 37 0
New York City, NY 382 I
Rochester, NYb I2 I
Charlotte, NC 14
Akron, OHb 13 0
Cincinnati, Crib 65 o . Replaced Index
Cleveland, cab 30 e 0
Dayton, OHb 45 I Replaced Index
Toledo, OHb 25 I
Oklahoma City, OK 15 I Discontinued Index
Portland, OR"1 20 e 0
Philadelphia, PA 94 I I Replaced Index
Pittsburgh, PA 82 I I
Chattanooga, TN 22 I
Memphis, TN 14 0
Nashville, TN 17 I I
Dallas, TX 21 I I
El Paso, TX 10
Houston, TX 76 I
Pasadena, TX 45
Fairfax Co., VAb 12 .
Seattle, WA 39 I I
Milwaukee, WI 25

 

aCity index is operated by State but is not part of State-wide index
system
bCity index is part of State-wide or regional index system

24

a measure of suspended particulate—OOH or total suspended
particulate (TSP)—and nine include 802. It is interesting to
note that the two newest indices, STARAQS and the Cana-
dian Indices, include all of the pollutant variables listed, in
addition to other indirect measures of ambient pollution
levels, visibility, and industrial emissions.

The variables included in US. city/county air pollution
indices are shown in Table V. If COH and high-volume sam-
pler measurements of TSP are lumped together as measures
of particulate matter, then particulate becomes the most
common air pollutant included by these agencies in their in-
dices. Of these 35 agencies, 33 (94%) include either COI-I or
TSP; COH is used by 22 agencies and TSP by 11 (Jacksonville,
FL uses both). The popularity of the COH measurement may
be due to the increasing use of telemetered air monitoring
networks which cannot readily handle TSP data determined
by the high volume sampler and to the shorter averaging time
(2 hr). CO and 802 are the next most common pollutants to
be included in these indices—26 agencies (74%) for each. The
next most popular pollutants are oxidant (20 agencies, 57%)
and N02 (16 agencies, 45%). Visibility is included in one
agency’s index, and particle scattering is the only variable
making up another agency‘s index.

When the air pollution indices used by states (or regions)
and Canadian provinces are examined (Table VI), a similar
pattern emerges. The most common pollutants are CO, 802,
and particulates (COI-I and TSP). The least common pollu-
tant is N02, with only two agencies—Ohio and the District of
Columbia—reporting it in their indices. The two Canadian
Province indices report different numbers of pollutants; the
smaller agency, Alberta, reports five air pollutants, while the
larger, Ontario, reports only two.

Results cl Clauflleatlon

The classification system described earlier was applied to
all the indices reviewed in this study.

Indices Reported in the Literature. Table I shows the
classification of the 11 indices reported in the literature. Eight
of the 11 use a combined calculation mode (Type 3), and seven
of these use a nonlinear (Type A) calculation method. Three
of the last four indices listed (i.e., excluding STARAQS) em-
ploy what has been termed a root-mean-square (RMS) index
equation. Although this equation tends to emphasize high
pollutant concentrations, it is so complex that it confounds
the interpretation of these concentrations, thereby confusing
the layman. This, and the general complexity of other “A3”
type index equations, may have contributed to the limited use
of these indices by local air pollution control agencies.

lndices Used by Agencies. Application of the index classi-
fication system to the 32 index systems reviewed in the survey
of air pollution control agencies revealed 14 basic types‘
(Table VII). To simplify comparison of these index types, they
are grouped according to their calculation method. Agencies
which either discontinued or replaced their indices did so ei-
ther because the indices were not consistent with air pollution
levels as perceived by the public, or to improve their old index
system, or to conform to state-wide index implementation.

Classification of the various indices used by the survey re-
spondents revealed a striking diversity and few clear patterns.
Of the 35 city/county agencies, 13 (37%) include five variables
in their index calculation (Table VIII). This is due mainly to
the fact that five agencies in Ohio and four in the Baltimore-
Washington, DC, area use indices incorporating five vari-
ables. In fact, each of these nine agencies used the 5BQB type
of index.

 

' A basic type refers to the calculation method and descriptor categories. but not to the
number of variables.

Journal of the Air Pollution Control Association

Table III. State-wide air pollution index systems

 

 

Agency
State Applicable City or County Sizea
Connecticut Bridgeport <10
Danbury —
Derby —
Enfield 7
Greenwich <10
Groton —
Hartford lOSb
New Britain -
New Haven 1 1
Stamford <10
Torrington 7
Waterbury —
District of Columbia Alexandria, VA <10
Arlington Co., VA <10
Fairfax Co., VA 12
Montgomery Co., MD 10
Prince Georges Co.. MD <10
Washington, DC 14b
Minnesota Duluth <10
Minneapolis 37b
Rochester <10
St. Paul 13
New jersey Ancora 7
Asbury Park 7
Atlantic City —
Bayonne 7
Burlington —
Camden —
Elizabeth <10
Freehold 7
Hackensack 7
Jersey City <l0
Morristown —
Newark —
Paterson -
Paulsboro —
Penns Grove 7
Perth Amboy <10
Phillipsburg —
Somerville —
Toms River 7
Trenton 1 70b
New York Albany 237b
Buffalo 44
Kingston 7
Mamaroneck —
Mineola-Eisenhower Park 36
New York City7Roosevelt Isl. 7
Niagara Falls <10
Rensselear <10
Rochester 12
Schenectady —
Syracuse 10
Utica —
Ohio Akron 13
Canton <10
Cincinnati 65
Cleveland 80
Columbus 228b
Dayton 45
Lorain <10
Mansfield <10
Painesville <10
Portsmouth <10
Steubenvillc <10
Toledo 25
Youngstown <10

Despite the heavy influence of these nine agencies on the
data, Table VII shows that the segmented linear function
(Type B) and the actual concentrations (Type D) are the most
popular calculation methods (37% and 32%, respectively).
Only four cities (11%) use a nonlinear index calculation,
suggesting a definite preference for the less complex calcula-
tion schemes. This preference is also evident in St. Paul, which
uses a nonlinear index (Type A: ORAQI) but in addition re-
ports actual concentration values (Type D). If St. Paul is re-
classified as Type D (shown in parentheses, Table VIII), then
only three (9%) of the agencies use the nonlinear type calcu—
lation method.

Table VIII shows that the “maximum” mode of calculation
was used by 15 (43%) of the agencies. Another 12 agencies
(34%) used the “individual” mode, thus indicating a prefer-
ence for uncombined (77%) over combined (23%) indices.
However, Jacksonville and St. Paul use both the individual
and combined mode. If they are reclassified as using the in-
dividual mode (as shown in parentheses, Table VIII), then
only six (17%) of the agencies use the combined mode and 83%
use the two uncombined modes.

Table VIII also shows that 25 agencies (71%) used index
descriptor categories based either on standards or on stan-
dards and episode criteria; the remaining one—third used ar-
bitrary categories.

The number of categories used in the indices is shown in the
histogram of Figure 4. There appears to be a definite prefer-
ence for three or four descriptor categories. However, there
does not appear to be any tendency toward using the same
words for the descriptor categories. A total of 44 different
words (Table IX) are used in the descriptor categories of 30
US. index systems. Indices containing the more common
words are shown in Table X. It is interesting to note that
certain words occur more frequently in multipollutant indices
and other words more frequently in indices involving only
particulate matter.

For the six states and two provinces, no detailed analysis
was possible due to small sample size. However, examination
of Table VI shows that all of the states report at least three
pollutants in their index; two states report all five NAAQS
pollutants. Since Minnesota reports individual pollutant
concentrations in addition to its combined index (ORAQI),
all state indices can be classified as using the segmented linear
or actual concentration calculation methods, while the cal-
culation mode is either individual or maximum. Thus, both
the States and cities make only limited use of the more com-
plex, nonlinear combined indices (Type A3). On the other
hand, in Canada there appears to be a preference for this type
of index, with both Alberta and Ontario using nonlinear
combined indices.

Summary of Preferred Index Characteristics. Combining

Table IV. Province-wide air pollution index systems.

 

Province Applicable city Agency sizea

 

Alberta Calgary

Edmonton 26"
Hamilton

Happy Valley

Sudbury

Toronto 70b
Welland

Windsor

Ontario

 

 

2‘Where no agency size is given. size is unknown.

bState agency.

May 1976

Volume 26, No. 5

aWhere no agency size is givenI size is unknown.
bProvince agency.

Table V. Characteristics of the 35 U.S. city/county air pollution indices.

 

 

 

Variablesa
Classi- No. of
fica- Descriptor
tion C0 SO, NO, 0" COH TSP Vis PS Categories

Anaheim, CA 30,8 0 o o 3
Los Angeles, CA 3D,8 o o o 3
San Francisco, CA 4D,C a o o o 6
Denver, CO 28,C o o 5
New Haven, CT 38,8 0 o o 5
Washington, DC 58,8 0 o o o o 7
Jacksonville. FL 6C3C o e I o a 0 None
Miami, FL 5C,C O o o o o 5
Tampa, FL 5A,A o o e o o 5
Atlanta. GA 3C3C o I 0 None
Chicago, IL 4D,B o 0 o o 4
Louisville, KY 5C38 o o o o o 4
Baltimore, MD 58,8 0 o o o o 8
Montgomery Co.. MD 58,8 0 o o o o 7
Detroit, MI 1A1C o 5
St. Paul, MN 3A38 o o c 4
Albany, NY SDIA o o o 3
Buffalo, NY 3DIA o o o 3
Mineola, NY 3D,A o o o 3
New York City, NY 5D,A o o a o o 4
Rochester, NY SD,A o o o 3
Akron, OH 58,8 0 o o o o 12
Cincinnati, OH 58,8 0 o o o o 12
Cleveland, OH 58,8 0 o o o o 12
Dayton. OH 58,8 0 a o o o 12
Toledo, OH 53,8 I o o o o 12
Portland, OR IDIC o 5
Philadelphia, PA 48,8 0 o o n 5
Pittsburgh, PA 2C,8 o o 6
Chattanooga, TN [BIG 0 4
Memphis, TN IA,C 0 4
Nashville, TN 1C,C o 4
Dallas, TX 2C,B O o 4
Fairfax C0,. VA 58,3 0 o o o o 7
Seattle, WA 28,8 0 o 3
Total 26 26 16 20 22 l l l I

 

aOx, oxidant; COH, coefficient of haze; TSP, total suspended particulate; Vis, visibility; PS,
particle scattering (integrating ncphelometer)

the results from the index classification (Table V) with the
analysis of descriptor words (Tables IX and X) enables one
to determine the most frequently encountered characteristics
of the indices currently used by air pollution control agencies
(Table XI). This summary is particularly useful when con-
sidering the possibility of developing a uniform national air
pollution index. Such an index, if it is to gain wide acceptance,
must have a structure which closely resembles that of the
greatest number of indices currently in use. Table XI shows

that such an index would be classified as a 5328 index. This
index would (1) be based on the five NAAQS pollutants (ex-
cluding hydrocarbons), (2) be a segmented linear function, (3)
be calculated using the maximum mode, and (4) have four
descriptor categories based on the NAAQS and Episode Cri-
teria, using the more frequently occurring words such as Good,
Moderate, Unhealthy(ful), and Unsatisfactory. These char-
acteristics were used to formulate the proposed uniform index
discussed in the Conclusion.

