ANALYSIS OF PHOTOCHEMICAL
OXIDANT AND PARTICULATE
POLLUTION PATTERNS IN NEW ENGLAND
USING REMOTE SENSING DATA
JUNE 1977
FINALREPORT
U.S. ENVIRONMENTAL PROTECTION AGENCY
REGION I
AIR BRANCH
-------
ANALYSIS OF PHOTOCHEMICAL
OXIDANT AND PARTICULATE
POLLUTION PATTERNS IN NEW ENGLAND
USING REMOTE SENSING DATA
JUNE 1977
FINALREPORT
U.S. ENVIRONMENTAL PROTECTION AGENCY
REGION I
AIR BRANCH
-------
TECHNICAL REPORT DATA
(Please read Itiztmdions on the reverse before completing)
1 REPORT NO
EPA 901/9-77-002
2.
4. TITLE AND SUBTITLE
\nalysis of photochemical oxidant and particulate pollu-
tion patterns in New England using remote sensing data
7 AUTHOR(S)
Ilinton J. Bowley, Joseph L. Horowitz, James C. Barnes
8. PERFORMING ORGANIZATION REPORT NO
P-2273
3 RECIPIENT'S ACCESSION-NO.
5 REPORT DATE
June 1977
6. PERFORMING ORGANIZATION CODE
9 PERFORMING ORGANIZATION NAME AND ADDRESS
Environmental Research £ Technology, Inc.
696 Virginia Road
oncord, Massachusetts 01742
10. PROGRAM ELEMENT NO.
11 CONTRACT/GRANT NO
68-02-2533
12. SPONSORING AGENCY NAME AND ADDRESS
Environmental Protection Agency
Region 1, Air Branch
JFK Federal Building
Boston. Massachusetts 02203
13. TYPE OF REPORT AND PERIOD COVERED
Final Report
14 SPONSORING AGENCY CODE
15 SUPPLEMENTARY NOTES
Abbreviated Title:
Analysis of Pollution Patterns in New
England Using Remote Sensing
16. ABSTRACT
Imagery from earth surveillance satellites is examined to assess the potential useful-
ness of satellite data for monitoring air pollutant patterns and defining the associa-
ted meteorological conditions in southern New England. Three high ozone situations,
one high total suspended particulate episode and one high sulfate episode are studied.
Imagery from one or more of the following satellite systems: Landsat, NOAA/VHRR, GOES,
and DMSP, ranging from the visible to the thermal infrared portions of the spectrum,
are visually interpreted for each case. Because of the limited data sample, it is
difficult to derive firm conclusions for particular imagery types or specified pollu-
tants. Nevertheless, the results of the study indicate that there is a good correla-
tion between certain types of high pollutant load (including ozone and sulfates) and
laze and smoke, with accompanying reductions in visibility. There is good evidence
that satellite imagery can display differences in reflectance (visible) or temperature
(thermal infrared) due to some combination of haze, smoke and atmospheric pollutant
load, on a regional basis. The greatest promise was shown by the DMSP thermal infrared
imagery in application to high sulfate situations.
KEY WORDS AND DOCUMENT ANALYSIS
jh.IDENTIFIERS/OPEN ENDED TERMS C. COSATI F leld/GrOUp
DESCRIPTORS
remote sensing
satellite
)ollution
regional pollution
ozone
)articulates
sulfates
3 DISTRIBUTION STATEMENT
Release Unlimited
19 SECURITY CLASS (Tins Report)
Unclassified
21. NO. OF PAGES
55
20 SECURITY CLASS (This page)
Unclassified
22. PRICE
EPA Form 2220-1 (9-73)
-------
EPA 901/9-77-002
ANALYSIS OF PHOTOCHEMICAL OXIDANT AND
PARTICULATE POLLUTION PATTERNS IN
NEW ENGLAND
USING REMOTE SENSING DATA
By
Environmental Research S Technology, Inc.
696 Virginia Road
Concord, Massachusetts 01742
CONTRACT NO. 68-02-2533
Mr. Valentine J. Descamps, Project Officer
FINAL REPORT
Prepared for
Environmental Protection Agency
Region I
Air Branch
JFK Federal Building
Boston, Massachusetts 02203
June 1977
-------
This air pollution report is issued by Region I, Environmental
Protection Agency, to assist state and local air pollution control agencies
in carrying out their program activities. Copies of this report may be
obtained, for a nominal cost, from the National Technical Information
Service, 5285 Port Royal Road, Springfield, Virginia 22151.
This report was furnished to the Environmental Protection Agency by
Environmental Research § Technology, Inc., Concord, Massachusetts in
fulfillment of EPA Contract No. 68-02-2533. This report has been
reviewed by Region I, Air Branch, EPA and approved for publication.
Approval does not signify that the contents necessarily reflect the
views and policies of the Environmental Protection Agency, nor does
mention of trade names or commercial products constitute endorsement or
recommendation for use.
Region I Publication No. EPA 901/9-77-002
-------
ENVIRONMENTAL RESEARCH & TECHNOLOGY, INC
FOREWORD
The investigation described herein was performed by Environmental
Research § Technology, Inc. (ERT) for Region 1, Air Branch of the Environ-
mental Protection Agency under Contract No. 68-02-2533.
The authors acknowledge the guidance provided by Mr. Valentine J.
Descamps, the EPA Project Officer. Data from the SAROAD data network
was provided by EPA, Region 1. The authors are extremely grateful to
Mr. Eugene Y. Tong of ERT for his assistance throughout this project and
many helpful suggestions, and for providing the regional sulfate data
used in the study.
