TVA
EPA
Tennessee
Valley
Authority
Office of Natural
Resources
Norris TN 37828
TVA/ONR/ARP-81/6
United States
Environmental Protection
Agency
Office of Environmental
Engineering and Technology
Washington DC 20460
EPA-600/7-81-114
July 1981
Research and Development
Remote Sensing of
Sulfur Dioxide
Effects on
Vegetation—Final
Report
Volume II.
Data
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development, U.S. Environmental
Protection Agency, have been grouped into nine series. These nine broad cate-
gories were established to facilitate further development and application of en-
vironmental technology. Elimination of traditional grouping was consciously
planned to foster technology transfer and a maximum interface in related fields.
The nine series are:
1. Environmental Health Effects Research
2. Environmental Protection Technology
3. Ecological Research
4. Environmental Monitoring
5. Socioeconomic Environmental Studies
6. Scientific and Technical Assessment Reports (STAR)
7. Interagency Energy-Environment Research and Development
8. "Special" Reports
9. Miscellaneous Reports
This report has been assigned to the INTERAGENCY ENERGY-ENVIRONMENT
RESEARCH AND DEVELOPMENT series. Reports in this series result from the
effort funded under the 17-agency Federal Energy/Environment Research and
Development Program. These studies relate to EPA's mission to protect the public
health and welfare from adverse effects of pollutants associated with energy sys-
tems. The goal of the Program is to assure the rapid development of domestic
energy supplies in an environmentally-compatible manner by providing the nec-
essary environmental data and control technology. Investigations include analy-
ses of the transport of energy-related pollutants and their health and ecological
effects; assessments of, and development of, control technologies for energy
systems; and integrated assessments of a wide range of energy-related environ-
mental issues.
This document is available to the public through the National Technical Informa-
tion Service, Springfield, Virginia 22161.
-------
TVA/ONR/ARP-81/6
EPA-600/7-81-114
July 1981
REMOTE SENSING OF SULFUR DIOXIDE EFFECTS ON VEGETATION
FINAL REPORT
VOLUME II - DATA
by
C. Daniel Sapp
Office of Natural Resources
Tennessee Valley Authority
Chattanooga, Tennessee 37401
Interagency Agreement EPA-IAG-D8-E721-DJ
Project No. E-AP 80 BDJ
Program Element No. INE 625C
Project Officer
James Stemmle
U.S. Environmental Protection Agency
401 M Street, SW.
Washington, DC 20460
Prepared for
OFFICE OF ENERGY, MINERALS, AND INDUSTRY
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
WASHINGTON, DC 20460
U.S. Environmental Protection Agency
-/"'".- ,., ;;.. Library (5PL-13)
-------
DISCLAIMER
This report was prepared by the Tennessee Valley Authority and has
been reviewed by the Office of Energy, Minerals, and Industry, U.S.
Environmental Protection Agency, and approved for publication. Approval
does not signify that the contents necessarily reflect the views and
policies of the Tennessee Valley Authority or the U.S. Environmental
Protection Agency, nor does mention of trade names or commercial products
constitute endorsement or recommendation for use.
11
-------
ABSTRACT
Three techniques for detecting and mapping sulfur-dioxide (862)
effects on the foliage of sensitive crops and trees near large, coal-fired
power plants were tested and evaluated. These techniques were spectrora-
diometry, photometric analysis of aerial photographs, and computer classi-
fication of airborne multispectral scanner data.
Spectroradiometry is a useful, ground-based technique for measuring
the changes in reflectance that accompany exposure of sensitive crops to
SC>2. Photometric analysis of aerial color-infrared photographs has some
practical advantages for measuring the reflectances of forest species or
for synoptic point-sampling of extensive areas; these tasks cannot be done
effectively by field crews. The relationships among reflectance, foliar
injury, and yield of crops are complex and are affected by many extraneous
variables such as canopy density. The SC>2 effects are easier to detect on
winter wheat than on soybeans, but in either case they cannot be consis-
tently detected by airborne remote sensors except under near-ideal condi-
tions when the injury is moderate to severe. Airborne multispectral
scanner data covering affected soybean fields were analyzed using three
computer-assisted classification procedures: unsupervised, supervised,
and pseudosupervised; the last method provided the best results. Landsat
imagery was also investigated, but the foliar effects of SQ^ were too
subtle to detect from orbit.
This report was submitted by the Tennessee Valley Authority, Office
of Natural Resources, in fulfillment of Energy Accomplishment Plan 80 BDJ
under terms of Interagency Agreement EPA-IAG-D8-E721-DJ with the Environ-
mental Protection Agency. Work was completed as of December 1980.
111
-------
CONTENTS
Page
Abstract iii
Figures vii
Tables x
Acknowledgment xii
1. Introduction 1
Background 1
Reflectance Properties and Physiological
Aspects of Vegetation 3
Reflectance and Vegetative Stress 6
Spatial Considerations 7
Literature Review 7
Purpose and Objectives 12
Scope 12
Hypothesis 14
2. Conclusions and Recommendations ~. . . . . 15
Conclusions 15
Recommendations 17
3. Laboratory Spectroradiometry 18
General 18
Instruments 18
Concepts of Radiance and Reflectance 21
Curve Normalization 21
Experimental Design 23
Plants 23
Scanning Procedure 24
Analysis of Soybeans 24
Analysis of Winter Wheat 25
Results and Discussion 26
Soybeans 26
Winter Wheat 31
Summary of Results 33
4. Field Spectroradiometry 37
General 37
Experimental Design 37
Plot Preparation and Planting 37
Exposure of Plots to S02 38
Observations of Foliar Injury 41
Scanning Procedure 42
Normalization of Spectral Curves 45
Results and Discussion 46
Soybeans 46
Winter Wheat 55
Summary of Results 66
IV
-------
CONTENTS
(Continued)
5. Interpretation and Analysis of Aerial Photographs .... 70
General 70
Methods and Instruments 70
Overflights 70
Photometric Analysis 75
Results and Discussion 79
General 79
Colbert Test 79
Field Conditions 79
Measurement and Comparison of
Reflectance 83
Johnsonville Test 84
Field Conditions 84
Measurement and Comparison of
Reflectance 88
Summary of Results 94
6. Analysis of Multispectral Scanner Data 96
General 96
Flight Lines and Sensor Characteristics 96
Ground Truth 102
Colbert Area 102
Shawnee Area 104
Data Reduction and Processing Procedures 104
Preprocessing Procedures 104
Processing Procedures 107
Pattern Recognition 107
Optimal Channel Selection Ill
Conventional Analysis Procedures 113
Procedures for Evaluating Classification
Accuracy 113
Results and Discussion 114
Optimal Altitudes 114
Colbert Test 115
Optimal MSS Channels 115
Unsupervised Classification 116
Supervised Classification 119
Shawnee Test 119
Optimal MSS Channels 119
Unsupervised Classification 121
Pseudosupervised Classification 124
Summary of Classification Results 135
Enhancement of Patterns of S02 Effects
Within Soybean Fields 135
v
-------
CONTENTS
(Continued)
References
Appendices
A Specifications of Optical Multichannel Analyzer
TSR System 144
B Spectral Curves, Laboratory Experiment 148
C Analysis of Variance, Laboratory Experiment 190
D Observations of Foliar Injury 195
E Spectral Curves, Experimental Plots 212
F Analysis of Variance, Experimental Plot 221
G Photometric Calibration 241
H NASA/ERL Letter Report 255
I Supervised Classification Procedure 257
J Pseudosupervised Classification Procedure 261
K Unsupervised Classification Procedure 264
VI
-------
FIGURES
Number Page
1 Reflectance of a typical leaf 4
2 Optical properties of a typical leaf 5
3 Effects of different spectroradiometer fields of
view on data characteristics 8
A Western part of Tennessee Valley showing four
steam plants 13
5 TSR system configuration 19
6 TSR system components in laboratory 20
7 Procedure for normalizing spectral curves of
radiance for laboratory experiments 22
8 Mean spectral curves for classes of chlorosis in
soybeans exposed to S02 28
9 Mean spectral curves for classes of necrosis in
soybeans exposed to SC-2 29
10 Mean spectral curves for classes of unaffected
and SC>2-affected winter wheat 34
11 Map of experimental plot of soybeans 39
12 Map of experimental plot of wheat 39
13 Aerial view of experimental plot of wheat 41
14 Mobile platform for spectroradiometry 43
15 Geometric configuration for scanning subplots .... 44
16 Procedure for normalizing spectral curves of
radiance for experimental field plots 46
17 Linear regression of yield on necrosis for
soybeans exposed to SC-2 47
18 Mean spectral curves for soybeans exposed to 862 . . 49
19 Mean spectral curves for S02-affected soybeans ... 52
20 Linear regressions of reflectance on necrosis
for S02-affected soybeans 54
21 Linear regression of reflectance curve area on
necrosis for winter wheat exposed to SQ% 56
22 Mean spectral curves for winter wheat exposed to
S02 58
23 Mean spectral curves for S02~affected winter
wheat 63
24 Linear regressions of reflectance on necrosis for
winter wheat 65
25 Linear regression of yield on necrosis for
S02-affected winter wheat 67
26 Linear regressions of reflectance on yield of
S02~affected winter wheat 68
27 Colbert Steam Plant area in northwestern Alabama . . 71
28 Johnsonville Steam Plant area in western Tennessee . 72
29 Shawnee Steam Plant area in western Kentucky .... 73
30 Widows Creek Steam Plant area in northeastern
Alabama 74
31 Grid for intensive sampling of soybean field near
Colbert Steam Plant 78
VII
-------
FIGURES
(Continued)
Number
32 Flight lines and S02-affected areas near Colbert
Steam Plant in 1977 80
33 Aerial CIR photograph showing Colbert area in 1977 . 81
34 Regression of IR/red ratio and foliar injury
levels for Colbert soybean fields photographed
in 1977 84
35 Thematic maps showing patterns in a soybean field
near Colbert Steam Plant in 1978 85
36 Computer-generated perspective views of a soybean
field near Colbert Steam Plant in 1978 86
37 Flight lines and S02-affected areas near
Johnsonville Steam Plant 89
38 Aerial CIR photograph of area near Johnsonville Steam
Plant 90
39 Statistical regressions of reflectance, injury
levels, and yield for Johnsonville soybean
fields 93
40 Selected S02-affected areas and flight lines near
Colbert Steam Plant in northwestern Alabama in
1977 98
41 Selected S02-affected areas and flight lines near
Shawnee Steam Plant in western Kentucky in 1978 . . 99
42 Basic components of a digital multispectral scanner
system 100
43 Configuration for airborne MSS data acquisition . . . 101
44 Distribution of S02-affected soybean fields near
Colbert Steam Plant in 1977 103
45 Distribution of S02-affected soybean fields near
Shawnee Steam Plant in 1978 105
46 Eight MSS channels depicting a scene near Colbert
Steam Plant 106
47 Data Analysis System at EMSL-LV 108
48 Functional tasks for processing MSS data on Data
Analysis System at EMSL-LV 109
49 Hypothetical two-dimensional plot of naturally
clustered spectral measurements of soybeans .... Ill
50 Unsupervised classification of MSS data for
Colbert area 117
51 Distribution of S02 effects within soybean fields
near Colbert Steam Plant 118
52 Unsupervised classification of MSS line 2 data near
Shawnee Steam Plant 122
53 Distribution of S02 effects within soybean fields
near Shawnee Steam Plant (MSS line 2,
unsupervised) 123
54 Pseudosupervised classification of MSS line 2 data
near Shawnee Steam Plant 127
Vlll
-------
FIGURES
(Continued)
Number
55 Distribution of SQ% effects within soybean fields
near Shawnee Steam Plant (MSS line 2,
pseudosupervised) .' 128
56 Pseudosupervised classification of MSS line 3
data near Shawnee Steam Plant 132
57 Distribution of S02 effects within soybean fields
near Shawnee Steam Plant (MSS line 3) 133
58 Enhanced MSS image (line 2) of area near Shawnee
Steam Plant showing S02-affected soybean fields . . 137
IX
-------
TABLES
Number Page
1 Data Classes and Foliar Effects of S02 on Soybeans . 27
2 Simple Correlation Coefficients (r) for Single-Band
Reflectance and Foliar Injury to Soybeans 31
3 Data Classes and Foliar Effects of S02 on Winter
Wheat 32
4 Simple Correlation Coefficients (r) for Single-Band
Reflectance and Foliar Injury to Winter Wheat ... 33
5 Reflectance Statistics for Classes of S02-Affected
Winter Wheat 35
6 S02 Dose, Injury, and Yield of Soybeans 47
7 Significant Differences in Reflectance between
S02-Exposed and Control Soybeans 51
8 Simple Correlation Coefficients (r) for Single-Band
Reflectance, Foliar Injury, and Yield of Soybeans . 53
9 Significant Differences in Reflectance and Yield
between S02-Affected and Unaffected Soybeans ... 55
10 Total Reflectance, S02 Dose, Foliar Injury, and
Yield of Winter Wheat 56
11 Significant Differences in Reflectance among
Four Classes of S02-Exposed Winter Wheat and
Control 57
12 Reflectance, Necrosis, S02 Concentrations, and
Yield for Winter Wheat by Dose Class 61
13 Reflectance, Necrosis, and Yield for Winter Wheat
by Injury Class 62
14 Simple Correlation Coefficients (r) between
Reflectance and Two Other Variables (Necrosis
and Yield) for Wheat 64
15 Significant Differences in Reflectance and Yield
among Four Necrosis Classes of S02~Affected
Winter Wheat 66
16 Reflectance and Foliar Injury for Colbert Soybean
Fields Photographed in 1977 83
17 Reflectance, Foliar Injury, and Yield for Soybean
Fields near Johnsonville 91
18 Coefficients of Determination (r2) for Reflectance,
S02 Injury, and Yield of Soybean Fields near
Johnsonville 92
19 Multispectral Scanner Channels 112
20 Optimal MSS Channels for Detecting and Classifying
S02-Affected Soybean Fields near Colbert Steam
Plant in 1977 115
21 Comparison of Field Observations of S02 Effects on
Soybeans and Results of MSS Unsupervised
Classification for Colbert Scene 2 120
22 Optimal MSS Channels for Detecting and Classifying
S02-Affected Soybean Fields near Shawnee Steam
Plant in. 1978 121
-------
TABLES
(Continued)
Number
23 Comparison of Field Observations of S02 Effects
on Soybeans and Results of MSS Unsupervised
Classification for Shawnee Flight Line 2 125
24 Comparison of Field Observations of S02 Effects on
Soybeans and Results of MSS Pseudosupervised
Classification for Shawnee Flight line 2 130
25 Comparison of Field Observations of S02 Effects on
Soybeans and Results of MSS Pseudosupervised
Classification for Shawnee Flight Line 3 134
26 Summary of Errors Using Three Procedures for
Detecting and Classifying S02 Effects on
Soybeans 136
27 Within-Field S02 Effects on Soybeans near Shawnee
Steam Plant in 1978 138
XI
-------
ACKNOWLEDGMENT
The EPA Project Officer for this research project is James Stemmle,
401 M Street, SW., Washington, DC. His contribution to the direction of
the research and his constructive review of the reported results are
appreciated. The TVA Project Director is Herbert C. Jones, Supervisor,
Air Quality Research Section, Air Resources Program, River Oaks Building,
Muscle Shoals, Alabama.
xn
-------
SECTION 1
INTRODUCTION
BACKGROUND
The effects on vegetation of sulfur dioxide (802) emissions from
large, coal-fired power plants have been recognized as a potential problem
for several decades. For measuring the intensity, geographical distribu-
tion, and areal extent of foliar injury, remote sensing offers numerous
advantages over traditional field-based surveillance. The traditional
method for gathering such information is to observe and record injury to
S02~sensitive indicator species such as ragweed and blackberry. Fixed
S02 monitoring stations are used to determine the spatial characteristics
•
of plume contact with the ground. Information from field observations and
monitors is often used to prepare maps illustrating the foliar injury from
S02 episodes.
Some problems exist with the traditional approach to surveying and
identifying S02 effects. The network of fixed S02 monitors around most
coal-fired power plants is often inadequate for mapping the exact limits
of the plume's contact with the ground. Field botanical surveillance is
usually restricted to readily accessible areas because of the requirement
to reconnoiter extensive areas quickly. The complex process of identifying
and recording symptoms of foliar injury requires highly trained personnel
who must distinguish between S02 effects and similar effects produced by
herbicides, lack of essential plant nutrients, and senescence.
Remote sensing can assist those engaged in field surveillance of S02
effects on crops and trees. The technique provides a permanent record on
film or magnetic tape. An airborne instrument platform can continuously
cover extensive areas, some of which may be inaccessible to field teams.
-------
-2-
The coverage can be essentially synoptic rather than spread over a period
of several days or more. Perhaps the greatest advantage of remote sensing
over field observations has to do with its greater objectivity and consist-
ency of measurements over time and space.
The state of the art of remote sensing requires that ground truth--
field observations—be gathered to support the analysis of the remotely
sensed data. Preliminary but detailed information must be gathered con-
cerning differences in spectral reflectance between the objects of interest
(affected foliage) and the background (unaffected foliage). Spectral
measurements may be made in the laboratory, in the field, or in both. If
a sufficient number of the resulting spectral curves is analyzed statis-
tically, the investigator can develop some sound generalizations about the
changes in spectral reflectance that occur after vegetation is stressed by
S02- Such information would allow the selection of appropriate sensor
configurations, films, filters, scanner channels and bandwidths, and other
options.
Although the traditional procedure for acquiring radiant energy data
in the field involves making measurements at discrete points with a porta-
ble spectroradiometer, comparable data may also be obtained indirectly by
making point measurements of the optical density of aerial photographs
covering the area. With appropriate calibrations, calculations, and cor-
rections of the data, either technique yields spectral reflectances that
may indicate S02-induced stress. Whereas spectroradiometry is primarily
a research method, aerial photographic radiometry--or photometry—can be
used operationally to determine the reflectance of agricultural crops and
forest species. Moreover, aerial photographs are permanent records of an
-------
-3-
air pollution episode that may be reexarained later if necessary. No ter-
rain is inaccessible to remote sensors. Vast areas may be reconnoitered
in hours rather than in weeks or months.
REFLECTANCE PROPERTIES AND PHYSIOLOGICAL ASPECTS OF VEGETATION
The first consideration of this effort was selection of an appropriate
region of the electromagnetic spectrum. For detecting the effects of air
pollution on vegetation, we selected the visible region, from 400 to 700
nanometers (run) wavelength; the near-infrared region, from 700 to 1300 nm;
and the thermal (far) infrared region from 8 to 14 micrometers (|Jm). This
selection was based partly on the inherent properties of the spectrum and
partly on the capabilities and availabilities of remote sensors.
Consider a typical reflectance spectrum of a leaf (Figure 1). Only
part of the incident energy is reflected; the remainder is absorbed or
transmitted (Figure 2). Beyond 1200 nm there is a zone of high absorption
by water.1 The visible region also shows high absorption of incoming radi-
ation; in this case it is caused by leaf pigments including chlorophylls,
carotenes, xanthophylls, and anthocyanins.2 Although chlorophylls govern
the reflectance in the visible spectrum, these pigments have no effect in
the infrared; they are completely transparent to infrared radiation.3
It is generally accepted that the high infrared reflectivity of
leaves is caused by their internal cellular structure. Some of the incom-
ing radiation is diffused and scattered through the cuticle and epidermis
to the mesophyll cells and air cavities inside the leaf. The radiation
is reflected and refracted many times within the leaf. When chlorophyll
and water are present in the leaf, the air cavities are filled and absorp-
tance of radiation predominates. This accounts for spectral bands of low
reflectance and transmittance by leaves.4
-------
-4-
100
80
60
o
cc
40
20
ABSORPTANCE
BY
PIGMENTS
0
400
TRANSMITTANCE
20
40
REFLECTANCE
LU
O
60
80
100
600 800 1000
WAVELENGTH (nm)
1200
Figure 1. Reflectance of a typical leaf in the visible and
near infrared.
-------
-5-
60 r
o
o
LlJ
ce
40 '
20 -
0
High Near-Infrared
Reflectance
400 600 -800 1000
WAVELENGTH (nm)
1200
Figure 2. Optical properties of a typical leaf
in the visible and near infrared.
-------
-6-
REFLECTANCE AND VEGETATIVE STRESS
When leaves are affected by disease and physiological stress, the
greatest change in reflectance occurs in the visible region rather than
the infrared because of the sensitivity of chlorophyll to physiological
disturbances.4 Moreover, changes in infrared reflectance are not very
reliable for indicating stress in plants.5 In advanced stages of sene-
scence, the infrared reflectance always decreases, most likely because of
the breakdown or deterioration of cell walls.6 What happens to visible
reflectance as a leaf is stressed? Wert measured ponderosa pine foliage
affected by oxidant air pollution and reported that visible reflectance
increased as chlorophyll content decreased.7 This agrees with empirical
data gathered for this report.
Our approach used the visible, the near-infrared, and the far-infrared
spectra. The ratio of near-infrared to red (IR/red) reflectance has
received considerable attention by many investigators, beginning with
Jordan8 in 1969, who used it to estimate biomass and leaf area index.
Colwell9'10 also found the ratio useful for estimating biomass. Others11'12'13
applied the IR/red ratio to Landsat image analysis for determining range
grassland biomass. The ratio is considered to be a measure of relative
"greenness" of vegetation.14 Thus, the implication is that the ratio is
associated with vegetative vigor or stress as well as canopy density (leaf
area index) and, roughly, photosynthetically active biomass. Because the
ratio is associated with so many other variables, it cannot be a perfect
indicator of stress; however, it appears to be the best available measure.
The actual reflectance curve of a species or variety of plant under
stress is not easily predicted. Visible reflectance generally increases
with stress, but the response of reflectance in the near-infrared is
-------
-7-
variable, although it eventually decreases in advanced senescence. In
remote sensing studies, the stress-causing agent cannot usually be identi-
fied without ground truth. Foliar markings, which indicate the identity
of the agent, cannot be resolved from the distances or altitudes at which
the sensor is operated. However, clusters of stressed plants can often be
distinguished from a background of normal plants.
SPATIAL CONSIDERATIONS
This study relates spectral measurements of reflectance from indi-
vidual leaves to the reflectance from entire canopies such as might be
imaged by remote sensors. The effects of the background (soil surface,
other vegetation, and shadows) must be considered. These normally reduce
the reflectance of a canopy below that of a single leaf.
The remote sensor field of view is another consideration. Figure 3
illustrates the effects of field of view on data characteristics. The
three inner concentric circles represent receptor apertures (angles) of
6 minutes, 20 minutes, and 1 degree, and the sensor would read a correct
spectral signature for the crop (soybeans). The outer circle (3-degree
aperture) would read a meaningless mixture of trees, cultural features,
and soybeans.
LITERATURE REVIEW
Although considerable data have been published to describe remote
sensing applications in plant disease and the detection of moisture stress,
very little information exists on the use of this tool for the detection
and study of air pollution effects on vegetation. Most of the reported
studies deal only with the use of aerial color-infrared photography.
There is some continuing work at the Environmental Protection Agency (EPA)
-------
-8-
Figure 3. Effects of different spectroradiometer fields of view on data
characteristics. The three inner concentric circles (represent-
ing apertures 6', 20', and 1°) would indicate a correct spectral
signature for the crop (soybeans), but the outer circle (3°)
would produce an integrated, meaningless mixture of reflectance
from trees, cultural features, and soybeans. (Infrared photo by
Tennessee Valley Authority, scale approximately 1:6000.)
-------
-9-
at Las Vegas and Corvallis, using Landsat and airborne raultispectral scan-
ner data for the detection of air pollution effects on vegetation.
Volume II of the Manual of Remote Sensing15 outlines some of the
remote sensing work dealing with air pollution effects on forest lands.
The works of Miller, et al,16 Heller,17 and Wert18 in 1969 are quoted.
These investigators studied the effects of oxidant air pollution on pon-
derosa pine foliage in the area of Los Angeles, California, using color
films. The best combination of film and scale for detecting the air pollu-
tion effects was reported as Anscochrome D/200 color film with a didymium
rare-earth filter, taken at large scale (1:1,600).
Zealer, Heller, Norick, and Wilkes19 published a report in 1971
covering the feasibility of using aerial color photography to detect and
evaluate S02 injury to timber stands in the vicinity of the Widows Creek
Steam Plant in northeast Alabama. These investigators reported that nor-
mal color film, exposed through a didymium filter at very large scales
(1:800 to 1:1,584), was the most accurate system, thus confirming the
results of the 1969 report by Heller.20 EPA asked the Tennessee Valley
Authority (TVA) to participate with the U.S. Forest Service in the 1971
study because TVA had identified visible SC>2 effects on the foliage of
trees in the vicinity of the Widows Creek Steam Plant and was conducting
intensive ground surveys to determine the extent and severity of the injury.