Table VI. Characteristics of the six State-wide and two Province-wide air pollution indices.

 

 

 

Variables No. of
Classi- ' Descriptor
fication CO SO, N0, Ox COH TSP Categories
State
finnecticut 38,8 0 o o 5
District of Columbia 58,8 0 o o o o 7
Minnesota 3A,8 o o o 4
New jersey 4D,A o o e o 4
New York SD,A o o o 3
Ohio 58,8 0 o o o o 12
Total 5 6 2 4 4 2
Province
Alberta 5A,C o o o o o 5
Ontario 2A,8 o o 5

 

26

Journal of the Air Pollution Control Association

Comments trom Agencies

Some insight into the reasons for the great variety of air
pollution indices was gained by examining the subjective views
of the survey respondents. Their comments, which were as-
sembled by the authors from notes taken during the informal
telephone conversations, are listed in greater detail in the
Compendium.1

Agencies Using Indices. As one might expect, most re-
spondents from agencies now using indices expressed satis-
faction with their own index, although one respondent ac—
knowledged that most people probably do not follow the index
in the newspapers. There was widespread opinion that the
numbers expressed by indices are not necessarily meaningful,
and one agency stated that its index was not developed on any
scientific basis and was not intended as such. In general, an
index was viewed as an informational tool designed to advise
the public as to the severity of air pollution levels, but not to
convey any information about the deleterious effects of air
pollution. Some agencies felt that the layman does not un-
derstand the technical language of air pollution, and thus
indices fulfill an important need by communicating with him
in a nontechnical way. However, in many cases the public may
not understand the index either. One agency using the com-
plex ORAQI index solved this problem by distributing a public
information bulletin on the subject. A number of the re-
spondents expressed concern about the lack of spatial rep-
resentativeness of index values. This concern apparently re-

Table V". The 14 basic types of indices and
their users.

 

 

Type Users

A,C Detroit, Oklahoma City,a Memphis

A3A Tampa

AJB Minnesota, Ontario

AJC Alberta

8,5 Baltimore,b Philadelphia.b Seattle,
Connecticut,b Ohio,c Washington,
DC

B}C Denver

C,C Nashville

C38 Louisville, Pittsburgh, Dallas

C,C jacksonville, Miami, Atlanta,
Phoenixa

DIA New Jersey, New York State

D,B Anaheim,b Los Angeles,b Chicago

DIC Portland, Chattanooga

D,A New York City

D B San Franciscob

 

aDiscontinued index.
b Replaced index.
CReplaced index in Cincinnati and Dayton.

fleets a lack of representativeness of the monitoring data used
to calculate the index.

Several agencies had found it necessary to make major
changes in their indices within the past 5 years. These agencies
indicated they had some difficulty overcoming the public
confusion which resulted from these changes, particularly
where the changes were great. One such agency stated that the
introduction, right now, of any new indices “would bury us."
Another agency changed its index from the combined form
to a maximum type because the combined index gave mis-
leading results—low values when one standard is violated and
other pollutants are low, or high values when all pollutants are
high but no standard is violated.

Agencies Not Using Indices. The diversity of opinion was
wider among agencies not presently using air pollution indices.
Respondents often expressed dismay about the large number
of air pollution indices in existence: With so many different

May 1976 Volume 26. No. 5

27

Table VI". Classification breakdown for the 35 U.S. city/county
air pollution indices.

 

Number of Variables No."l Percenta
l 5 14
2 4 l l
3 9 26
4 3 9
5 13 37
6 l 3
Calculation Method
A. Nonlinear 4 (3) ll (9)
B. Segmented Linear 13 37
C. Linear 7 20
D. Actual Concentrations ll (12) 32 (34)
Calculation Mode
1. Individual 12 (l4) 54 (40)
2. Maximum I5 43
3. Combined 8 (6) 23(17)
Descriptor Categories
A. Standards 6 17
B. Standards and Episode Criteria 19 54
C. Arbitrary 10 29

 

2‘Numbers in parentheses show the results when indices can be
classified two ways (see text)

air quality indices around, “people really can get confused
when they move from city to city.“ Many respondents felt it
better to familiarize the public with the scientific notation for
pollutant concentrations—ug/m“, ppm—than to teach the
public how to understand an index. Some agencies felt they
would adopt an index if they could find one that was under-
standable and agreeable to all, but that there is not yet a suf-
ficient scientific basis for such an index. Some of the reasons
offered for not adopting an air pollution index were purely
practical—not enough monitoring data to implement the
index or not enough staff to compute the index. The problem
of finding an index that is consistent with the public‘s per-
ception of air quality was cited frequently as a reason for not
using an index. One agency used ORAQI for several years but
found it did not correlate well with observed air quality and
resulted in many complaints; thus, it was abandoned:

We soon found that [index values] did not relate to
what the public saw. Many times, mountains surround-
ing [the city] would be barely visible during winter
months and yet the index would register in the light air
pollution range. This resulted in considerable contro—
versy concerning the validity of the index with the gen-
eral public and the news media. We went to consider-

.—
\JQK’O

m

147.

Number 01 agencies

Op—Nwhw

 

   

l 4 -' l I l I
1236567891011121314
Number of categories

Figure 4. Histogam of the number 01 descriptor categories used
in the 35 US. city and county air pollution indices.

 

 

 

 

 

able length to explain that the index was a combination
of a number of pollutant levels and that the gaseous
pollutants could be low and there could still be suffi-
cient fine suspended particulates in the air to obscure
visibility, resulting in a low combined index. After one
year of increasing problems with the combined index,
which included several attempts” to modify it, we dis-
continued its use.

Another agency suggested that the decision of whether or
not to use an index probably depends on the level of public
education, and that any approach is satisfactory as long as the
public fully understands it. Another agency felt that it was
preferable to forecast tomorrow’s pollutant values (or index
values) than to 'announce today’s or yesterday’s values, as
many indices are designed to do. A common reason given for
not using an index had to do with the lack of clear under»
standing of indices by the public, leading to confusion and
misunderstanding. One agency noted that most of the current
indices. because they are arbitrary, are really not interpreta~
ble. As such, this respondent felt most indices represent a
“nonunderstandable, nondimensional number." One agency
discontinued its index because the news media “sensation-
alized" it by reading more into the index than was intended:
”The index is now at 80 and when it gets to 100 you will have
to start worrying."

Criteria for a Uniform Index. Although some respondents
expressed much skepticism about the understandability and
meaningfulness of air pollution indices in general, particularly
about combined air pollution indices, there was general
agreement that prevailing confusion might be reduced if the
Federal Government were to develop, endorse, and support
a single, uniform air pollution index. Only two dissenting
arguments were offered against this concept: one respondent
felt that adoption of a uniform index would make comparisons
of air pollution levels in different cities too easy, creating
discord among the cities. Another respondent felt that each
index should be tailored to the public it serves and therefore
should be different in different cities. From the comments,
it appears that one obstacle to adopting any standardized
index is the fact that many cities already using indices would
have difficulty changing them to conform to some new index
system. The process of reeducating the public poses a difficult
problem which would need to be addressed. Another serious
problem lies in formulating an index that is truly satisfactory
to these agencies and to their public. From the viewpoints

Table IX. Frequency distribution of the 44
words used for descriptor categories.

 

Heavy l 1 Normal 2
Good 10 Stage I 2
Light 9 Stage 2 2
Moderate 7 Stage 3 2
Unhealthylful) 7 Very Heavy 2
Unsatisfactory 7 Warning 2
Emergency 6 Above Average 1
Poor 5 Acute 1
Alert 4 Average 1
Extremely heavy 4 Below Average 1
Fair 4 Endangerment 1
Medium 4 Extremely Poor 1
Satisfactory 4 Harmful 1
Severe 4 High 1
Clean 3 Red Alert 1
Extremely light 3 Significant 1
Very Poor 3 Slight l
Acceptable 2 Very Dangerous 1
Dangerous 2 Very Good 1
Excellent 2 Very Light 1
Hazardous 2 Watch 1
Low 2 Yellow Alert 1

 

28

expressed by respondents, it appears that any uniform air
pollution index would have to possess the following desirable
features:

1. Easily understood by the public. The index should trans-
form the scientific concentration units of each pollutant into
a nondimensional number which is easily understood by the
public. To facilitate comprehension of the index, the basis for
the transformation should be identical for each pollutant and
should relate to a nationally uniform set of air quality stan-
dards.

2. Include major pollutants and capable of including future
pollutants. The index should include all those pollutants
which have been identified as major pollutants by the Federal
Government and for which nationally uniform standards and
episode criteria have been established. Furthermore, the
structure of the index should be flexible so that when new
pollutant standards are set they can be included in the index
without modifying its basic form.

Table. X. lndices using the more commonly occurring words in
descriptor categories.

 

Multipollutant lndices lndices for Particulate Matter Only

 

New York City Detroit, Memphis, Oklahoma City3

Good Extremely Light
Acceptable Light
Unsatisfactory Medium
Unhealthy Heavy
Extremely Heavy
Minnesota, New jersey Chattanooga
Good Light
Satisfactory Moderate
Unsatisfactory Heavy
Unhealthful Alert
Miami
Good
Normal
Moderate
Heavy
Severe

 

aDiscontinued index.

3. Relate to ambient air quality standards and goals. The
index should relate to the NAAQS which have been estab-
lished for CO, S02, TSP, and 01.19 The index values should
also be an indicator of the relationship between air pollution
levels and these national air quality goals.

4. Relate to Federal Episode Criteria. The recommended
Federal Alert, Warning, and Emergency Criteria,20 and the
Significant Harm Levels,21 which have been established for
CO, 802, TSP, SOz-TSP product, N02, and 0,, should be an
additional basis for the index. This will provide a uniform
system of both public information and agency administrative
procedures during air pollution episodes.

5. Calculated in a simple manner using reasonable as-
sumptions. Pollutant concentrations should be easily con-
vertible into their corresponding index values using simple
equations or appropriate plots of the index values vs. pollutant
concentration. The index equation and/or curves should be
based on the relationships described in Criteria 3 and 4.

6. Based on a reasonable scientific‘premise. Because the index
should relate to the NAAQS and Federal Episode Criteria
(Criteria 3 and 4), its basis would be as solid as the scientific

basis on which the standards and episode criteria were es»
tablished.

Journal of the Air Pollution Control Association

Table X1. Summary of preferred structural characteristics of the
indices in 35 agencies.