-------
ENVIRONMENTAL RESEARCH* TECHNOLOGY, INC
TABLE OF CONTENTS
Page
FOREWORD v
LIST OF ILLUSTRATIONS i*
1. INTRODUCTION 1
1.1 Purpose of Study 1
1.2 Review of Previous Studies 1
1.3 Description of Satellite Data 2
1.3.1 Landsat 2
1.3.2 NOAA/VHRR (Very High Resolution Radiometer) 3
1.3.3 GOES (Geostationary Operational Environmental
Satellite) 3
1.3.4 DMSP (Defense Meteorological Satellite Program) 4
2. ANALYSIS PROCEDURES - PHOTOCHEMICAL OXIDANTS/OZONE 5
2.1 Data Sample 5
2.1.1 Case 1, 20-23 May 1975 6
2.1.2 Case 2, 23-24 June 1975 12
2.1.3 Case 3, 23 July 1975 13
3. ANALYSIS PROCEDURES - TOTAL SUSPENDED PARTICULATES (TSP) 23
3.1 Data Sample 23
3.2 Data Analysis for 6 May 1976 23
4. ANALYSIS PROCEDURE - SULFATES (S04) 27
4.1 Data Sample 27
4.2 Comparison of 8-10 July 1974 High Sulfate Episode
with Correlative Satellite Data and Meteorological
Conditions 28
5. SUMMARY AND CONCLUSIONS 37
5.1 Discussion of Results 37
5.1.1 Photochemical Oxidants/Ozone 37
5.1.2 Total Suspended Particulates (TSP) 38
5.1.3 Sulfates (SO.) 38
4
5.2 Conclusions and Recommendations 39
6. REFERENCES 43
vn
-------
ENVIRONMENTAL RESEARCH & TECHNOLOGY, INC
LIST OF ILLUSTRATIONS
Figure Page
2-1 Analysis of Ozone Concentration over Southern
New England at 1500 GMT, 20 May 1975 8
2-2 Mosaic of Landsat-1, MSS-4 Imagery Viewing
Southern New England at 1440 GMT, 20 May 1975 9
2-3 Analysis of Ozone Concentration over Southern
New England at 1400 GMT, 23 May 1975 10
2-4 NOAA/VHRR Daytime Visible Image Viewing the
Northeastern United States at 1401 GMT,
23 May 1975 11
2-5 Map Displaying the Extent of Reduced Visibilities
in Haze and Smoke over the Northeast at 1600 GMT,
23 May 1975 14
2-6 Analysis o-f Ozone Concentration over Southern
New England at 1800 GMT, 23 May 1975 15
2-7 Map Displaying the Extent of Reduced Visibilities
in Haze and Smoke over the Northeast at 1600 GMT,
23 June 1975 17
2-8 Map Displaying the Extent of Reduced Visibilities
in Haze and Smoke over the Northeast at 1600 GMT,
24 June 1975 18
2-9 NOAA/VHRR Daytime Visible Image Viewing the
Northeastern United States at 1439 GMT, 24 June
1975. A Region of Low Visibility in Haze and
Smoke Located over this Region is Easily Detected
due to Illumination by the Sun. 19
2-10 Analysis of Ozone Concentration over Southern
New England at 1500 GMT, 23 July 1975 20
2-11 Landsat-2, MSS-4 Image Viewing Connecticut and
Eastern Long Island Sound at 1455 GMT, 23 July 1975 21
IX
-------
ENVIRONMENTAL RESEARCH & TECHNO LOGY, INC
LIST OF ILLUSTRATIONS (Continued)
Figure Page
3-1 Analysis of 24-Hour Average Values of Total
Suspended Particulates over Southern
New England on 6 May 1976 24
3-2 Landsat-2, MSS-4 Image Viewing Connecticut and
South Central Massachusetts at 1450 GMT,
6 May 1976 25
4-1 Analysis of 24-Hour Average Values of Sulfates
over the Northeastern United States on
8 July 1974 30
4-2 DMSP Daytime Visible-to-Near Infrared Image
Viewing the Northeastern United States on
8 July 1974 31
4-3 DMSP Daytime Thermal Infrared Image Viewing the
Northeastern United States on 8 July 1974 32
4-4 Analysis of 24-Hour Average Values of Sulfates
over the Northeastern United States on
9 July 1974 33
4-5 Enlargement of Portion of DMSP Daytime Visible-
to-Near Infrared Image Viewing the Northeastern
United States on 9 July 1974 34
4-6 Enlargement of Portion of DMSP Daytime Thermal
Infrared Image Viewing the Northeastern United
States on 9 July 1974, with Sulfate Analysis
Superimposed 35
-------
ENVIRONMENTAL RESEARCH & TECHNOLOGY, INC
1. INTRODUCTION
1.1 Purpose of Study
The purpose of this investigation was to examine the potential use-
fulness of satellite data for monitoring air pollutant patterns and for
defining the associated meteorological conditions in southern New
England. The principal objective of Task 1 was to conduct analyses of
Landsat and meteorological satellite data (NOAA/VHRR and GOES) during
three episodes of high photochemical oxidant levels over southern New
England as reported by the EPA's "SAROAD" data network.
The completion of Task 2 was initially directed toward analyses of
Landsat data during two episodes of high total suspended particulate
(TSP) levels over New England as reported by the "SAROAD" network.
However, problems such as cloud contamination and isolated high TSP
readings suggesting influence of local fugitive dust rather than overall
high TSP events became apparent; therefore, it was mutually agreed (ERT
and EPA Project Officer) that further analyses of TSP episodes would be
replaced by analysis of one or two incidents of high sulfate concentra-
tions over the Northeast utilizing meteorological satellite data.
1.2 Review of Previous Studies
The ability to detect certain types of air pollution patterns from
satellite imagery has been noted by a number of investigators. McLellan
(1971) successfully correlated a series of ATS-3 satellite images of
reflected radiation from the City of Los Angeles at 40 minute intervals
on 23 April, 1968 with various ground based meteorological data. The
analysis of the data indicated the feasibility of satellite monitoring
of short-term atmospheric pollution variation over large urban areas.
Prospero, et al. (1970) showed the feasibility of satellite monitoring
of global scale movements of natural pollution. In this study, a series
of ATS-3 images were correlated with aircraft and ship sampling data of
dust being transported by air currents from the arid and semi-arid
regions of North Africa across the Atlantic to the Caribbean during the
summer of 1969.
-------
ENVIRONMENTAL MSEARCH & TECHNOLOGY, INC
Recently, the utilization of improved satellite imagery as a means
for monitoring environmental quality has been further realized. Lyons
and Northouse (1973) showed that the interaction of air pollution and
meteorological conditions over the Great Lakes led to inadvertant
weather modification detectable in Landsat imagery. Griggs (1973) found
that radiance measurements over water could be used successfully to
calculate the vertical aerosol burden to an accuracy approaching ±10%.
In a more recent investigation, Brown and Karn (1976) examined the
usefulness of Landsat imagery for detecting sources of air pollution
resulting from coal utilization in the Ohio and Monongahela River
valleys south of Pittsburgh, Pennsylvania. This area is heavily indus-
trialized and includes many coal-fired power-generating stations. In
this investigation, large billowy steam plumes from specific point
sources were identified in Landsat imagery.
In addition, Lyons and Husar (1976) reported the likely detection
of a synoptic-scale sulfate air pollution episode over the central and
eastern United States in SMS/GOES visible images. It was shown that
image "haziness" correlated well with midday surface visibility reports,
which in turn appeared to correlate with sulfate aerosol concentrations.
Unfortunately, Lyons and Husar were not able to show the regional extent
of the high sulfate eposode, probably because the available sulfate
measurements were only for a limited area.
1.3 Description of Satellite Data
The satellite data obtained for the completion of this study
program included imagery from the Landsat, NOAA/VHRR, GOES and DMSP
satellite systems. The wide variety of imagery provided by these
various experimental and operational spacecraft are defined in the
following paragraphs.
1.3.1 Landsat
The initial spacecraft in the experimental Landsat series was
placed into orbit at a height of 900 km in late July 1972 and was
called at that time ERTS-1, the Earth Resources Technology Satellite.
The second spacecraft, Landsat-2, was placed into operation in January
-------
ENVIRONMENTAL RESEARCH & TECHNOLOGY, INC
1975. Each satellite repeats coverage of the same area only once every
18 days; however, Landsat-2 was placed into orbit such that coverage of
the same area was being provided by one of these spacecraft every nine
days. A recent orbital adjustment has resulted in repeat coverage on a
six- and 12-day cycle.