Some published findings describe attempts to use satellite data for
detecting zinc smelter air pollution damage and alteration to vegetation.
Fritz and Pennypacker21 were unable to distinguish an affected eastern
white pine stand from a healthy stand because of insufficient spatial
resolution of the Landsat MSS system. An earlier report by Wiegand22
describes successful application of similar satellite data to differentiate
-------
-10-
chlorotic (iron deficient) and normal grain sorghum plants. Affected areas
as small as 1.1 hectares (ha) could be detected using digital techniques
including the ratio of channel 5 (red) to channel 7 (near-infrared). Fritz
and Pennypacker1s negative results seem to contradict the earlier findings
of Wiegand, indicating that more work needs to be done using satellite
platforms for detection of S02 effects. NASA concludes that the success-
ful use of Landsat imagery to detect SC>2 effects depends on the severity
of the air pollution effects (Dr. Cartmill, personal communication, 1976).
Regardless, the limits of detection need to be determined and defined.
A few publications describe the use of aerial color-infrared photog-
raphy from aircraft to detect and study SC>2 effects on vegetation.
Murtha23 describes a method whereby S0£ damage to forest lands can be
evaluated from flying heights up to 12,200 meters (m) and at image scales
as small as 1:160,000. It was not possible, however, to delineate zones
of "light injury" using such techniques.
German scientists have used aerial photography for detecting air
pollution injury to trees. Color-infrared film at large scale (1:5,000)
is favored for the task. East German investigators have used spectrozonal
(multispectral) aerial photography at scales ranging from 1:5,000 to
1:8,000).24
Walker and Dahm25 developed a method for measuring environmental
stress. Aerial color-infrared film was used to detect moisture stress,
air pollution-induced stress, and disease-induced injury or damage to
vegetation. A patented "photometric" calibration technique accounts for
the variables affecting the image scene.
Schott, et al.,26 recently published a report (not dated) that described
an aerial photographic technique for measuring vegetative stress from SC>2.
-------
-11-
The authors correlated infrared reflectance, measured from the photograph,
with stress. They recommend that photographs be obtained during the
maximum vigor period. The final report describes several problems that
prevented the acquisition of definitive results. These problems included
lack of photography during the appropriate part of the growth cycle, lack
of sensitometric data, and presence of senescent plants.
Some studies have focused upon laboratory use of photography for the
detection of air pollution effects on vegetation. Pell and Brock recently
completed a fumigation chamber study of hybrid poplar response to ozone.27
Special filters and a densitometer were used for the photographic experi-
ment. The investigators state that the infrared photographs showed areas
of injury that were not otherwise visible to the observer. Other labora-
tory work involving photographing diseased plants using color-infrared,
black-and-white infrared, and other photographic films has been published.
These papers are too varied to summarize, but many excellent papers deal-
ing with the topic are published in the Journal of the Biological Photo-
graphic Association, such as one authored by Jackson, entitled "Detection
of Plant Disease Symptoms by Infrared."28 The Eastman Kodak publication
M-28 on infrared photography also has technical merit.29
A basis for interpretation of thermal infrared imagery for detection
of moisture stress in vegetation is given by Rohde and Olson.3 Their
results could perhaps be extended to detecting vegetative stress induced
by air pollution. Their idea is that stressed vegetation should be several
degrees warmer than adjacent healthy vegetation in early afternoon when
transpiration rates are at maximum, and that thermal infrared scanning
can detect this difference.
-------
-12-
Finally, another publication should be mentioned although its intent
is not to deal with the photographic technique at all. This is a pic-
torial atlas by Jacobson and Hill which is helpful in recognizing air
pollution injury to vegetation. The book contains pictures of affected
leaves and descriptions of the characteristics of injury and damage. The
color shifts described have a direct application to remote sensing.31
PURPOSE AND OBJECTIVES
The purpose of this five-year study was to analyze and evaluate
remote sensing techniques to detect and map the effects of S02 emissions
from large, coal-burning power plants on the foliage of sensitive crops
and trees. The objectives were to test, refine, and develop remote sens-
ing instrumentation and techniques for this purpose. Ground-based, air-
borne, and satellite platforms were used for gathering data.
SCOPE
The scope of the project included four coal-fired power plant sites,
several experimental plots, and several species of vegetation. These sites
were: the Shawnee Steam Plant in western Kentucky; the Widows Creek Steam
Plant in northeastern Alabama; the Colbert Steam Plant in northwestern
Alabama; and the Johnsonville Steam Plant in western Tennessee (Figure 4).
The experimental plots were located in Muscle Shoals, Alabama, and near
Widows Creek Steam Plant. Laboratory-based spectroradiometric experiments
were performed on soybeans, wheat, and cotton at a TVA facility in Muscle
Shoals. The digital image analysis was done at EPA's Environmental Moni-
toring Systems Laboratory in Las Vegas, Nevada (EMSL-LV), and at TVA's
Mapping Services Branch in Chattanooga, Tennessee. Photometric analysis
of aerial photographs showing S02~affected soybean fields was performed
at Calspan's Advanced Technology Center in Buffalo, New York.
-------
-13-
ILL
.BOUNDARY OF TENNESSEE RIVER WATERSHED
10 0
SCALE
50
100km
Figure 4. Western part of Tennessee Valley showing four steam plants.
-------
-14-
Investigations began with the following list of S02-sensitive crop
and tree species:
Common Name Scientific Name
1. Soybeans Glycine max (L.) Merr.
2. Winter wheat Triticum aestivum
3. Cotton Gossypium hirsatim
4. Virginia pine Pinus virginiana
5. Loblolly pine Pinus taeda
6. White pine Pinus strobus
7. Shortleaf pine Pinus echinata
8. Hickory Garya spT
9. Northern red oak Quercus rubra (L.)
10. Southern red oak Quercus falcata Michx
As the project progressed, its scope was narrowed to exclude the
hardwoods (numbers 8, 9, and 10), because S02~affected stands of these
trees were never encountered. Some affected pine stands were found and
studied, but the injury was light and discontinuous. Affected wheat and
cotton fields were never found, so SC>2 injury to plots containing these
species was induced experimentally for study. Affected soybeans were
studied in the laboratory, in experimental plots outdoors, and in the
field under natural conditions using several remote-sensing techniques.
HYPOTHESIS
The hypothesis of this research is that there is a relationship
between the reflectance of the plants and foliar injury from sulfur
dioxide. If such a relationship exists, it will be characterized in
detail, because there would be a theoretical basis for using remote sen-
sors to detect and map the distribution of S02~affected plants. Possible
relationships between reflectance and other variables including crop
yield, canopy density, stage of growth, soil moisture, and weediness will
also be investigated whenever data are adequate. Appropriate statistical
tests will be used to measure the strength of associations between and
among variables.
-------
-15-
SECTION 2
CONCLUSIONS AND RECOMMENDATIONS
CONCLUSIONS
Spectroradiometry is a useful technique for measuring the changes in
visible and near-infrared (IR) reflectance that may accompany S02~induced
injury to the foliage of sensitive row crops. The technique has proven
impractical for measuring reflectances of mature trees because of the dif-
ficulty in scanning such large objects systematically from specific angles.
Spectroradiometry is practical for providing data before performing remote-
sensor overflights so that the instrument systems can be designed and
calibrated to detect the stress.
Complex relationships exist between reflectance and foliar injury to
soybeans and winter wheat. Laboratory-based scanning experiments indicate
that the total (visible plus IR) spectrum is associated with the necrosis
symptom but not chlorosis in soybean foliage. The IR spectrum alone shows
no association with either symptom, but the IR/red ratio does correlate
significantly (a=.05) with necrosis in soybeans. In scans of wheat, the
total visible spectrum, as well as the green and red bands, shows close
relationships with chlorosis and necrosis. IR scans of wheat were not
made for the laboratory experiment.
Analysis of data from scans of experimental field plots shows no
relationship between reflectance and the observed symptoms of injury to
the soybeans. The natural range of variation in reflectance of the
foliage is evidently greater than the effects induced by S02- For wheat,
however, there is an association between reflectance and observed necrosis
and chlorosis in wheat. The relationship includes the total (visible and
IR) spectrum, as well as the individual green, red, and IR bands.
-------
-16-
The crop yield variable is related to the reflectance of wheat but
not of soybeans. The red band and the IR/red reflectance ratio of wheat
is closely associated with yield.
Aerial overflights of S02-affected soybean fields near three coal-
fired power plants produced color-infrared photographs from which reflect-
ances were obtained photometrically. When the reflectance measurements
were compared with field observations of foliar injury, the results showed
that single-band reflectance (green, red, and IR bands) was not statis-
tically related to foliar injury. Comparison of the IR/red ratio and
injury showed conflicting results. The green, red, and IR bands were each
highly related to yield, but the IR/red ratio was not.
Analysis of multispectral scanner data indicates it is possible to
differentiate S02-affected soybean fields from unaffected fields when the
canopies are continuous (uniform and dense) and the foliar effects are at
least moderate to severe. A "pseudosupervised" classification procedure
provides greater accuracy than two other procedures, "unsupervised" and
"supervised," and it can distinguish moderately to severely affected soy-
beans from unaffected soybeans with errors ranging from 11 to 24 percent.
Experience with three aircraft altitudes for acquiring scanner data
indicates that 1800 m flying height is superior to 500 and 3660 m for
detecting and mapping moderate to severe chlorosis symptoms on the foliage
of row crops. Light chlorosis may be below the threshold of detection of
an aerial multispectral scanner or camera system regardless of platform
altitude, and is certainly too subtle to be detected by these sensors in
orbital vehicles such as Landsat. The results also indicate that unless
the S02 effects on foliage are severe enough to result in necrosis, they
will not be detectable from any operational altitude (>150 m above ground
-------
-17-
level) by remote sensors. Even when necrosis is present, detection may
be possible only under ideal conditions when the influence of masking
variables (weediness, non-S02~related stress, etc.) is absent.
RECOMMENDATIONS
Spectroradiometry should be an integral part of flight planning for
remote sensor data acquisition whenever the goal is to detect and map
stressed crops. The scanning technique is useful for measuring the
spectral differences between target and background so that success in
detecting stressed vegetation can be predicted. It also has value as a
laboratory instrument for quantifying levels of foliar injury.
Photometric analysis of aerial photographs is a potential alternative
to field-based scanning with a spectroradiometer. It could be useful if
reflectances need to be sampled at many points or over an extensive area.
Our negative findings were probably a result of extraneous variables con-
trolling reflectance. The color-infrared film type should normally be
used instead of conventional color film, because infrared film shows the
patterns of stress better and provides superior penetration of atmospheric
haze.
Multispectral scanning from aircraft should be used, if appropriate,
to detect and map S02-related injury to row crops whenever the foliar
symptoms are at least moderate and consist primarily of necrosis, and the
canopy is continuous, dense, and weed-free. These conditions are quite
restrictive because S02 effects in the field are usually subtle and con-
sist mainly of chlorosis. These effects cannot be detected consistently
with currently available airborne or spaceborne remote sensors.
-------
-18-
SECTION 3
LABORATORY SPECTRORADIOMETRY
GENERAL
Spectroradiometry provides information on the spectral reflectance
characteristics of objects. This section describes the results of a series
of laboratory experiments in which spectroradiometry was used to obtain
spectral curves of visible and near-infrared reflectance of S02-affected
soybean and wheat plants. Some S02-affected cotton plants were scanned,
but technical problems with instruments invalidated the data and the
experiment could not be repeated because of scheduling conflicts.
The following section describes attempts to verify the findings using
data from experimental field plots. The laboratory research included grow-
ing uniform groups of plants in a greenhouse; exposing them in a controlled
fashion to S02 in a laboratory exposure chamber; systematically observing
the foliar effects of S02; scanning the plants with a spectroradiometer
while controlling as many of the variables as possible; and finally, sta-
tistically analyzing the results to make some sound generalizations about
the spectral changes that plants undergo when they are affected by S02-
INSTRUMENTS
Two different types of spectroradiometers were used to scan the plants.
The project began with a conventional instrument and later acquired a new,
state-of-the-art multichannel telespectroradiometer (TSR). Both TSR's
provided spectral curves of reflectance versus wavelength.
The conventional TSR was used to scan the wheat plants. Its slow
0>20-second) scan speed was not a problem in the laboratory. Spectral
coverage ranged from 400 to 700 nanometers (run) in the visible channel.
-------
-19-
A separate, cooled optical head is used to scan the red and infrared spec-
trum from 600 to 1000 nm. Unfortunately, the two heads cannot be used
simultaneously. Output is in radiance units and the system has a 4-nm
spectral resolution. The telescope field of view (FOV) was either 1 or
3 degrees. The same telescope and FOV's were used with the other TSR
described below.
The TSR used to scan the soybeans has an optical multichannel detec-
tor. It is a microprocessor-controlled optical system that TVA adapted
to remote sensing in the laboratory and field. Application of optical
multichannel TSR's to the natural sciences has been very rare.32 The TVA
system is called an Optical Multichannel Analyzer (OMA-2) manufactured by
Princeton Applied Research Corporation.33 Figure 5 illustrates the con-
figuration of the system; Appendix A provides details and specifications
of the components. Scanning is essentially instantaneous (total scan
time =0.7 milliseconds), and data storage is on flexible disc. Figure 6
shows the components of the OMA-2 system.
-FIBER OPTIC CABLE
TELESCOPE
RECEPTOR
"* *• UbltCTOR
POLYCHHUMA 1 OH IJbltCIOR * > CONTROLLER
,, PARALLEL I/O
HEAD
COMPUTER
DEC LSI -I:
16-BIT PARALLEL
•*—
—
— *-
—
FLEXIBLE
DISC
LDRPVE 1
KEYBOARD
X-Y
RECORDER
Figure 5. TSR system configuration.
-------
-20-
Figure 6. TSR system components in laboratory.
The telescope is normally mounted on a tripod.
The OMA-2 operation is unique because it accumulates scans, integrat-
ing them over a preselected time frame or number of scans. Such accumula-
tion improves the signal-to-noise ratio and averages out short-term
variations caused by movement of foliage or inadvertent movements of the
instrument platform during scanning. The shape of a spectral curve from
an ordinary slow-scan TSR would be distorted by these short-term movements.
The vidicon detector of the OMA-2 is sensitive to a broad range of
visible and near-infrared wavelengths (Appendix A provides the spectral
response curve). The coverage of our system is restricted to a spectrum
that is 337 run wide, but the center of the scanned spectrum can be posi-
tioned anywhere within the range of detector sensitivity. For the soy-
beans, we selected a range of 430 to 767 nm, which included the green
chlorophyll peak, the red chlorophyll absorptance region, and part of the
near-infrared reflectance plateau. The spectral resolution of the system
-------
-21-
is about 2-4 nra, being limited in our case by the size of the input slit
to the spectrometer, the diameter of the fiber optics bundles, and our
success in aligning the optical components before each set of measurements.
The intensity scale (y-axis) of the OMA-2 cathode-ray tube display
reads in counts per channel. For each sample, a total of 500 discrete
channels was scanned at the 140-|Js/channel rate. The y-axis was calibrated
by scanning a lamp having a known spectrum. Background subtraction, sta-
tistical grouping, and averaging of curves were performed on the console
keyboard, after which the curves were stored on flexible discs.
CONCEPTS OF RADIANCE AND REFLECTANCE
The term "reflectance" describes the ratio of radiant energy reflected
from a surface to that incident upon it. Reflectance cannot be measured
directly, but can be computed from measurements of radiant energy. The
shape of the radiance curve obtained is a function of three wavelength-
dependent parameters: (1) variation of the intensity of illumination;
(2) variation of the reflected energy; and (3) variation in sensitivity
of the detector. If parameter (2) is to be isolated from parameters (1)
and (3), a standard surface with an established reflectance curve should
be used. This curve may then provide the basis for normalizing radiance
curves (Figure 7).
CURVE NORMALIZATION
Spectral curves were normalized to convert them from radiant energy
units to reflectance units. The instrument used to scan the soybeans had
an integral LSI-11 minicomputer, which is used to store, process, and
manipulate the curves. The TSR used to scan the wheat has an analog output.
These curves were digitized separately on the graphics tablet of a
-------
-22-
(O
-*-»
o
cc
o>
o
c
to
T3
CO
cc
o
c
03
o:
CD
O
c
RR
NASA GRAY
TARGET
SPECTRUM
TVA SCAN
OF NASA
TARGET
SPECTRUM
OF TVA
REFERENCE
STANDARD
VEGETATION
SPECTRUM
Wavelength
cu
o
Wavelength
NORMALIZED
VEGETATION
SPECTRUM
(1) One-time procedure to generate absolute curve (C):
C = (RR • A) /R
(2) Subsequent procedure to produce normalized curves of vegetation (NV)
NV = (V ' C) /RR
Figure 7. Procedure for normalizing spectral curves of radiance for
laboratory experiments.
-------
-23-
Tektronix 4014 graphics terminal; they were then normalized, averaged,
and otherwise manipulated and stored using an IBM 370 computer.
As Figure 7 illustrates, the scanning procedure incorporated a stan-
dard 18-percent gray reflecting surface, which was one of a set of panels
borrowed from NASA. This NASA gray target has a reflectance curve sup-
plied with it. For convenience, a small (20- by 25-cm), 18-percent gray
card was calibrated against the NASA gray target for day-to-day use in
the laboratory and field. The gray card—the TVA reference standard—was
scanned along with the vegetation to provide a basis for curve normaliza-
tion. Reference standards and calibration curves are illustrated in
Appendix B.
EXPERIMENTAL DESIGN
Plants
Soybean and winter wheat plants were grown to maturity in a green-
house. Each step in the procedure was carefully controlled to assure
uniform groups of plants. Registered seed guaranteed varietal purity.
The soil mixture consisted of specific proportions of Decatur loam, sand,
and peat, with nitrogen, phosphorus, and potassium nutrients added and
the pH adjusted to 6.5. The photoperiod was maintained at 14 h of light
by timers on the greenhouse lights. Watering was carefully controlled on
a pot-by-pot basis.
Groups of plants were exposed to 50% in a controlled exposure chamber
for specific doses to create the foliar effects. After exposure, the
plants were returned to the greenhouse, and the effects gradually appeared.
Observations of chlorosis and necrosis were made a week after exposure to
assure that all the effects had developed.
-------
-24-
Scanning Procedure
All scanning was done in the laboratory under carefully controlled
illumination. Individual soybean leaflets and wheat canopies were scanned.
The lamps were carefully oriented to minimize the specular (glare) compo-
nent of light into the telescope. This was done for wheat and soybeans
by maintaining a 45-degree angle between the telescope axis and the
tungsten-halogen (3200 kelvins) lamps. The range (distance between tele-
scope and plant) was determined by selecting a telescope FOV of 1 or 3
degrees and filling the FOV with foliage to minimize the contribution of
the background. Overhead views of the wheat canopies were obtained by
orienting the top of the plant container so that it was perpendicular to
the telescope axis. When individual soybean leaflets were scanned, their
surfaces were mounted flat on a board that was perpendicular to the tele-
scope axis. Soybean leaflets were detached and scanned immediately after
the observations were recorded. The wheat plants were scanned, pot-by-pot,
after observation was completed. The experiments are discussed below in
greater detail.
Analysis of Soybeans
Groups of soybeans were exposed to one of several SOg doses that were
chosen to create a wide range of foliar effects. The peak concentrations
and durations of exposure were (1) 15720 pg/m3 for 0.50 and 0.67 h and
(2) 10480 [Jg/m3 for 0.75 and 1.25 h. A total of 153 leaflets was scanned.
The selection was made by determining which nodes were affected on the
most severely affected plants and then including the leaflets growing on
these nodes on all plants. The unaffected leaflets on those nodes were
also included. The spectral reflectance curves were grouped in broad
classes by level of injury. A mean curve was computed for each class or
-------
-25-
combination of classes to show trends in the relationship between foliar
effects and reflectance. Analysis of variance was used to determine
whether differences among classes were statistically significant. This
analysis was done for specific wavelengths: blue (450 nm), green (550 nm),
red (650 nm), and infrared (750 nm). The statistical analysis included
the two components of foliar injury, chlorosis and necrosis. F-tests were
used to determine significance of differences in reflectance between
injury classes at significance levels of a=.05 and ot=.10.34
Analysis of Winter Wheat
Winter wheat plants (Coker variety) were grown in containers in the
greenhouse using the same closely controlled procedures as described for
soybeans. The mature plants were exposed to five levels of S(>2 in the
controlled exposure chamber to create a wide range of foliar effects. The
S02 concentrations (all 3-h averages) used were 3930, 5240, 6550, 7860,
and 9170 pg/m3. After exposure, the plants were returned to the greenhouse
to allow the effects to develop. A week later, the plants were observed
systematically, pot-by-pot, to obtain data on the chlorosis and necrosis
components. Effects ranged from none for the control group to "very
severe" for the groups receiving 6550, 7860, and 9170 pg/m3 S02.
When observations were complete, spectral scanning began. For this
task, the Gamma Scientific TSR was used. The analog curves of visible
radiance were digitized, normalized, averaged, and plotted on a Tektronix
4014 computer terminal with digitizer graphics tablet. The near-infrared
(IR) curves had to be obtained with a separate optical head. No useful
IR scans of wheat were obtained because of problems in calibration. These
problems were solved later by substituting the OMA-2 TSR for the Gamma
-------
-26-
Scientific model. The IR curves obtained for this study were not statis-
tically analyzed because apparent variation was slight—typically less
than 2 percent (Appendix B).
Correlation-regression analysis and one-way analysis of variance
(ANOVA) were performed on the wheat. The latter procedure focused on
differences among injury classes. F-tests were used to determine sig-
nificance (a=.05).35 This analysis was done for three specific
wavelengths--450 nm, 550 nm, and 650 nm.
RESULTS AND DISCUSSION
Soybeans
Descriptive statistics for foliar injury (chlorosis and necrosis)
were obtained for soybeans (Table 1). The class boundaries were selected
to conform to traditional breakpoints used in field surveillance of SO^
effects. These chlorosis classes were "unaffected (0%)," "light (1-10%),"
"moderate (11-25%)," "severe (26-50%)," and "very severe (>50%)." For
necrosis the "severe" and "very severe" classes were combined and repre-
sented all levels above 25 percent.
The individual reflectance curves for soybeans were arithmetically
averaged to produce a mean curve for each class of chlorosis and necrosis
(Figures 8 and 9).
Some of the averaged reflectance curves for chlorosis (Figure 8) show
more separation than others. Chlorosis class 4 (very severe injury) has
the highest visible and lowest IR reflectance of all. However, the other
classes are not separable at any wavelength.
The averaged reflectance curves for necrosis show more differences
among classes than those for chlorosis (Figure 9). The greatest separa-
tion is in the red wavelengths (chlorophyll absorption band), at about
-------
TABLE 1. DATA CLASSES AND FOLIAR EFFECTS OF S02 ON SOYBEANS
Chlorosis
Qualitative level
of injury
None
Light
Moderate
Severe
Very severe
Class
0
1
2
3
4
Range
(W
0
1-10
11-25
26-50
>50
Mean injury
(%)
0
6.5
15.6
39.3
69.3
No. of
leaflets
7
52
20
29
26
134
Qualitative level
of injury
None
Light
Moderate
Severe and
very severe
Necrosis
Class
0
1
2
3
Range
(%)
0
1-10
11-25
>25
Mean injury No. of
(%) leaflets
0 7
6.9 15
21.7 3
50.0 _3
28
N3
-^1
I
-------
§5
LlJ
o
o
80-
70-
60-
50-
40-
30-
20-
10-
0
CURVE
CLASS
CONTROL 0
1 1
ce. 80-
70-
60-
50-
40-
30-
20-
10-
0
INJURY(%)
0
1-10
CURVE CLASS INJURY(%)
CONTROL 0
Control
and 3
Control
450 500 550 600 650 700 750
CURVE CLASS
CONTROL 0
2 2
INJURY(%)
CURVE CLASS INJURY(%)
CONTROL 0
4 4
Control
Control
450 500 550 600 650 700 750
WAVELENGTH(nm)
Figure 8. Mean spectral curves for classes of chlorosis in soybeans exposed to S(>2.
i
K3
00
I
-------
80-
70-
CURVE CLASS
CONTROL 0
1 1
INJURY(%)
0
1-10
CC 80-
70-
60-
50-
40-
30-
20-
10-
0
CURVE CLASS INJURY(%)
CONTROL 0 0
2 2 11-25
Control
450 500 550 600 650 700
CURVE CLASS INJURY(%)
CONTROL 0 0
3 3 >25
Control
750 450 500 550 600 650
WAVELENGTH(nm)
700 750
Figure 9. Mean spectral curves for classes of necrosis in soybeans exposed to SC>2.