 

 

Structural Number of
Characteristic Preferred Structure Agencies Percent
Number of variables 5 13 37
Calculation Method B. Segmented linear 13 37
Calculation Mode 2. Maximum 15 43
Descriptor Categories B. Standards and 19 54
Episode Criteria
Number 4 8 23
3 7 20
Words Good 10 36 3
Moderate 7 263
Unhealthy(ful) 7 223
Unsatisfactory 7 213

 

aPercent use in 28 indices.

7. Not inconsistent with perceived air pollution levels. Index
values become inconsistent with perceived air pollution levels
when, in combined or aggregated indices, high pollutant
concentrations are “masked” by lower concentrations. To
circumvent this problem, the index should calculate separate
values for each pollutant, thus enabling each index value to
be reported separately, if desired.

8. Spatially meaningful. The spatial meaningfulness of an
index depends on how the data are selected or manipulated
prior to the index computation. For example, the pollutant
concentration may be averaged over several monitoring sta-
tins, or the highest pollutant concentration from any station
in a city may be selected for the index calculation. The latter
method should be used because it prevents the masking of
high pollutant concentrations which occurs in the averaging
process and allows index values to be associated with locations
of specific monitoring sites.

9. Exhibit day-to-day variation. The structure of the index
should allow for noticeable variation from day to day (and
hourly if desired). When the index is calculated for each pol-
lutant separately (Criterion 7) and the maximum value re-
ported, variation also is possible in the pollutant reported.
Combined indices do not possess these important character—
istics. Reporting the maximum actual pollutant concentra-
tions or the ratio of the pollutant concentration to the NAAQS
gives pollutant variation. However, because both of these
systems report the index value as a fraction, neither exhibits
the large numerical variation necessary for the values to be
clearly understood by the public.

10. Can be forecast a day in advance (optional). When the
index is based on the NAAQS and episode criteria (Criteria
3 and 4), the pollutant concentrations are determined using

Table Xll. Breakpoints for SUAQI.

the Federal Reference Methods (FRM). However, the FEM
methods for $02 and TSP require 24 hr averaging times,
preventing the index from being reported more frequently
than once every 24 hr. More current information, or even index
forecasts, are of much greater interest and use to the public.
More current index values can be obtained using other
equivalent FRM monitoring methods having shorter aver-
aging times. A 24 hr index forecast will be difficult to make
without extensive meteorological data, but qualitative index
forecasting is possible using the National Weather Service’s
Operation Manual for Air Pollution Weather Forecasts.22 This
system can provide an 18 hr forecast of index values using the
following word descriptors: “remain the same," “decrease,"
or “increase."

Conclusion

The findings of this review are that a great diversity and
lack of consistency exist in the way air quality conditions are
reported to the public by means of air pollution indices. States,
provinces, and US. cities use daily informational indices
which differ strikingly from each other and differ from the
more complex, long—term trend indices that appear in the
scientific literature. 0f the 55 US. metropolitan air pollution
control agencies surveyed, 35 (64%) routinely use some form
of air pollution index. The metropolitan indices are simpler
in form than those appearing in the literature, but there are
dramatic differences from city to city. The index classification
system reveals l4 basically different index types among the
35 US. cities and six states currently using daily air pollution
indices. With two minor exceptions, when descriptor words
are taken into account, no two indices are exactly the some.
Each has a different mathematical formulation and a different
meaning to the public. Thus, an index value of 100 reported
in Washington, DC means something entirely different from
a value of 100 reported in Cleveland, OH.

The Federal Government is currently making a serious at-
tempt to standardize the methods by which air pollutants are
monitored. While air monitoring networks require substantial
public investments both in personnel and equipment, air
pollution indices are one of the most important ways by which
the public receives its air quality information from these
networks. Thus, lack of uniformity among different indices
creates serious problems. Not only does the diversity raise
questions about the meaningfulness of today's indices, but an
interested member of the public who travels from city to city
will readily become confused about air pollution levels in each
city. Informed members of the public, when they wish to
compare air pollution levels in different cities, cannot do so
using today’s air pollution indices. Regulatory officials, with
so many disparate indices in existence, are unable readily to
use these numbers to draw a national picture of air pollution
levels and trends. Further, the diversity itself suggests that

 

 

 

 

CO SOI TSP SO2 X TSP N0, Ox
SUAQI ppm ppm us/m3 (us/m3)“ ppm ppm
Averaging time (hr)
Federal level NA 8 24 24 NA 1 l

50% of primary NAAQS 50 4.5 0.07 1503 b b 0.04
Primary NAAQS 100 9 0.14 260 b b 0.08
Alert 200 15 0.3 375 65,000 0.6 0.10
Warning 300 30 0.6 625 261,000 1.2 0.40
Emergency 400 40 0.8 875 893,000 1.6 0.50
Significant harm 500 50 1.0 1,000 490,000 2.0 0.60

 

a Secondary NAAQS.

an NAAQS exists—SUAQI is not reported below the Alert level.

May 1976 Volume 26, No. 5

29

no consistent scientific rationale has been employed in de-
veloping these indices. Thus, the present lack of uniformity
among air pollution indices has at least three undesirable
consequences: (1) it creates potential confusion, (2) it raises
questions of technical validity, and (3) it prevents the indices
from being used to gain insight into national air pollution
problems and changes in these problems over time.

A potential solution to these problem is to develop a uni-
form air pollution index that could be applied in many dif—
ferent cities. An index which meets the 10 criteria above and
matches preferred characteristics of the indices currently in
use (Table XI) would have the following structural charac-
teristics:

- Includes six variables (CO, 802, TSP, 802 X TSP, N02,
0, ).

o Incorporates a segmented linear function.

Calculated using a maximum mode.

Uses the NAAQS, Episode Criteria, and Significant Harm
Levels as breakpoints.

- Includes four descriptor categories.

Figure 5 shows the segmented linear functions for such an
index, the Standardized Urban Air Quality Index (SUAQI),
which appears in the Compendium,“ Table XII gives the
SUAQI breakpoints. Four descriptor categories are used in
this index: 0—50, Good; 51—100, Satisfactory; 101—199, Un-
healthful; 200—500, Hazardous.

The SUAQI index shown here forms the basis for the Pol-
lutant Standards Index (PSI) now under consideration by
theFederal Government.” An early version of PSI proposed
by the authors” was nearly identical to SUAQI and contained
the above four descriptor categories. At the time of this writ-
ing, PSI has evolved to contain six descriptor categories, with
health effects language included for each category.10

We believe that the SUAQI/PSI formulation offers the
greatest chance for acceptance by state and local air pollution
control agencies. Adoption of a nationally uniform air pollu-
tion index should (1) facilitate comparisons of air quality levels
in different cities, (2) reduce confusion among residents of
cities about daily air quality levels, and (3) provide policy
makers with a uniform measure for evaluating the impact of
regulatory actions.

Acknowledgments

The authors appreciate the cooperation of the many state
and local air pollution control agencies that responded to our
survey, particularly those agencies that took the time to review
and comment on the Compendium and SUAQI. We also are
grateful for the guidance and support received from Dr. James
J. Reisa of the President’s Council on Environmental Quality.

 

 

l
0 250 ya/m3 TSP 500 750 1000 1250
0.2 ppm 502 or Ox 0.4 0.6 0.8 1.0
10 ppm CO 20 30 40 50
0.4 ppm NO; 0.8 1.2 1.6 2.0
262 393 524 655

131x103(;4 m3)2
SOZXTSP 8/

Figure 5. The SUAQI index functions.

30

We wish to thank Ms. Karen Shrum for typing the manuscript
and Ms. Irene Kiefer for editorial assistance.

References

1. G. C. Thom and W. R. Ott, “Air Pollution Indicea: A Compendium
and Assessment of Indices Used in the United States and Cana-
da," Joint Study sponsored b the Council on Environmental
Quality and the United States nvironmental Protection Agency,
Dec. 1975.

2. E. L. Stockton, “Uniform air quality index," J, Air Poll. Control

Assoc. 25: 978 (1975).

. W. F. Hunt, W. M. Cox, W. R. Ott, and G. C. Thom, “A Common
Air Quality Reporting Format: Precursor to an Air Quality
Index,” presented at the Fifth Annual Environmental Engi-
neering and Science Conference, Louisville, KY, March 1975.

. M. Reidy and C. Dziewulski, “Homogeneous Dissemination of

Ambient Air Quality Levels to the News Media and the Public

or the Need to Eschew Obfuscation," Paper No. 78-352, presented

at the 66th Annual Meeting of the Air Pollution Control Associ-

ation, June 1973.

“Planning for Environmental Indices," Report of the Planning

Committee on Environmental Studies Board, National Research

Council, National Academy of Sciences, Feb. 1975.

6. L. R. Babcock, “A combined ollution index for measurement

of total air pollution,” J. Air oll. Control Assoc. 20: 653 (1970)
. A. Thomas, L. R. Babcock, and W. B. Schults, “Oak Rid e Air

Quality Index,“ ORNL-NSF-EP-B, Oak Ridge National abo-

ratory, Oak Ridge, TN, Sept. 1971.

8. C. A. Bisselle. S. H. Lubore, and R. P. Pikul, “The National En-
vironmental Indices: Air Quality and Outdoor Recreation,"
MTR<6159, The Mitre Corporation, McLean, VA, 1972.

9. G. C. Thom and W. R. Ott, “Progress toward a uniform air pol~

lution index," J, Air Poll. Control Assoc. 25: 1157 (1975).

. “Guideline for Public Reportin of Daily Air Quality—Pollutant
Standards Index (PS1)," OAQ S No. 12-044, U.S. EPA, Office
of Air Quality Planning and Standards, Research Triangle Park,
NC, in preparation.

. L. Shenfeld, “Note on Ontario’s air pollution index and alert

system,” J. Air Poll. Control Assoc. 20: 612 (1970).

“M.U9.§8.)C. index tells Detroiters how dirty air is," Air Eng. 10(6):

28 (1 .

M. H. Green, “An air pollution index based on sulfur dioxide and

smoke shade," J. Air Poll. Control Assoc. 11: 703 (1966).

T. A. Rich, “Air pollution studies aided by overall air pollution

index," Environ. Sci. Technol. 1:796 (1967).

J. C. Fensterstock, K. Goodman, G. M. Duggan, and W. S. Baker,

“The Development and Utilization of an Air Quality Index,"

Paper No. 69-73, presented at the 62nd Annual Meeting of the

Air Pollution Control Association, June 1969.

T. L. Miller, “Short Time Averaging Relationships to Air Quality

Standards (STARAQS)—A Predictive Air Quality Index Model

for Use by Air Pollution Control Agencies," Paper No. 73-351,

presented at the 66th Annual Meeting of the Air Pollution Control

Association, June 1973.

. H. Inhaber, "A set of an gested air quality indices for Canada,"

Atmos. Environ. 9: 353 1975).

1973—1974 Directory of Governmental Air Pollution Agencies,

Air Pollution Control Association, Pittsburgh, PA, 1973.

. National primary and secondary ambient air quality standards,
40 CFR 50.