The Landsat Multispectral Scanner Subsystem (MSS) observes in four
spectral bands within the visible-to-near infrared portions of the
spectrum (MSS-4, 0.5-0.6 urn; MSS-5, 0.6-0.7 urn; MSS-6, 0.7-0.8 vim; and
MSS-7, 0.8-1.1 ym) . Therefore, Landsat observations are available only
during periods of adequate solar illumination. Landsat views an area
approximately 185 km on a side, per image frame, as it proceeds along
its orbital path, and has a resolution of 70-100 meters. The scale of
the standard Landsat 9 inch x 9 inch prints is 1:1 million.
1.3.2 NOAA/VHRR (Very High Resolution Radiometer)
The VHRR sensor has been flown on the NOAA-2, NOAA-3, NOAA-4, and
NOAA-5 spacecraft, so that the resulting data have been available on a
nearly continuous basis (one or two observations per day) since January
1973. The VHRR is a two-channel radiometer, with one channel sensitive
to reflective solar radiation in the 0.5-0.7 ym wavelengths, and the
other channel measuring the thermal radiation emitted by clouds and the
earth's surface in the infrared 10.5-12.5 ym wavelengths. The spatial
resolution of the VHRR is about 1 km. The scale of the standard size
prints is approximately 1:10 million.
The VHRR is designed primarily for direct readout use as an opera-
tional meteorological satellite. The area that can be observed when the
satellite passes directly overhead is a strip about 2,200 km wide and
more than 5,000 km long. This allows broad overall synoptic weather
patterns to be determined from each individual scene.
1.3.3 GOES (Geostationary Operational Environmental Satellite)
Another type of operational meteorological satellite data is
provided by the GOES system. Launched on 16 May 1974, GOES-1 is sta-
tioned over the equator in an earth synchronous orbit in a position to
view the eastern United States and Atlantic Ocean. Because of its earth
-------
ENVIRONMENTAL RESEARCH & TECHNOLOGY. INC
synchronous orbit, GOES provides coverage of the earth's disc on an
essentially continuous basis, with data being collected every one-half
hour. Thus, cloud motions can be tracked and storm development can be
monitored; using the cloud motions, it is possible to derive low-level
wind flow. As with the VHRR, the GOES sensor system carries a thermal
infrared channel for obtaining nighttime data. The visible channel
radiometer operates in the 0.55 to 0.70 ym wavelength at a spatial
resolution of 1 km, whereas the infrared channel operates-in the 10.5 to
12.5 ym wavelength at a resolution of 7 km.
1.3.4 DMSP (Defense Meteorological Satellite Program)
The Air Force DMSP is an operational satellite system useful for
providing meteorological and oceanographic data. These polar orbiting,
sun synchronous weather satellites are positioned 900 km above the
earth. In the operational mode, the DMSP satellites provide imagery
every 6 hours over any spot on earth. The imagery includes the visual
to near infrared and thermal infrared spectral intervals over a 3,000 km
wide swath below the satellite.
Real time meteorological data within the acquisition range of the
satellite signal, approximately a radius of 2,800 km of the receiving
station, are provided to military sites. These sites are scattered
around the globe and on board U. S. Navy aircraft carriers. The Air
Force Global Weather Center (AFGWC) at Offutt Air Force Base, Nebraska,
receives stored data of global coverage; these data have been available
to nonmilitary users since 1972, and are archived at the University of
Wisconsin. The visible to near infrared wavelength channel has a
spectral interval of 0.4 to 1.1 ym with a spatial resolution of 0.6 km;
the thermal infrared channel operates in the 8 to 13 ym wavelength with
a spatial resolution of 3 km. Scale of the standard prints is
1:10 million.
-------
ENVIRONMENTAL RESEARCH & TECHNOLOGY, INC
2. ANALYSIS PROCEDURES - PHOTOCHEMICAL OXIDANTS/OZONE
2.1 Data Sample
The three episodes of high photochemical oxidant levels over
southern New England selected for analysis were initially included on a
listing of high episodes supplied by the EPA-Region 1, Air Branch. Each
of the episodes was examined to determine the availability.of useful,
cloud-free Landsat imagery. A computer listing supplied by the EROS
Data Center, Sioux Falls, South Dakota, showed that useful Landsat data
were available for only two of the episodes. The correlative satellite
data available for the other episode selected for analysis included only
the NOAA/VHRR and GOES imagery.
The three episodes of high photochemical oxidant levels analyzed in
this investigation consisted of the following spring and summertime
periods of 1975: 20-23 May; 23-24 June; and 23-24 July. Correlative
Landsat imagery were available for 20 May and 23 July. The MSS-4
70 mm positive transparencies for these dates were obtained from the
U. S. Army Corps of Engineers (Waltham, Massachusetts) browse file, and
standard size, 9 inch x 9 inch (1:1 million scale), as well as enlarged
(approximately 1:500,000 scale), positive paper prints were processed in
ERT's photographic laboratory. The MSS-4 visible band was used in each
case because past experience has shown that this is the most useful band
for detecting pollution and haze. Much of the evidence of this has come
from investigators of earth surface features who have found these
features to be more susceptible to obscuration by atmospheric conditions
in Band 4 than in any other band. Also, all available NOAA/VHRR (visible
and infrared) and GOES (visible) imagery viewing the New England region
during the periods of each episode were obtained from the NOAA-Satellite
Data Services Branch, World Weather Building, Camp Springs, Maryland.
Listings of the EPA's 1975 SAROAD data for Massachusetts, Connecticut
and Rhode Island were obtained from the EPA's facility in Lexington,
Massachusetts. The ozone concentrations measured closest to the times
of the satellite images were plotted and analyzed on area base maps.
The comparisons of maximum ozone concentrations with the correlative
-------
ENVIRONMENTAL RESEARCH & TECHNOLOGY, INC
satellite imagery and associated meteorological conditions for each of
the episodes are discussed in the following sections.
2.1.1 Case 1, 20-23 May 1975
Hourly ozone concentrations at times corresponding to the readout
times of the correlative Landsat and NOAA/VHRR satellite images O1500 GMT,
1000 LSI), generally exceeded the national ambient air quality primary
and secondary standard of 0.080 ppm over broad areas of southern New
England. However, analyses of ozone concentrations during this period
revealed a definite diurnal trend with maximum concentrations occurring
during the mid to late afternoon hours. Analysis of the ozone concen-
trations over southern New England at 1500 GMT on 20 May is shown in
Figure 2-1. Maximum concentrations, exceeding 0.140 ppm, are located in
extreme southwestern and southeastern Connecticut, while another rela-
tively high concentration, exceeding 0.120 ppm, is located immediately
north and east of Boston.
A mosaic of two Landsat, MSS-4, images viewing this region at
1440 GMT on this date is shown in Figure 2-2. In this image, no varia-
tions in brightness are detectable over the land area that could be
attributed to the high ozone concentrations. However, in the original
photographic prints used in the analysis some increased brightness
levels [likely associated with reported low visibilities in haze (less
than 5 miles) and possibly, high ozone] are observed over the darker
water areas of Long Island Sound and to the east of Boston and Cape Ann.
Over the land area, any varying degree of image brightness due to air
pollution density is likely obscured by the varying brightness levels of
ground-feature reflectivities (most notably, the dense urban developments
which appear highly reflective in the images), No useful NOAA or GOES
imagery was available for this date.