-------
-30-
650 nm. An ordered sequence is apparent in the curves, with curve 0
(control) having the lowest reflectance at all wavelengths. Class 3
(>25 percent necrosis) shows relatively high green and red reflectance
and relatively low IR reflectance, as predicted. In fact, the IR reflect-
ance is lower for class 3 than for all other classes.
The possibility of a relationship between the total area (cm2) beneath
each spectral curve and injury to soybeans was investigated. Correlation
and regression analysis revealed a coefficient of determination (r2) of
only 0.22 between curve area and percent chlorosis, but a higher r2 value
(0.83) was found between curve area and percent necrosis. The regression
showed that curve area increased with necrosis. Since the shape of a
curve can change without affecting its total area, shifts in reflectance
at various wavelengths or regions of the curve were examined.
The relative utility of four spectral bands (blue, green, red, and
IR) for detecting foliar injury was examined by an analysis of variance
using the reflectance data. The objective of the statistical procedure
was to determine whether statistically significant differences in reflec-
tance existed between pairs of injury classes. The classes were based
on ranges of chlorosis and necrosis. In addition, each spectral band and
the IR/red ratio were analyzed separately.
Some significant results were obtained from the comparisons of classes.
For chlorosis, a significant difference (at the less stringent level of
a=.10, but not at a=.05) was found in the green band between the unaffected
and lightly chlorotic classes (combined) and the severely chlorotic class.
For necrosis, significant (a=.05) differences were found in the red and
green bands between class 0 (unaffected) soybeans and affected classes 2
-------
-31-
and 3 combined (necrosis >10 percent). A significant (a=.05) difference
was also found in the IR/red ratio between the unaffected and affected
soybeans.
The strength of the relationships between reflectance and foliar
injury is indicated by the correlation coefficients listed in Table 2.
TABLE 2. SIMPLE CORRELATION COEFFICIENTS (r) FOR SINGLE-BAND
REFLECTANCE AND FOLIAR INJURY TO SOYBEANS
Symptom Blue Green Red IR IR/Red
Chlorosis +0.20 +0.72 +0.36 -0.10 -0.32
Necrosis +0.89 +0.92 +0.98 0.00 -0.94
Underlined coefficients are significant, a=.05
The table shows that the correlation coefficients for chlorosis,
except for green reflectance, are below 0.50; for necrosis the visible
reflectance bands, especially red, correlate much higher. IR reflectance
showed no significant relationship to either chlorosis or necrosis.
The IR/red reflectance ratios listed in Table 2 should be an indirect
indicator of stress in foliage, as explained previously. The low r (-0.32)
between chlorosis and IR/red indicates essentially no relationship between
these variables. However, there is a strong relationship (r=-0.94) between
necrosis and the ratio. With increasing necrosis, the IR component of the
ratio decreases and the red component increases, thus bringing the ratio
value down, closer to unity.
Winter Wheat
The wheat was divided into six groups, five of which were exposed to
different concentrations of S02 for 3 h in the chamber. One group was
-------
-32-
used for control. The range of foliar symptoms was broad, consisting
primarily of necrosis (Table 3).
TABLE 3. DATA CLASSES AND FOLIAR EFFECTS OF S02 ON WINTER WHEAT
Qualitative
level of
effects
None
Light
Moderate
Very severe
Very severe
Very severe
Foliar Injury
Dose
Class
0
1
2
3
4
5
No. of
pots
6
4
2
4
3
2
Mean
chlorosis
(%)
0
1
1
4
5
5
Mean
necrosis
(%)
0
2
22
50
56
76
S02
concentration
3-h avg.
(Mg/m3)
0
(control)
3930
5240
6550
7860
9170
Q
Where light injury is <10 percent; moderate, 11-25 percent; severe,
26-50 percent; and very severe, >50 percent.
The scanning procedure yielded normalized curves which showed an
increase in overall reflectance that corresponded to increasing foliar
injury (Figure 9). The possibility of a relationship between the total
area (cm2) beneath each spectral curve and injury as well as S02 dose was
investigated. Correlation and regression analysis revealed an r2 of 0.72
between curve area and percent chlorosis in wheat. A closer association
(r2=0.85) emerged between curve area and percent necrosis. All relation-
ships were positive; curve area increased with injury level. Next, curve
area and S02 concentration (|jg/m3) were compared, yielding an r2 of 0.85.
This relationship was also positive.
-------
-33-
The change in reflectance in particular wavelength regions was also
studied. The red (650 ±4 nm) reflectance increased with increasing stress;
this rise was particularly evident at moderate and severe levels of stress.
Green (550 ±4 nm) peak reflectance also increased, but was more evident
at light levels of stress (Figure 10).
Statistical analysis of the blue, green, and red reflectance curves
for winter wheat yielded correlation coefficients ranging between +0.73
and +0.90 (Table 4). IR reflectances of wheat were not measured because
of problems in instrument calibration.
TABLE 4. SIMPLE CORRELATION COEFFICIENTS (r) FOR
SINGLE-BAND REFLECTANCE AND FOLIAR INJURY TO WINTER WHEAT
Reflectance
Symptom Blue Green Red
Chlorosis +0.83 +0.90 +0.81
Necrosis +0.73 +0.83 +0.85
Underlined coefficients are significant, a=.05
One-way analysis of variance was performed on the reflectance and
injury data for wheat. The chlorosis component was divided into three
classes, and the necrosis component was divided into four classes, based
on the level of severity of injury (Table 5). The results showed that
significant (F-test, a=.05) differences existed among all classes for
chlorosis and necrosis. The trends are readily apparent in means for the
blue, green, and red reflectances listed in the table. Appendix C lists
the results of the analysis of variance.
Summary of Results for Soybeans and Winter Wheat
The relationship between spectral reflectance and foliar injury from
S02 is complex. It was analyzed by separating injury into its components—
-------
35
30-
25-
20i
o
10-
5-
Curve Class Foliar Injury
None (Control)
Light
Moderate
Very Severe
Very Severe
Very Severe
350 400 450 500 550 600
WAVELENGTH (nm)
650
700
750
Figure 10. Mean spectral curves for classes of unaffected and S02-affected winter wheat.
See Table 3 for S02 doses and levels of foliar injury.
-------
TABLE 5. REFLECTANCE STATISTICS FOR CLASSES OF
S02-AFFECTED WINTER WHEAT
Reflectance
1
2
3
4
Necrosis
class
(0% injury)
(1-25%)
(26-50%)
050%)
Mean
Blue
6.4
7.3
7.9
8.7
(%)
Green
16
17
21
23
.2
.6
.1
.7
Standard Deviation Chlorosis
Red
10.
12.
21.
24.
3
4
6
3
Blue
.44
.17
.37
.32
Green
4.14
2.89
6.78
.25
Red class"
3.86 1 (0% injury)
9.62 2 (0.1-1%)
9.68 3 (>1%)
1.84
Reflectance
Mean (%) Standard Deviation
Blue Green Red Blue
6.4 16.2 10.3 .44
7.7 20.2 19.0 .36
8.4 21.9 20.3 .78
Green Red
4.14 3.86
10.22 33.57
7.20 32.76
For necrosis, 1-25% represents light/moderate injury; 26-50%, severe; and >50%, very severe.
For chlorosis, all three classes represent very light injury.
Ui
I
-------
-36-
chlorosis and necorsis--and reflectance into bands within the visible and
near-infrared spectra. Results indicate that, for soybeans, only green-
band reflectance could differentiate between severely chlorotic plants and
unaffected or lightly chlorotic plants. Red-band reflectance could differ-
entiate between necrotic and unaffected soybeans. The IR/red ratio was
highly correlated (r=-0.94) with necrosis, but was not related to chlorosis
in soybeans.
Scans of winter wheat yielded more uniformly positive results. Total
visible reflectance as well as individual wavelength bands could be used
to distinguish the SQ% effects. Three classes of chlorosis and four
classes of necrosis, based on severity, could be distinguished by their
visible reflectance characteristics.
-------
-37-
SECTION 4
FIELD SPECTRORADIOMETRY
GENERAL
This section of the report describes an extension of research to the
field. Experimental plots of soybeans and winter wheat were grown to the
seed-filling stage of development, exposed to several controlled doses of
S02, and observed systematically to determine foliar effects. Then the
plots were scanned row by row with the OMA-2 TSR. The resulting spectral
curves were statistically analyzed to arrive at some soundly based gen-
eralizations on the changes in spectral reflectance that occur when these
plants are stressed by $0%. Results indicate the potential of reflectance-
measuring remote sensors for detecting and mapping the effects of SO^
emissions on these crops.
EXPERIMENTAL DESIGN
Plot Preparation and Planting
Soybeans (Glycine Max [L.] Merr. var Essex) and winter wheat (Triticum
aestivum [L.] var Coker 68-15) were planted in separate 0.40-ha experimen-
tal plots. Care was taken to ensure uniformity of growth throughout the
plot. Registered seed guaranteed varietal purity, fertilizer was spread,
and herbicide was applied uniformly at time of planting. The soil had a
high clay content but was uniform within each plot. The slope of the
plot was slight but was adequate to provide good drainage.
Wheat was planted during early October 1978. The rows were drilled
in pairs 18 cm apart, each pair being separated by 76 cm (center-to-center).
At the time of S02 exposure, late April and early May 1979, coinciding with
the beginning of head-filling, no variations in foliage density, color, or
plant height were apparent.
-------
-38-
Soybeans were planted during early May 1979 in rows spaced 76 cm
apart. As with the wheat, no variations in foliage density, color, or
height were apparent on any plants used for 802 fumigation. The 802
exposures were done during August at the pod-filling stage of development.
Exposure of Plots to Sulfur Dioxide
The wheat and soybeans were exposed to S02 at the stage of growth
when the effects of the pollutant would theoretically have the greatest
effect on yield of the crop. For the wheat, the critical time is the
heading stage; for the soybeans, it is the pod-filling stage.
The exposures were effected in situ by dividing the plots into arrays
of subplots, the wheat having five rows 6 m long, and the soybeans having
three rows 6 m long. The two outer rows of the five in the wheat subplots
served as buffers. A similar buffer function was served by plants on
either end of each row beyond the 6-m segments to be exposed to the pollu-
tant. Figures 11 and 12 illustrate the layout of the two experimental
plots. The exposed subplots were paired with unexposed (control) subplots;
this pairing was done to minimize locational influences on the results.
Treated plots were exposed to either 2620, 5240, 7860, or 10480 |Jg/m3
of anhydrous 862 for two hours. The pollutant was introduced at the intake
of a large blower which mixed and dispensed the desired SOg concentration
through four 30-cm-diameter perforated, inflatable plastic tubes laid under
the foliage, between the rows. In each case, the dose was changed by
adjusting the concentration. The method was developed by TVA for another
project which required a charcoal filter air pollution exclusion system to
remove or exclude 802-polluted air from test plots located downwind of
coal-fired power plants.36 For the present experiments, the filters were
removed from the electric blowers and 802 was injected into the airstream
-------
-39-
B
A4
NS
Ao
C4
f
N
SCALE IN METERS
LEGEND
A = SO 2 Affected Soybean Subplot
B = Buffer Soybeans
C = Control Soybean Subplot
NS = Subplots Not Scanned
Figure 11. Map of experimental plot of soybeans showing
division into subplots.
B
C14
A16
C
NS
A15
B
R
B
A14
C
NS
CIS
A12
A13
Cll
C12
All
R
B
B
3)
m
m
z
I
0
C
CO
m
LEGEND
A= Affected Wheat Subplc
B= Buffer Wheat Plants
C= Control Wheat Subplo
NS= Subplots Not Scanned
036
1 1 — 1 1 1
SCALE IN METERS
A10
CIO
A9
C9
C8
A8
C7 A7 C6
A6 C5 A5 B
B
NS
C
NS
A4
C4
A3 C3 A2
C2 Al Cl B
Figure 12. Map of experimental plot of wheat showing
division into subplots.
-------
-40-
and forced through the perforated tubes and onto the foliage. The concen-
trations within the tube and outside it, at the midpoint of the plant
canopy within the subplot, were monitored with a Teco SO^ monitor during
exposure.
In addition to 862 exposure, other parameters were considered.
Meteorological conditions at the time of exposure to air pollution can
influence the severity of foliar effects.37 Therefore, atmospheric humid-
ity and temperature were monitored during the experiments. A Li-Cor
diffusion-resistance porometer was used to verify that the stomata of the
soybean leaflets were open before the gas was dispensed. The plot was
watered the afternoon before every experiment. The applications usually
began at approximately 10 a.m. and ended about 4 p.m., so that four sub-
plots, representing one replication, were exposed each day.
Access to the subplots for the experimental equipment was provided
by straight swaths across the plot. Figure 13 shows the wheat plot and
the swaths, equipment, and instrumentation used for the controlled expo-
sure experiment. Portable fiberglass fences 1.2 m high were moved from
subplot to subplot to enclose the rows during exposure to the pollutant.
The purpose of these barriers was to prevent the horizontal dispersion of
the gas beyond the particular subplot and to reduce the effects of wind
below 1.2 m height.
-------
-41-
Figure 13. Aerial view of experimental plot of wheat showing
equipment and instrumentation for controlled SC>2
exposure.
Observations of Foliar Injury
The foliar effects (symptoms) of SC>2 on vegetation typically emerge
slowly, taking from several hours to several days to develop completely.
The total time required for development depends on the level of dose,
temporal characteristics of the concentration, and several other factors
that vary from species to species.38 Therefore, a week was permitted to
elapse after exposure before the observations of foliar injury were made
and recorded. Injury estimates using the L x A method (Appendix D) were
made on ten randomly selected plants on each 6-m row. For winter wheat,
the upper three of six leaves on each tiller were always selected for
observation. (Senescence of the lower leaves was present on the soybeans
and wheat.) The measurements for the ten plants were averaged, providing
mean values for chlorosis and necrosis for each of the three rows within
a particular subplot. For comparative purposes, the three row figures
were averaged to yield a mean percentage of chlorosis or necrosis for the
subplot. This procedure of observation was followed for soybeans and wheat.
-------
-42-
Scanning Procedure
The plots were scanned as soon as practical before or after observa-
tions of foliar injury were made. Scanning required sunny weather, a
condition that in summer was most likely to occur during mornings, so the
scans were acquired during the 10:00 a.m. to 12:00 a.m. period. Several
days were required to complete the scanning.
A van with an aluminum platform on top provided a convenient place
for the operator and TSR (Figure 14). A helicopter was tested as a scan-
ning platform but was judged unsatisfactory and inconvenient, so the van
was used. Power for the instrument was provided by a portable, 1700-watt
alternator which was shock-mounted on the platform. The voltage gener-
ated by the unit was controlled with a regulated transformer and solid-
state surge eliminator. Vibrational effects on the optics were minimized
with shock mounts.
The angle of view from the telescope to the target is an important
factor in any spectral scanning experiment.39 These angles were held
constant during scanning by controlling the orientation of the van.
Figure 15 shows the angular configuration and distances used. After each
scan set was completed, the van was advanced one row for the next scan set.
The NASA gray target was also scanned at the same angle and distance each
time. A scan set consisted of 50 scans of a row of plants and 50 scans of
the NASA gray target. The telescope was swiveled horizontally through 72
degrees after scanning the target, and then the vegetation was scanned.
Minor adjustments in the vertical angle had to be made to compensate for
topography.
Another important factor in any spectral scanning experiment is the
angle of illumination.40 Solar effects must be known and minimized or
-------
-43-
is inside the van.
^^
-------
-44-
NASA Gray Target
46m
Figure 15. Geometric configuration for scanning subplots,
-------
-45-
compensated for. The van was oriented so that the sun was behind it and
the plant canopy was fully illuminated on the side that was being scanned.
Variations in illumination that occurred during scanning were compensated
for by scanning the NASA gray target immediately before scanning the vege-
tation and normalizing each scan set according to the following procedure.
Normalization of Spectral Curves
As in the laboratory-based experiments, the spectral curves were nor-
malized to convert from radiant energy units (counts per photon) to percent
reflectance. This conversion was necessary to ensure comparability among
curves that had been acquired under different conditions of illumination.
The conversion to percent reflectance involved scanning the NASA gray
target and then immediately scanning the vegetation. No delay between the
two scans was allowed because a change in illumination during the interval
would introduce error. Normalization consisted of dividing the vegetation
curve by the previous measurement of the NASA gray target and multiplying
the resultant curve by the "true" reflectance curve for the target (Figure
16). Note that the actual NASA gray target was scanned instead of a TVA
reference standard (gray card).
Normalization, background subtraction, statistical grouping, and
averaging of curves were done from the console keyboard after scanning was
complete. Then the curves were stored on flexible discs for output later
to the x-y plotter.
-------
-46-
A
Wavelength
R
Wavelength
V —
SPECTRUM of
NASA GRAY
TARGET
TVA SCAN of
NASA GRAY
TARGET
VEGETATION
SPECTRUM
NORMALIZED
VEGETATION
ff NV SPECTRUM
I
s.
Wavelength
NV=VA/R ~\
Wavelength
Figure 16. Procedure for normalizing spectral curves of radiance
for experimental field plots.
RESULTS AMD DISCUSSION
Soybeans
Data from seven subplots of soybeans were analyzed statistically to
determine relationships among reflectance, SC>2 dose (concentration x time),
and foliar injury (chlorosis and necrosis). Reflectance was the dependent
variable. First, the relationship between dose and injury was examined;
then the relationship between dose and reflectance was studied; finally,
the relationships between reflectance and injury, and between reflectance
and yield were determined and characterized. The subplot measurements
are listed in Appendix D.
The relationship between SC-2 dose and observed levels of foliar injury
to soybeans is complex, as the preceding laboratory studies have shown.
Plots of dose versus chlorosis and dose versus necrosis revealed no rela-
tionships (Table 6). No relationship was evident between either dose and
-------
-47-
TABLE 6. S02 DOSE, INJURY, AND YIELD OF SOYBEANS
Dose
Class
1
(control)
2
3
4
5
S02
Concentration
(2-h avg)
0 pg/m3
2620
5240
7860
10480
No. of
Subplots
3
1
1
1
1
No. of
Rows
8
3
3
3
2
Mean
Chlorosis
(%)*
0
25.0
14.3
41.0
11.5
Mean
Necrosis
(%)3
0
0
45.0
0
30.0
Mean
Yield
(kg/ha)
2730
2596
2234
2759
2352
Where light effects are classified as <10 percent; moderate effects,
11-25 percent; and severe effects, >25 percent. See Appendix D for
method of computation.
either dose and yield or chlorosis and yield, but a significant (a=.05)
negative correlation (r=-0.97) between necrosis and yield was indicated
(Figure 17).
3000 n
2500 -
2000
SOYBEANS
10
20 30
MEAN NECROSIS
40
50
Figure 17. Linear regression of yield on necrosis for
soybeans exposed to S02-
-------
-48-
Next, the individual reflectance curves (shown in Appendix E) for
the rows within subplots were sorted and grouped by S02 dose, then aver-
aged to provide one mean curve for each of the four classes (Figures 18a,
b, c, and d).
Correlation coefficients were calculated between total reflectance
(the total area under the mean reflectance curve) and dose, injury, and
yield. These statistics revealed no significant (a=.05) relationships.
Since the shape of a curve can change without affecting its total area,
shifts in reflectance at various wavelengths or regions of the curve were
examined.
Consider the control and dose class 2 (Figure 18a). There is an
obvious difference in visible reflectance around 500 to 550 nm, but
otherwise the curves are similar. Now compare control with dose class 3
(Figure 18b): Prominent differences exist in the red (~ 650 nm) and IR
(~ 750 nm) regions. With exposure, the red reflectance has increased and
the IR has decreased.
Next, compare control and dose class 4 (Figures I8c). Some decrease
in IR reflectance has occurred, but there is very little difference in
the visible bands. Finally, compare control and dose class 5
(Figure 18d): Some increase in red reflectance has occurred, but the
major change is a decrease in IR reflectance.
An analysis of variance was performed to compare control to the
four dose classes combined (Table 7). Three spectral regions were
sampled at their central wavelengths to provide representative data for
analysis. Accordingly, green reflectance was measured at 550 nm; red,
at 650 nm; and IR, at 750 nm. The measurements are listed in Appendix E.
-------
Reflectance (%)
H-
•s
i-i
(D
01
oo
pi
3
n g
O (D
3 Pi
o w
en i-i
O Pi
M
en
p4 n
o c
a H
3 <
(D
t-h CO
O
H Hi
O
n i-i
O
3 en
T3 O
pi V!
i-i a4
H- CD
CD PI
O 3
p en
(D
X
-a
o
en
fD
fD
3
OQ
Ol-
o
o-
o
en
(ji-
O
en
O-
o
en
CJl-
o
o-
o
CJl —
o
o
I
o
I
o
I
OJ
o
I
o
I
o
I
o
I
o
I
hO
o
I
CO
o
I
o
I
o
I
o
CO
m
>
z
CO
01-
o
o-
o
C71
O —
O
CJl —
o
o-
o
CJl-
o
I
.p-
en
o
N3
-------
Reflectance (%)
H-
OQ
c
l-i
00
n
n g
o ro
H
O
fD O
M rt
CO l-i
O (to
M
CD
04 n
o c
3 i-t
3
Ol •
O
CJ1
o-
o
on-
o
CTl
o-
o
01
Ol •
o
o
I
ro
o
co
o
o
I
Ol
o
I
Ul
o
-------
-51-
TABLE 7. SIGNIFICANT DIFFERENCES IN REFLECTANCE
BETWEEN S02-EXPOSED AND CONTROL SOYBEANS
Reflectance
Dose Classes Green Red IR IR/Red
Control (unexposed) versus 0.87 3.07 4.58 6.95
S02-exposed soybeans
Underlined values are significant, F-test, ot=.05.
Table 7 indicates that the differences in green- and red-band reflec-
tance for the dose classes are not statistically significant (F-test,
a=.05). However, the differences in IR and IR/red reflectance are
significant at this level.
Perhaps the most important comparison in this study is reflectance
versus foliar injury. Appendix D shows that the affected subplots have
higher reflectance in the green, higher in the red, lower in the IR, and
a lower IR/red ratio. The yield of the affected subplots is also lower.
These comparisons reveal general trends only. The mean curves in Figures
19a and b were obtained by grouping the reflectance measurements into
three classes: control (unaffected soybeans), chlorotic soybeans, and
necrotic soybeans. The mean curve for chlorotic soybeans (Figure 19a)
shows lower IR reflectance than that for the control. The curve for
necrotic soybeans (Figure 19b) shows greater differences: the IR reflec-
tance is lower and the green and red reflectances are higher than that
for the control.
Table 8 reveals that the visible reflectance variations correlate
significantly (a=.05) with necrosis, but not with chlorosis. No signifi-
cant correlation between reflectance and yield is indicated.
-------
-52-
50 -
40-
20-
10-
(a)
SOYBEANS
Control
Chlorotic
450
500
550
600
750
40 -
30-
20-
(b)
SOYBEANS
Control
Necrotic
450 500 550 600 65
Wavelength (nm)
Figures 19a and b.
Mean spectral curves for S02~affected soybeans.
Control also shown for comparison.
-------
-53-
TABLE 8. SIMPLE CORRELATION COEFFICIENTS (r) FOR SINGLE-BAND
REFLECTANCE, FOLIAR INJURY, AND YIELD OF SOYBEANS
Reflectance
Symptom
Chlorosis
Necrosis
Yield
Green
+ .47
+ .83
+ .00
Red
+ .57
+ .97
+ .33
IR
-.62
-.65
+ .55
IR/Red
-.68
-.84
+ .51
Underlined coefficients are significant (a=.05).
Regressions of reflectance and foliar injury further characterize the
significant relationships between these variables (Figures 20a, b, and c).
The regression of necrosis and green reflectance shows a positive associa-
tion (Figure 20a). The same is true for necrosis and red reflectance
(Figure 20b). On the other hand, the correlation between necrosis and the
IR/red ratio is negative (Figure 20c). The IR regression was not signifi-
cant and is not shown.
To determine whether the differences in reflectance between S02~
affected soybeans and control soybeans were significant, an analysis of
variance was performed. Table 9 summarizes the results and Appendix F
provides the details. There was a significant (a=.05) difference in IR
and IR/red reflectance between the chlorotic soybeans and unaffected
(control soybeans. There were also significant differences in red
reflectance, IR reflectance, and the IR/red ratio between necrotic and
unaffected soybeans.
There were no significant differences in yield between any of the
classes of affected and unaffected soybeans. These findings agree with
the insignificance of the r coefficients for yield versus reflectance
(Table 8).
-------
-54-
14 -
» '3'
~ 12-1
£ 11
5 10
UJ
i 9 •
UJ
04 8 •
7
(a)
GREEN
= 0.69
20 40 60 80 100
NECROSIS (%)
14 -
_ 13 -
£ 12.
5 10-
=! 9 -
LU '
" 8-
7
r2 = 0.94
0 20 40 60 80 100
NECROSIS(%)
(c)
6 -
5 -
o
i 4-
IR/RED
K^
X^>*
\r »K«
^J*..