. Requirements for preparation, ado tion, and submittal of im-

plementation plans, Appendix L— xam le regulations for pre-

vention of air pollution emergency epis es, 40 CFR 51.

Requirements for preparation, adoption, and submittal of im-

plementation plans, prevention of air pollution emergency epi-

sodes, 40 CFR 51.16.

National Weather Service, Operations Manual, Air Pollution

Weather Forecasts, WSOM Issuance 75-15, Part C, Chapter 30,

July 16, 1975.

A Recommended Air Pollution Index [or Standardized Daily

Reporting, Report of the Federal Task Force on Air Quality In-

dicators, Council on Environmental Quality, in preparation.

G. C. Thom and W. R. Ott, “A proposed uniform air pollution

index,” Atmos. Environ. 10: 261 (1976).

6G

.a.

S"

>1

1

._.

12.
13.
14.
15.

16.

_.
9°

2 ,

.—

2 .

IO

23.

24.

 

Dr. Ott is Senior Systems Analyst for uality Assurance
with the Monitoring Technology Division, ffice of Research
and Development, U.S. Environmental Protection Agency,
401 M Street, S.W., Washington, DC 20460. Mr. Thom is a
doctoral de ree candidate in chemistry at The American
University, ashington, DC. He currently serves as a consul~
tant to the President’s Council on Environmental Quality,
722 Jackson Place, N,W., Washington, DC 20006.

 

 

 

Journal of the Air Pollution Control Association

APPENDIX B

GUIDELINE FOR PUBLIC REPORTING OF DAILY
AIR QUALITY--POLLUTANT STANDARDS INDEX (PSI)*

Preface
1. Executive Summary
2. Introduction
3. The EPA Recommended Daily Indicator--Pollutant Standards
Index (PSI)
3.1 Number of Pollutants
3.2 Calculation Method
3.3 Descriptor Categories
4. Reporting Procedures
4.1 Reporting the Index
4.2 Reporting the Federal Episode Criteria
4.3 Forecasting the Index
4.4 Flexible Media Reporting
5. Monitoring Requirements
5.1 Need for Monitoring Uniformity
5.2 Network Considerations
5.3 Measurement Practices and Reporting Frequencies
6. References
7. Appendix A. Information Needs for Forecasting PSI

*Prepared_by EPA Working Group To Develop an Air Quality

Index, August 1976.

31

PREFACE

The U. S. Environmental Protection Agency's recommended "Pollutant
Standards Index" (PSI) is the result of a joint effort on the part of
EPA's Offices of Research and Development, Air and Waste Management, and
Planning and Management. The guideline was prepared by the EPA Working
Group to Develop an Air Quality Index in response to a request from the
Federal Interagency Task Force on Air Quality Indicators of which EPA is
a member. The Federal Task Force, chaired by the Council on Environmental
Quality, was created as a result of a joint EPA/CEQ report1 which pointed
out existing problems resulting from the present diversity of indices used
in the United States and Canada.

This guideline suggests the use of the Pollutant Standards Index
(PSI) for those local and state air pollution control agencies wishing to
report an air quality index on a daily basis. The PSI places maximum
emphasis on protecting the public health; that is, it advises the public
of any possible adverse health effects due to pollution. In order to err
on the side of public safety, the index stresses reporting on the basis
of the stations with the highest pollutant concentrations and assumes that
other unsampled portions of the community will also experience high con-
centrations. In addition, its emphasis is upon acute health effects
occurring over very short time periods (24 hours or less) rather than
chronic effects occurring over months or years. It is not intended for,
and should not be used for, ranking urban areas in terms of the severity
of their air pollution problems. Such rankings require the use of many
other kinds of environmental data not incorporated in this index.

Finally, Appendix A discusses the meteorological information needs
of forecasting relative index changes. This was prepared by personnel
from the National Oceanic and Atmospheric Administration.

32

1. EXECUTIVE SUMMARY

This guideline suggests the use of the Pollutant Standards Index
(PSI) for those local and state air pollution control agencies wishing to
report an air quality index on a daily basis. The
document also includes appropriate monitoring and reporting guidance. The
guideline is the result of an earlier study1 showing that of all the air
quality indices in use today, no two are exactly the same. A potentially
serious problem of public confusion can occur in regions where neighboring
states and cities use different indices. The PSI also responds to the
request of several state and local agencies that the U. S. Environmental
Protection Agency provide them with a reconmended uniform air quality index.

The recommended index incorporates five pollutants--carbon monoxide,
sulfur dioxide, total suspended particulate, photochemical oxidants, and
nitrogen dioxide-—for which there are short-term (24 hours or less) health—
related National Ambient Air Quality Standards (NAAQS),2 and/or Federal Episodr
Criteria,3'5 and Significant Harm Levels.3’4’6 A sixth variable—-the product of
total suspended particulate and sulfur dioxide--is computed and is included in
the index equation. This variable and also nitrogen dioxide are treated
differently than the other pollutants because they have no short-term
NAAQS. Therefore, they are reported when they exceed the Federal Episode
Criteria and Significant Harm Levels. Because of the basic design of the
index, any further pollutant requiring NAAQS, Federal Episode Criteria, and
Significant Harm Levels can be readily added.

The index uses a "segmented linear function"* to convert each air
pollutant concentration into a normalized number. The NAAQS for each
pollutant corresponds to PSI=lOO, and the Significant Harm Level corresponds
to PSI=500.

At a minimum, PSI reports the pollutant with the highest index value
of all the pollutants being monitored, a dimensionless number, and a
descriptor word. On days when two or more pollutants violate their

 

*

A segmented linear function consists of two or more straight lines,
drawn between successive coordinates (”breakpoints") where each line may
have a different slope.

33

respective NAAQS, eacn of the pollutants should be reported. Five
descriptor words have been chosen to characterize daily air quality: "good,"
“moderate,“ ”unhealthful," "very unhealthful," and "hazardous." In

addition, for each descriptor word, generalized health effects and cautionary
statements are provided for use when the air is characterized as "unhealthful"
or worse.

For large metropolitan areas comprised of many smaller cities and
suburbs where significant air quality differences may exist, the air
pollution control agency may wish to report separate index values for
each community. This has the advantage of showing the public how air
pollution varies over the larger metropolitan area. The pollutants would
be monitored at population-oriented locations where the maximum concen-
tration for the particular pollutant is expected to occur, and the public
within each community would be made aware of the worst air quality to
which it is exposed.

Further guidance is given on the measurements practices and monitor
siting considerations (Section 5).

PSI should not be used to rank cities. An evaluation of PSI in
eight cities7’8 illustrated the difficulties of attempting to Compare air
quality levels in different cities using this or any other index. PSI is
designed for the daily reporting of air quality to advise the public of
potentially acute, but not chronic health effects. To properly rank the
air pollution problems in different cities, one should rely not just on
air quality data, but should include all data on population characteristics,
daily population mobility, transportation patterns, industrial composition,
emission inventories, meteorological factors, and the spatial represena
tativeness of air monitoring sites. A correct ranking should also consider
the number of people actually exposed to various concentrations, as well
as the frequency and duration of their exposure.

Adoption of PSI should reduce the confusion due to the existence of
many indices. PSI has several advantages: (1) it is simple and can be
-easily understood by the public, (2) it can accommodate new pollutants,

34

(3) it is based on a reasonable scientific premise, (4) it relates to
NAAQS, Federal Episode Criteria, and Significant Harm Levels, (5) it
exhibits day-to-day variations, and (6) a qualitative trend in the index
can be forecast for periods up to a day in advance, especially during
episodic conditions.
2. INTRODUCTION

A major area of concern in the field of air pollution control is how
to best report daily air quality to the public. A recent CEO/EPA Report1
indicates that of the 55 largest U. S. metropolitan air pollution control
agencies, 33 use an air pollution index. In addition, five states and
two Canadian Provinces operate state-wide (or Province-wide) index systems.
With two minor exceptions, no two indices were found to be exactly the same.
The public confusion generated by the use of so many indices is particularly
evident in bordering states using different indices. Therefore, there is
a need to develop a uniform index to report the daily status of air

pollution.

A recent paper9

emphasizes the need for a truly meaningful index to
have a sound scientific basis. The paper suggests that such an index
be based on the relationship between pollutant concentration and adverse
health (welfare) effects-—that is, a "damage function ” Unfortunately, it is
an extremely complex undertaking to relate measured air pollutant concen-
trations to the many diverse effects of air pollution--for example, aggra-
vation of disease in susceptible people, increased incidence of respiratory
illness in healthy persons, impairment of human motor function, reduced
visibility, corrosion of materials, and soiling of buildings. Arriving
at an air quality standard for a given pollutant--which is just one point
in a damage function-—has required vast quantities of data, medical
advisory conmittees, detailed epidemiological studies, and other extensive
research. The air quality criteria documents published for the major air
pollutantslo'14 reflect the complexity of the process.

The recent paper9 also emphasizes the importance of an index

accounting for the adverse effects associated with combinations of

35

pollutants--that is, synergism. For example, the criteria document on
sulfur oxidesH states that adverse health effects attributable to sulfur
oxides are intensified in the presence of particulate matter. Under-
standing synergistic effects adds greatly to the problem of obtaining a
truly meaningful air quality index. These problems stress the need for
additional research to develop pollutant—related damage functions that
take into account synergistic effects on health and welfare.

As an interim solution to these problems, this guideline recommends
a uniform index to report daily air quality, along with appropriate
monitoring guidance, This index will serve until a more meaningful air
quality index can be created. If adopted, a uniform index should end the
confusion associated with the use of many varied indices.

3. THE EPA RECOMMENDED DAILY INDICATOR--POLLUTANT STANDARDS INDEX (PSI)

The Pollutant Standards Index (PSI or w) is the result of a joint
effort by EPA's Offices of: Research and Development, Air and Waste
Management, and Planning and Management. Its evolution has included
formulation of several candidate index structures,]5’16 and the index
has undergone an extensive review process involving state and local
air pollution control agencies, public organizations, and media repre-
sentatives.

The recent CEO/EPA compendium of air pollution indices1 developed an
“index classification system” to analyze and compare the various indices
used by state, Provincial, and local agencies. Indices were categorized
according to four criteria: (1) number of pollutant variables measured,
(2) calculation method used to compute the index, (3) descriptor categories
reported with the index, and (4) method of reporting (whether it is
“combined," "maximum," or "individual").

The report found that the greatest number of the indices in use1
incorporate five of the six National Ambient Air Quality Standard (NAAQS)
pollutants (hydrocarbons are excluded because there are no direct health
veffects associated with the pollutant. It is controlled because it is a

36

precursor to the formation of photochemical oxidants.); (2) use a
segmented linear function*; (3) are based on the maximum of one of the
pollutant variables; and (4) use three to five descriptor categories.

In the following sections, the structure of PSI is presented accord-
ing to the "index classification system” categories.