Isolated high concentrations of ozone at a somewhat lesser peak
value (but still >0.080 ppm) remained over portions of the region during
21 and 22 May; however, no features were detectable in either the NOAA/
VHRR or GOES imagery on these dates because of cloud contamination.
Analysis of the SAROAD ozone data for 1400 GMT on 23 May (Figure 2-3)
showed a broad band of maximum concentration O0.120 ppm) located in
west central Connecticut, and an isolated small pocket O'O.IOO ppm)
-------
ENVIRONMENT*!- RESEARCH & TECHNOLOGY INC
located just to the west of Boston. The correlative NOAA/VHRR visible
image (1401 GMT) for this date is shown in Figure 2-4. Although dense,
offshore fog is observed surrounding the southern New England coastal
regions from Long Island Sound into the southern Gulf of Maine, no
features associated with the maximum ozone concentrations are observed
on the image. The band of higher reflectance, located near the area of
highest ozone concentration in Connecticut, is again associated with
dense urban development in the Connecticut River Valley.
Examination of the meteorological conditions during this period
showed a generally light, southerly flow of warm, moist, tropical air
associated with a near stationary ridge of high pressure extending from
well off the middle Atlantic coastal region, southwestward across the
entire southeastern United States. The importance of heat and moisture
becomes more apparent when examining the late morning surface tempera-
tures and corresponding dew point temperature readings. During the
beginning of the period (20 and 21 May), late morning surface tempera-
tures ranged between 80° and 85 °F across southern New England except for
extreme southern coastal sections where temperatures remained in the
mid-70's due to the effect of the light southerly wind flow from the
cooler ocean waters. Corresponding dewpoint temperatures were generally
in the middle 60°F range at these times.
The slight reduction in peak ozone concentrations observed on
22 May was likely the result of a weak "back door" cold front that moved
westward across New England into eastern New York State during the
morning hours, before retreating eastward again as a warm front early on
the 23rd. During the early afternoon hours on 23 May, when the highest
ozone concentrations for the entire period were observed, surface
temperatures ranged between the upper 80's to low 90's, with dewpoint
temperatures ranging from the upper 60's to low 70's.
Throughout the period, broadscale haziness resulting in reduced
visibilities accompanied the high ozone concentrations. This finding is
consistent with a report by Husar, et al. (1976), in which he mentions
that broad areas of haziness and reduced visibilities accompanied high
mid-day ozone concentrations that exceeded the national ambient air
quality standard in the St. Louis metropolitan area during summertime
periods of 1975. With this in mind, area base maps of skycover, winds,
-------
060
Figure 2-1 Analysis of ozone concentration (ppm) over southern New
England at 1500 GMT, 20 May 1975
-------
ENVIRONMENTAL RESEARCH & TECHNOLOGY INC
Figure 2-2 Mosaic of Landsat-1, MSS-4 imagery viewing southern New
England at 1440 GMT, 20 May 1975
IO
en
-------
ENVIRONMENTAL RESEARCH & TECHNOLOGY, INC
.060
.080
. .060
Figure 2-3 Analysis of ozone concentration (ppm) over southern New
England at 1400 GMT, 23 May 1975
10
-------
ENVIRONMENTAL RESEARCH & TECHNOLOGY
FiS^e 2-4 NOAA/VHRR daytime
United States at
"» -«heastern
11
-------
ENVIRONMENTAL RESEARCH 4 TECHNOLOGY INC
and visibility were plotted from the weather reports for the period of
18 May through 23 May. The 1600 GMT hourly reports were chosen to
reduce the effect on visibilities by early morning fog conditions.
Analyses of the reported visibilities suggested an overall long range
transport of haziness and smoke into southern New England from the New
Jersey, Maryland and eastern Pennsylvania area during the period, as
well as some probable in situ development over southern New York State
and western Long Island. The extent of minimum visibilities of from
1.5 to 2 statute miles in haze and smoke for much of Connecticut on
23 May, shown in Figure 2-5, compares well with the extent of maximum
ozone levels of from 0.150 to 0.205 ppm reported for this region at
1800 GMT on this date (Figure 2-6).
2.1.2 Case 2, 23-24 June 1975
The ozone concentrations as reported by the SAROAD data base for
Massachusetts, Connecticut and Rhode Island at the corresponding image
times (NOAA/VHRR - 1438 GMT, and GOES - 1800 GMT) for the period of
23 and 24 June 1975 were plotted on area base maps and analyzed. No
Landsat imagery was available for either of these days. These analyses,
as in the previous case, showed a marked increase in ozone concentration
at the later time period (1800 GMT), especially on 23 June when maximum
levels increased nearly 200%, from about 0.100 ppm to 0.280 ppm over the
4-hour period.
A plot of skycover, winds and visibility for 1600 GMT for 22, 23,
and 24 June suggested some long range transport of pollutants into
southern New England from New York State and Pennsylvania, as well as
probable in situ pollution development. Areas of minimum visibilities
of 2 to 4 miles in haze reported for much of Connecticut on 23 and
24 June (Figures 2-7 and 2-8) correlate well with the locations of
maximum ozone concentrations.
With one exception, no evidence of the maximum ozone concentrations
over southern New England was detectable on the NOAA/VHRR or GOES
imagery of either day. The exception is the NOAA-4 visible image of
24 June, shown in Figure 2-9. This image was acquired along a satellite
track positioned far enough to the west to display a low sun angle along
the middle Atlantic coast from about Cape Hatteras to southern New
12
-------
ENVIRONMENTAL RESEARCH S TECHNOLOGY, INC
England. Therefore, the area of low visibility in haze and smoke
(2-4 miles) located over New Jersey, Pennsylvania, southeastern New
York State and Connecticut is easily detected due to illumination by the
sun. The extent of the low visibility area observed over Connecticut
agrees well with the location of the maximum ozone concentration as
reported by the SAROAD network. Of all of the satellite imagery examined,
this NOAA image revealed the only detectable evidence of high pollution
concentration associated with peak ozone concentrations over land areas.
Sun angle apparently plays an extremely important role in the ability of
satellites to detect high pollution concentrations and associated low
visibilities in the visible portions of the spectrum.
Examination of the overall synoptic meteorological pattern during
this period revealed the•dominance of a warm, humid maritime tropical
air mass throughout the entire region. A near-stationary high pressure
system (anticyclone) was positioned just off the middle Atlantic coastal
region resulting in the flow of warm, moist air into New England on
light to moderate (10-15 knots) southwesterly winds. Daytime high
temperature readings of from the upper 80° to low 90°F range were
experienced over southern and central New England on each of the days,
and dewpoint temperatures climbed into the low 70°F range.
2.1.3 Case 3, 23 July 1975
The hourly ozone concentrations closest to the time of correlative
Landsat imagery (1500 GMT) on this date are shown in Figure 2-10. The
analysis of the ozone distribution reveals a band of maximum concen-
tration (>0.120 ppm) extending from extreme southwestern Connecticut
northward to near Waterbury with amounts in excess of 0.080 ppm extending
as far north as Hartford. Other areas of southern New England reported
amounts generally between 0.050 and 0.070 ppm.