1 = 0.71
0 20 40 60 80 100
NECROSIS (%)
Figures 20a, b, and c. Linear regressions of reflectance on necrosis
for SOg-affected soybeans.
-------
-55-
TABLE 9. SIGNIFICANT DIFFERENCES IN REFLECTANCE AND
YIELD BETWEEN S02-AFFECTED AND UNAFFECTED SOYBEANS
Class
Test
Statistic
Reflectance
Green Red IR IR/Red
Yield
Chlorotic versus
unaffected soybeans
Necrotic versus
unaffected soybeans
(neither necrotic
nor chlorotic)
0.87 2.28 6.94 6.93
2.80
3.42 13.12 19.73 33.02
3.69
Underlined values are significant, F-test, a=.05.
Winter Wheat
The 30 subplots of winter wheat were grouped by SC-2 dose class and
foliar injury class to relate changes in spectral reflectance to the
classes. Appendix D lists the subplot measurements. A comparison of the
subplots shows that the exposed ones have higher red reflectance, lower
IR reflectance, a lower IR/red ratio, lower yield, and, of course, a higher
percentage of necrosis. Correlation and regression analyses were done to
compare total reflectance (area under spectral curve) with dose, necrosis,
and yield (Table 10). The only significant (a=.05) relationship between
percent area under the curve and any of these three variables was with
necrosis (r2=0.66); as necrosis increased, so did curve area (Figure 21).
Next, the dose classes were examined with respect to the three individual
spectral bands (green, red, and IR) and the IR/red ratio. Four 2-h con-
centrations of the pollutant and the control formed five dose classes.
-------
-56-
TABLE 10. TOTAL REFLECTANCE, S02 DOSE, FOLIAR INJURY, AND
YIELD OF WINTER WHEAT
Dose Class
Concentration
(2-h avg)
Reflectance
Curve Area (%);
Yield (kg/ha)
1 (control) 0 (Jg/™3
2 2620
3 5240
4 7860
5 10480
Necrosis Mean Range of
Class Necrosis (%) Necrosis (%)
1 7.7 <10
2 13.8 11-25
3 33.8 26-50
4 68.5 >50
31.29
29.02
26.95
28.49
31.02
Reflectance
Curve Area (%)a
21.05
24.70
25.79
26.82
3094
3325
2479
2241
1894
Yield (kg/ha)
3044
3128
3135
2229
Percentage of rectangular plotted area filled by the curve.
s
30 -i
25 -
20
WINTER WHEAT
r2 = 0.66
10 20 30 40 50 60 70
NECROSIS {%)
Figure 21. Linear regression of reflectance curve
area on necrosis for winter wheat exposed
to S02
-------
-57-
The individual reflectance curves for each subplot were sorted and grouped
by dose, then averaged to provide one mean curve for each of the five
classes (Figures 22a, b, c, and d). The curves are plotted with the con-
trol (dose class 1) to display the relative differences. In Figure 22a
(control and dose class 2) only minor differences are apparent; the
exposed wheat has lower overall reflectance than the control. Figure 22b
(control and dose class 3) shows that the green and IR reflectances of the
exposed wheat are considerably lower, and the red reflectance is slightly
higher than the control. Figure 22c (control and class 4) shows a response
that is not markedly different than that of class 3. However, Figure 22d
(control and class 5) shows a remarkable difference in the visible and IR
wavelengths. The visible reflectance of the exposed wheat has increased
markedly and the IR reflectance has decreased slightly when compared with
control. An overall flattening of the spectral curve is evident. This
flattening is a typical response for the reflectance of stressed vegetation.41
The five dose classes were compared using analysis of variance
(Table 11). The table indicates that the differences among each of the
five means of red, IR, and IR/red reflectance are statistically signifi-
cant (ot=.05). However, differences among the five means of green
reflectance are not significant.
TABLE 11. SIGNIFICANT DIFFERENCES IN REFLECTANCE AMONG
FOUR CLASSES OF S02-EXPOSED WINTER WHEAT AND CONTROL
Test Reflectance
Statistic Green Red IR IR/Red
F 1.37 3.82 9.25 14.71
Underlined values are significant, F-test, a=.05.
-------
-58-
CD
o
c
3
o
cc
35-
30-
25-
20-
15-
10-
5-
(Control)
(2620
for 2h)
450
500
550
600
650
700
—r~
750
35-
30-
25-
20-
(b)
WHEAT
Class 1 (Control)
Class 3 (5240 /xg/m3 for 2h)
15-
10-
750
Wavelength (nm)
Figures 16a and b. Mean spectral curves for winter wheat
exposed to S02. Control also shown
for comparison.
-------
-59-
CO
-4-»
o
o>
CE
35-
30-
25-
20-
15-
10 -
5-
(C)
WHEAT
Class 1 (Control)
Class 4 (7860 ^g/m3 for 2h)
450
I
500
T
550
I
600
I
650
700
750
35-
30-
25-
20 -
15 -
10 -
5 -
(d)
WHEAT
Class 1 (Control)
Class 5 (10480 ^.g/m3 for 2h)
450
500
550
600
650
700
750
Figures 16c and d.
Wavelength (nm)
Mean spectral curves for winter wheat
exposed to S0~. Control also shown
for comparison.
-------
-60-
The relationship between S02 dose and observed levels of foliar injury
is not linear; in this experiment the necrosis symptoms began slowly at low
dose levels, increased more rapidly at moderate levels, and then tapered
off at high dose levels (Table 12). These trends may apply only to this
particular data set. This statistical analysis focuses only upon necrosis
because chlorosis was negligible in the plot (Appendix D). Characteristi-
cally, wheat responds to S02 by becoming necrotic rather than chlorotic.
In the study of foliar injury and reflectance, the necrosis percent-
ages were grouped into four classes (Table 13). The injury classes and
dose classes do not correspond because of the great variability in response
of individual plants to the pollutant. The breakpoints for the classes of
necrosis are the traditional boundaries used by field biologists. Four
mean reflectance curves, each representing a necrosis class, were produced
by grouping and averaging the individual curves (Figures 23a, b, and c).
The potential of reflectance measurements for detecting foliar injury
was determined by statistical analysis of necrosis data and three reflec-
tance bands (green, red, and IR), as well as the IR/red ratio. The objec-
tive of the statistical procedure was to determine whether statistically
significant differences in reflectance existed among the injury classes.
These classes were based on ranges of necrosis. Finally, statistical
tests concerning the relationships among reflectance, necrosis, and crop
yield were conducted.
Table 14 lists the relationships between reflectance and the two vari-
ables, necrosis and yield. The red band and IR/red ratio of reflectance
seem to be most useful for detecting either necrosis or yield reduction.
-------
TABLE 12. REFLECTANCE, NECROSIS, S02 CONCENTRATIONS, AND YIELD FOR WINTER WHEAT BY DOSE CLASS
S02
Dose Concentration
Class (2-h avg)
1 0 pg/m3
2 2620
3 5240
4 7860
5 10480
Mean
Necrosis
(%) ±S.D.
8.5±1.9a
19.4±6.7
58.1±17.5
67.0116.0
76.6±15.5
Mean Reflectance (%)
Green
10.83+1.31
10.5511.15
9.7610.60
10.60+0.68
10.8111.61
Red
10.3911.26
10.58+1.79
11.2710.58
11.6011.43
12.15+1.71
+S.D.
IR
32.16+5.83
29.1914.80
25.1613.82
32.0719.10
24.0315.37
Mean
IR/Red +S.D.
3.1310.36
2.8610.73
2.2310.30
2.84+0.99
1.9610.16
Mean
Yield
( kg/ha) IS. D.
30941436
33251396
2479179
2241+570
18941763
Necrosis in Class 1 (control) is associated with natural senescence that occurred after scanning but before
observations were made.
-------
TABLE 13. REFLECTANCE, NECROSIS, AND YIELD FOR WINTER WHEAT BY INJURY CLASS
Injury
Class
1
2
3
4
Mean
Necrosis
1S.D.
7.7+1.2
13.8+3.8
33.814.8
68.5116.6
Mean Reflectance (%) +S.D.
Green Red IR
10.7411.48 10.3511.47 30.0315.50
10.8610.96 10.7111.32 33.37+5.68
10.50+0.09 10.4010.76 32.2414.22
10.3911.21 11.72+1.43 27.0017.72
Mean
IR/Red 1S.D.
2.9310.54
3.1910.76
3.1510.64
2.32+0.74
Mean Yield
(kg/ha) 1S.D.
31091511
31281228
30071642
21901627
Class 1
2
3
4
- <10% necrosis
- 11-25 %
- 26-50 %
- >50 %
-------
-63-
35-
30-
10-
(a)
WHEAT
Class 1 (< 10% Necrosis)
Class 2 (11-25%)
—T~
650
—r~
700
~I
750
450
—I
500
~1
550
600
35-
30-
25-
20-
10-
(b)
WHEAT
Class 1 (^10% Necrosis)
Class 3 (26-50%)
450
600
700
35-
20-
(C)
WHEAT
Class 1 (< 10% Necrosis)
Class 4 (>50%)
I
450
I
500
550
650
Wavelength (nm)
700
750
Figures 23a, b, and c. Mean spectral curves for S02-affected winter wheat,
-------
-64-
TABLE 14. SIMPLE CORRELATION COEFFICIENTS (r) BETWEEN REFLECTANCE AND
TWO OTHER VARIABLES (NECROSIS AND YIELD) FOR WHEAT
Reflectance
Symptom Green Red IR IR/Red
Necrosis -.06 +.59 -.53 -.71
Yield +.52 -.63 +.45 +.67
Underlined coefficients are significant, a=.05.
The trends of the relationships are also noteworthy (Figures 24a, b,
and c). Red reflectance increases and IR reflectance decreases as percent
necrosis rises. The IR/red ratio decreases as necrosis rises. These
findings conform to theory.42
-------
-65-
11.5-
0 10 20 30 40 50 60 70 80
NECROSIS (%)
35-i
g 30^
UJ
<
LU
=: 25 H
20
r2=0.28
0 10 20 30 40 50 60 70 80
NECROSIS (%)
4
3
2 •
2 = 0.50
0 10 20 30 40 50 60 70 80
NECROSIS (%)
Figures 24a, b, and c. Linear regressions of reflectance on necrosis
for winter wheat.
-------
-66-
Yield of wheat is related to necrosis and reflectance. Yield and
necrosis are inversely related (Figure 25). Green and IR reflectances
and the IR/red ratio increase with yield (Figures 26a, b, and c), but
red reflectance decreases as yield increases. These findings also conform
to theory.43
Results of a one-way analysis of variance are summarized in Table 15.
When the four injury classes were compared, there were significant (a=.05)
differences between the means of red and IR reflectance, as well as the
IR/red ratio and yield. The means of the green band reflectance for the
necrosis classes were not significantly different. Details are presented
in Appendix F.
TABLE 15. SIGNIFICANT DIFFERENCES IN REFLECTANCE AND YIELD
AMONG FOUR NECROSIS CLASSES3 OF S02-AFFECTED WINTER WHEAT
Test
Statistic
F
Reflectance
Green
0.80
Red
4.58
IR
6.37
IR/Red
15.92
Yield
18.09
o
Class 1 £10% necrosis
2 11-25%
3 26-50%
4 >50%
Underlined values are significant (F-test, a=.05)
Summary of Results
An analysis of variance showed statistically significant differences
in IR reflectance and the IR/red ratio when chlorotic soybeans were com-
pared with unaffected soybeans. Significant differences in red and IR
reflectance as well as the ratio were found when necrotic soybeans were
compared with unaffected soybeans. The affected soybeans had higher
green and red but lower IR reflectance and a lower IR/red ratio than the
-------
3500 -|
3000 -
2500 -
2000
-67-
20 40 60 80 100
NECROSIS (%)
Figure 25. Linear regression of yield on necrosis for
S02~affected winter wheat.
-------
-68-
11.5-
«s n.o H
g
LLJ
.=? '0.5 H
10.0
RED
= 0.40
2000 2500 3000 3500
YIELD (Kg/ha)
35-1
30H
I—
VJ
LU
25 H
20
IR
2 = 0.20
2000 2500 3000 3500
YIELD (Kg/ha)
11.5-
= 11.0-
S 10.5-
10.0
(b)
GREEN
2000 2500 3000 3500
YIELD (Kg/ha)
3 -
2-
(d)
IR/RED
r2 = 0.45
1
2000 2500 3000 3500
YIELD (Kg/ha)
Figures 26a, b, c, and d. Linear regressions of reflectance on yield of
S02-affected winter wheat.
-------
-69-
unaffected soybeans. No relationship was found between reflectance and
yield. The statistical tests were conducted^, at the 95 percent confidence
level.
Statistical analysis of the reflectance data from the wheat plot gave
more positive results than those obtained from the soybean plot. An analy-
sis of variance revealed statistically significant differences in red and
IR reflectance and the IR/red ratio, but not green, when four necrosis
classes were compared. No chlorosis was observed. The necrotic wheat
had higher red reflectance, lower IR, and a lower IR/red ratio than the
unaffected wheat. Yield decreased as necrosis increased. The statistical
tests were conducted at the 95 percent confidence level.
-------
-70-
SECTION 5
INTERPRETATION AND ANALYSIS OF AERIAL PHOTOGRAPHS
GENERAL
During the first year of this project, the foliar effects of S02 on
crop and tree species were detected, mapped, and studied qualitatively by
conventional stereoscopic interpretation of aerial photographs. These
results were not consistent because many physical variables other than
foliar effects controlled the variations in exposure. Therefore, a method
of photometric analysis was used to measure and calculate the reflectances
of vegetation and other objects in aerial color-infrared (CIR) photographs.
The calibration technique was especially valuable when photographs of a
different flight line, altitude, or date were compared.
METHODS AND INSTRUMENTS
Overflights
Overflights of areas near four of TVA's 12 coal-fired power plants
were performed during the 1977 and 1978 growing seasons when the foliar
effects of S02 were still visible to ground observers. The overflights
covered soybeans growing near Colbert Steam Plant in northwestern Alabama,
Johnsonville Steam Plant in western Tennessee, Shawnee Steam Plant in
western Kentucky; and soybeans, winter wheat, and pines growing near
Widows Creek Steam Plant in northeastern Alabama (Figures 27, 28, 29, and
30). Altitudes for the overflights ranged from 500 to 1800 m above ground
level (AGL). Large-format aerial mapping cameras equipped with 152-mm
focal-length lenses were used to expose Kodak Aerochrome Infrared 2443
(CIR) false-color reversal film and Kodak Ektachrome MS Aerographic 2448
-------
Figure 27. Colbert Steam Plant area in northwestern Alabama.
-------
I
•-J
Figure 28. Johnsonville Steam Plant area in western Tennessee.
-------
Rogland
Shawnee Steam
Bandana ^-r—.^ _ Plant
ossington
I ITTT71 I I I lh/>
CO
Figure 29. Shawnee Steam Plant area in western Kentucky.
-------
/*.
WIDOWS
CREEK
STEAM
PLANT
Figure 30. Widows Creek Steam Plant area in northeastern Alabama.
-------
-75-
true-color reversal film. The resulting range of image scales varied from
about 1:3000 for the low-altitude runs to 1:12,000 for the high-altitude
runs.
TVA used a Wild RC-8 camera over Johnsonville and Widows Creek, and
EPA used Wild RC-10's over Johnsonville, Colbert, Shawnee, and Widows Creek.
EPA also flew a Daedelus DS-1260 multispectral scanner over Colbert Steam
Plant in 1977 and Shawnee Steam Plant in western Kentucky in 1978.
In July 1978 two soybean fields near Colbert were photographed for
an intensive study. One field was SC^-affected and one was not. TVA's
Wild RC-8 mapping camera was used for basic CIR coverage and a cluster of
three Hasselblad 70-mm cameras was used for black-and-white multiband
photography. The Hasselblads had normal 80-mm focal-length lenses. The
70-mm multiband coverage covered the visible and near-IR spectral regions,
and the cameras were fitted with narrow-band interference filters: a
green (550 ±23 nm) filter, a red (650 ±23 nm) filter, and a visibly opaque
(720 ±13 nm) filter.
Photometric Analysis
The primary objective of photometric analysis was to detect SC>2 effects
in soybean fields by measuring differences in spectral reflectance. A sec-
ondary objective was to relate the reflectance measurements to yield of
soybeans. To do this, systematic errors in the photographs were measured
and eliminated through an image calibration process. The errors resulted
from film processing, atmospheric effects, and variation in illumination.
The reflectance patterns and trends were then compared with ground-truth
data on 862 effects to ascertain whether any relationships existed.
Photometric analysis involved use of the Scene Color Standard (SCS)
technique. Specific details of the SCS technique have been published.44
-------
-76-
The basics of photometric analysis have been described by Lillesand in an
introductory text.45 Lillesand calls the technique photographic radiometry.
The SCS procedure permits determination of the atmospheric and illumination
variables directly from the photograph. One advantage is that no a_ priori
knowledge of the reflectances of ground objects is necessary, because all
the information for calculating reflectances is available from the photo-
graph itself. Reflectances of specific objects, such as individual soy-
bean plants on large-scale images or integrated spot canopy measurements
on small-scale images, may be calculated from spot density measurements.
For the analysis of 1977 photography, two densitometer apertures, 1.0 mm
and 150 [im, were used. The 1-mm densitometer aperture was adequate for
canopy sampling on the l:12,000-scale photographs, where the spot covered
a ground area 12 m in diameter. The 150-(jm aperture was used to make
density measurements in shadow areas. For the analysis of the 1978
Colbert photographs, a 150-|Jm aperture was used for density measurements
of the canopy from the 1:12,000-scale CIR transparencies. The ground spot
for sampling was thus 1.8 m in diameter. Also measured in the calibration
process were images of asphalt surfaces (roofs and roads) and bare soil.
Reflectances from these surfaces are relatively constant temporally and
can be used to subtract the effects of atmosphere and illumination.
Reflectances of other objects can be measured to an accuracy of about 5
percent of their true values.46 This performance is comparable with that
expected from a field radiometer.
The image calibration procedure required rigorous control of varia-
bles. A density step wedge was processed with the film so that processing
effects could be measured. This procedure used sensitometric curves (D-log
E) that plot density against relative log exposure values (Appendix G).
-------
-77-
Tri-band (analytical) densities were then measured directly from the film
and converted to changes in relative exposure.
Reflectances can be obtained from each of the three spectral bands
comprising color film, whether the emulsion is true color or CIR. The
spectral coverage of a particular band is determined by the sensitivity
of that component of the film emulsion. For CIR, the emulsions are sensi-
tive to either green, red, or near-infrared wavelengths.47 Both densito-
meters were equipped with selectable filters (nos. 92, 93, 94, and 106),
enabling the operator to measure density in any one or all three emulsion
layers.
Reflectance values from one band may be divided by reflectance values
from other bands to provide a ratio that may be a sensitive indicator of
stress. Three simple ratios were calculated in this study: IR to red;
IR to green; and red to green. The IR/red ratio approaches unity as the
curve flattens with stress from S02 or some other agent.
For the 1978 Colbert test, a 21-ha affected field was sampled system-
atically with a 30-m spacing between points (Figure 31). The 196 points
were then identified on the CIR photographs so that optical densities of
the points could be measured with a microdensitometer and converted to
reflectance. Statistical correlations were done on the IBM 370 computer
using a SAS 76 program, and graphics displays were made on a Tektronix
4014 graphics terminal. The sampled parameters for this intensive study
were red, green, IR, and IR/red reflectance; plant height, percent chloro-
sis (both L and A components of index); and field elevation. A Kelsh
stereo plotter was used to construct a detailed topographic map of the
field.
-------
COLBERT AFFECTED FIELD 1500
SYSTEMATIC SAMPLING NETWORK
JULY 17, 1978
30X30 METER GRID
I
~^j
00
" s>
Figure 31. Grid for intensive sampling of soybean field near Colbert Steam Plant.
-------
-79-
RESULTS AND DISCUSSION
General
The photometric analysis technique was tested with data from
Colbert, Johnsonville, and Shawnee steam plants determine whether it
could be used to derive spectral reflectances at sample points within
selected S02-affected soybean fields. Ground-truth data describing the
effects were available from all sites, including a field near Colbert
that was intensively sampled.
Colbert Test
Field Conditions--
Exposures to S02 caused visible foliar injury to soybeans in the
Colbert project area on August 3 and August 26, 1977 (Figure 32). EPA
was asked to perform an overflight after the first exposure; but because
of concurrent requests, the aircraft did not arrive until August 29.
Drought-induced senescence and growth of the soybean canopy and weeds in
areas of adequate moisture tended to dilute the S02 effects. Effects
from the August 26 injury were still fresh. An attempt was made to dis-
tinguish and separate S02-related stress from these natural effects
through photometric analysis of the aerial photographs.
Analysis of the Colbert photographs taken in 1977 focused on five
soybean fields that fell within a single frame (Figure 33). Four of the
fields were affected by S02 and one was unaffected. The photographs also
show a set of six test panels (arrow) for calibrating reflectance in the
photometric analysis procedure. The soybeans were mature (generally 7 to
10 nodes high) and had stopped showing new growth but were not yet senes-
cent. Some areas were infested with cockleburs [Xanthium strumarium (L.)]
The soybean canopies were generally continuous, with only a few areas of
-------
COLBERT STEAM PLANT
I
00
o
Figure 32 Flight lines and S02-affected areas near Colbert Steam Plant in 1977,
Boxed numbers locate fixed S02 monitoring stations. Overlapped
patterns indicate areas of multiple exposures of plants to S02.
-------
-81-
Figure 33. Aerial CIR photograph showing Colbert area in 1977. Letters
identify selected soybean fields discussed in report.
Scale 1:12,000.
-------
-82-
soil showing from overhead. The effects of S02 exposure consisted of
light levels of chlorosis, but no necrosis.
Analysis of the Colbert photographs taken in 1978 focused on two
fields of mature Essex soybeans, one of which was affected by S02- The
chlorosis was about a week old and ranged from very light to moderate in
intensity. There was no necrosis. The affected field covered 21 ha, and
had a uniform canopy and no weed infestations (Appendix D). The soil in
some of the low areas was damp.
Measurement and Comparison of Reflectance--
For the 1977 test, the sample points were systematically selected
within the fields and the optical density of the film at each point was
measured with the microdensitometer. The densities were then converted
to reflectance. Exceptions to uniform sampling were made when required
by irregular field boundaries. No relationships between the reflectance
statistics and foliar injury levels were indicated. The S02-affected
soybeans showed lower IR and higher red reflectance; this finding was in
accordance with theory.48
A comparison of the injury levels with the reflectance ratios for the
five fields showed that the field with the highest mean ratio (A) was
unaffected by SC^; field B, with 2 percent injury, had the next highest
mean ratio; and field C, with the highest level of injury, had the lowest
mean ratio (Table 16). It appears then, that the lower the IR/red ratio,
the greater the S02 effects. Figure 34 further illustrates the relation-
ship between injury and IR/red. Field D was infested with weeds and had
a discontinuous canopy; this heterogeneity was reflected in the high stan-
dard deviation (1.11) for its reflectance ratio. A low standard deviation
would indicate a homogeneous canopy and few weeds.
-------
TABLE 16. REFLECTANCE AND FOLIAR INJURY FOR COLBERT SOYBEAN FIELDS
PHOTOGRAPHED IN 1977
Field
designation
A
B
C
D
E
Green (%)
Mean
4.68
6.02
7.53
5.70
4.57
S.D.
0.27
0.25
0.63
0.03
1.69
Red(%) IR(%)
Mean
4.67
5.43
9.05
5.62
5.27
S.D.
0.11
0.37
1.54
0.08
0.91
Mean
26.93
28.98
27.23
25.57
25.90
S.D.
1.51
1.04
1.93
3.78
2.39
IR/Red
Mean
5.77
5.34
3.01
4.55
4.91
S.D.
0.34
0.68
0.49
1.11
0.57
Observed
injury (%)
0
2
4C
2
2
Keyed to Figure 31.
LxA method (Appendix D).
"Estimated because of conflicting data.
oo
OJ
-------
-84-
o
LU
ot
4.0-
r2 =0.85
0.0
2.0
INJURY (%)
4.0
Figure 34.
Regression of IR/red ratio and foliar
injury levels for Colbert soybean fields
photographed in 1977.
For the 1978 Colbert intensive study, data from 196 sample points
within an affected field were compared to determine the relationships
between reflectance and three other parameters: chlorosis, plant height,
and elevation of the field. These data were mapped to discern correspon-
dence in patterns (Figures 35a through c). Statistical regressions of
reflectance versus chlorosis, and chlorosis versus elevation were calcu-
lated. None of these relationships was significant (or=.05), and the r
coefficients were all below 0.25 (Appendix D). When all data points with
soybean plant heights less than eight nodes and chlorosis <1 percent were
excluded, the change in r2 was minimal. A comparison of three-dimensional
plots of the data shows little similarity between or among the variations
in reflectance, chlorosis, and plant height (Figures 36a through e).