3.l Number of Pollutants
PSI includes five pollutants: carbon monoxide (CO), sulfur dioxide

(502), total suspended particulate matter (TSP), photochemical oxidant (03)
and nitrogen dioxide (N02). Primary (that is, health related) NAAQS, and/or
Federal Episode Criteria, and Significant Harm Levels exist for all five.
In addition, one pollutant product TSPxSOZ is included becags: it has

’ As with
N02, which has no short-term primary NAAQS, the product is reported when
the Federal Episode or Significant Harm Levels are exceeded. Finally,
because of the structure of the index, any pollutant identified in the
future for which NAAQS, Federal Episode Criteria, and Significant Harm
Levels are adopted can be added without modifying the basic form of the
index.

both Federal Episode Criteria and a Significant Harm Level.

3.2 Calculation Method

A segmented linear function is used relating actual air pollution
concentrations to a normalized number. For example, PSI (w) equals l00
when the NAAQS for each pollutant is reached, while u equals 500 when the
Significant Harm Level for each pollutant is reached. The normalized
number should be easier for the general public to understand because it
does not require one to know specific NAAQS concentrations or the many
different Federal Epsiode and Significant Harm Levels.

The index breakpoints are listed in metric units (Table l) and in
parts per million (Table 2). The first breakpoint separates the descriptor

 

*

A segmented linear function consists of two or more straignt lines,
drawn between successive coordinates ("breakpoints") where each line may
have a different slope.

37

82

 

 

TABLE 1. Breakpoints for PSI (¢) in Metric Units
PSI TSP S02 C0 03 N02
Va1ue ug/m3 ug/m3 TSPxSOz mg/m3 ug/m3 ug/m3
Breakpoints (w) 24—hr. 24-hr. (ug/m3)2 8 hours l-hr. l—hr.
50:; of primary short- 50 75a 80a b 5.0 80 b
term NAAQS
Primary short-term NAAQS 100 260 365 b 10.0 160 b
A1ert Leve1 200 375 800 65x103 17.0 400c 1130
Warning Leve1 300 625 1600 261x103 34.0 800 2260
Emergency Leve1 400 875 2100 393x103 46.0 1000 3000
Significant Harm Leve1 500 1000 2620 490x103 57.5 1200 3750

 

aAnnua1 primary NAAQS.

bNo index va1ue reported at concentration 1eve1s be1ow those specified by the A1ert

1eve1 criteria.

c .
.For the PSI index 400 ug/m3 appears to be a more consistent breakpoint between the
descriptor words ”unhea1thfu1" and "very unhea1thfu1” than the

200 ug/m3.

03 A1ert Leve1 of

68

TABLE 2. Breakpoints for PSI (w) in Parts Per Mi11ion

 

 

PSI 502 TSPXSOZ c0 03 N02
Breakpoints Va1ue 24-hr. (pg/m3x 8 hours 1-hr. 1-hr.
(w) ppm)
50% of primary NAAQS 50 .036 b 4.5 0.04 b
Primary NAAQS 100 .14 b 9.0 0.08 b
A1ert Leve1 200 .30 22.727 15.0 0.20 C 0.60
Warning Leve1 300 .60 91.259 30.0 0.40 1.20
Emergency Leve1 400 .80 137.413 40.0 0.50 1.60
Significant Harm Leve1 500 1.00 171.329 50.0 0.60 2.00

 

aAnnua1 primary NAAQS.

bNo index va1ue reported at concentration 1eve1s be1ow those specified by the

A1ert Leve1 criteria.

CFor the PSI index 0.2 ppm appears to be a more consistent breakpoint between the descriptor

words Punhea‘thfu1" and ”very unhea1thfu1“ than the

0.1 ppm.

03 A1ert Leve1 of

categories ”good" and ”moderate.” For C0 and 03, the first breakpoint was
chosen at 50 percent of the primary NAAQSs. In the case of TSP and $02,
concentrations equal to their respective primary annual NAAQS were chosen
because the frequent occurrence of values greater than these concentrations
could lead to violations of their respective annual NAAQS. In an area where
a violation of either the annual primary TSP or SO2 standard occurs, approxi-
mately 50 percent or more of the days will thus be classified as ”moderate“
or worse. This approach minimizes the potential for public confusion which
might arise from a preponderance of days reported as “good," followed by
the report that the annual health-related standard has been violated.
The breakpoints between the primary NAAQS and Significant Harm
Levels are somewhat arbitrarily set at the Federal Episode Alert, Warning,
and Emergency Levels, except for oxidants. In the case of oxidant, 400 ug/m3
was used as the PSI breakpoint for the descriptor words ”unhealthful”
and "very unhealthful“ because it appears to be more consistent with the
descriptor words than the suggested administrative Alert level of 200 ug/m3.*
Figures 1 through 5 show the segmented linear function for each of the
NAAQS pollutants, and Eigure 6 showsthe function for the product of TSP and
$02. If NAAQS for new pollutants are adopted in the future, they can be
accommodated by drawing a new segmented linear function.

3.3 Descriptor Categories

PSI is primarily a health related index as shown by the descriptor
words: ”good,” ”moderate," "unhealthful,” “very unhealthful,“ and "hazardous,”
(Table 3). The breakpoints used to separate these descriptor words are
somewhat arbitrary. 0n the basis of health effects data above, it is not
possible to establish a sharp demarcation between any two descriptor words.
However, when the five pollutants were examined in the context of
severity of health effects, their NAAQS and EPA suggested
administrative Alert, Warning, and Emergency levels tended to
provide convenient breakpoints, except for the oxidant Alert

3
level, which was replaced with 400pg/m , as discused earlier.

 

*Several air pollution control agencies are using 400 ug/m
instead of 200 pg/m3 as their Alert level with concurrence
by the Environmental Protection Agency.

40

H7

TABLE 3. COMPARISON OF PSI VALUES WITH POLLUTANT CONCENTRATIONS, DESCRIPTOR WORDS
GENERALIZED HEALTH EFFECTS, AND CAUTIONARY STATEMENTS

 

 

POLLUTANT LEVELS

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

TSP 802 CO 03 N02 HEALTH
INDEX AIR QUALITY (24—hour), (24-hour), (8-hour), (1-hour). (1»hourl, EFFECT
VALUE LEVEL ug/m3 uglm3 mg/m3 ug/m3 pig/mal DESCRIPTOR GENERAL HEALTH EFFECTS CAUTIONARY STATEMENTS
——- 500 ———5'GzLFF:§AANT 1000 2520 57.5 1200 3750
Premature death of ill and elderly. All persons should remain indoors,
Healthy people will experience ad- keeping wrndows and doors closed.
verse symptoms that aflect their All persons should Minimize physi-
normal activity. cal exertion and avord traffic.
—-400 —— EMERGENCY 875 2100 46.0 1000 3000—— HAZARDOUS
Premature onset of certain diseases Elderly and persons with existing
in addition to significant agyava- diseases should stay indoors and
(ton of symptoms and decreased avord physical exertion. General
exercise tolerance in healthy persons. population should avoid outdoor '
actmty.
~— 300 ——WARNING 625 1600 34.0 800 2260
Significant aggavation of symptoms Elderly and persons with existing
VERY and decreased exercise tolerance in heart or lung disease should stay
UNHEALTHFUL persons with heart or lung disease, indoors and reduce physical
wash widespread symptoms in the activniy.
healthy population.
— 200 —— ALERT 375 800 17.0 400‘: 1130
Mild awavation of symptoms in Persons with existing heart or ‘
susceptible persons, with irritation respiratory ailments should reduce
UNHEALTHFUL symptoms in the healthy popula» physical exertion and outdoor
tron. actsvrlv
>——‘100 —— NAAOS 260 365 10.0 160 a
; MODERATE
l
i——- 50 -——-50% 0F NAAOS —l— 755 Bob 5.0 80 a
l
I GOOD
L— 0 0 0 0 n a

 

 

 

 

 

 

 

 

 

aN0 index values reported at concentration levels below those specrfied by "Alert Level" criteria.

bAnnual primary NAAQS.

c400 yg/m3 was used instead ol the 03 Alert Level 01200 ug/m3 (see text).

 

 

 

 

 

 

 

 

 

 

 

 

     

 

 

 

 

 

 

 

 

     
 

 

 

 

m - m ;
su I I I I
y-—I —‘ F— -
tuucncv muctncv
m -—- uvu. a — M - mm m —-
5 mama 3 mama
1' m— uvn u -—- E Jfllr— mm m -
ALERT nu!
ma ._ uvu 11 — m — uvu m ""
._ —I _ _
"II —- W mm" mun: — Inn —- m mm" ..
MAME
-— snmmuuuu — ~7’ ”mun. rammv "‘
nuns ' .—
. I l I I I . I . l I I
o In In n In a u I m m m no mo nu
CAIIDI IDIOIIUI (Ia-1 luflflmfl AVIIAGI). Did-1 “BK-DH! IAI'ICUIATI HAITI! Rum IUNNIDI IVIIABK LIID’)
“‘° I I I I I '"
_ ——I '_ _.
m IMEIBENCV
«a __ mama um. um ._ ’_ “V“ Ion —
- 3 manna
* Intanmcv _._ _
E 390— um. um __ 2 1'“ mm. m _
m _. Am" uvn m ._ 1" — m '—
n— —. b. I—
m III
M —- N5 mun" nuns - ’— nu“)? ’—
" ...n m nuns .—
AMIUAL IIIMAIV NAADS
. I I I L I ., IL I I l I
a no me me am mu m ‘“ m m “M '1“
Inn: A a
SULFUR DIOXIDE [luau RUNNING AVIR‘GUJIVN’ ‘ I, u AVER‘CELHVI
miners“
m -- mu _
an — .—
5 __
5 1°. '— _ g mum
E an — um. .—
All"
I” “ mm nu -- nut
m - _.
um. I:
m _. _
m — .—
. I I I I I
I ma mo am «no ma no. u L I I I I
I no. no m m In at
Imam mono! In... "man. an!
rout wrung unncuun I so; nun-Hunt unusual"). mum
Figures 1—6. PSI Functlons
. . I, . .
Source: Env1ronmental Protectlon Agency, Guldellnes for

Public Reporting of Daily Air Quality —— Pollutant

Standards Index"

42

Air pollution levels between the short-term primary NAAQS and the Alert
level for TSP, 502, and CO and 400 ug/m3 for 03 are deemed "unhealthful,"
because mild aggravation of respiratory symptoms in susceptible persons and
irritation symptoms in the healthy population occur at some point above the
short-term primary NAAQS and at and below the Alert levels for TSP, 502, and

C0 and 400 ug/m3 for 03.10'13 N02 is not reported until concentrations exceed

the Alert level because no short-term NAAQS has been established.14 Air
pollution concentrations above the Alert level but below the Warning level
are classified as "very unhealthful," while concentrations above the Warning
level are "hazardous.”

These classifications are related to generalized health effects and
appropriate cautionary statements (Table 3).17 A single set of generalized
health effects and cautionary statements is indicated for the descriptor
words "unhealthful" and "very unhealthful." The “hazardous" category has
two sets of generalized health effects and cautionary statements. The
first set is reported when the index exceeds 300 and the second when the
index exceeds 400 indicating the increasing severity of the air pollution
levels.