The Landsat MSS-4 image viewing the area on 23 July is shown in
Figure 2-11. Again, as discovered in Case 1, the Landsat imagery
reveals no indication of the band of maximum ozone levels over land
areas; however, evidence of the reduced visibilities in haze being
reported from Bridgeport, Connecticut, eastward to Block Island, Rhode
Island, is readily detectable (in the original photographic prints) over
the darker water area of Long Island Sound. Reduced visibilities in
haze and smoke have accompanied each of the high photochemical oxidant
episodes being examined.
13
-------
ENVIRONMENTAL RESEARCH &TECHNOLOGY INC
Figure 2-5
Map displaying the extent of reduced visibilities (statute
miles) in haze and smoke over the northeast at 1600 GMT,
23 May 1975
o
o
14
-------
ENVIRONMENTAL RESEARCH &TECHNOLOGY INC
.080
080
.080
.060
Figure 2-6 Analysis of ozone concentration (ppm) over southern New
England at 1800 GMT, 23 May 1975
15
-------
ENVIRONMENTAL RESEARCH & TECHNOLOGY INC
Figure 2-7 Map displaying the extent of reduced visibilities (statute
miles) in haze and smoke over the Northeast at 1600 GMT,
23 June 1975
17
-------
ENVIRONMENTAL RESEARCH 8.TECHNOLOGY INC
Figure 2-8 Map displaying the extent of reduced visibilities (statute
miles) in haze and smoke over the Northeast at 1600 GMT,
24 June 1975
18
-------
ENVIRONMENTAL RESEARCH & TECHNOLOGY INC
Figure 2-9 NOAA/VHRR daytime visible image viewing the northeastern
United States at 1439 GMT, 24 June 1975. A region of low
visibility in haze and smoke located over this region is
easily detected due to illumination by the sun.
in
o
19
-------
ENVIRONMENTAL RESEARCH & TECHNOLOGY, INC
/080]
ft/
.060
060
.060
.060
Figure 2-10
Analysis of ozone concentration (ppm) over southern New
England at 1500 GMT, 23 July 1975
20
-------
ENVIRONMENTAL RESEARCH & TECHNOLOGY IMC
Figure 2-11 Landsat-2, MSS-4 image viewing Connecticut and eastern Lon£
Island Sound at 1455 GMT, 23 July 1975
21
-------
ENVIRONMENTAL RESEARCH 4TECHNOLOGY, INC
The synoptic meteorological conditions associated with this episode
were also analogous to the previous two cases. Examination of the
23 July surface charts showed a nearly stationary ridge of high pressure
centered well to the southwest near Virginia. Therefore, a warm,
moist flow of air on light to moderate southwesterly surface winds was
experienced over southern New England. Maximum surface temperatures
reached the low 80's across extreme southern coastal sections and the
upper 80°F range over northern Connecticut and Massachusetts. Corres-
ponding dew point temperatures ranged from the mid to upper 60°F level
throughout the region.
22
-------
ENVIRONMENTAL RESEARCH S, TECHNOLOGY, INC
3. ANALYSIS PROCEDURES - TOTAL SUSPENDED PARTICULATES (TSP)
3.1 Data Sample
Examination of correlative Landsat coverage during a number of
periods of high total suspended particulate (TSP) levels over southern
New England during 1974 and 1975 showed that no data were acquired by
Landsat during any of the episodes. Therefore, a computer query search
was requested from the EROS Data Center, Sioux Falls, South Dakota, for
a listing of Landsat coverage of central and southern New England for
all dates with less than 50% cloud cover. A check of this computer
listing for dates of useful Landsat coverage showed that correlative
SAROAD TSP data [recorded every six days) was available for only two of
the dates, 4 April and 6 May 1976. A check of the 1976 SAROAD data,
obtained from the EPA facility in Lexington, revealed low TSP levels on
4 April; however, three small pockets of relatively high TSP levels
(exceeding national ambient air quality secondary standard of 150 yg/m
for 24-hour average) were found to exist over southern New England on
6 May 1976.
Although no Landsat coverage was available for two dates of reportedly
high TSP concentrations over southern New England (24 January and
18 April 1975), NOAA/VHRR and GOES imagery were obtained to determine if
the lower resolution of these satellites was at all useful for detecting
the high TSP events. Unfortunately, the imagery was not useful because
of cloud contamination over the region on each of these dates.
The 24-hour average values of TSP were plotted and analyzed on area
base maps for 6 May 1976, the one case for which correlative Landsat
data existed. A comparison of the TSP analysis with the Landsat satellite
data is discussed in the following section.
3.2 Data Analysis for 6 May 1976
Analysis of the 24-hour average values of TSP as reported by the
EPA's SAROAD data network is shown in Figure 3-1. This analysis reveals
that the highest concentrations were confined to isolated small areas in
central Connecticut, near Springfield, Massachusetts, and over Providence,
3
Rhode Island, where values generally exceeded 150 yg/m .
23
-------
ENVIRONMENTAL RESEARCH S TECHNOLOGY INC
Figure 3-1 Analysis of 24-hour average values of total suspended
particulates (ug/m^) over southern New England on 6 May
1976
24
-------
ENVIRONMENTAL RESEARCH & TECHNOLOGY INC
42° *
Figure 3-2 Landsat-2, MSS-4 image viewing Connecticut and south cen-
tral Massachusetts at 1450 GMT, 6 May 1976
25
-------
ENVIRONMENTAL RESEARCH & TECHNOLOGY INC
The correlative Landsat image (1450 GMT), shown in Figure 3-2,
views the areas of high TSP levels located in central Connecticut and
near Springfield, Massachusetts. This image shows a thin veil of high
cloudiness (cirrostratus) obscuring much of the area of high TSP
3
(>200 yg/m ) located over Springfield. Examination of the 1500 GMT
hourly weather data revealed that most stations in southern New England
were reporting high, thin clouds. The small pocket of high TSP
(>150 yg/m ) located over central Connecticut lies under generally clear
skies; however, no distinct reflectance boundaries associated with the
maximum TSP concentration are observed in the image.
Although the isolated pockets of high TSP on this date fall within
a narrow band of relatively high TSP levels O100 yg/m ), which extends
northward across Connecticut and then to the east-northeast across
Massachusetts (Figure 3-1), it is possible that the high readings
reflect the actual influence of local fugitive dust (such as from roadways
and gravel areas) rather than the ambient TSP conditions. Also, the
precise times of the occurrence of high TSP levels on this date, which
contributed to the overall high 24-hour average values, could have taken
place much later in the day, and not near the time of the Landsat image.
Analysis of the 1500 GMT hourly surface weather data revealed a
cold front advancing slowly southeastward across northern areas of New
England at the time of the Landsat image, with a moderate (15 to 20 knots)
south to southwest surface wind flow regime over southern New England.
This surface wind flow pattern agrees remarkably well with the orienta-
tion of the band of maximum TSP levels (>100 yg/m ) observed in Figure 3-1,
Further examination of the ensuing hourly weather information showed
that this wind pattern prevailed throughout much of the remainder of the
day.
The correlative NOAA/VHRR and GOES imagery on this date revealed no
features suggesting association with the locations of the maximum TSP
levels.