Johnsonville Test
Field Conditions--
Two incidents of SQ% injury to vegetation occurred in the photographed
Johnsonville area during July 1977. The effects were classified generally
-------
-85-
(b)
SHORT PLANTS < 7 NODES
(c)
TOPOGRAPHY
2 FOOT (06m) CONTOURS
Figures 35a-c. Thematic maps showing patterns in a soybean
field near Colbert Steam Plant in 1978.
-------
-86-
(a)
KFLECTMttC. 9KB*
(b)
KPUCTMKC DO IKCTML
Figures 36a-c.
Computer-generated perspective views of a soybean field
near Colbert Steam Plant in 1978. The vertical axis repre-
sents reflectance, and the horizontal axes represent
locations of sample points. (Compare with Figure 35.)
-------
-87-
(d)
KMCHT CNLOMMIf
(e)
PUWT WHHT
Figures 36d, e.
Computer-generated perspective views of a soybean field
near Colbert Steam Plant in 1978. In these graphs the ver-
tical axis represents observed chlorosis (d) and plant
height (e). The horizontal axes represent locations of
sample points in the field. (Compare with Figures 35 and
36.)
-------
-88-
as light to moderate. The earliest incident occurred on July 3 in an area
northwest of the plant (Figure 37). The effects persisted and were photo-
graphed by EPA on July 21. TVA acquired ground truth and obtained a dupli-
cate copy of the film from the EPA Vint Hill Farms Station. Another SC-2
incident occurred on July 23 in the same general area. Injury to soybeans
was still visible in the field on August 2, the date of the TVA overflight.
The photometric analysis focused first on 15 soybean fields in the
Johnsonville area where the SOz plume contacted the crop. The number was
later reduced to nine to exclude fields of immature soybeans. Figure 38
shows some of the fields.
As with the Colbert data analysis, this analysis involved the field-
to-field variations in reflectance ratios; the relationship between the
ratios and the S02 injury levels; and the relationships among reflectance,
S02 injury levels, and yield.
Measurement and Comparison of Reflectance--
Reflectances from soybean fields located northwest of the Johnsonville
Steam Plant were compared to determine whether they were related to foliar
injury levels. The S02 effects ranged from very light to severe, providing
a full scale for study. The affected area contained no-till fields, tilled
(plowed) fields, and barren (unplanted but recently plowed) fields. Some
fields contained mature soybeans, whereas others had young plants. Some
fields had been planted twice. Weeds, mainly cockleburs, were prevalent
in some of the no-till fields. Some no-till fields contained wheat
stubble between the rows of bean plants.
The image densities were measured, and the values were converted to
percentage reflectance. Means, standard deviations, and ratios were com-
puted (Table 17 and Appendix G). Coefficients of determination were
-------
***».
*e
3?
' *1
e>
-------
-90-
Figure 38. Aerial CIR photograph of area near Johnsonville Steam Plant.
Letters identify selected fields discussed in report. Photo
taken by EPA in 1977.
-------
TABLE 17. REFLECTANCE, FOLIAR INJURY, AND YIELD FOR SOYBEAN FIELDS NEAR JOHNSONVILLE
Reflectance
Field
Designation
F
G
H
I
J
K
L
Bare soil
IR
Mean
21.1
23.2
24.8
24.0
17.9
22.6
25.7
20.6
(%)
S.D.
3.7
4.5
2.8
4.3
1.7
1.8
1.0
1.9
Red
Mean
6.4
6.5
7.3
6.8
6.6
6.7
4.6
17.1
(%) Green (%)
S.D.
1.0
0.9
0.8
0.7
0.6
0.6
0.8
2.0
Mean
7.2
7.5
7.9
7.8
7.0
7.8
4.7
13.7
S.D.
0.8
1.3
0.6
0.6
0.7
0.5
0.6
1.0
(
IR/Red
Mean
3.3
3.6
3.4
3.5
2.7
3.4
5.6
1.2
S.D.
0.66
0.73
0.55
0.65
0.35
0.46
0.95
0.10
)bserved level
of injury
(%)
34.8
44.6
20.2
30.8
20.0
22.0
14.5
-
bu/acre
22.5
23.4
28.0
28.5
22.4
24.9
31.0
-
Yield
m3/ha
0.300
0.312
0.373
0.380
0.299
0.332
0.413
0.000
kg/ha
3676b
3823b
4574
4656
3659
4068
5064C
0
L x A method (Appendix D).
Yield = 1/2 actual reported value due to double planting.
£
Tilled field; all others are no-till.
-------
-92-
calculated for reflectance versus S02 injury, and reflectance versus
yield (Table 18). The IR/red ratio correlated significantly (a=.05) with
injury, although the positive direction of the relationship was not what
was expected (Figure 39a). None of the single-band reflectances had any
relationship with injury.
TABLE 18. COEFFICIENTS OF DETERMINATION (r2) FOR REFLECTANCE,
S02 INJURY, AND YIELD OF SOYBEAN FIELDS NEAR JOHNSONVILLE3
Parameter
Injury
Yield
Green
0.02
0.74
Red
0.33
0.65
IR
0.04
0.61
IR/Red
0.32
0.23
T - b
Injury
0.07
o
Underlined coefficients are significant (a=.05).
Percent injury for fields calculated by the L x A method (Appendix D)
Yield data were collected in fields at harvest.
The best indicator of yield was green reflectance (Figure 39b),
although the red and IR reflectances (Figures 39c and d) were also good.
In fact, all relationships between single-band reflectance and yield were
significant (a=.05). However, the IR/red ratio and yield did not corre-
late significantly at this level (Figure 39e). Yield also actually
increased with injury (Figure 39f). A possible explanation for this
relationship is related to canopy density. The density variable, as
measured from overhead photographs, is associated with stage of growth,
availability of soil moisture, soil fertility, and many other factors
that affect plant conditions before, during, and after an incident of
exposure to S02- The more dense canopy is probably associated with
higher yield. Injury level was apparently not a sufficiently powerful
factor to overcome the density (and, therefore, yield) factor.
-------
-93-
LU
ce
3.0-
r2 = 0.32
(a)
2OO 30.0 40.0
INJURY (%)
5000-
ra
.c
4000-
r2 = 0.74
72 7.6
GREEN (%)
5000-
ro
.c
^.
00
4000-
r2 = 0.65
6.4
6'8
RED (%)
7:2
5000-
4000
r2 = 0.61
16.0 20.0 24.0
IR (%)
5000-
UJ
>-
14000-
r2 = 0.23
28
3.2
IR/RED
3.6
5000-
UJ
>-
! 4000-
r2 = 0.07
(f)
20.0 300 40.0
INJURY (%)
Figures 39a through f. Statistical regressions of reflectance,
injury levels, and yield for Johnsonville
soybean fields.
-------
-94-
The relationship between the effects of power plant emissions and
productivity of crops is not well known. In general, crop yields are not
affected by SC>2 exposure unless visible foliar effects occur. Evidently,
over 5 percent of the leaf area must be affected to measurably reduce
yield.48 Common practice for estimating yield reduction involves field
sampling to determine the percentage of leaf area destroyed, and applying
an empirically or theoretically derived factor to calculate loss. The
difficulty is that many uncontrollable cultural, edaphic, and climatic
factors also affect yield. Certainly, the stage of growth at the time of
exposure to S02 is one of the more important factors. Significant reduc-
tions in yield caused by 862 exposures during the pod-filling stage of
growth in soybeans have been documented and related to the amount of foliar
chlorosis.49 However, other exposures to soybeans during the prebloom
stage did not significantly reduce yield.50
SUMMARY OF RESULTS
The Colbert data showed no significant correlations between single-
band reflectance and foliar injury. The IR/red ratio provided better
separation of affected and unaffected soybean fields, as the affected
soybeans had lower IR and higher red reflectance. The ratio correlated
significantly with injury levels. The lower the ratio, the greater the
injury. A low standard deviation for reflectance indicated a homo-
geneous canopy with few weeds.
The Johnsonville data showed conflicting results. The IR/red ratio
increased with injury, rather than decreased. The ratio did correlate
significantly with injury in spite of the direction of the relationship
being the reverse of what was theorized. All single-band reflectances
correlated significantly with yield, although the ratio did not.
-------
-95-
Extraneous variations in canopy density associated with stage of growth
may have caused these inconsistencies and may also have reversed the
direction of the relationship between IR/red and injury.
-------
-96-
SECTION 6
ANALYSIS OF MULTISPECTRAL SCANNER DATA
GENERAL
Automated classification of digital data from airborne multispectral
scanners plays an increasingly prominent role in remote sensing and sup-
plements traditional techniques of photo interpretation. Three times
since 1975 TVA has arranged MSS overflights of S02-affected soybean fields.
The first overflight, which covered the Shawnee Steam Plant area near
Paducah, Kentucky, was conducted in 1975 by NASA/Earth Resource Laboratory
(ERL), Slidell, Louisiana, using their RS-18 scanner. Despite acquisition
of comprehensive ground truth the results of the analysis were negative
because of diverse farming practices and differing stages of crop growth
(Appendix H). The second scanner overflight was conducted in 1977 by
EMSL-LV with an eleven-channel scanner over the Colbert Steam Plant area
in northwestern Alabama. The third overflight was conducted in 1978 by
EMSL-LV with the same scanner over the Shawnee Steam Plant area. Analyses
of the 1977 and 1978 data are covered in this report. For comparison,
several data classification procedures were used to produce different
classification images (maps) for comparison. There are many ways to
classify MSS data and each problem seems to require a different procedure.
Our objective was to determine which of the available classification pro-
cedures would provide the best map of the S02 effects. Ground truth would
determine accuracy.
FLIGHT LINES AND SENSOR CHARACTERISTICS
The remote sensing overflights of the Colbert Steam Plant site and
of the Shawnee Steam Plant site were conducted on August 29, 1977, and on
-------
-97-
August 15, 1978, respectively. The same equipment and sensors were used
at both locations. Timing of the overflights was critical. Optimal con-
ditions were clear weather, a high sun angle, a full soybean canopy in
most of the fields, and SOz effects that were visible to the field observer.
These conditions were satisfied at both sites in late August when the crop
had stopped putting on new growth, but was not yet senescent. Figures 40
and 41 show the flight lines for data acquisition and the areas of plume
contact with the vegetation near the two sites.
The MSS system was a Daedelus DS-1260, 11-channel scanner (Figure 42).
It had a visible band, 10-channel spectrometer and a thermal infrared
detector. The scanner was roll-stabilized and its scan rate was synchro-
nized with aircraft ground speed, which for these overflights was approxi-
mately 300 km/h (160 knots). A Wild-Heerbrugg RC-10 mapping camera with
a 152-mm focal-length lens and Kodak Aerochrome Infrared type 2443 (CIR)
film was used concurrently for aerial photography. The nominal width of
photo coverage was about 20 percent narrower than the MSS coverage.
Nominal flying heights specified for the sites were 500 and 1800 m
AGL. Unsatisfactory results had been obtained during analysis of MSS data
of Shawnee acquired by NASA from 3600 m AGL in 1975, so lower altitudes
were used in an attempt to improve the quality of the data.
Flying height and certain sensor characteristics determine the image
resolution of a scanner. At 500 m, the Daedelus scanner yields a picture
element (pixel) with a diameter of about 1 m on the ground. From 1800 m,
the pixel diameter is about 4 m. These pixel dimensions are determined
by the instantaneous field of view (IFOV) of the scanner; for the Daedelus
system this figure is 2.5 milliradians.51 Figure 43 shows the flight
configuration and illustrates that the width of coverage along the flight
-------
-------
SHAWNEE
STEAM
PLANT
Figure 41. Selected S02~affected areas and flight lines near Shawnee Steam Plant in
western Kentucky in 1978.
-------
\
o
o
Figure 42. Basic components of a digital multispectral scanner system, showing scan head,
power distributor/reference source controller, scanner control console, and
digitizer. Tape recorder not shown (photo courtesy Daedelus Enterprises, Inc.)
-------
PIXEL (IFOV)
Figure 43. Configuration for airborne MSS data acquisition.
-------
-102-
line is proportional to flying height and total field of view. The gated
scan angle of the scanner is 86°, which yields a nominal swath width of
933 m (ground distance) from 500 m flying height and 3359 m from 1800 m
flying height.
Flight line orientations were dictated by the distribution of the
S02~affected fields. For efficiency, as many fields as possible were
scanned per flight line. Each flight line was classified independently.
GROUND TRUTH
Field surveillance biologists from TVA and personnel from EMSL-LV
gathered data on SOz injury to soybeans and other vegetation at Shawnee.
Colbert ground truth was gathered by TVA personnel only.
Colbert Area
The affected area was located west of Florence, Alabama, about
three to seven km north of the power plant (Figure 44). The vegetation
injury occurred from July 20 to August 31, 1977. Evidence of multiple
exposures was confirmed. There were four 50% monitoring stations in the
immediate area. About 1620 ha were involved in the survey of S02 effects
near Colbert.
The overflight was made on August 29, 1977. The soybeans had stopped
putting on new growth by that time, so there was no concern about possible
masking effects of new foliage over the affected leaves. However, natural
senescence was of concern. Accelerated by drought, these symptoms began
to occur in late August. Senescence appears as widespread chlorosis, but
a field observer can discriminate between senescence and S02~induced
chlorosis.
-------
LEGEND
^—AFFECTED AREA LIMITS
Q53 SOYBEAN FIELD AND NUMBER
— 2 — FLIGHT LINE AND NUMBER
0 S02 MONITOR LOCATION
o
u>
I
Figure 44. Distribution of S02~affected, continuous-canopy soybean fields near Colbert Steam
Plant in 1977.
-------
-104-
Shawnee Area
For Shawnee, the SC>2 concentrations, winds, and apparent age of
effects indicated that vegetation was affected on August 7, 1978, in three
areas (A, B, and C on Figure 45). The survey was conducted on August 7-11.
Additional observations were made later in the month; in particular, new
ground truth was gathered on the day of the overflight, August 15.
The most severe effects were observed in area A (1113 ha) located
3.4 to 11.9 km east-southeast of the source. Light to moderate effects
were noted in both area B (870 ha) located 3 to 7.5 km south-southeast,
and area C (121 ha) located 5.0 to 7.0 km southeast of the source. Within
these three areas, 857 ha of soybeans were affected. The map (Figure 45)
shows both affected arid unaffected soybean fields, the unaffected fields
being located outside the areas of plume contact A, B, and C. Many unaf-
fected soybean fields beyond the immediate area of interest were not
surveyed and, therefore, are not plotted on the map.
DATA REDUCTION AND PROCESSING PROCEDURES
Preprocessing Procedures
Preprocessing of the MSS data involved three steps. First, the raw
data tapes were decommutated and reformatted to become computer-compatible
tapes (CCT's). Next, the tapes were checked for data anomalies, including
missing data and recording errors. Finally, the quality of the data was
checked by inspecting hard copy or displaying single-channel images on
the color television monitor of the console. Figure 46 shows the eight
decommutated channels for an agricultural scene near Colbert Steam Plant.
-------
SHWNEE
STEAM
PLANT
LEGEND
AFFECTED AREA LIMITS
SOYBEAN FIELD AND NUMBER
FLIGHT LINE AND NUMBER
METROPOLIS
'/
LOW »•
ALTITUDE
MONITOR LOCATION
AND NUMBER
CLASSIFIED
MSS
SCENE
I .5 0
CLASSIFIED
MSS
•CENE
o
Cn
I
Figure 45. Distribution of 803-affected, continuous-canopy soybean fields near Shawnee Steam Plant
in 1978. Fields without numbers had discontinuous canopies.
-------
-106-
2 0.42-0.45 Blue
4 0.50-.055 Green
6 0.60-0.65 Red
7 6.65-0.70 Red
0.70-0.79 Near IR
9 0.80-0.89 Near IR
10 0.92-1.10 Near IR
11 8.00-14.00 Thermal
IR
Figure 46. Eight MSS channels depicting a scene near Colbert Steam Plant.
-------
-107-
Processing Procedures
Pattern Recognition—
The CCT's were processed on the EMSL-LV Data Analysis System (DAS)
(Figure 47). The operation was organized as tasks, each being composed
of one or more software programs (Figure 48). The heart of the processing
operation is the pattern recognition block.52 Spectral pattern recogni-
tion is a numerical process for simplifying the complex patterns inherent
in MSS data. The process results in a limited number of spectrally sepa-
rable, discrete classes of data. A decision strategy is used by the
computer to assign pixels to classes.
In this project, the MSS data were classified independently using
three procedures: supervised, pseudosupervised, and unsupervised (Appen-
dices I, J, and K). The supervised procedure heavily used analyst and
computer time, and it also required a priori knowledge of the scene that
had to be gained through in situ measurements and interpretation of aerial
photographs acquired along with the MSS data. Ground truth was used to
program the computer to recognize S02~affected soybeans, unaffected soy-
beans, and background vegetative and nonvegetative cover.
The pseudosupervised classification procedure is a modification of
the conventional supervised procedure. It was used to analyze the
Shawnee data acquired in 1978, and is a modified maximum likelihood
classifier for four MSS channels (MAXL4). It uses a very fast clustering
program in combination with a minimal amount of information provided by
the analyst to train the computer to recognize certain phenomena such as
soybean fields. This information is provided for small ground areas and
called training samples. These areas have relatively homogeneous spectral
-------
OPERATOR'S TERMINAL
AND CARD READER
9 TRACK MAGNETIC
TAPE DRIVES
PLAYBACK SYSTEM
AND CENTRAL COMPUTER
o
00
I
INTERACTIVE DISPLAY SYSTEM
COLOR FILM RECORDER
Figure 47. Data Analysis System at EMSL-LV (EPA photopraph).
-------
-109-
/SENSOR\
FILM
RECORDER
OUTPUT
DATA
TRANSFORMATION
AND
PREPROCESSING
COMPUTER
COMPATIBL
TAPE
INTERACTIVE
DISPLAY
SYSTEM
1
X
1
1
1
1
<
I
PATTERN RECOGNITION
•TRAINING SAMPLE SELECTION
•STATISTICS COMPUTATIONS
• CHANNEL SELECTION
•CLASSIFICATION
•CHANNEL RATIOING
1
1
_ _/ CLASSIFIED
TAPES
Figure 48. Functional tasks for processing MSS data on Data Analysis
System at EMSL-LV.
-------
-110-
characteristics. The CCT's were manipulated initially under SEARCH, an
unsupervised trainer for the maximum likelihood classifier which generates
up to 49 classes from which statistics are computed. The statistics are
then stored in a file, and training samples are then selected using a
combination of SEARCH and manually determined data.53 The rest of the
classification procedure is automated.
An unsupervised classification procedure (UNSUP) was also used to
process digital data from selected flight lines over Shawnee and Colbert.
The procedure removes the analyst from the data processing loop, since no
training samples are used, so the need for a priori knowledge of the scene
is not great.54
However, an evaluation of the accuracy of the final classified
product always requires ground truth. Whereas the supervised and
pseudosupervised (MAXL4) classifiers use informational classes, the
unsupervised classifier uses spectral classes. Once the spectral
classes are determined, the analyst checks them for utility. A typical
problem with the unsupervised classifier is that two or more types of
phenomena to be separated may have similar spectral characteristics (e.g.,
soybeans and weeds). Unsupervised classifiers may involve n-dimensional
cluster analysis as illustrated hypothetically in Figure 49. In cluster
analysis, the determination of cluster centers and boundaries affects the
results; these determinations are especially influential when dealing with
multidimensional data sets. Iterative vector computations facilitate these
operations.55
-------
-111-
S02-AFFECTED SOYBEANS
•UNAFFECTED SOYBEANS
DECISION BOUNDARIES
MSS CHANNEL 9 ( INFRARED)
Figure 49. Hypothetical two-dimensional plot of naturally clustered
spectral measurements of soybeans.
Optimal Channel Selection--
When the number of channels that can be processed simultaneously on
the computer is less than the number of channels acquired during the over-
flight, a method for selecting the best channels for classifying the phe-
nomena of interest is needed. For our analysis of Colbert and Shawnee
data, the 10 original channels were reduced to 8 (Table 19), the maximum
number the computer could accept. The thermal infrared channel from
Shawnee was one of those eliminated because it had shown little util-
ity during the previous analysis of Colbert. The unsupervised classifier
used all eight preselected channels, but these were reduced to four for
supervised and pseudosupervised classifications. Previous analyses of
airborne MSS data have shown no advantage in processing more than four to
six channels.56
-------
-112-
TABLE 19. MULTISPECTRAL SCANNER CHANNELS3
Channel
1
2b
3C
4b'C
5C
6b,c
7b,c
8b,c
9*,c
10b'c
llb
12
Wavelength
(Mm)
Not used
0.42-0.45
0.45-0.49
0.50-0.55
0.55-0.60
0.60-0.65
0.65-0.70
0.70-0.79
0,80-0.89
0.92-1.10
8.00-14.0
Not used
Spectral region
Blue
Blue
Green
Green
Red
Red
Near-IR
Near-IR
Near-IR
Thermal IR
a
Specifications from Daedelus Enterprises, Incorporated.
Channels used for Colbert in 1977.
Channels used for Shawnee in 1978.
-------
-113-
The optimal channel selection procedure used a program entitled
SEPARATION which calculated the interclass distance separation to
derive the optimal four channels and displayed results in a matrix
format.57 This procedure was followed for each flight line of data.
Conventional Analysis Procedures—
The aerial CIR photographs were interpreted stereoscopically, and
field identifications, field boundaries, and other information were iden-
tified on registered overlays. Soybean fields were examined for uniform-
ity of canopy density, closure, weediness, and presence of visible effects
from SC>2. The photointerpreted information supported field observations
and selection of training samples.
Procedures for Evaluating Classification Accuracy
To evaluate the classifications, the color-coded patterns of the MSS
scene were compared with ground truth. During interactive processing, a
color table was set up so that certain colors represented specific land
cover types, such as affected or unaffected soybeans. Our evaluation
measured how closely the depicted pattern of colors corresponded to actual
land cover patterns as determined by field observation.
In MSS data classification there are errors of omission and errors
of commission. The first error results in underclassification and the
second, in overclassification of the phenomena of interest.
The evaluation of accuracy of the MSS classifications began by mea-
suring the areas of specific coded colors in the scene with a planimeter.
Pixel counts were not used because they were not obtainable on a field-by-
field basis. (The shapes of the fields were generally irregular.) The
area figures were recorded and tabulated along with the percentage of area
affected in each field. This percentage is the P factor of the LxAxP index
-------
-114-
(Appendix D) used in the field to describe foliar injury. Then the areas
of each field were multiplied by P, yielding measurements for the propor-
tion of affected soybeans in each field. All area measurements were con-
verted from image-based units to hectares. For example, an accurate
classification of a particular soybean field known to be 50 percent
affected by SC-2 would exist if half of the imaged area within the field
appeared in the color selected to represent affected soybeans. The loca-
tion, pattern, and clustering of the affected pixels within the delineated
field boundaries were not considered in this evaluation.
After the measurements were tabulated on a field-by-field basis, they
were summed. The measured (color-coded) areas and actual field-derived
values were then compared to determine the accuracy of the classification.
Accuracy was measured by computing the classification error E, a percent-
age representing the error of omission or commission. The formula used
is
to —i ml
p/0/\ _ I <* "M „ -1 AA
.Cil/oJ — X 1UU
a
where a is the actual observed area of S02-affected soybeans expressed
as a percentage of total field area, and m is the measured area of the
appropriate coded color representing S02-affected soybeans within the
field. Factor m is an image-derived area measurement.
RESULTS AND DISCUSSION
Optimal Altitudes
The MSS lines over Colbert and Shawnee were flown at 1800 m and 500
m AGL. Unsupervised classification of the lower altitude data from
Colbert indicated that they provided no improved accuracy, so analysis
-------
-115-
of it was discontinued. A low-altitude line generates more data per
kilometer flown and is therefore more costly and time-consuming to
analyze.
Colbert Test
Optimal MSS Channels--
Computer algorithms were used to compute divergence matrices showing
optimal separation of data classes and maximum divergence among individual
areas (agricultural fields). The matrices provided a means for selecting
the best four channels from eight (Table 20). The results of the two
TABLE 20. OPTIMAL MSS CHANNELS FOR DETECTING AND CLASSIFYING
S02-AFFECTED SOYBEAN FIELDS NEAR COLBERT STEAM PLANT IN 1977
MSS channel.
Procedure designation
1 4
7
8
9
2 4
6
7
8
Wavelength
(Mm)
0.50-0.55
0.65-0.70
0.70-0.79
0.80-0.89
0.50-0.55
0.60-0.65
0.65-0.70
0.70-0.79
Spectral
region
Green
Red
Near-IR
Near-IR
Green
Red
Red
Near-IR
a
Procedures discussed in text.