‘ In the case of TSP and $02, short—term secondary air quality standards
also exist below their primary NAAQS. Secondary standards are designed to
protect against the adverse effects of pollution on the public welfare (animals
vegetation, materials, visibility, etc.), According to PSI, if their short-
term secondary NAAQSs are violated, the concentrations would be classified as
"moderate" or worse. While this descriptor word is valid from a health view-
point, the air quality is unsatisfactory from the standpoint of welfare
effects. Because PSI is a health-related index, the user may wish to report
on the possible welfare effects when either the short-term TSP or $02 NAAQS
is violated.

4. REPORTING PROCEDURES

PSI has been designed to be as flexible as possible in allowing air
pollution control agencies to decide for themselves the information to
include in their reports to the various media. This section examines the
recommended method of reporting the index, the reporting of the Federal
Episode Criteria, and the concept of flexible media reporting.

43

4.l Reporting the Index
Since each pollutant is examined separately by comparing its measured

concentration with the NAAQS, the Episode Levels, and the Significant Harm
Level, each pollutant can be reported separately. At the minimum, the

pollutant with the highest index value should be reported to advise the
public of the worst air pollution to which it is exposed. On days when
two or more pollutants violate their respective NAAQS——that is, have
PSI values greater than lOO--then each of the pollutants should be reported.
The index values of the other pollutants may also be reported for complete-
ness. When the air pollution level is reported as “unhealthful,” “very
unhealthful," or ”hazardous," cautionary statements should also be used.
In addition, the generalized health effects can be used.

Users of PSI may wish to report on the health effects of each pollu-
tant individually, thereby providing more detailed language on each
pollutant than is available in Table 3. In preparing such information for

the public, the user is encouraged to seek appropriate medical advice
and to consult the literaturem'14

4.2 Reporting the Federal Episode Criteria

When the Federal Episode Levels for each pollutant are exceeded, the
user should report the administrative actions associated with the Alert,
Warning, or Emergency Levels. The issuance of administrative actions
depends, of coarse, upon the forecast of meteorological conditions affecting
future pollution levels.

Issuance of administrative actions also apply to the product of TSP
and $02, which has both Federal Episode Criteria and Significant Harm
Levels.2’3 Although available health effects information has not been
codified to tie the descriptor words to the product of TSP and $02, the
product is included for purposes of administrative completeness.

4.3. Forecasting the Index

The forecasting of a quantitative index for periods up to a day in advanc<
would be difficult without extensive meteorological data and specialized exper
tise that some air pollution control agencies may not possess. However,
qualitative index forecasting is practicable using the National Weather

44

Service's Air Pollution Weather Forecast Program.]8’19 With this weather

information, along with available emissions and air quality trend data,
agencies can develop techniques or procedures to forecast the relative
change in the index by using the following word descriptors: No signifi-
cant change, decrease, or increase. The principal responsibility for
obtaining the necessary emission and air quality information lies with
the air pollution control agency using the index. The air pollution con-
trol agency would integrate the meteorological information and apply the
predictive methods to generate the forecast. The information needs for
forecasting relative index changes is discussed further in Appendix A.

4.4 Flexible Media Reporting

The index has been designed to be as flexible as possible in reporting
the status of air quality to the public. Either short or long reports can
be used. For television, the report could read, “Today the air pollution
index is 50, the air quality is good." However, when the air pollution
becomes unhealthful, then several possible reports could be considered
for television, the news media, or telephone recordings. For example,
when oxidant pollution reaches a concentration of 280 ug/m3 (0.14 ppm), the
report could take several different forms.

(1) Today, the air pollution index is 150. The air is "unhealthful."
The pollutant 03 is responsible.

(2) An air pollution alert has (or has not) been called based on
the forecast for the remainder of the day (and/or) tomorrow.

(3) Repeat the above and add the following cautionary statements:
”Persons with existing heart or respiratory ailments should reduce physical
exertion and outdoor activity."

(4) The report could include everything said in (l), (2), and (3) and
then add that ”unhealthful" air can cause "mild aggravation of symptoms
in susceptible persons, with irritation symptoms in the healthy population."

(5) Finally, the report could conclude with the forecast of tomorrow's
air pollution level, such as "no change in the air pollution level is
expected.”

Table 3 should be referred to in preparing the air pollution status
reporttm theinmlic. Figure 7 illustrates the above ozone example

45

“RHEALTHFUL

 

PSI = 150
POLLUTANT: Oxidants
TODAY’S HEALTH IMPLICATIONS: Respiratory ailment

and heart disease patients should reduce exertion
and outdoor activity.

FORECAST: No change.

Figure 7. Example of possible PSI report for television.

46

by showing possible reports for television and newspaper. The
figure provides essential information, indicating the PSI value,
the critical pollutant, the health implications for the public,
and the next day's forecast. Each of the descriptor categories
has been given equal weight. The information is displayed so

that it can be presented as rapidly as possible in an easy-to-
understand format.

5. MONITORING REQUIREMENTS

5.1 Need for Monitoring Uniformity

 

In order for PSI to be readily accepted, the data used in calculating
the index must be comparable from site to site within a region. Since
these data are to be obtained at existing air monitoring sites, certain
easily implementable practices can eliminate considerable variability in
the data. Among these are using: (1) uniformity of site types--that is,
residential, commercial, etc.; (2) Federal Reference Methods or their
equivalent; (3) standardizing sampling height and probe exposure; and (4)
good housekeeping and quality control procedures to provide high quality
data.

5.2 Network Considerations
Air pollution control agencies need not undertake additional monitoring
requirements in the implementation of PSI, but can simply select sites
from their existing network. The sites selected, however, should generally
meet two basic criteria: (1) sites should be representative of population
exposure-~that is, not unduly influenced by a single emission point or
background-oriented, and (2) sites should be located in areas of maximum
concentration for the pollutant of interest, but should not be unduly in-
fluenced by any single source. Areas suitable for monitoring, by pollutant are
- TSP - populated areas substantially downwind of large sources or in
the midst of numerous area sources.
- $02 - populated areas substantially downwind of large sources or in
the midst of numerous area sources.

47

- C0 - densely populated, high-traffic volume areas, including areas
in the center city.

- 03 - populated areas substantially downwind of areas of maximum
hydrocarbon emissions density, such as the central business district.
The site should be 100 meters or more removed from major traffic
arteries or parking lots.

- N02 - populated areas downwind of areas of high traffic density.

If a pollutant(s) is (are) measured at several locations within a
metropolitan area, it would be desirable (if possible) to base the index on
the site showing the highest reading on a given day. This would mean that
different sites would be used on different days.

For large metropolitan areas comprised of many smaller cities and
suburbs where significant air quality differences may exist, the air
pollution control agency may wish to report separate index values for
each community. This has the additional advantages of showing the public
how air pollution varies over the larger metropolitan area. Furthermore, for
example, the photochemical pollutants tend to be higher in the suburban fringe.

5.3 Measurement Practices and Reporting Frequencies

5.3.1. Use of Federal Reference Methods

Since PSI is based on the NAAQS, the Federal Reference Methods (FRM) or
equivalent should be used where possible. Such methods are consistent with
the averaging time of the primary standards. Further, continuous methods

should be used, where possible, to facilitate the reporting of the index
numbers two or three times per day.

5.3.2 Carbon Monoxide, Nitrogen Dioxide, and Ozone

The FRM for C0 is based on the nondispersive infrared measurement
principle. The proposed method for N02 and the existing method for 03
employ the chemiluminescence measurement principle and give continuous
data. A FRM or equivalent method for CO, N02, and 03 must alga meet
performance specifications set forth in the Federal Register.

5.3.3 Sulfur Dioxide

The FRM for $02 is the pararosaniline 24-hour bubbler method. The
solution may be analyzed automatically or manually at the central labora—
tory. Serious logistics problems can arise if an index number must be

48

calculated from multiple sites two or three times per day. Fortunately,
there are procedures for designating continuous SO2 analyzers as equivalent
to the FRM,20
Therefore, the use of the continuous SO2 analyzer is recommended to

collect the data used in the index. If one is not available, then a
pararosaniline 24-hour bubbler method can be used if several precautions
are taken. To prevent deterioration in the sample, the sample should be
collected at ambient temperature or no warmer than l5°C if ambient tempera-
tures are below freezing.' The sample should then be analyzed as soon as
possible, with no later than a six-hour.delay from end of sampling to
analysis.

and from these 24-hour running averages are easily obtained.

5.3.4 Total Suspended Particulate

The FRM for TSP uses a high-volume sampler and specifies a midnight-
to-midnight 24-hour sample followed by a 24-hour equilibration at a
relative humidity less than 50 percent. This leads to a two-day delay
in the reported value. .For index reporting, the simpleuzmodification to
the FRM is to make thezsampling time more convenient—-that is, 8 a.m.-to—
8 a.m. or noon-to—noon, etc. The sample could be weighed immediately to
provide a TSP value for the index. Later a true value could be calculated
after the recommended equilibration time of 24 hours. A study in EPA
Region IV has shown that the true TSP values are usually within 10
percent of the values measured immediately after collection.21 The true
value would be recorded as the correct one, reported to the National Aero-
metric Data Bank, and used to calculate annual averages and maxima.

5.3.4.1 Staggered high-volume sampler measurements

During episode conditions, the air pollution control agency may
find it necessary to infonn the public of existing conditions two or three
times per day. Therefore, several high—volume samplers could run for
24 hour periods staggered every 4 to 6 hours throughout the episode.
The sample could be weighed immediately, and that weight used in
deciding what action should be taken concerning the possible

49

emergency. Then the filter would be equilibrated for 24 hours and
reweighed.

5.3.4.2 Alternative measurements

The paper tape sampler and the integrating nephelometer can be used
to indicate the need for overlapping high-volume sampler measurements.
The paper tape sampler has been used in most previous indices and has both
Federal Episode Criteria and a Significant Harm Level. The Coefficient
of Haze (COH) value from the paper tape sampler, however, is poorly
correlated with TSP levels. In addition, the paper tape sampler has not
been determined to be an "equivalent method” to the FRM. Therefore, its
use should be limited to index repOrting and must not be used to determine
compliance with the NAAQS for particulate matter.

A newer instrument relatively untested in routine field applications
is the integrating nephelometer. It measures the scattering of light
from small particles and correlates well with visibility and TSP measure-
ments. Both the paper tape sampler and the nephelometer can produce a
running 24-hour value which can be used as a qualitative indicator of TSP
loadings in the atmosphere.

5.3.5 Frequency of Reporting and Appropriate Averaging Times

The frequency of reporting is left up to the agency, within these
suggested ranges. It may be desirable to report the index once a day
but probably not more than three times per day. Because the high-volume
sampler has a 24-hour averaging period, agencies might consider operating
two or more high-volume samplers at the same station but with off-set
time periods, ending between 8 a.m. and 6 p.m. to provide reporting infor-
mation during the most desirable period.