26
-------
ENVIRONMENTAL RESEARCH S TECHNOLOGY, INC
4. ANALYSIS PROCEDURE - SULFATES (SO ~)
4.1 Data Sample
Because of the difficulty of identifying useful correlative Landsat
imagery, plus possible TSP sampler location problems affecting the com-
pletion of the Task 2 phase of this investigation (see Section 3), a
modification of the original work statement was requested. It was
subsequently agreed by the EPA Project Officer that further analysis of
TSP episodes would be deleted. Instead, an analysis of one or two
incidents of high sulfate concentrations over the Northeast would be
undertaken with comparisons being conducted with correlative meteoro-
logical satellite imagery. The rationale for examining regional sulfate
episodes with corresponding satellite-observed haze layers was provided
by several recent regional air pollution studies. Husar, et al. (1976),
for example, found that large masses of low-visibility air can be tracked
as entities for several days, using only the routine synoptic observations.
Furthermore, there was evidence that the areally extensive turbidity
reflected the existence of a large number of sulfate aerosols, whose
particle size distribution shows a peak between 0.05 and 1 pm (see, for
example, Wilson, et al., 1976). The correlation between atmospheric
visibility (and turbidity) and sulfate concentration, both of which
exhibit regional characteristics, leads therefore to the present attempt
to find a qualitative correspondence via the remotely-sensed medium.
The case selected for analysis originated from ERT's data base
which is derived from the Electric Power Research Institute, Sulfate
Regional Experiment (EPRI/SURE) program, several Tennessee Valley
Authority (TVA) stations, as well as from the National Air Sampling
Network (NASN) stations. The sulfate measurements taken at all stations
consist of 24-hour average values obtained by chemical analysis of high-
volume particulate filters. This high sulfate episode over the north-
eastern United States, for which data existed in the ERT data base and
correlative satellite imagery was found to exist, occurred during the
period of 8-10 July 1974. The comparison of this high sulfate episode
with the corresponding satellite data and associated meteorological
conditions is discussed in the following section.
27
-------
ENVIRONMENTAL RESEARCH a TECHNO LOGY, INC
4.2 Comparison of 8-10 July 1974 High Sulfate Episode with Correlative
Satellite Data and Meteorological Conditions
Analysis of sulfate levels during this period, as derived from the
ERT data base, showed that this episode, which actually began about
4 July over the Ohio Valley, reached a peak on 8 and 9 July over West
Virginia, Pennsylvania and New York. By 10 July, the geographical peak
had subsided over the Ohio Valley, but a new episode had developed along
the middle Atlantic coastal region northward into southern New England.
Figure 4-1 shows the analysis of 24-hour average sulfate levels
over the region on 8 July. Concentrations as high as 40 yg/m are
observed over southwestern Pennsylvania and the Ohio-West Virginia
border. The four statute mile visibility isopleth is also indicated and
shows a generally good agreement with the location of the maximum sul-
fate levels.
The DMSP daytime visible-to-near infrared channel image and cor-
responding daytime thermal infrared image of 8 July are shown in
Figures 4-2 and 4-3, respectively. The visible-to-near infrared image
shows mostly cloud-free conditions over the region of high sulfate
concentrations but does not indicate any variations in reflectance that
could be attributed to the high sulfate concentrations. In the thermal
infrared, however, an overall increase in brightness is observed over
the region. This increase in brightness (lower temperatures) suggests
that the radiometer is measuring, at least in part, the emittance of
the top of the pollutant/haze/smoke layer rather than the surface. In
addition, the daytime surface temperatures beneath this layer averaged
4°F to 8°F lower than in surrounding, unobscured areas, thus con-
tributing to the depressed thermal response in the imagery.
The analysis of the 24-hour average sulfate levels over the region
on 9 July, shown in Figure 4-4, reveals an overall increase in maximum
concentrations with levels as high as 75 yg/m over western Pennsylvania
3
and 48 yg/m over West Virginia. The four statute mile visibility
isopleth is also indicated and again shows a generally good agreement
with the location of maximum sulfate concentrations.
28
-------
ENVIRONMENTAL RESEARCH S TECHNOLOGY. INC
Enlargements of portions of the DMSP daytime visible-to-near infrared
band and corresponding daytime thermal infrared (sulfate analysis super-
imposed) images of 9 July are shown in Figures 4-5 and 4-6. The visible-
to-near infrared image displays small areas of scattered to broken
stratocumulus clouds located over portions of New York, Pennsylvania,
West Virginia, Ohio and Kentucky but, again, does not display any patterns
associated with the area of high sulfates. These clouds also appear
very bright (cold) in the corresponding thermal infrared image; in
addition, the entire area of maximum sulfate concentrations appears
somewhat brighter (colder) on this image than the surrouding cloud-free
areas, suggesting once again the presence of high pollution concen-
trations and low visibilities. In this instance, the brighter pattern
in the thermal infrared and the reduced visibilities (four mile visibility
isopleth, Figure 4-4) both extend into Michigan and Wisconsin, sug-
gesting that high sulfate concentrations may also have been present in
those areas on this date; corroborative ground sulfate measurements were
unavailable for these two states.
Analysis of NOAA/VHRR nighttime thermal infrared imagery for 8 and
9 July also showed increased brightness (lower temperatures) within the
areas of maximum sulfate concentrations; however, the extremely high
contrast of these images made them unsuitable for reproduction, and so
they are not shown here.
The meteorological conditions during this period may be summarized
as the dominance of a warm and humid, maritime tropical air mass through-
out the entire region. On 8 July, a slow-moving cold front extending
from Quebec to Minnesota was advancing slowly southward. By 10 July
when the mid-Atlantic coast episode occurred, the cold front was situated
along an east-west line extending from central New England to Minnesota.
During the early part of the period of 8 and 9 July, the large-
scale surface wind flow was light southerly or southwesterly, rarely
reaching 10 knots at any reporting station. The fairly persistent wind
flow from the south and southwest would qualitatively indicate large-
scale transport of pollutants toward the northeast. However, in this
episode, the time required for northeastward transport of pollutants
would have been so long that pollutant advection alone cannot fully
explain the regionality of the sulfate episode. There was probably
some in situ pollution development.