See Table 19 for the 8 channels considered for Colbert.
-------
-116-
procedures are similar, as the green, red, and near-IR channels were
chosen each time. The blue and thermal IR channels were rejected; these
two spectral regions have inherently low contrast with respect to
vegetation.
Unsupervised Classification--
The UNSUP classification procedure was only partly successful in map-
ping SC>2 effects. The classification (Figure 50) was studied to determine
how closely its patterns corresponded to ground truth. Specifically
examined were the separation of soybean fields from all other land cover
without consideration of SC>2 effects and the separation of S02~affected
soybean fields from unaffected soybean fields.
The colors of the unsupervised classification were coded during
interactive processing so that dark red represented affected soybeans and
medium red represented unaffected soybeans.
The entire scene, including the soybean fields and areas outside
the fields, was examined first. Since the dark and medium red areas out-
side the actual fields are not soybeans, they represent overclassifica-
tion. The overclassification error is 7.2 percent, a figure obtained by
dividing the area of total red color (78.4 ha) by the total scene area
(1081.8 ha) and multiplying by 100. Field checking identified these
erroneous red areas as patches of weeds in otherwise defoliated cotton
fields. The 7.2-percent error was judged to be acceptable so we
attempted to distinguish S02-affected soybeans from unaffected soybeans.
Evaluation of the accuracy of this classification requires considering
the patterns within the soybean fields. The MSS scene includes 18 soy-
bean fields (Figure 51). All areas of dark and medium red color, repre-
senting affected and unaffected soybeans, respectively, were measured
-------
Figure 50. Unsupervised classification of MSS data for
Colbert area. See Figure 51 for distribution
of soybean fields and SC>2 effects.
Unaffected Soybeans
SC>2-Affected Soybeans
-------
-118-
-18
-10
a,
I km
Figure 51. Distribution of SC>2 effects (screened pattern) within dense-
canopy soybean fields near Colbert Steam Plant. Keyed to
unsupervised classification (Figure 50).
-------
-119-
and the numbers were converted to equivalent ground area in hectares.
The ground truth was obtained by multiplying the area of each field by
an observation of the proportion of soybeans affected in the field, the
result being a figure representing the area of affected soybeans for
each field.
Comparison of the image colors and ground truth revealed an over-
classification of 142 percent for the unsupervised procedure (Table 21).
In summary, the unsupervised classification procedure could separate
soybeans from other land cover with an error of commission of 7.2 percent.
It could not, however, distinguish S02~affected soybeans from unaffected
soybeans. The difficulty is attributed to the low levels of foliar injury
prevailing at Colbert during the 1977 overflight. The highest percentage
of area affected in any dense-canopy soybean field was 12 percent (Table 21)
Supervised Classification--
The supervised procedure was unsuccessful in separating soybeans
from other field crops because of spectral overlap of the classes. These
results are considered inconclusive because they were not tested on a full
range of foliar effects.
Shawnee Test
Optimal MSS Channels--
As stated previously, the best channels for classifying soybeans
were selected with the SEARCH algorithm. The unsupervised classifier
used all eight decommutated channels (numbers 3 through 10, Table 19),
but the pseudosupervised classifier required reducing the number of
channels to four.
-------
TABLE 21. COMPARISON OF FIELD OBSERVATIONS OF S02 EFFECTS ON SOYBEANS AND
RESULTS OF MSS UNSUPERVISED CLASSIFICATION FOR COLBERT SCENE 2
Ground truth
Field
identification
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
TOTAL
Classification
MSS classification
Field Proportion Actual Measured
area of field affected affected
(ha) affected (%)3 area (ha) area (ha)
15.26
5.64
2.09
6.27
1.25
10.66
1.88
14.63
1.05
12.54
2.72
4.18
6.48
3.76
3.14
4.39
9.82
16.72
la - ml
error —
a
1.6
2.7
4.8
4.8
4.0
0.6
0.6
12.0
2.0
4.0
2.7
0.4
2.4
0.2
0.0
0.6
1.6
1.0
x 100 =
0.24
0.15
0.10
0.30
0.05
0.06
0.01
1.76
0.02
0.50
0.07
0.02
0.16
0.01
0.00
0.03
0.16
0.17
a = 3.81
(3.81 - 9.22|
3.81 x
0.21
0.00
0.63
0.00
0.00
0.00
0.00
0.21
0.00
0.84
0.21
0.21
0.21
1.05
0.00
0.84
4.81
0.00
m = 9.22
100 - 142.0 percent
Measured
unaffected
area (ha)
15.05
5.64
1.46
6.27
l.r 25
10.66
1.88
14.42
1.05
11.70
2.51
3.97
6.27
2.71
3.14
3.55
5.01
16.72
overclassified
Error
(%)
12.5
100.0
530.0
100.0
100.0
100.0
100.0
88.1
100.0
68.0
200.0
950.0
31.3
10,400.0
0.0
2,700.0
2,906.3
100.0
Ni
O
I
factor of LxAxP index (Appendix D).
-------
-121-
Table 22 lists the channels selected for the north-south and east-
west flight lines. A blue channel was selected for the north-south
lines instead of green. Earlier selections using Colbert data preferred
green to blue. SEARCH is influenced by the spectral characteristics of
the scanned terrain and features, and an abundance of water in the scene
will alter the result from that obtained if the scene were all land.
Part of the Ohio River was included in the Shawnee flight lines and this
may have biased the selection.
TABLE 22. OPTIMAL MSS CHANNELS FOR DETECTING AND CLASSIFYING
S02-AFFECTED SOYBEAN FIELDS NEAR SHAWNEE STEAM PLANT IN 1978
MSS channel Flight
designation Line
3 2,3
7
8
9
5 5
7
8
9
Wavelength
0.45-0.49
0.65-0.70
0.70-0.79
0.80-0.89
0.55-0.60
0.65-0.70
0.70-0.79
0.80-0.89
Spectral
region
Blue
Red
Near-IR
Near-IR
Green
Red
Near-IR
Near-IR
MSS channels 3 through 10 considered. Procedure for channel selection
accomplished by calculating interclass distance separation.
Unsupervised Classification--
The unsupervised classifier (UNSUP) was also applied to flight line
2 data. Only the dense-canopy fields were included in the evaluation. A
total of 32 dense-canopy soybean fields, 25 of which were affected by S02,
are depicted in the scene (Figures 52 and 53).
-------
-122-
Unaffected
Soybeans
Unaffected
Soybeans
S02-Affected
Soybeans
Figure 52. Unsupervised classification of MSS line 2 data near Shawnee
Steam Plant. See Figure 53 for distribution of soybean
fields and S02 effects.
-------
-123-
Figure 53. Distribution of 862 effects (screened pattern) within dense-
canopy soybean fields near Shawnee Steam Plant. Keyed to MSS
flight line 2, unsupervised classification (Figure 52).
-------
-124-
The classification scheme was moderately successful in differentiat-
ing soybean fields from other land cover. The green-colored areas appear-
ing where soybeans did not exist totaled 242.5 ha. Dividing this by the
total scene area analyzed (3192.3 ha) yielded a classification error of
7.6 percent.
The UNSUP classifier could not differentiate S02-affected soybeans
from unaffected soybeans. For this evaluation, the analyst focused only
upon the patterns appearing within the soybean fields. Table 23 lists the
fields and relevant data. The measured light- and medium-green areas,
representing affected soybeans in the scene, totaled 223.1 ha. According
to ground truth, the actual total affected area in the scene was 110.8 ha,
so the classification error is 101.4 percent.
Pseudosupervised Classification--
The MAXL4 pseudosupervised classifier was applied to east-west
flight line 5 and north-south flight lines 2 and 3. The classification
of line 5 data was eventually abandoned because of complications asso-
ciated with variations in canopy density within the soybean fields
there. This discussion describes results from lines 2 and 3 only.
The classification from flight line 2 contains 26 dense-canopy soy-
bean fields, 21 of which were affected by 862 (Figures 54 and 55). The
area of each of the colors was measured to evaluate the overall perfor-
mance of the classifier. To determine the extent of overclassification,
if any, we measured the total area of the colors representing soybeans
where the crop did not exist. For this scene, the classification error
was 5.1 percent, a figure derived by dividing the total area of misidenti-
fied soybeans (109.1 ha) by the total scene area analyzed (2120.6 ha).
-------
TABLE 23. COMPARISON OF FIELD OBSERVATIONS OF S02 EFFECTS ON SOYBEANS AND
RESULTS OF MSS UNSUPERVISED CLASSIFICATION FOR SHAWNEE FLIGHT LINE 2
Field
identification
1
2
3
4
5
6
7
8
9
10
11
12
13
16
17
18
19
22
23
25
26
27
28
29
30
31
32
33
34
Field
area
(ha)
11.7
17.6
24.3
6.9
8.5
5.3
32.4
14.2
8.1
25.1
8.9
40.5
16.2
4.5
4.9
2.8
6.5
5.3
1.6
2.0
4.0
8.1
4.1
5.3
13.0
12.1
1.6
3.2
7.3
Ground truth
Proportion
of field
affected (%)a
20
0
40
40
40
60
50
0
50
50
40
0
50
50
40
30
70
90
20
0
0
0
40
80
30
50
70
60
90
MSS classification
Actual
affected
area (ha)
2.3
0.0
9.7
2.8
3.4
3.2
16.2
0.0
4.1
12.6
3.6
0.0
8.1
2.3
2.0
0.8
4.6
4.8
0.3
0.0
0.0
0.0
1.6
4.2
3.9
6.1
1.1
1.9
6.6
Measured
affected
area (ha)
11.7
17.6
21.8
5.9
8.5
5.3
32.4
2.5
6.8
10.9
7.5
21.8
6.8
3.4
4.9
2.8
5.9
5.1
1.6
0.8
2.5
8.1
4.1
5.3
5.1
0.0
0.0
0.0
7.3
Measured
unaffected
area (ha)
0.0
0.0
2.5
1.0
0.0
0.0
0.0
11.7
1.3
14.2
1.4
18.7
9.4
1.1
0.0
0.0
0.6
0.2
0.0
1.2
1.5
0.0
0.0
0.0
7.9
12.1
1.6
3.2
0.0
Error
(%)
408.7
-
124.7
110.7
150.0
65.6
100.0
-
65.9
13.5
108.3
-
16.0
47.8
145.0
250.0
216.7
6.3
433.3
-
-
-
156.3
26.2
30.8
100.0
100.0
100.0
10.6
N5
-------
TABLE 23 (Continued)
Field
identification
35
36
TOTALS
Classification
Field
area
(ha)
6.5
2.8
|a -
a
Ground truth
Proportion
of field
affected (%)a
40
70
m| inft IllO
MSS classification
Actual
affected
area (ha)
2.6
2.0
a = 110.8
.8-223.11
110.8
Measured
affected
area (ha)
5.9
0.8
m = 223.1
101.4 percent
Measured
unaffected
area (ha)
0.6
2.0
overclassified
Error
126.9
60.0
Factor P of LxAxP index (Appendix D).
-------
-127-
S02~Affected Soybeans
Unaffected Soybeans
Unaffected Soybeans
Pasture, other Grass,
Incomplete Canopy
Soybeans
Figure 54. Pseudosupervised classification of MSS line 2 data near
Shawnee Steam Plant. For distribution of SC>2 effects
within soybean fields, see Figure 55.
-------
-128-
Figure 55. Distribution of 862 effects (screened pattern) within dense-
canopy soybean fields near Shawnee Steam Plant. Keyed to
flight line 2, pseudosupervised classification (Figure 54).
-------
-129-
The primary objective of this study was to determine how well the
MAXL4 classifier could differentiate S02-affected soybeans from unaffected
soybeans. For this determination, only the colors appearing within the
delineated field boundaries were considered.' Table 24 lists the fields,
their actual affected proportions, the affected areas measured from the
classified image, and the totals where appropriate. According to ground
truth, the total area of S02-affected soybeans was 102.1 ha. The light
green color representing S02-affected soybeans on the image covers 77.2
ha. The error is therefore 24.4 percent underclassified.
In summary, the pseudosupervised classification of line 2 data has
resulted in an overclassification of soybeans, without regard to SQ%
effects, of about 5 percent. The attempt to differentiate S0£-affected
soybeans and unaffected soybeans resulted in an underclassification
error of about 24 percent.
The pseudosupervised classification from flight line 3 depicts 22
dense-canopy soybean fields, 19 of which were affected by SC>2 (Figures 56
and 57, and Table 25). The correspondence of the color-coded patterns to
the actual distribution of soybeans was excellent, the area of misidenti-
fication being only 20.2 ha and the total analyzed scene area being 3207.4
ha. The error of commission, 0.6 percent, represents a slight overclassi-
fication of soybeans without regard to S02 effects.
The separation of S02-affected soybeans and unaffected soybeans
with the pseudosupervised classifier was good for line 3 data. The
total area coded as soybeans (medium yellow) in the scene is 81.7 ha.
The actual area determined by field observations was 73.4 ha, so the
classification error is 11.3 percent.
-------
TABLE 24. COMPARISON OF FIELD OBSERVATIONS OF S02 EFFECTS ON SOYBEANS AND
RESULTS OF MSS PSEUDOSUPERVISED CLASSIFICATION FOR SHAWNEE FLIGHT LINE 2
Field
identification
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
Field
area
(ha)
11.7
17.6
24.3
6.9
8.5
5.3
32.4
14.2
8.1
25.1
8.9
40.5
16.2
30.4
2.0
4.5
4.9
2.8
6.5
1.6
2.8
5.3
1.6
Ground truth
Proportion
of field
affected (%)a
20
0
40
40
40
60
50
0
50
50
40
0
50
30
40
50
40
30
70
80
80
90
20
MSS classification
Actual
affected
area (ha)
2.3
0.0
9.7
2.8
3.4
3.2
16.2
0.0
4.1
12.6
3.6
0.0
8.1
9.1
0.8
2.3
2.0
0.8
4.6
1.3
2.2
4.8
0.3
Measured
affected
area (ha)
2.5
1.7
13.4
2.5
6.8
5.1
1.7
0.0
0.8
11.7
6.8
1.7
12.6
0.0
0.9
4.5
1.7
0.4
0.8
0.3
0.0
0.7
0.0
Measured
unaffected
area (ha)
9.2
15.9
10.9
4.4
1.7
0.2
30.7
14.2
7.3
13.4
2.1
38.8
3.6
30.4
1.1
0.0
3.2
2.4
5.7
1.3
2.8
4.6
1.6
Error
(%)
8.7
-
38.1
10.7
100.0
59.4
89.5
0.0
80.5
7.1
88.9
-
55.6
100.0
12.5
95.7
15.0
50.0
82.6
76.9
100.0
85.4
100.0
-------
TABLE 24 (Continued)
Ground truth MSS classification
Field Proportion Actual Measured Measured
Field area of field affected affected unaffected
identification (ha) affected (%) area (ha) area (ha) area (ha)
24 12.6 30
25 2.0 0
26 4.0 0
TOTALS
a
3.8 0.0 12.6
0.0 0.0 2.0
0.0 0.0 4.0
a = 102.1 m = 77.2
102.1 - 77.21 _ „ , _
•inn -t ~ 24.4 percent underclassified
Error
(%)
100.0
0.0
0.0
Factor P of LxAxP index (Appendix D).
-------
-132-
Unaffected Soybeans
S02-Affected Soybeans
S02~Affected Soybeans
Weedy Soybeans
Figure 56. Pseudosupervised classification of MSS line 3 data near
Shawnee Steam Plant. For distribution of S02 effects
within soybean fields, see Figure 57.
-------
-133-
Figure 57. Distribution of SQ% effects (screened pattern) within dense-
canopy soybean fields at Shawnee Steam Plant. Keyed to flight
line 3, pseudosupervised classification (Figure 56).
-------
TABLE 25. COMPARISON OF FIELD OBSERVATIONS OF S02 EFFECTS ON SOYBEANS AND
RESULTS OF MSS PSEUDOSUPERVISED CLASSIFICATION FOR SHAWNEE FLIGHT LINE 3
Ground truth MSS classification
Field Proportion Actual Measured Measured
Field area of field affected affected unaffected
identification (ha) affected (%)a area (ha) area (ha) area (ha)
1 11.7 20
2 17.6 0
3 24.3 40
5 8.5 40
6 5.3 60
9 8.1 50
10 25.1 50
11 8.9 40
12 40.5 0
13 16.2 50
15 2.0 40
16 4.5 50
17 4.9 40
18 2.8 30
20 1.6 80
21 2.8 80
27 8.1 0
28 4.1 40
29 5.3 80
34 7.3 90
35 6.5 40
36 2.8 70
TOTALS
la - m|
Classification error x 100
a
2.3 6.8 4.9
0.0 9.2 8.4
9.7 5.1 19.2
3.4 8.5 0.0
3.2 4.2 1.1
4.1 5.1 3.0
12.6 19.2 5.9
3.6 3.4 5.5
0.0 0.0 40.5
8.1 2.6 13.6
0.8 0.0 2.0
2.3 4.5 0.0
2.0 3.9 1.0
0.8 2.8 0.0
1.3 0.0 1.6
2.2 0.0 2.8
0.0 2.5 5.6
1.6 0.0 4.1
4.2 1.3 4.0
6.6 1.3 6.0
2.6 0.0 6.5
2.0 1.3 1.5
a = 73.4 m = 81.7
73.4 - 81.71 , n 0 ^ T -f- A
11 -\ n r* Y" p r* T*i i o"vr*T~p 1 i*"**"1! TT r*n
^_ x i J- J- • *J LJCA. t-C 11 L- L/VC4_V_._LCl^OJ-J-J-C;U
15 . 4
Error
195.7
-
47.4
150.0
31.2
24.4
52.4
5.6
0.0
67.9
100.0
95.7
95.0
250.0
100.0
100.0
-
100.0
69.0
80.3
100.0
35.0
aFactor P of LxAxP index (Appendix D).
-------
-135-
In summary, the pseudosupervised classification of line 3 data resulted
in less than 1 percent overclassification of soybeans without regard to 862
effects. The attempt to differentiate S02~affected soybeans and unaffected
soybeans resulted in an underclassification of about 11 percent.
Summary of Classification Results
The results of the evaluation of the three procedures for classifica-
tion of the MSS data are summarized in Table 26. The pseudosupervised
classifier is superior to the others, but it should be kept in mind that
the Colbert test dealt with low-level SC>2 effects, whereas the Shawnee
test dealt with a wide range of foliar injury. The supervised classifier
may have yielded better results had it been available for analyzing
moderately or severely injured soybeans.
Enhancement of Patterns of 862 Effects Within Soybean Fields
The classifications previously described dealt only with field-
to-field differences in S02 effects. Within-field variations should
also be considered, but verification of the patterns in the field
was not possible because of time constraints on the ground observers
at Shawnee. Still, sufficient field data was available to verify the
patterns in a few areas.
The I2S Image Processing System at TVA's Mapping Services Branch
in Chattanooga was used to enhance and display selected scenes of MSS
data covering the Shawnee area. A density level-slicing procedure was
used to display the background in monochrome and the S02-effects in orange
(Figure 58). The correspondence of patterns is fairly close in some
fields (compare Figures 58 and 55). Only those soybean fields that were
in the scene and had dense, mature canopies were considered. Table 27
-------
TABLE 26. SUMMARY OF ERRORS USING THREE PROCEDURES
FOR DETECTING AND CLASSIFYING S02 EFFECTS ON SOYBEANS
Task
Site with
very light to light S02 effects
Unsupervised Supervised
Site with
moderate to severe S02 effects
Unsupervised Pseudosupervised
Separation of soybeans
from other land cover
Separation of S02-affected
from unaffected soybeans
+7.2% *
+142.0% *
+7.6%
+101.4%
+5.1% (line 2);
+0.6% (line 3)
-24.4% (line 2);
+11.3% (line 3)
^Inconclusive results, error not determined.
+ indicates overclassification.
- indicates underclassification.
Zero percent would indicate no error.
-------
-137-
Figure 58. Enhanced MSS image (line 2) of area near Shawnee Steam Plant
showing S02~affected soybean fields. Dense-canopy soybean
fields are numbered, the designations corresponding to those
in text and Table 28. Light toned pattern indicates chlorosis.
Numbered points are sites of observations (Table 27).
-------
-138-
TABLE 27. WITHIN-FIELD S02 EFFECTS ON SOYBEANS
NEAR SHAWNEE STEAM PLANT IN 1978
Field
0
Designation
4
15
16
17
18
22
25
26
Data
Point
1
2
3
4
1
2
3
1
2
3
4
5
1
2
3
1
2
3
1
2
3
1
2
3
4
5
1
2
3
4
5
Chlorosis
(%)
3.0
1.2
1.4
1.6
1.8
4.8
6.0
3.2
4.0
6.0
7.2
10.0
2.4
2.0
2.7
1.2
1.8
1.8
25.0
24.0
13.5
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
Mean Overall Level
Chlorosis (%) of Chlorosis
1.8 Very Light
4.2 Very Light
6 . 1 Light
2.4 Very Light
1.6 Very Light
20.8 Severe
0.0 Unaffected
0.0 Unaffected
aField number keyed to flight line 2 maps presented previously
(Figure 55).
LxAxP index (Appendix D).
-------
-139-
lists the individual chlorosis observations for each point and a mean
chlorosis value for each field. The data are interpreted as follows:
Field 4: Contradictory results. The field is affected and the
image shows effects, but the patterns do not coincide.
Field 15: Good correspondence of image pattern with ground truth.
Field 16: Contradictory results. The field is apparently affected,
but the image shows it to be unaffected.
Field 17: Fairly good correspondence. The data point showing the
highest level of chlorosis falls within the orange area, and the
points outside the orange area have lower levels of chlorosis. All
of the field is apparently affected to some degree.
Field 18: General correspondence. All of the field was apparently
chlorotic and the image shows it as orange.
Field 22: Correspondence uncertain. The image indicates presence
of spotty effects, but the ground truth neither confirms nor denies
it. Ground truth does indicate some severe effects.
Field 25: Corresponds. No effects in field or on image.
Field 26: Corresponds. No effects in field or on image.
These comparisons indicate that either the field observations need
to be made differently or in more detail, or that the enhanced patterns
do not represent the effects. It would not generally be practical to
increase the density of observations within each field because of time
constraints on field surveillance teams. The problem is probably that
the scanner cannot detect very light and light chlorosis.
-------
-140-
REFERENCES
1. Allen, W. A. and A. J. Richardson. Interaction of Light with a
Plant Canopy. J. Opt. Soc. Amer., 58:1023-28, 1968.
2. Gates, D. M., H. J. Keegan, J. C. Schleter, and V. R. Weidner.
Spectral Properties of Plants. Appl. Opt., 4(1):11-19, 1965.
3. Knipling, H. B. Physical and Physiological Basis for the Reflectance
of Visible and Near-Infrared Radiation from Vegetation. Remote
Sensing of Environ., 1:155-159, 1970.
4. Ibid.
5. Ibid.
6. Knipling, H. B. Leaf Reflectance and Image Formation on Color-Infrared
Film, in Johnson, P. L., ed. , Remote Sensing in Ecology. University
of Georgia Press, Athens, p. 20, 1969.
7. Wert, S. L. A System for Using Remote Sensing Techniques to Detect
and Evaluate Air Pollution Effects on Forest Stands. Proc., Sixth
International Symposium on Remote Sensing of Environ., Ann Arbor,
Michigan, p. 174, 1973.
8. Jordan, C. F. Derivation of Leaf Area Index from Quality of Light
on the Forest Floor. Ecology 50 (4) 663-666, 1969.
9. Colwell, J. E. Bidirectional Spectral Reflectance of Grass Canopies
for Determination of Above Ground Standing Biomass. Ph.D. Thesis,
Univ. Michigan, 174pp., 1973.
10. Colwell, J. E. Vegetation Canopy Reflectance. Remote Sensing of
Environ., 3:175-183, 1974.
11. Rouse, J. W., R. H. Haas, J. A. Schell, and D. W. Deering. Monitoring
Vegetation Systems in the Great Plains with ERTS. Third ERTS Symposium,
NASA SP-351 I, 309-317, 1973.
12. Rouse, J. W., R. H. Haas, J. A. Schell, D. W. Deering, and J. C. Harlan.
Monitoring the Vernal Advancement and Retrogradation (Greenwave Effect)
of Natural Vegetation. NASA/GSFC Type III final report, Greenbelt,
Maine, 371 pp, 1974.
13. Johnson, G. R. Remote Estimation of Herbaceous Biomass. M.S. Thesis,
Colorado State Univ., Fort Collins, 120 pp., 1976.
14. Tucker, C. J. Red and Photographic Infrared Linear Combinations for
Monitoring Vegetation. NASA/GSFC Tech Memo 79620, Greenbelt,
Maryland, p. 26, 1978.