If the agency desires, the paper tape sampler or integrating nephelo-
meter could be used in conjunction with the high—volume sampler to provide

50

estimates of the most recent ambient particulate loading. Thus used, the

paper tape sampler provides some guidance on whether or not to undertake
more intensive measurements during high air pollution levels.

Appropriate averaging times for which the index should be tabulated

and reported for each pollutant are:

6.

y“:
.

- TSP - TSP values taken with the high-volume sampler are discrete

24-hour values. Monitoring and data collection should be on a
schedule consistent with the agency's need to report the air quality
index. Other overlapping times may be employed by those agencies
wishing to report more than one index value per day.

- $02 - The suggested reporting value is the most current 24-hour

running average since the last reporting period.

- C0 - Although there are two standards for CO (8 hours and l hour),

the 8-hour standard is usually considered the limiting one and will
be the one violated in the vast majority of cases. The most
current 8-hour running average since the last reporting should be
used. In addition, the agency could also report the index value
associated with the highest 8—hour average during the reporting
period.

- 03 — The suggested reporting value for O3 is the highest hourly

value since the last reporting period. The reporting periods are
usually 24 hours or shorter.

~ N02 - Although the standard for N02 is an annual one, there are

hourly values associated with episode criteria; therefore, using

the highest hourly value since the last reporting period is recom-
mended.

REFERENCES

Thom, Gary, and Wayne R. Ott. Compendium Analysis, and Review of
United States and Canadian Air Pollution Indices, joint study by the
U. S. Environmental Protection Agency and the Council on Environmental
Quality, December 1975.

Federal Register, Vol. 36, April 30, l97l, pp. 8l86-820l.

51

l0.

ll.

12.

13.

14.

15.

16.

17.

Federal Register, Vol. 36, November 25, l97l, pp. 22390-224l4.
Federal Register, Vol. 36, December 17, l97l, p. 24002.
Federal Register, Vol. 36, March l3, l974, p. 9672.

Federal Register, Vol. 40, August 20, l975, pp. 36330-36333.

Ott, Wayne R. and William F. Hunt., Jr. "A Quantitative Evaluation

of the Daily Air Pollution Index Proposed by the U. S. Environmental
Protection Agency.” Presented at the 69th Annual Meeting of the Air
Pollution Control Association, Portland, Oregon, June 1976.

Hunt, William F., Jr., and Wayne R. Ott. Pollutant Standards Index
(PSI) Evaluation Study, Joint Office of Air and Waste Management and

Research and Development Report, U. 5. Environmental Protection Agency,
April 1976.

 

Hunt, William F., Jr., William M. Cox, Wayne R. Ott, and Gary Thom.
A Common Air Quality Reporting Format, Precursor to an Air Quality
Index, presented at the Fifth Annual Environmental Engineering and
Science Conference, Louisville, Kentucky, March 3—4, l975.

Air Quality Criteria for Particulate Matter, USDHEW, PHS, CPEHS,
NAPCA, WaShington, D.C., January I969, No. AP-49.

Air Quality Criteria for Sulfur Oxides, USDHEW, PHS, CPEHS, NAPCA,
Washington, D.C., January 1969, No. AP-SO.

Air Quality Criteria for Carbon Monoxide, USDHEW, PHS, CPEHS,
Washington, D.C., March l970, No. AP-62.

Air Quality Criteria for Photochemical Oxidants, USDHEW, PHS, CPEHS,
Washington, D.C., March l970, No. AP-63.

Air Quality Criteria for Nitrogen Dioxide, EPA, APCO, Washington,
D.C., January l97l, No. AP-84.

Thom, G.C. and W.R. Ott, Atmospheric Environment, lg, 26l(1976).
Thom, G.C., W.R. Ott, W.F. Hunt, and J.B. Moran. "A Recommended
Standard Air Pollution Index,” presented at l7lst National Meeting
of the American Chemical Society, New York, N.Y., April l976.

Knelson, John H., U. S. Environmental Protection Agency, memorandum

to Raymond Smith, U. 5. Environmental Protection Agency, December l5,
1975.

52

18.

19.

20.
21.

National Weather Service, Operations Manual, Air Pollution Weather
Forecasts,.WSOM Issuance 75-l3, Part C, Chapter 30, April 1975.

National Weather Service, Technical Procedures Bulletin No. 122: Air
Stagnation Guidance for Facsimile and Teletype (3rd Edition),

October 21, 1974. (Supersedes previous TPB's Nos. 52, 58, and 69).
Federal Register, Vol. 40, February 18, l975, pp. 7049-7070.

Helms, G.F. U. 5. Environmental Protection Agency, Region IV,
Atlanta, Georgia. Personal communication, December 1975.

53

APPENDIX A

INFORMATION NEEDS FOR FORECASTING PSI

INTRODUCTION

The information needed to qualitatively forecast the
Pollutant Standards Index (PSI) is of two types: (1) pollutant-
related and (2) meteorological. The pollutant-related
information may include data on source locations, physical
source characteristics and emissions, atmospheric—physio-
chemical transformation processes, and actual air quality
measurements and trends. Meteorological information that may be
included are data on synOptic weather features, on meteorological
parameters indicative of the dispersive capability of the
lower atmosphere, and of the photochemical potential. It
might also include information on the effect of local
terrain complexities upon meteorological parameters.

Together, pollutant-related and meteorological information
form the input to locally tailored predictive techniques

such as mathematical models, statistically derived methods,
or other techniques that may be applied along with subjective
judgment to some degree.

The necessary pollutant-related information is to be
obtained by the air pollution control (APC) agency having
local responsibility for issuing the Index. The National
Weather Service (NWS) is the primary agenCy supplying the
needs of APC agencies for meteorological information. NWS ser-
vices include issuance of advisories on air pollution
potential and air stagnations. However, some APC agencies
and/or their consultants may also collect and interpret

meteorological information to supplement that available
from the NWS.

GENERAL DATA NEEDS

The types and amounts of pollutant-related information
needed will vary depending on the particular pollutant(s)
of concern and the source to monitoring site configurations
in the particular geographical area. For example, in the
Los Angeles Basin, photochemical oxidant is the primary
pollutant of concern and since precursor sources (mainly
mobile) are widespread, the potential for maximum impact
exists over a rather large area. In contrast, in Pittsburgh
and Birmingham where suspended particulate matter from

54

industrial ferrous emissions will most likely cause elevated
pollutant levels, the maximum impact will probably be more
localized; thus, pollutant-related information may not

have to be as extensive. It is also important to know the
diurnal, weekly, and seasonal characteristics of emissions.
For instance, carbon monoxide concentrations are closely
associated both spatially and temporally with automobile
emissions. Typical diurnal patterns reflect morning and
evening peaks in vehicular traffic. High concentrations may
shift weekly in response to changes in workday versus weekend
automotive travel patterns. Seasonal patterns may shift

in some areas with vacation travel.

Generally, an up—to-date emissions inventory should be
available for communities where PSI is to be utilized in
order to adequately assess the source to monitoring site
impact relationships. For point sources (usually > 100 tons/
year of a pollutant) information should include the source
location, pollutants emitted, emission rates, and stack
parameters. Area source data, including lesser point
emissions, are not normally as specific. Available area
emissions, in tons per year, are usually quantified by city
or county. Vehicular emissions may be estimated by com-
bining local traffic pattern information with documented
vehicle—fleet emissions rates. These emissions data are
available from the EPA National Emissions Data System
(NEDS), state planning agencies, and private sources. It
may be necessary to supplement these data with emissions
information affecting the various temporal cycles; for
instance, information on the normal operating schedules of
large point sources and on traffic volume cycles in con-
gested areas.

Trends in the concentrations of pollutants can also
be useful in predicting the PSI. Trend information
might include the day—to-day variation in peak
hourly values or 24-hour averages. Trends data should
always be evaluated relative to changes taking place or
anticipated in emissions or meteorology. Persistence of a
trend would especially aid in arriving at the PSI forecast
if no definitive changes in emissions or meteorological
features are indicated. Interpretations of trends infor-
mation, on a day—to—day basis, require care and experience
because of the fluctuations that for varied reasons tend
to occur about a mean trend.

The types of meteorological information that could be
used for forecasting the PSI have been rather well
defined through past experience with forecasting
methods developed in support of air pollution control activities.

55

This support has largely dealt with forecasting indices
and episodic conditions. The meterological features and
parameters that are most often utilized in forecasting
air quality indices at the present time are:

- Character and Movement of Air Masses and Fronts

- Areas of Air Mass Subsidence

- Incidence, Intensity, and Height of Inversions

- Mixing Layer Height

- Prevailing Wind Direction

- Mean Wind Speed (Surface and Mixing Layer)

' Ventilation (Mixing Layer Mean Wind Speed x Mixing Heighi
- Precipitation

- Temperature

- Total Sky Cover

Of course, the emphasis placed on particular features and
parameters listed above will vary with location and pollu-
tant(s) of concern.

NWS INFORMATION AND SUPPORT SERVICES

The NWS operates a comprehensive Air Pollution Weather
Forecast Program. The program is administered from NWS
National and Regional Headquarters with operational program
elements at the National Meteorological Center (NMC) and
local Weather Service Forecast Offices (WSFO's). Details
of the program are contained in the NWS Operations Manual
and Technical Procedures Bulletins. This program generates
a variety of national, regional, and local air pollution
weather forecast products which are issued to the public,
to control agencies, or to both, as appropriate.

The NMC is responsible for providing the large—scale
meteorological guidance used by field offices in the pre—
paration of advisories and other products which are parti-

cularized and tailored to specific geographic areas to user
requirements.

The air pollution weather products of NMC are comprised
of the following elements:

a. Forecast Air stagnation Charts. Issued every
morning on facsimile, these four panel computer

based charts depict expected areas of atmospheric
stagnation (Figure l).

b. Air Stagnation Narrative. This plain language

teletype meSSage describing the Air Stagnation
Charts, is issued every morning.

56

L9

   
 
  
  

   
 
  
  

Forecast Area of
Air Stagnation Valid in
12 Hours

Forecast Area of
Air Stagnation Valid in
24 Hours

Forecast Area of Ai Forecast Area of Ai
stagnation Va11d 1H 36 Stagnation Valid in
Hours 48 Hours

Figure 1. Sample of Stagnation Chart sent on facsimile‘ depicting significant areas of
large scaie stagnation. Shaded area indicates area of 1arge-scaie stagnation.
hatched area indicates area that is under 1arge-sca1e stagnation on all four
nanei

c. Air Stagnation Data. This computer derived tele-
type message currently consists of today's mixing

height and transport wind speeds for selected NW5
stations.