29
-------
ENVIRONMENTAL RESEARCH & TECHNOLOGY INC
' — ' CTl
i — (
t/5
0> X
4-> i-l
OS p
oo
to
tn cu
CD -P
3 cti
i— I 4-1
ni en
-P
-H
f-i 0>
3 +->
O t/>
f. rt
I
f-l
4-1 O
O C
•H X
(/> -P
X
I-H fn
OJ 0)
p: >
<: o
0)
fn
M
30
-------
ENVIRONMENTAL RESEARCH S TECHNOLOGY INC
0
•M
bo
•H
0
•H
>
0
cd ^3"
• H CTi
rH
0 X
rt '3
4-1
C 00
• H
r< O
rt
0 t/i
C 0
I 4->
O CTj
I (/)
0
,0 0
•H 4->
(/) -H
•H C
> 3
0 C
• H 0
X w
as n)
13 0
CL, +->
S o
Q C
CM
I
n-
31
-------
ENVIRONMENTAL RESEARCH & TECHNOLOGY INC
f-l
o
0)
fi
bO
•H
(D
•H
0> h-
M O1
-O 3
0) 1-5
fn
rt oo
^H
MH C
C O
• H
03 4->
E rt
H 4->
0) CO
e
•H
oj C
t3 fn
-------
ENVIRONMENTAL RESEARCH & TECHNOLOGY IMC
e
M rf
3 r-~
v — ' CTi
i — l
l/l
3 rt
T— ( +->
03 CO
>
T3
CD 0)
M 4->
03 -H
fn C
03
fn
CD
f, 03
l 0)
•* ,fi
CXI 4-1
^
4n O
O C
i/l
X
>
O
3
tx>
• H
tL,
-------
ENVIRONMENTAL RESEARCH ^TECHNOLOGY INC
o
fn
oj t/)
CD (L)
C 4-1
i rt
O 4->
4-1 CO
I
H
t/> C
•H 3
c
0) fn
G CD
•H 4-1
4-1 t/)
X oj
oj 0
O, (H
co o
O
C
C
•H -H
4-> S
fn CJ
O -H
O M
nj
4,S^
CD
S 13
CD
LO
I
bo
•H
P-,
34
-------
ENVIRONMENTAL RESEARCH &TECHNOLOGY INC
Figure 4-6 Enlargement of portion of DMSP daytime thermal infrared
image viewing the northeastern United States on 9 July
1974, with sulfate analysis superimposed
35
-------
ENVIRONMENTAL RESEARCH » TECHNOLOGY INC
Highest sulfate concentrations were, for the most part, found
within the warm, moist air south of the cold front. On each day during
the period, the dewpoint temperatures reached 70°F as far north as the
mid-Atlantic states along the coast and into the upper midwest region
west of the Appalachians. In addition, visibilities were generally less
than four miles in haze and smoke over much of the region.
36
-------
ENVIRONMENTAL RESEARCH & TECHNOLOGY, INC
5. SUMMARY AND CONCLUSIONS
5.1 Discussion of Results
5.1.1 Photochemical Oxidants/Ozone
Analysis of three high photochemical oxidant episodes over southern
New England indicates a definite diurnal trend with the maximum ozone
concentrations occurring in the mid to late afternoon, and minimum
levels generally confined to the late night and early morning hours.
Unfortunately, the mid-morning readout times of the correlative Landsat
images did not coincide with these afternoon peak levels; nevertheless,
the ozone concentrations at the time of the satellite passage (1400-
1500 GMT) generally exceeded the national ambient air quality primary
and secondary standards of 0.080 ppm over much of southern New England.
Throughout each of the periods examined, broad scale haziness
resulting in reduced visibilities accompanied the high ozone events.
Analyses of the locations of minimum visibilities in haze and smoke
across southern New England were found to compare well with the loca-
tions of peak ozone concentrations.
Although, for the most part, the areas of the maximum ozone con-
centrations and haziness were not discernible in the correlative Landsat
imagery, some increased reflectance levels were observed over the darker
water areas of Long Island Sound and to the east of Boston near coastal
stations that were reporting low visibilities. Over the land areas, any
varying degree of image reflectance due to air pollution density or low
level haze layer was not discernible due to the highly variable bright-
ness levels associated with various ground-feature reflectances, espe-
cially the higher reflectances associated with dense urban development.
Examination of the correlative NOAA/VHRR and GOES satellite data
also revealed no features suggesting a relationship to locations of peak
ozone concentrations, with one notable exception. In this instance, a
NOAA-4 daytime visible image was acquired along a satellite track far
enough to the west to display a low sun angle along the mid-Atlantic
coastal region northward into southern New England. This resulted in an
area of low visibility in haze and smoke (and possibly high ozone con-
centration) of two to four miles over New Jersey, Pennsylvania, southern
37
-------
ENVIRONMENTAL RESEARCH S TECHNOLOGY INC
New York and Connecticut being easily detected on the imagery due to
illumination by the sun. Sun angle apparently plays a significant role
in the capability of satellites to observe high pollutant concentrations
and associated low visibilities in the visible portions of the spectrum.
Analysis of the meteorological conditions accompanying each of the
three high ozone episodes indicates a strong similarity in overall
synoptic pattern. In each case, the episode was accompanied by the
dominance of warm, humid tropical air throughout the region. In each
case, a nearly stationary high pressure system (anticyclone) was cen-
tered over the mid-Atlantic states resulting in the flow of warm, moist
air into New England on light to moderate south to southwesterly surface
winds. Daytime high temperature readings were generally reported in the
mid-80°F to low 90°F range, while corresponding dew point temperatures
ranged from the upper 60°F to low 70°F level.
5.1.2 Total Suspended Particulates (TSP)
A lack of correlative Landsat coverage during times of high TSP
events over southern New England, as well as cloud contamination when
other satellite coverage was available, together with a lack of con-
fidence in the representativeness of a number of the important TSP data
readings severely limited this phase of the investigation.
Analysis of one correlative Landsat image that provided coverage
during a period of isolated high TSP levels reported over southern New
England revealed no distinct boundaries to suggest any association with
the location of the maximum TSP concentrations. NOAA/VHRR and GOES
imagery for the same period also showed no evidence of high TSP concen-
trations. It appears likely, however, that the isolated high levels
were associated with localized conditions at the location sites of the
sensors rather than with a widespread pollution episode, and thus no
real test of the value of satellite imagery for high TSP episodes was
accomplished in this study program.
5.1.3 Sulfates (SO )
The final phase of this investigation involved the detection of
high sulfate episodes in correlative meteorological satellite imagery.
Although the areas of high sulfate concentrations could not be detected
38
-------
ENVIRONMENTAL RESEARCH & TECHNOLOGY, INC
in the DMSP visible-to-near infrared channel imagery at the time of a
high sulfate episode over the Northeast (8-10 July 1974), both the
daytime and nighttime DMSP thermal infrared imagery during this period
showed an overall increase in brightness (lower temperatures) within
regions of peak sulfate concentrations. The lower temperatures observed
within the area of high sulfate levels, as compared to other areas,
suggest that the satellite sensor was viewing, at least in part, the top
of the pollutant/haze/smoke layer rather than the surface. In addition,
the daytime surface temperatures beneath this layer averaged 4°F to
8°F lower than in surrounding, unobscured areas, thus contributing to
the depressed thermal response in the imagery.
The dominance of a warm, humid, maritime tropical air mass through-
out the region of high sulfate levels was observed throughout the
period. A persistent, light surface wind flow pattern from the south
and southwest indicated large-scale transport of pollutants toward the
northeast, as well as in situ development of high pollution concentrations.
5.2 Conclusions and Recommendations
Various types of satellite imagery are capable of detecting the
atmospheric conditions associated with some regionalized air pollutant
concentration situations. The current study looked at Landsat,
NOAA/VHRR, GOES and DMSP imagery products, including portions of the
visible, near infrared, and thermal infrared sectors of the spectrum.
The small sample size of each imagery type and differing concentrations
of various pollutants does not allow any definitive conclusion to be
drawn regarding the specific effectiveness of a particular imagery
type for a specified pollutant. However, some important general observa-
tions may be derived from this study, as follows:
1) It is apparent that there is a good correlation between certain
types of high pollutant load (including ozone and sulfates) and
haze and smoke, with accompanying reductions in visibility.