-------
-141-
15. Reeves, R. 6., A. Anson, and D. Landen, Manual of Remote Sensing,
Vol. 2, Am. Soc. Photogramm., Falls Church, Virginia, 1975, p. 1407.
16. Miller, P. R., J. R. Parameter, Jr., B. H. Flick, and C. W. Martinez.
Ozone Dosage Response of Ponderosa Pine Seedlings. Air Pollution
Control Assoc. Jour., 19:435-438, 1969.
17. Heller, R. C. Large-Scale Color Photo Assessment of Smog-Damaged Pines.
Proc., Am. Soc. Photogramm. and Soc. Photog. Sci. and Eng., New York:
85-98, 1969.
18. Wert, S. L. A System for Using Remote Sensing Techniques to Detect
and Evaluate Air Pollution Effects on Forest Stands. Proc., Sixth
Internat. Symposium on Remote Sensing of Environ., Ann Arbor,
Michigan, 1169-78, 1969.
19. Zealer, K. A., R. C. Heller, N. X. Norwick, and M. Wilkes,
The Feasibility of Using Color Aerial Photography to Detect and
Evaluate Sulphur Dioxide Injury to Timber Stands. U.S. Forest Service,
Berkeley, California, November 1971.
20. Heller, R. C., op. cit.
21. Fritz, E. L., and S. P. Pennypacker. Attempts to Use Satellite Data
to Detect Vegetative Damage and Alteration Caused by Air and Soil
Pollutants. Phytopathology, 65(10):1056-60, 1975.
22. Wiegand, C. L. Reflectance of Vegetation, Soil, and Water. U.S.
Dept. Agr., Agr. Res. Serv., Progress Report Type II, E74-10265,
Weslaco, Texas, 59 pp., 1974.
23 Murtha, P. A. S0% Forest Damage Delineation on High-Altitude
Photographs. Canada Centre for Remote Sensing, Proc., First Canadian
Symposium on Remote Sensing, Ottawa, Canada, 71-82, 1972.
24. Reeves, R. G., A. Anson, and D. Landen. op cit.
25. Walker, J. E., and D. B. Dahm. Measuring Environmental Stress.
Environ. Sci. and Tech., 9(8):714-719, 1975.
26. Schott, J. R., D. W. Gaucher, and J. E. Walker. Aerial Photographic
Technique for Measuring Vegetation Stress from Sulfur Dioxide.
Calspan Corporation Report YB-5967-M-1. Buffalo, New York, No date,
17 pp.
27. Pell, E., and R. Brock. Spatial and Spectral Mapping of the Response
of Vegetation to Air Pollutants (Summary), Center for Air Environment
Studies, The Pennsylvania State University, Annual Report, 1974-75.
28. Jackson, R. Detection of Plant Disease Symptoms by Infrared,
J. Biol. Phot. Assoc. 32(2):45-58, 1964.
29. Eastman Kodak Company. Applied Infrared Photography, Pub. No. M-28.
Rochester, New York, 88 pp., 1972
-------
-142-
30. Rohde, W. G. and C. E. Olson, Jr. Detecting Tree Moisture Stress.
Photogramm. Eng. 36(6):561-566, 1970.
31. Jacobson, J. S. and A. C. Hill (eds.). Recognition of Air Pollution
Injury to Vegetation: A Pictorial Atlas. Air Pollution Control Assoc.
Inf. Report. 1, 1970.
32. Talmi, Y. Application of Optical Multichannel Spectrometric Detectors.
American Laboratory, p. 79, March 1978.
33. Princeton Applied Research Corporation. OMA-2 Model 1215 Operations
Manual. Princeton, New Jersey, 158 pp., 1978.
34. Barr, A. J., J. H. Goodnight, J. P. Soil, and J. T. Helwig. A User's
Guide to SAS76. Sparks Press, Raleigh, North Carolina, p. 275, 1976.
35. Ibid.
36. Jones, H. C., N. L. Lacasse, W. S. Liggett, and F. P. Weatherford.
Experimental Air Exclusion System for Field Studies of S02 Effects
on Crop Productivity. EPA-600/7-77-122, 67 pp, 1977.
37. Ibid., p. 1-5.
38. Ibid.
39. Holmes, R. A. Field Spectroscopy, in Remote Sensing with Special
Reference to Agriculture and Forestry. Washington, D.C.: National
Academy of Sciences, p. 298-308, 1970.
40. Ibid.
41 Murtha, op cit., p. 1149.
42. Ibid.
43. Ibid.
44. Piech, K. R., and J. E.Walker. Interpretation of Soils. Photogramm.
Eng., 38(l):87-94, 1974.
45. Lillesand, T. M. An Introduction to Photographic Radiometry and
and Spectral Pattern Recognition. State University of New York,
Syracuse, 1976, p. 1.
46. Walker, J. E., Personal Communication, 1977.
47. Eastman Kodak Company. Infrared and Ultraviolet Photography.
Part 2, Applied Infrared Photography, Tech. Pub. M-27/28-H.
Rochester, New York, 1972.
48. Barrett, T. W., and H. M. Benedict. Sulfur Dioxide, in Recogni-
tion of Air Pollution Injury to Vegetation. Air Pollution Control
Association, Pittsburgh, Pennsylvania, pp. C1-C17, 1970.
-------
-143-
49. Jones, H. C., F. P. Weatherford, W. S. Liggett, Jr., and J. R.
Cunningham. Effect of Foliar Injury Caused by Exposure to Sulfur
Dioxide on Yield of Soybeans - Results of a Large Scale Field
Investigation. Division of Environmental Planning, Tennessee Valley
Authority, Muscle Shoals, Alabama. Manuscript in preparation.
50. Jones, H. C., J. R. Cunningham, S. B. McLaughlin, N. T. Lee, and
S. S. Ray. Investigation of Alleged Air Pollution Effects on Yield
of Soybeans in the Vicinity of the Shawnee Steam Plant. E-EB-73-3,
Division of Environmental Planning, Tennessee Valley Authority,
Muscle Shoals, Alabama, 36 pp., 1973.
51. Daedelus Enterprises, Inc., Daedelus DS-1260 Multispectral Scanner,
Ann Arbor, Michigan, undated brochure.
52. Anderson, J. E., and C. E. Tanner. Remote Monitoring of Coal Strip
Mine Rehabilitation. U.S. Environmental Protection Agency, Report
EPA-600/7-78-149, p. 8, 1978.
53. Tanner, C. E. Unpublished Documentation. Lockheed Electronics
Company, Las Vegas, Nevada, 1979.
54. Pooley, J. Unpublished Program Documentation for Unsupervised
Sequential Cluster Program (UNSUP), NASA/Earth Resource Laboratory,
Slidell, Louisiana, 11 pp. including update, 1976.
55. Su, M. Y. The Composite Sequential Clustering Technique for
Analysis of Multispectral Scanner Data. NASA Contractor Report
CR-128999. Marshall Space Flight Center, Alabama, 1-1, 1972.
56. Bauer, M. E., J. E. Cipra, P. E. Anuta, and J. B. Etheridge. Identi-
fication and Area Estimation of Agricultural Crops by Computer
Classification of Landsat MSS Data. Remote Sensing of Environ.,
8(1), p. 90, 1979.
57. Pearson, R. Unpublished Program Documentation for Optimum Channel
Selection, NASA/Earth Resources Laboratory, Slidell, Louisiana, date
unknown.
58. Tanner, C. E., op cit.
59. Pooley, J. , pj> cit.
-------
APPENDIX A
SPECIFICATIONS OF OPTICAL MULTICHANNEL ANALYZER TSR SYSTEM
-------
-145-
SPECIFICATIONS OF OPTICAL MULTICHANNEL TSR SYSTEM
SYSTEM DESCRIPTION
The OMA-2 Optical Multichannel Analyzer is a microprocessor-
controlled multichannel optical detector and visual display system
consisting of the 1215 console, 1216 detector controller, 1252E detector,
and components adapted by TVA including a telescope, a fiber-optic cable
and faceplate, and an x-y recorder.
Component
Console, Model 1215
Front Panel
Rear panel
Memory system
RAM data storage
Disc storage
Arithmetic functions
Power input
Dimensions
Unit weight
Spectrometer, Model 1225
(Jobin Yvon M25
grating monochromator)
Specifications
84-key keyboard, 41-cra diagonal display,
flexible disc drive (IBM compatible
format).
Analog recorder output.
Monostore Xl/Planar LSI-11 16K x 16
memory assembly by Monolithic Systems
Corporation; signal interface through
DEC LSI-11 I/O bus.
4K single precision, 2K double precision.
Operating system plus 100 spectral curves
of 500 double precision points each.
+, -, T, natural log, and decimal log; con-
stants, powers, and roots on full curves;
arithmetic functions are formatted in
algebraic notation with parenthesis capa-
bility for seven reference curves and
recursive operation up to disc capacity.
115/230 V ± 10 percent, 50/60 Hz, 450 watts.
44.86 cm W x 72.72 cm D x 39.27 cm H.
50 kg
Focal length 0.25 m, aperture f/3,
holographic grating, 152.65 grooves/mm,
blazed at 2 pm, input slits, 2 available,
0.25 mm or 0.90 mm wide.
-------
-146-
Cproponent
Detector, Model 1252E
(IR-enhanced)
Detector controller,
Model 1216
Fiber optic interface,
Model 1225Q
X-Y recorder
Telescope
Calibration lamps
Specifications
Silicon-vidicon target, sensitivity
2400 photons/count at 600 nm, background
noise 1.5 counts rms max., full-scale
16,383 counts/channel/frame, dynamic range
1 x 104 rain., linearity as a function of
intensity ± 1 percent, scanned area
12.5 x 12.5 mm, useful spectral range 350
to 1100 nm (scan width restricted to 337
nm band)
16-bit computer peripheral, frame
scan time 10 to 70 ms, channel time 20 to
140 ps, number of channels scanned 2 to
512/track, number of tracks 1 to 256,
power input 115/230 V ± 10 percent, 50/60
Hz, 14 watts, dimensions 44.2 cm W x 46.7
cm D x 13 cm H, unit weight 15.5 kg.
152.4 cm L x 0.1 mm D (input slit, output
circle), quartz fiber optics bundles,
adaptor faceplate connects cable to
polychromator.
Hewlett-Packard Model 2D-2
Gamma Scientific, Inc., Model 2020-31 f/2.8,
focal length 190 mm, selectable angles of
view 3°, 1°, 20', 6'.
Pen-Ray Krypton 760-nm line; Pen-Ray
low-pressure mercury vapor 436-, 546-,
579-nm lines.
-------
-147-
10-3 T
O
£
O
O
10
-5
200
400 600 800
WAVELENGTH (nm)
1000
1200
Figure A-l. Typical spectral response of Model 1252E infrared-enhanced
silicon-vidicon detector (data from Princeton Applied
Research Corporation).
-------
APPENDIX B
SPECTRAL CURVES, LABORATORY EXPERIMENT
-------
-149-
450 500 550 600 650 700 750
WAVELENGTH (nm)
SPECTRUM Soybeans areram
S02 CONC .'j.72.0
FOLIAR INJURY: TOTAL.
CHLOROSIS _
NECROSB_3£_%
Y-AXIS COUNTS AT 767nm -tell
450 SCO 550 600 650 700 750
WAVELENGTH (nm)
SPECTRUM Sovb«on» aecnan MZ
SO; CONG /572.0 uo/m3 a a h
FOLIAR IN JURY: TOTAL >s %
CHLOROSIS >" %
NECROSIS !_%
Y-AXIS COUNTS AT 767nm
450 500 550 600 650 700 750
WAVELENGTH (nm)
SPECTRUM Sovfa«an« RECORD
SOj CONC }^^^G Hg/nfl .50
FOLIAR INJURY: TOTAL
CHLOROSIS
NECROSIS %
Y-AXIS COUNTS AT 767nm 4TP"
450 500 550 600 650 700 750
WAVELENGTH (nm)
SPECTRUM ffjuj ft~/
SOj CONC _
FOLIAR INJURY:
RECORD
TOTAL
CHLOROSIS %
NECROSIS %
Y-AXIS COUNTS AT 767nm_t3i2la
450 500 550 600 650 700 750
WAVELENGTH (nm)
SPECTRUM Sovbaani RECORD 11JS
S02 CONC (5-7IQ ng/m3__i£__h
FOLIAR INJURY: TOTAL O %
CHLOROSIS %
NECROSIS %
Y-AXIS COUNTS AT 767nm o??> t^H
450 500 550 600 650 700 750
WAVELENGTH (nm)
SPECTRUM SnvbMns RECORD ft/h
SOj CONC '^72-0 Mo/m3 .so h
FOLIAR INJURY: TOTAL O %
CHLOROSIS %
NECROSIS %
Y-AXIS COUNTS AT 767nm ,
Figure B-l. Individual Spectral Curves - Soybeans
-------
-150-
450 500 550 600 650 700 750
WAVELENGTH (nm)
RECORD_i!t/2_
SPECTRUM Soybeans
S02 CONC _
FOLIAR INJURY:
.60 h
450 500 550 6OO 650 700
WAVELENGTH (nm)
SPECTRUM
CONC
750
TOTAL.
FOLIAR INJURY :
CHLOROSIS
NECROSIS & %
Y-AXIS COUNTS AT 767nrn °?C.O?87
ug/m3
TOTAL
CHLOROSIS.
NECROSIS_
-70
_%
Y-AXIS COUNTS AT 767 nm
450 500 550 600 650 700
WAVELENGTH (nm)
SPECTRUM Soybeans RECORD.
S02 CONC >57ZD MQ/m3
FOLIAR INJURY: TOTAL _
750
450 500 550 600 650 700 750
WAVELENGTH (nm)
SP€CTRUM Soybeans
.SP
FOLIAR INJURY:
TOTAL
CHLOROSIS.
NECROSIS _
Y-AXIS COUNTS AT 767nm.
%
NECROSIS
Y-AXIS COUNTS AT 767nm
450 500 550 600 650
WAVELENGTH (nm)
700 750
SPECTRUM Soybeans
S02 CONC _
FOLIAR INJURY:
RECORD_JMI
Mg/m3 •_$
TOTAL Q_
CHLOROSIS
NECROSIS
455 5&5 550 600 650 700 750
WAVELENGTH (nm)
SPECTRUM &r*y f>s»,l RECORD IH/ll
S02 CONC Mg/m3
FOLIAR INJURY: TOTAL <
Y-AXIS COUNTS AT 767nm
CHLOROSIS
NECROSIS
Y-AXIS COUNTS AT 767nm
_%
_%
-------
-151-
450 500 550 600 650 700
WAVELENGTH (nm)
SPECTRUM Sovbeons RECORD-Ml
S02 CONG )6?zO
750
FOLIAR INJURY:
TOTAL.
30
CHLOROSIS 3Q%
NECROSIS __i£_%
Y-AXIS COUNTS AT 767nm_i7 3.0 ng/m3.
. f,^ h
FOLIAR IN JURY:
TOTAL 6 a %
CHLOROSIS_«£_%
NECROSIS,
Y-AXIS COUNTS AT 767nm
450 500 550 600 650 700
WAVELENGTH (nm)
750
SPECTRUM Soybeans
S02 CONC__L£I££.
RECORD
.60
FOLIAR INJURY:
TOTAL
CHLOROSIS
NECROSIS
Y-AXIS COUNTS AT 767nrru
SPECTRUM Sflyjjjjans-
S02 CONC ISi20
FOLIAR INJURY:
550 600 650 700 750~
WAVELENGTH (nm)
RECORD M/li-
Ha/m3 . 5 o
TOTAL £_
./o
%
CHLOROSIS 6
NECROSIS %
Y-AXIS COUNTS AT 767nm a^O 811
-------
-152-
450 500 550 600 650 700 750
WAVELENGTH(nm)
SPECTRUM .6*jf./W RFfnan 14/30
S02 CONC
FOLIAR INJURY:
TOTAL.
CHLOROSIS.
NECROSIS _
Y-AXIS COUNTS AT 767nm
_ Vo
_%
_%
.so
450 500 550 600 650 700 750
WAVELENGTH (nm)
SPECTRUM Soybeans
S02 CONC
FOLIAR INJURY:
Y-AXIS COUNTS AT
RECORD.
Mg/m3
TOTAL __i£_%
CHLOROSIS_i£_%
NECROSIS __i£.%
450 500 550 600 650
WAVELENGTH (nm)
700 750
SPECTRUM
S02 CONC 1
FOLIAR INJURY:
RECORDJi/ii.
ng/m3
TOTAL
CHLOROSIS
NECROSIS
-to
.%
.%
450 500 550 600 650 700 750
WAVELENGTH (nm)
SPECTRUM Sovbeons RFmRn M/21
902 CONC /S7M> ug/m3 .50
FOLIAR IN JURY:
10
TOTAL.
CHLOROSIS_i£_%
NECROSIS %
Y-AXIS COUNTS AT 767nm
450 500 550 600 650 700 750
WAVELENGTH (nm)
S02 CONC
CHLOROSIS_£fL%
NECROSIS__j£_%
Y-AXIS COUNTS AT 767nm
A Soybeans
1 57^-0
IURY:
RECORD iy/1.\
Mg/m3 . 5o
TOTAL
FOLIAR INJURY:
RECORD ±U
ng/m3_
TOTAL __If_
.50
Y-AXIS COUNTS AT 767nm
CHLOROSIS.
NECROSIS _
Y-AXIS COUNTS AT 767nm.
10 %
-------
-153-
450 500" 550 600 650
WAVELENGTH(nm)
SPECTRUM Soybeans
S02 CONC _
FOLIAR INJURY:
700 750
.So
TOTAL.
CHLOROSIS.
NECROSIS _
Y-AXIS COUNTS AT 767nm_22jJ2il_
SPECTRUM.
S02 CONC _
RECORD
FOLIAR INJURY:
TOTAL.
CHLOROSIS.
NECROSIS _
Y-AXIS COUNTS AT 767nm
450 5OO 55O 6OO 650
WAVELENGTH (nm)
SPECTRUM Soybeans
S02 CONC
FOLIAR INJURY:
RECORD yy
ng /m ^ ,
TOTAL
CHLOROSIS
NECROSIS
Y-AXIS COUNTS AT 767nm
450 500 550 600 650 700 750
WAVELENGTH (nm)
_To
_%
700 750
450 5OO 550 600 650 700 750
WAVELENGTH (nm)
SPECTRUM Soybeans
g/m3_
•50
FOLIAR INJURY:
10
TOTAL
CHLOROSIS ,o %
NECROSIS %
Y-AXIS COUNTS AT 767nm
450 500 550 600 650 700 750
WAVELENGTH (nm)
SPECTRUM Soybeans
S02 CONC__J1Z£2-
RECORD
fa?
FOLIAR INJURY:
TOTAL
CHLOROSIS.
NECROSIS_
Y-AXIS COUNTS AT 767nm_
600 650 700
WAVELENGTH (nm)
750
SPECTRUM Soybeans
S02 CONC ;672-Q
FOLIAR INJURY:
RECORD M/A!
HQ/m3 -<«? h
TOTAL 5 %
CHLOROSIS 5 %
NECROSIS %
Y-AXIS COUNTS AT 767nm.
-------
-154-
450 500 550 600 650 700 750
WAVELENGTH(mm)
SPECTRUM Gteu. panel RF(DRD 1^/32.
SOg CONC Hg/fa3 h
FOLIAR INJURY: TOTAL %
CHLOROSIS %
NECROSIS %
Y-AXIS COUNTS AT 767nm_
450 500 550 600 650 700 750
WAVELENGTH (nm)
SPECTRUM Soybeans
S02 CONC /O'-fBo ng/m3_
FOLIAR INJURY: TOTAL .
.75
CHLOROSIS_i£_%
NECROSIS %
Y-AXIS COUNTS AT 767nm .
450 500 550 600 650 700 750
WAVELENGTH (nm)
SPECTRUM GvyfUutl RECORD 11/l(r
S02 CONC Mg/m3 h
FOLIAR INJURY: TOTAL %
CHLOROSIS %
NECROSIS %
Y-AXIS COUNTS AT 767nm
450 500 550 600 650 700
WAVELENGTH (nm)
SPECTRUM Soybeans
SOg CONC 10* 8
-------
-155-
450 500 550 60O 650 700 750
WAVELENGTH(nm)
RECORD Ji/iL__
HgAn3 ,7*5 h
SPECTRUM Soybeans
SOo CONC lb
450 500 550 600 650
WAVELENGTH (nm)
700 750
SPECTRUM Soybeans RECORD]
SO? CONC ni 1,0 Mg/m3
FOLIAR INJURY: TOTAL _
,7-j
CHLOROSIS
NECROSIS
Y-AXIS COUNTS AT 767nm °)\jtOlfl
450 500 550 600 650 700
WAVELENGTH (nm)
750
SPECTRUM.
S02 CONC L£
FOLIAR INJURY:
Soybeans
TOTAL *° %
CHLOROS1S_1£_%
NECROSIS %
Y-AXIS COUNTS AT 767nm
450 500 550 600 650 700
WAVELENGTH (nm)
750
SPECTRUM Sovbaans
S02 CONC it)'1^0
RECORD
FOLIAR INJURY:
TOTAL
CHLOROSIS.
NECROSIS_
Y-AXIS COUNTS AT 767nm_
45O
550 600 650 700
WAVELENGTH (nm)
750
SPECTRUM Soybeans RECORD.
S02 CONC l°'t%o Mg/m3
FOLIAR INJURY: TOTAL _
IS %
CHLOROSIS ^ %
NECROSIS %
Y-AXIS COUNTS AT 767nm .
-------
-156-
4SO 500 550 600 650 700 750
WAVELENGTH (nm)
panel. RECORDlitte_
Hg/fo3 _ h
SPECTRUM
S02 CONG
FOLIAR INJURY:
TOTAL
450 500 550 600 650
WAVELENGTH (nm)
SPECTRUM Soybaons
SOgCONC /o^gg
FOLIAR INJURY:
TOO 750
.75"
CHLOROSIS.
NECROSIS _
Y-AXIS COUNTS AT 767nm
TOTAL 40 %
CHLOROSIS 30 %
NECROSIS lO %
Y-AXIS COUNTS AT 767nm
450 500 550 60O 650 700 750
WAVELENGTH (nm)
SPECTRUM Soybaons RECORD.
SOg CONC.
450 500 550 600 650 700 750
WAVELENGTH (nm)
SPECTRUM Sovbaona
S02 CONC )t>4&>
RECORD
.75" h
FOLIAR INJURY:
TOTAL
FOLIAR INJURY:
CHLOROSIS
N
Y-AXIS COUNTS AT 767nm
NECROSIS 5 %
TOTAL i
CHLOROSIS_
NECROSIS
Y-AXIS COUNTS AT 767nm_221!d=L_
li%
450 500 550 600 650
WAVELENGTH (nm)
700 750
550 600 650 700~
WAVELENGTH (nm)
750
SPECTRUM
S02 CONC
FOLIAR INJURY:
RECORD m/tt
ng/m3
TOTAL
CHLOROSIS
NECROSIS
SPECTRUM Soybaans
S02 CONC
FOLIAR INJURY:
RECORD H/
-------
-157-
450 500 550 600 650 700 750
WAVELENGTH(nm)
SPECTRUM Soybeans RECORD ^/$" , ,
SOg CONG iQ^ftO Hq/m3 i7g" h
FOLIAR INJURY: TOTAL Oo %
CHLOROSIS _2£L%
NECROSE %
Y-AXIS COUNTS AT
450 500 550 600 650 700
WAVELENGTH (nm)
SPECTRUM &ffy Panel RECORD
CONG _ ng/m3
750
FOLIAR INJURY:
TOTAL.
CHLOROSIS %
NECROSIS %
Y-AXIS COUNTS AT 767nm.
450 500 550 600 650
WAVELENGTH (nm)
700 750
SPECTRUM Soybeans
S02 CONC
FOLIAR INJURY:
RECORDS44
/O
TOTAL.
CHLOROSIS_L£_%
NECROSIS %
Y-AXIS COUNTS AT 767nm
450 500 550 600 650 700 750
WAVELENGTH (nm)
SPECTRUM Soybeans RECORCU£&— __
CONC
g/m3
• TT h
FOLIAR INJURY:
Y-AXIS COUNTS AT
70
TOTAL.
CHLOROSIS_70_%
NECROSIS %
450 500 550 600 650 700 750
WAVELENGTH (nm)
SPECTRUM Soybeans RECORD lt)&3
S02 CONC lOVftO Mg/m3__j2£___h
FOLIAR INJURY: TOTAL frO %
CHLOROSIS^fcO.%
NECROSIS %
Y-AXIS COUNTS AT 767nm_
600 650 700
WAVELENGTH (nm)
750"
SPECTRUM Soybeans
SOoCONC /ovftO
FOLIAR INJURY:
RECORD W/5S
Mg/m3 . 7-T h
TOTAL 30 %
CHLOROSIS 3o %
NECROSIS.