The WSFO's have responsibility for local forecast
products within designated geographic boundaries, including

the issuance of the following three basic air pollution
products:

a. Air Stagnation Advisories (ASA). Issued to the
public and control agencies when locally established
critical values of transport wind, mixing height,
and ventilation are forecast to be reached and
cenditions are expected to persist for at least
36 hours, causing probable significant decrease
in air quality.

b. Special Dispersion Statements. A special product
issued only to control agencies when a potential
air pollution situation is determined by an NWS
forecaster to exist but no ASA will be issued because
such an issuance would not be in the public interest.

c. Dispersion Outlooks. A routine product issued by
all WSFO's where it has been determined that the
APC needs routine meteorological information to
facilitate day—to—day operations and adequate
manpower is available at the WSFO. The format,
content, and issuance times of this product is
determined by the WSFO and APC. The Dispersion
Outlook is issued only to the APC.

Occasionally, air pollution episodes of public concern
may occur during non-stagnant situations. These involve
predesignated episode levels that require control actions
to improve the air quality condition. In these situations,
the WSFO provides the appropriate government agencies with
the meteorological support necessary for pollution control
or abatement procedures.

In conjunction with these services, the NWS provides
supplemental, low—level upper air soundings at designated
stations upon request from agencies and/or WSFO's. This
program which provides for greater spatial and temporal
detail on dispersion conditions, eSpecially during episodes
or potential episodes, is available for several cities.

These locations are listed below, together with the sounding
scheduled:

58

Location Program

Birmingham, Ala. 1 per day routine week day, weekend and
second daily observation on call

Charleston, W. Va. 1 per day routine week day, weekend and
second daily observation on call

Chicago, Ill. 1 per day routine week day, weekend and
second daily observation on call

El Monte, Ca. 2 per day routine week days except
occasionally omit afternoon soundings
on well ventilated days

Houston, Tex. l per day routine week day, weekends
and second daily sounding on call

Los Angeles, Ca. 2 per day, 7 days a week
Philadelphia, Penn. all observations on call

Additionally, special low—level soundings are available on an on—
call basis at the regular upper air observation facilities near
Denver, Co., New York, N.Y., Oakland, Ca., Pittsburgh, Pa., and
Washington, D.C. An aircraft sounding is available at Sacramento,
Ca. Through a Cooperative effort, state APC agencies take
soundings as needed in Seattle, Boston, Portland, Ore., and

San Jose, Cai These are taken at special facilities that were
established by the NWS.

The NWS has, up until recently, not been too closely involved
nationwide in predicting conditions conducive to buildup of photo-
chemical pollutants. Because of recent interest and increasing
demand for such information, the NWS is in the process of evaluating
possible techniques with the objective of modifying or adding to
current air pollution weather forecast products and services.

DEVELOPMENT OF PREDICTION METHODOLOGY

The available services and information briefly described above
form the basis for developing a local community procedure for making
local qualitative forecasts of the PSI. These forecasts can be
reasonably made for periods up to a day in advance in terms of fig
Significant Change, Increase, or Decrease. It is advisable for
agencies planning to use the index along with a forecast procedure
to have personnel on their staffs familiar with meteorological data
and how these data may be applied in development of index prediction
methodology.

 

59

Considering the wealth of information available from
the NWS, it seems logical that the issuance of an index
forecast should be scheduled at intervals complementary to
operations at the NWS. This would allow the APC agency to
have the advantage of the most current NMC weather products
and WSFO air pollution forecast services. In addition, it
would encourage further cooperation and support of the local
NWS facility. However, while it can be expected that NWS
meteorologists will be able to closely coordinate with a
local agency in arriving at index change predictions during
potential or actual episodic conditions, they will most likely
not be able to give such attention to routine day-to-day
forecasting of the index. Also, NWS personnel would not be ex-
pected to have detailed knowledge of pollutant-related factors.

Where an APC agency may have developed the expertise
necessary to make quantitative predictions of the PSI for
the following day, they should be encouraged to make these
predictions. HOWever, it should be cautioned that making
quantitative predictions of air quality or air quality
indices should not be attempted without a reasonable expec-
tation of success based on well—tested techniques. Other-
wise, a less than satisfactory forecast record could result,
which would tend to have an adverse effect on public accept-
ance of the PSI.

Mathematical air quality simulatiOn models have to date
not been used to any appreciable extent in index prediction.
Because of their relative complexity, cost of modifying for
local use, and time and expense that may be involved in making
day-to-day predictions, their use for predicting the index
qualitatively will initially be limited. However, where APC
agencies may progress to the point of making quantitative
forecasts, the use of models may become necessary. A
listing and brief description of possible air quality models
that could be applied are contained in OAQPS Guideline No.
1.2-031.

CURRENT USE OF METEOROLOGICAL INFORMATION IN INDEX PREDICTION

Approximately half of the 25 local agencies currently
issuing air pollution indices make forecasts of their index
a day in advance. Of these, only one third have meteoro~
logists on their staffs, while the remainder rely upon NWS
meteorologists for interpretation of meteorological data.
Three of the local agencies were selected to serve as examples
of how varying degrees of meteorological information can be
incorporated into air quality index forecasting.

60

One of the more sophisticated forecast techniques, the
Air Pollution Dispersal Index, was developed six years a o
by the State of Colorado Department of Health in Denver. A
forecast is issued each morning for four time periods, a.m.
today, p.m. today, a.m. tomorrow, and p.m. tomorrow. The
technique developed by department meteorologists is based
upon concepts of mixing heights and wind speed discussed
by Holzworth in AP-lOl, and employs a nomogram of wind
speed vs. mixing heights, with isolines of constant venti-
lation factor values serving to demarcate four diSpersion
categories. These categories are:

Ventilation Factor (m2/sec)

 

 

(Wind Speed x Mixing Height) Associated Dispersion
< 2000 Bad
> 2000 to 4000 Fair
> 4000 to 6000 Good
> 6000 Excellent

The mixing heights used for the "today" forecast are deter-
mined from a plot of the Denver morning upper air sounding,
the morning minimum surface temperature at Stapleton Airport
plus 3° to 4°C, and the forecast afternoon maximum temperature.
The "tomorrow" mixing heights are determined from the fore—
cast 24—hour minimum and 36-hour maximum temperature, and a
forecast of the Sounding using locally-tailored analytical
techniques. All tranSport wind speeds are derived from
either observed or forecast NWS data. Critical factors in
Denver are the typical low-level morning inversions which
{serve to deteriorate air quality and the occurrence or
forecast of rain or snow which automatically leads to a
forecast of improving air quality.

The City of Philadelphia Department of Public Health4
uses general meteorological conditions and a NWS Air Stag-
nation Index to predict the Philadelphia Air Quality Index.
The local agency receives meteorological information twice
daily from the Philadelphia NWS office. Parameters of most
concern are wind speed, gustiness and the likelihood of a
frontal passage with its associated turbulent mixing. Wind
direction is not a vital concern since emission sources in
the city are relatively Well distributed in all directions.
Specifically, the Air Stagnation Index is formulated from
the algebraic sum of several weighted meteorological para-
meters as shown in Table 1. To determine the index value,
the weights associated with each observed parameter are
summed. When at least one of the meteorological values is

m

l

 

 

 

 

 

TABLE 1. Air Stagnation Check Sheet
Meteorological Value
parameters categories Weights
morning > 10 STOP
wind speed i 10 > 8 —1
(knots) : 8 > 6 +1
: 6 +2
afternoon and > 11 STOP
evening : 11 > 9 -1
wind speed < 9 > 6 +1
(knots) E 6 +2
morning > 1500 STOP
mixing height < 1500 > 750 -1
(meters) 3 750 > 500 O
afternoon > 8000 STOP
ventilation factor 2 8000 > 6000 -2
(meter /sec) 5 6000 > 4000 0
i 4000 +1

 

1

Philadelphia Forecast Office

National Weather Service

National Oceanic and Atmospheric Administration.
U.S. Department of Commerce

Sum of weights

—1, -2, —3
0
+1
+2, +3

associated with a "STOP," excellent diSpersion is forecast.

Otherwise, dispersion is forecast according to the following
scheme:

Forecast dispersion

good

marginally good
marginally poor

62

poor

However,'due to the nature of the Philadelphia Air Quality
Index, a dramatic change in dispersion is required to effect
a change in the index values.

The Department of Public Health in Dallas4 uses meteoro-
logical data in a very qualitative manner. The general
weather situation is examined daily with primary importance
directed toward stagnating high pressure systems, cold frontal
passages, and prevailing wind direction. NMC trajectory
analysis data, surface Weather patterns, and prognostic
charts are used in a non-rigorous manner. For example,
geographical plots of smoke and haze reports are occasionally
used to determine the area extent and approach of pollutants
due to large scale circulation patterns.

Improving conditions are forecast with the occurrence
of precipitation, a frontal passage, and increasing Wind
speed. Deteriorating air quality is predicted when trajec-
tories persist from local or more distant sources or sources
areas.

REFERENCES

National Weather Service, Operations Manual, Air Pollution
Weather Forecasts, WSOM Issuance 75—13, Part C, Chapter 30,
April 1975.

National Weather Service, Technical Procedures Bulletin
No. 122: Air Stagnation Guidance for Facsimile and
Teletype (3rd Edition), October 21, 1974. (Supersedes
previous TPB's Nos. 52, 58, and 69.)

U.S. Environmental Protection Agency, Guidelines for Air
Quality Maintenance Planning and Analysis, Volume 12:
Applying Air Quality Models to Air Quality Maintenance
Areas, EPA—450/4-74-012, September 1974 (OAQPS No. 1.2—031),
Research Triangle Park, N.C.

Thom, G., and Wayne R. Ott, "Compendium Analysis, and Review
of United States and Canadian Air Pollution Indices,"
Joint Study by the U.S. Environmental Protection Agency and

the Council on Environmental Quality, Washington, D.C.,
December 1975.

Holzworth, G.C., "Mixing Heights, Wind Speeds, and Potential
for Urban Air Pollution Throughout the Contiguous United
States," U.S.jEnvironmenta1 Protection Agency, Research
Triangle Park, N.C. January 1972 (AP-101).

63

MEMBERS OF THE FEDERAL TASK FORCE ON AIR QUALITY INDICATORS

Council on Environmental Quality .................. James J. Reisa*

(Chairman of the Task Force)
Kay H. Jones
Gary C. Thom**
(Technical Secretary of the Task Force)

Environmental Protection Agency ................... Raymond Smith*
(Office of Air and Waste Management) Robert Neligan
William Hunt
(Office of Planning and Management) ............. Barry C. Korb*
(Office of Research and Development) ............ Wayne R. Ott*
John H. Knelson

John Moran

Department of Commerce ............................ George S. Gordon*
(Office of Environmental Affairs)

(National Bureau of Standards) .................. James R. McNesby*
(National Oceanic and Atmospheric

Administration) ............................. Isaac Van Der Hoven*
E. L. Martinez

*Official representative
**Consultant

65

:7 H, S_ (EOV'I‘IHNEVIPIN'I' PRINTING OFFICE: 1‘I7n 0 - 200-662

 

 

 

" \\.

 

[incl-iiiiiimflill

(020200535