There is good evidence that satellite imagery can display
differences in reflectance (visible) or temperature (thermal
infrared) due to some combination of haze, smoke and atmo-
spheric pollutant load, on a regional basis. It does not
39
-------
ENVIRONMENTAL RESEARCH & TECHNOLOGY, INC
appear possible, at this time, to differentiate between
pollutant concentrations and haze and/or smoke utilizing
only the satellite imagery;
2) It has been found that some types of imagery hold more promise
than others for helping to delineate the regional extent of
pollution-associated conditions. Landsat visible imagery
tends to exhibit so much variability in reflectance levels
over land areas that differences in atmospheric conditions,
other than cloudiness, are difficult to observe. On the other
hand, differences among land features are far less pronounced
on some of the smaller scale visible and thermal infrared
imagery (NOAA/VHRR and DMSP), due to significantly lower
resolutions. These image types provide much better backgrounds
on which to detect reflectance and thermal differences resulting
from high atmospheric pollution episodes. In addition, sun
angle seems to play a significant role in the detectability of
pollutant-associated conditions in the visible portions of
the spectrum. Of the various combinations of type of episode
and imagery analyzed during this project, the greatest
promise was shown by the DMSP thermal infrared imagery in
application to high sulfate situations;
3) The value of satellite systems for monitoring the regional
transport of pollutants ultimately depends on their capability
for at least daily repetitive observation, as well as the
ability of these systems to temporally coincide their obser-
vations with the usual time of day of highest pollutant
levels.
It is suggested that further study be made to allow additional
specificity and quantification of these conclusions. Since the current
investigation was heavily Landsat-oriented, ERT recommends that a
further in-depth investigation be made of the smaller scale types of
imagery (>1 km resolution and less) such as NOAA/VHRR, DMSP and GOES for
which much greater promise has been shown in the detection and moni-
toring of pollution-associated situations. A suggested approach to such
40
-------
ENVIRONMENTAL RESEARCH S TECHNOLOGY, INC
a study would be to take several high sulfate episodes over the eastern
United States and determine the relative capability of each specific
satellite system for delineating and monitoring the movement of the
pollutant-laden air. ERT would utilize its extensive and unique data
base which is derived from the EPRI/SURE program, several Tennessee
Valley Authority (TVA) stations, and the National Air Sampling Network
(NASN) stations. As part of this study, a further investigation of the
influence of factors such as sun angle upon detection of pollutants
within the visible bands of satellite sensors could be undertaken. For
this, it would be possible to utilize the GOES imagery, which is available
on a half-hourly basis throughout the day. It is also suggested to
combine with this program a theoretical analysis of the physics of
detecting pollutants from space, to help substantiate the findings of
the imagery/field data oriented studies.
An additional topic for investigation could include analysis of
selected digitized, computer-compatible data from the NOAA/VHRR and
Landsat sensors. This would allow much finer separation of reflec-
tance and thermal levels on the imagery than by visual interpretation.
No analysis of digitized data was possible under the scope of the
current study. Some further means of extracting more useful information
from the Landsat imagery could also be assessed, such as comparing
imagery from different MSS spectral bands, perhaps using ratioing
techniques, or processing the Landsat imagery photographically in such a
way as to subdue earth surface features and enhance atmospheric features.
Since only a very limited sample was available in this study,
further investigation should be carried out to assess the utility of
satellite observations to detect high TSP episodes. As one possible
study, correlative meteorological satellite imagery (NOAA/VHRR and GOES)
and available Landsat data obtained during the recent, severe dust
storms originating over the widespread drought areas of the southwestern
United States (late February 1977) would be analyzed to determine the
extent to which the high TSP levels are detectable in the various satellite
data. ERT already has seen significant evidence of the dust clouds on
both NOAA/VHRR and GOES imagery of this period. Attempts would be made
41
-------
ENVIRONMENTAL RESEARCH & TECHNOLOGY, INC
to obtain sufficient ground truth TSP measurements for comparison with
the imagery. An investigation of this type would allow identification
of the levels and distribution of TSP necessary for detection by earth
observational satellites, as well as determination of to what extent
the advection and dispersion characteristics can be monitored by
satellite imagery viewing land areas.
The conclusions of the current study demonstrate the vast potential
that exists in the realm of satellite remote sensing for detection and
monitoring of regional pollutant concentrations. Sufficient evidence
has been accumulated here to suggest that the Environmental Protection
Agency (EPA) continue to maintain interest, and play a significant
role, in the development of the specifications for the use of present
and future earth surveillance satellite systems. Only by such con-
tinued awareness and involvement will EPA ensure that the potential of
these systems will be available in the proper modes and further
developed, as appropriate, for EPA operational purposes.
42
-------
ENVIRONMENTAL RESEARCH & TECHNOLOGY, INC
6. REFERENCES
Brown, F. R., and F. S. Karn, 1976: "Air Pollution from the Ohio River
and Monongahela River Valleys," ERTS-1, A New Window on our Planet,
U. S. Geological Survey, Professional Paper 929, pp. 261-265.
Griggs, M., 1973: "A Method to Measure the Atmospheric Aerosol Content
Using ERTS Data," Third ERTS Symposium, Paper E2, pp. 1505-1518.
Husar, R. B., N. V. Gillani, J. D. Husar, C. C. Paley, and P. N. Turcu,
1976: "Long Range Transport of Pollutants Observed Through Visibility
Contour Maps, Weather Maps and Trajectory Analysis," Proc. Third
Symposium on Atmospheric Turbulence, Diffusion and Air Quality,
October 19-22, 1976, Raleigh, North Carolina, American Meteorological
Society, pp. 344-347.
Lyons, W. A., and R. A. Northouse, 1973: "Use of ERTS-1 Imagery in Air
Pollution and Mesometeorological Studies Around the Great Lakes,"
Third ERTS Symoposium, Paper El, pp. 1491-1504.
Lyons, W. A., and R. B. Husar, 1976: "SMS/GOES Visible Images Detect a
Synoptic-Scale Air Pollution Episode," Monthly Weather Review,
Vol. 104, pp. 1623-1626.
McClellan, A., IV, 1971: "Satellite Remote Sensing of Large Scale Atmo-
spheric Pollution," Proc. Second International Clean Air Congress,
Academic Press, New York, p. 1408.
Prospero, J. M., E. Bernatti, C. Schubert, and T. N. Carlson, 1970: "Dust
in the Caribbean Atmosphere Traced to an African Dust Storm,"
Earth and Planetary Science Ltrs., Vol. 9, p. 287.
Tong, E. Y., G. Battel, and R. B. Batchelder, 1976: "Case Studies of
Atmospheric Sulfate Distribution over the Eastern United States,"
Proc. Fifteenth Purdue Air Quality Conference, November 8-9, 1976,
Indianapolis, Indiana.
Wilson, W. E., R. J. Charlson, R. B. Husar, K. T. Whitby, and D. Blumenthal,
1976: "Su_lfates in the Atmosphere", Paper 76-30-06, 69th Annual
Meeting, Air Pollut. Control Assoc., Portland, Ore., 20 pp.
43
------- |