Y-AXIS COUNTS AT 767nm.
-------
-158-
450 500 550 600 650 700 750
WAVELENGTH (nm)
RECORDJiL^j£_
SPECTRUM
S02CONC
FOLIAR INJURY:
TOTAL
CHLOROSIS
NECROSIS
Y-AXIS COUNTS AT 767nm
%
%
450 500 550 600 650 700
WAVELENGTH (nm)
SPECTRUM Soybeans
S02 CONG
FOLIAR INJURY:
750
RECORD.
Mg/m3 21
TOTAL k£
CHLOROSIS.
NECROSIS_
Y-AXIS COUNTS AT 767nm
SPECTRUM
S02 CONC
FOLIAR INJURY;
Y-AXIS COUNTS AT
450 500 550 600 650 700 750
WAVELENGTH (nm)
RECORD.
Mg/m3
TOTAL
CHLOROSIS.
NECROSIS _
_/o
.%
450 500 ,550 600 650
WAVELENGTH (nm)
700 750
SPECTRUM Soybeans
CONC lotto
RFf.npn
Mg/m3___17£
FOLIAR INJURY:
TOTAL __Zi_%
2£_%
CHLOROSIS
NECROSIS
Y-AXIS COUNTS AT 767nm
450
500 550 600 650 700
WAVELENGTH (nm)
750
SPECTRUM Sovbaans
so2 CONC
RECORD
. 7 ?
FOLIAR INJURY:
TOTAL
CHLOROSIS.
NECROSIS_
%
Y-AXIS COUNTS AT 767nm
550 600 650 700
WAVELENGTH (nm)
750
SPECTRUM Soybeans
CONC_
FOLIAR INJURY:
RECORD
Mg/m3
TOTAI [0.
7"T
CHLOROSIS.
NECROSIS _
Y-AXIS COUNTS AT 767nm ,
.%
-------
-159-
450 500 550 600 650 700 750
WAVELENGTH(nm)
SPECTRUM Soybeans
S02 CONC in 2j
^
S02 CONC
FOLIAR INJURY:
RECORDjWL
TOTAL.
CHLOROSIS.
NECROSIS _
Y-AXIS COUNTS AT 767nm
450 500 550 600 650
WAVELENGTH (nm)
700 750
SPECTR UM Soybeans
S02 CONC .«H20
FOLIAR INJURY:
RECORD
. 75
TOTAL 10 %
CHLOROSIS_ZL%
450 500
550 600 650
WAVELENGTH (nm)
700 750
SPECTRUM Soybeans
902 CONC ' Q4 80
FOLIAR IN JURY:
TOTAL ___££_%
CHLOROSIS_fce_%
NECROSIS %
Y-AXIS COUNTS AT 767nm
450 500 550 600 650 700
WAVELENGTH (nm)
SPECTRUM Soybeans RECORD^!
502'
750
FOLIAR INJURY:
TOTAL
CHLOROSIS_i£L%
NECROSIS
Y-AXIS COUNTS AT 767nny37053S
550 600 650 700
WAVELENGTH (nm)
750
SPECTRUM Soybeans
S02 CONC _
FOLIAR INJURY:
RECORD l^jdr?
Hg/m3 , 7 r
TOTAL ^(7
CHLOROSIS_J&_%
Y-AXIS COUNTS AT
Y-AXIS COUNTS AT 767nm
-------
-160-
450 500 550 600 650 700
WAVELENGTH (nm)
SPECTRUM^/y^^W RECORD It) to
SOo CONG
750
FOLIAR INJURY:
MgAi3
TOTAL
CHLOROSIS
NECROSIS
Y-AXIS COUNTS AT 767nm_Wo£H
450 500 550 600 650 700 750
WAVELENGTH (nm)
SPECTRUM Soybeans RECORD .l*f/-?ft
S02 CONC J»+3o Mg/m^ ?"> h
FOLIAR INJURY:
TOTAL.
CHLOROSIS
NECROSIS
Y-AXIS COUNTS AT 767nm
%
450 500 550 600 650 700 750
WAVELENGTH (nm)
SPECTRUM GfJUfltonfl RECORD
SOz CONC
FOLIAR INJURY:
ng/m3
TOTAL _
CHLOROSIS.
NECROSIS _
. vo
_%
Y-AXIS COUNTS AT 767nrn
450 500 550 600 650 700 750
WAVELENGTH (nm)
SPECTRUM Soybeans
902CONC '^90 MO/m3_ • 7?
FOLIAR INJURY:
TOTAL
CHLOROSIS
NECROSIS
Y-AXIS COUNTS AT 767nm3ft?.0ll
450 500 550 600 650 700
WAVELENGTH (nm)
SPCCTRUM Soybeans RECORD,
S02 CONC_
FOLIAR INJURY: TOTAL
750
CHLOROSIS.
NECROSIS_
Y-AXIS COUNTS AT 767nm AlA i
%
450 500 550 600 650 TOO"
WAVELENGTH (nm)
750
SPECTRUM Soybeans
S02 CONC
FOLIAR INJURY:
RECORD
TOTAL 10 °.
CHLOROSIS 30 °
Y-AXIS COUNTS AT 767nm
-------
-161-
450 500 550 600 650 700 750
WAVELENGTH(nm)
SPECTRUM Soybeons RECORD It^
SOo CONC /QfSO Hg/n3 , T^ h
FOLIAR INJURY: TOTAL -5 %
CHLOROSIS _JL_%
NE CROSS %
Y-AXIS COUNTS AT 767nm_ '
450 500 550 600 650 700
WAVELENGTH (nm)
750
SPECTRUM
S02 CONC
FOLIAR INJURY:
RECORD Jfhto
Hg/m3
TOTAL
CHLOROSIS.
NECROSIS_
Y-AXIS COUNTS AT 767nm
450 500 550 600 650
WAVELENGTH {nm)
700 750
SPECTRUM j,
S02 CONC
FOLIAR INJURY:
RECORD
rig/m'
TOTAL
CHLOROSIS,
NECROSIS _
Y-AXIS COUNTS AT 767nm
450 500 550 600 650 700 750
WAVELENGTH (nm)
SPECTRUM Soybeans
902CONC
FOLIAR INJURY:
RECORD_aj2L_
nq/rr? . ? T
TOTAL so %
CHLOROSIS_l£_%
NECROSIS %
Y-AXIS COUNTS AT 767nm
450 500 550 600 650 700 750
WAVELENGTH (nm)
SPECTRUM.
S02 CONC_
RECORD.
FOLIAR INJURY:
TOTAL
CHLOROSIS %
NECROSIS %
Y-AXIS COUNTS AT 767nm
450 5&5 550 600 650 700 750~
WAVELENGTH (nm)
SPECTRUM ^-s-
S02 CONC
FOLIAR INJURY:
RECORD.
TOTAL
CHLOROSIS.
NECROSIS _
Y-AXIS COUNTS AT 767nm
-------
-162-
450 500 550 600 650 700
WAVELENGTH (nm)
SPECTRUM Soybeans RECORDI5/J
SOo CONC
750
FOLIAR INJURY:
H0/hl3 .75
TOTAL
CHLOROSIS
NECROSIS
Y-AXIS COUNTS AT 767nm.
450 500 550 600 650 700 750
WAVELENGTH (nm)
SPECTRUM Soybeans RECORD 15/3
SOg CONC /o*/8o MO/m3 . ?<" h
FOLIAR INJURY:
TOTAL
CHLOROSIS_2L_%
NECROSIS _L_%
Y-AXIS COUNTS AT 767nm 3
450 500 550 600 650 700 750
WAVELENGTH (nm)
SPECTRUM Soybeans
S02 CONC
FOLIAR INJURY:
RECORD 6/6
Mq/m3 ,7<" h
TOTAL go %
CHLOROSIS gO %
NECROSIS %
Y-AXIS COUNTS AT 767nm
450 500 550 600 650 700 750
WAVELENGTH (nm)
SPECTRUM Soybeans RECORQ_j5/2__
SOg CONC /oygO Mq/m3 .if
FOLIAR IN JURY:
TOTAL 3S- %
CHLOROSIS_2£_%
Y-AXIS COUNTS AT 767nm
450 500 550 600 650 700
WAVELENGTH (nm)
SPECTRUM
S02 CONC
FOLIAR INJURY:
750
)
550 600 650 700
WAVELENGTH (nm)
750
SPECTRUM Soybeans
S02 CONC
FOLIAR INJURY:
RECORD lS/1*
TOTAL 80 %
CHLOROSIS_&L%
_%
NECROSIS.
Y-AXIS COUNTS AT 767nm.
-------
-163-
450 500 550 600 650 700
WAVELENGTH (nm)
750
SPECTRUM Soybeans
S02 CONC /o^fto
FOLIAR INJURY:
RECORD
TOTAL 10 %
CHLOROSIS 40 %
NECROSIS %
Y-AXIS COUNTS AT 767nm ato&H
450 500 550 600 650 700 750
WAVELENGTH (nm)
SPECTRUM Soybeans
S02 CONC /o4go
FOLIAR INJURY:
RECORD 1 5 | 'I
TOTAL .
CHLOROSIS /O %
NECROSIS %
Y-AXIS COUNTS AT 767nm.
450 500 550 600 650 700 750
WAVELENGTH (nm)
SPECTRUM Soybeans
S02 CONC
FOLIAR INJURY:
Y-AXIS COUNTS AT
RECORD
TOTAL /0 %
CHLOROSIS_iO_%
NECROSIS %
450 500 550 600 650 700
WAVELENGTH (nm)
•/ RECORD
750
SPECTRUM (Sfy*
CONC.
FOLIAR INJURY:
TOTAL
CHLOROSIS.
NECROSIS_
Y-AXIS COUNTS AT 767nm
450 500 550 600 650 700
WAVELENGTH {nm)
750
SPECTRUM Soybeans
SQ2 CONC /
-------
-164-
450 500 550 600 650 700
WAVELENGTH(nm)
SPECTRUM Sovbeons
S02 CONC 10*130
FOLIAR INJURY:
750
RECORD.
Hg/m3 .-lnd RECORD 4^
S02 CONC Hg/m3
FOLIAR INJURY: TOTAL
CHLOROSIS.
NECROSIS _
Y-AXIS COUNTS AT 767nm
750
_vo
.%
_%
-------
-165-
450 500 550 600 650 700 750
WAVELENGTH (nm)
SPECTRUM Soybeans RFCORn
S02 CONG l-biXQ Hg/fci3
FOLIAR INJURY: TOTAL
.(tl
CHLOROSIS _£_%
NECROSIS _ %
Y-AXIS COUNTS AT 767nm
450 500 550 600 650 700
WAVELENGTH (nm)
750
SPECTRUM
S02 CONG
FOLIAR INJURY:
RECORD
H a/m3
TOTAL
CHLOROSISjL
NECROSIS
Y-AXIS COUNTS AT 767nm
450 500 550 600 650
WAVELENGTH (nm)
700 750
SPECTRUM Soybeans
S02 CONC
FOLIAR INJURY:
RECORD
TOTAL
CHLOROSIS.
NECROSIS _
10
Y-AXIS COUNTS AT 767nm
450 5OO 550 600 650 TOO 750
WAVELENGTH (nm)
SPECTRUM Soybeans RFCORn
902CONC__iiZl£.
.fa?
FOLIAR IN JURY:
TOTAL.
CHLOROSIS_S _ %
NECROSIS
%
Y-AXIS COUNTS AT 767nm
450 500 550 600 650 700
WAVELENGTH (nm)
750
SPECTRUM
S02 CONC
RECORD
FOLIAR INJURY:
TOTAL
CHLOROSIS.
NECROSIS_
Y-AXIS COUNTS AT 767nm (tAfT\\
_ vo
.%
500 550 600 650 700
WAVELENGTH (nm)
SPECTRUM Soybeans RECORD
750
SC^CONC C->72.-c
FOLIAR INJURY;
1 Mg/m3 • t>7
TOTAL * 5
CHLOROSIS_£_
' h
%
_%
NECROSIS.
Y-AXIS COUNTS AT 767nm.
-------
-166-
450 500 550 600 650 700
WAVELENGTH(nm)
SPECTRUM Sovbeons RECORD
S02 CONC _
750
450 500 550 600 650 700 750
WAVELENGTH (nm)
SPECTRUM .fto
S02 CONC I'lTi-O Hg/m3 , i/7 h
FOLIAR INJURY: TOTAL __££_%
CHLOROSIS io %
NECROSIS %
Y-AXIS COUNTS AT 767nm_£2H-Lk!a
FOLIAR INJURY:
500 550 600 650 700 750
WAVELENGTH (nm)
RECORO_l£Jil__
na/m3 ,(*? _1
TOTAL /6 %
450
SPECTRUM Sovbaans
S02 CONC
450 500 550 600 650 TOO 750
WAVELENGTH (nm)
RECORD,
Mg/m3
TOTAL
SPECTRUM
S02 CONC
FOLIAR INJURY:
CHLOROSIS JT %
NECROSIS %
Y-AXIS COUNTS AT 767nm
CHLOROSIS.
NECROSIS _
Y-AXIS COUNTS AT 767nm
-------
-167-
450 500 550 600 650 700 750
WAVELENGTH(nm)
SPECTRUM Soybeans RECORDAhi—
FOLIAR INJURY: TOTAL to %
CHLOROSIS _j£_%
NECROSIS %
Y-AXIS COUNTS AT 767nm 3^\f>(fl
450 500 550 600 650 700 750
WAVELENGTH (nm)
SPECTRUM Soybeans
SOg CONC
FOLIAR INJURY:
Y-AXIS COUNTS AT
RECORD
.L>7 h
TOTAL.
If)
CHLOROSIS_/£_%
NECROSIS
450 500 550 600 650 700
WAVELENGTH (nm)
750
SPECTRUM Soybeans
S02 CONC
FOLIAR INJURY:
RECORDJi/iL
10
TOTAL.
CHLOROSIS_i{L%
NECROSIS %
Y-AXIS COUNTS AT 767nm
450 500 550 600 650
WAVELENGTH (nm)
SPECTRUM Soybeans
S02CONC___/
FOLIAR IN JURY:
700 750
Y-AXIS COUNTS AT 767nm
450 500 550 600 650 700
WAVELENGTH (nm)
750
SPECTRUM.
S02 CONC_
RECORD 1S
FOLIAR INJURY:
TOTAL
CHLOROSIS
NECROSIS
Y-AXIS COUNTS AT 767nm Mel ^
>50 SOO 550 600 650 700
WAVELENGTH (nm)
750
SPECTRUM Soybeans
S02 CONC
FOLIAR INJURY:
RECORD
.to?
J T
TOTAL
CHLOROSIS 2
-------
-168-
450 500 550 600 650 700
WAVELENGTH(nm)
750
SPECTRUM Soybeans
S02 CONC 75H-Z.O
FOLIAR INJURY:
RFCQBD
TOTAL 3 5 %
CHLOROSIS _iS_%
NECROSIS %
Y-AXIS COUNTS AT 767nm3,L>?>iai
450 500 550 600 650 700
WAVELENGTH (nm)
750
SPECTRUM Soybeans
S02 CONC /5T *-°
FOLIAR INJURY:
RECORD j5jbti
MO/m3 • (f"7
TOTAL
CHLOROSIS,
NECROSIS _
Y-AXIS COUNTS AT 767nm
450 500 550 600 650
WAVELENGTH (nm)
700 750
SPECTRUM Soybeans
S02 CONC
FOLIAR INJURY:
RECORD ft [Hi,
rig/m3__Jf7_
TOTAL /P_
CHLOROSIS '0 %
NECROSIS _ %
Y-AXIS COUNTS AT 767nm
450 500 550 600 650 700 750
WAVELENGTH (nm)
SPECTRUM igfoy p*nfi RFCORn is/go
SOg CONC ng/m3 I
FOLIAR INJURY:
TOTAL
CHLOROSIS.
NECROSIS_
%
%
Y-AXIS COUNTS AT 767nm
450 500 550 600 650 700
WAVELENGTH (nm)
750
SPECTRUM Soybeans
S02CQNC
RECORD
FOLIAR INJURY:
TOTAL lo %
CHLOROSIS 10 %
NECROSIS %
Y-AXIS COUNTS AT 767nm^?3673
450 500 550 600 650 700 750
WAVELENGTH (nm)
SPECTRUM ^
CONC
FOLIAR INJURY:
RECORD \*i\
Mg/m3
TOTAL
CHLOROSIS
NECROSIS _
Y-AXIS COUNTS AT 767 nm
_/o
_%
_%
-------
-169-
450 500 550 600 650 700 750
WAVELENGTH(nm)
SPECTRUM Sovbeons
S02 CONC i SIZ.Q
FOLIAR INJURY: TOTAL 5 %
CHLOROSIS _!_%
NECROSIS %
Y-AXIS COUNTS AT 767nm_
450 500 550 600 650 700 750
WAVELENGTH (nm)
SPECTRUM Soybeans
S02 CONC _
FOLIAR INJURY:
RECORD lf>Jn
Ha/m3 • fa7
TOTAL 10 %
CHLOROSIS /0 %
NECROSIS %
Y-AXIS COUNTS AT 767nm
_l_
_L
450 500 550 600 650
WAVELENGTH (nm)
700 750
SPECTRUM Soybeans
S02 CONC _
FOLIAR INJURY:
RECORD
TOTAL /& %
CHLOROSIS /* %
Y-AXIS COUNTS AT 767nm
JL
JL
J_
450 500 550 600 650 700 750
WAVELENGTH (nm)
SPECTRUM Soybeans
SOgCONC /S-7ZQ
FOLIAR IN JURY:
• d£*r/ RECORD tij
S02 CONC Hg/m3
FOLIAR INJURY: TOTAL
CHLOROSIS.
NECROSIS_
Y-AXIS COUNTS AT 767nm
450 500 550 600 650 700
WAVELENGTH (nm)
75O
SPECTRUM Soybeans
CONC__LV7_2C
FOLIAR INJURY:
RECORD \-S[^0
Hq/m3 . la 7
TOTAL.
CHLOROSIS
NECROSIS
Y-AXIS COUNTS AT 767nm
-------
-170-
450 500 550 600 650 700 750
WAVELENGTH (nm)
SPECTRUM Soybeans RECORDJ^U _
SOoCONC
•(el
FOLIAR INJURY:
TOTAL 10 %
CHLOROSIS 10 %
NECROSIS
Y-AXIS COUNTS AT 767nm.
450 500 550 600 650 700 750
WAVELENGTH (nm)
SPECTRUM Soybeans
S02 CONC lo
ng/m3 _ • 7
TOTAL
A %
CHLOROSIS_5_%
NECROSIS %
Y-AXIS COUNTS AT 767nm
450 500 550 600 650 700 750
WAVELENGTH (nm)
SPECTRUM Soybeans
S02 CONC io«jflO
FOLIAR INJURY:
RECORD
. 7 f h
TOTAL _
CHLOROSIS
NECROSIS
10 %
'o
.%
450 5OO 550 600 650
WAVELENGTH (nm)
700 750
SPECTRUM
SO2 CONC
FOLIAR IN JURY:
Hg/m3
TOTAL
CHLOROSIS
NECROSIS
Y-AXIS COUNTS AT 767nm
450 500 550 600 650 700
WAVELENGTH (nm)
SPECTRUM Soybeans
S02 CONC.
750
FOLIAR INJURY:
RECORD.
Hg/m3 . ? f
TOTAL.
A %
CHLOROSIS __i_%
Y-AXIS COUNTS AT 767nm
450
550 600 650 700
WAVELENGTH (nm)
750
Y-AXIS COUNTS AT 767nm
SPECTRUM tptyup*
SOo CONC
FOLIAR INJURY:
Y-AXIS COUNTS AT
WfJ. RECORD _ii|:
ng/m3
TOTAL
CHLOROSIS
NECROSIS
767nm OfeOfO
~*
-------
-171-
450 500 550 600 650 700 750
WAVELENGTH(nm)
SPECTRUM Soybeans RFmRni5/.<,fl
SOo CONC /(W80 HQ/m3 .7T h
FOLIAR INJURY: TOTAL 5Q %
CHLOROSIS So %
NECROSIS %
Y-AXIS COUNTS AT 767nm .3^87ft
450 500 550 600 650 700 750
WAVELENGTH (nm)
SPECTRUM Griif &nel RECORD i6|i,n
S02 CONC _ Hg/m3 _ h
FOLIAR INJURY: TOTAL _ %
CHLOROSIS
NECROSIS
Y-AXIS COUNTS AT 767nm
%
450
500 550 600 650
WAVELENGTH (nm)
700 750
SPECTRUM Soybeans
S02 CONC
FOLIAR INJURY:
RECORD fe^m>
na/m3 , 7f
TOTAL AQ
CHLOROSIS 50
NECROSIS
Y-AXIS COUNTS AT 767nm
450 500 550 600 650 700 750
WAVELENGTH (nm)
SPECTRUM Soybeans
SO2CONC
FOLIAR INJURY:
RECORD
TOTAL
CHLOROSIS,
NECROSIS_
»o
Y-AXIS COUNTS AT 767nm
450 500 550 600 650 700 750
WAVELENGTH (nm)
SPECTRUM Soybeans
CONC /05|(p
|(p|
.75"
FOLIAR INJURY:
TOTAL
CHLOROSIS
NECROSIS.
Y-AXIS COUNTS AT 767nmJVT^MI
550 600 650 700 750
WAVELENGTH (nm)
SPECTRUM Soybeans
S02 CONC
FOLIAR INJURY:
RECORD 4i)j£3_
TOTAL
CHLOROSIS.
NECROSIS _
Y-AXIS COUNTS AT 767nm
%
-------
-172-
450 500 550 600 650 700 750
WAVELENGTH (nm)
/ RECORD
SOZCONC
FOLIAR INJURY:
TOTAL
NECROSIS.
Y-AXIS COUNTS AT 767nm
450 5OO 550 600 650 700 750
WAVELENGTH (nm)
SPECTRUM SoybBons
CONG
RECORD
FOLIAR INJURY:
TOTAL 10 %
CHLOROSIS_!L_%
NECROSIS %
Y-AXIS COUNTS AT 767nm.
450 500 550 600 650 700
WAVELENGTH (nm)
750
SPECTRUM
S02 CONC
FOLIAR INJURY:
RECORD
iig/m3
TOTAL
NECROSIS _
Y-AXIS COUNTS AT 767nm_0%!k.
450 500 550 600 650 700 750
WAVELENGTH (nm)
SPECTRUM Soybeans
S02CONC_
FOLIAR INJURY:
ng/m3_
TOTAL.
CHLOROSIS.
NECROSIS_
Y-AXIS COUNTS AT 767nm \\tO
450 500 550 600 650 700 750
WAVELENGTH (nm)
SPECTRUM Soybeans
S02 CONC /o«/90
RECORD \t»|vt1
FOLIAR INJURY:
450
550 600 650 700
WAVELENGTH (nm)
750
SPECTRUM Soybeans
SC>2 CONC
FOLIAR INJURY:
RECORD _
Hg/m3 7b
TOTAL f 5
CHLOROSIS_l£_%
NECROSIS %
Y-AXIS COUNTS AT 767nm.
-------
-173-
450 500 550 600 650 700 750
WAVELENGTH (nm)
SPECTRUM Soybeans RFrDRn
SOo CONC /n^tio HtfArfl '7S~ h
FOLIAR INJURY: TOTAL 3 o %
CHLOROSIS _14_%
NECROSB %
Y-AXIS COUNTS AT 767nm
450 500 550 600 650 700 750
WAVELENGTH (nm)
SPECTRUM
S02 CONC
RECORD
FOLIAR INJURY:
TOTAL.
CHLOROSIS.
NECROSIS_
Y-AXIS COUNTS AT 767nm'Vbii3_
450
500 550 600 650
WAVELENGTH (nm)
700 750
SPECTRUM Soybeans
S02 CONC
FOLIAR INJURY:
RECORD
Mg/m3___z£__h
TOTAL __1£_%
CHLOROSIS ia%
NECROSIS %
Y-AXIS COUNTS AT 767nm
450 500 550 600 650 700 750
WAVELENGTH (nm)
SPECTRUM Soybeans RECORD_J^/2I_
SQgCQNC
Mg/m3
FOLIAR INJURY:
TOTAL
CHLOROSIS
NECROSIS
Y-AXIS COUNTS AT 767nm 33.09k?
450 500 550 600 650 700 750
WAVELENGTH (nm)
SPECTRUM Sovbaojis
S02 CONC
RECORD
FOLIAR INJURY:
TOTAL
CHLOROSIS tjO %
NECROSIS %
Y-AXIS COUNTS AT 767nm_
.' ' • •
550 600 650 700
WAVELENGTH (nm)
750
SPECTRUM Soybggns
CONC_/£li£.
FOLIAR INJURY:
RECORD JS]2£_
Mg/m3_iZf__
TOTAL f-o
CHLOROSIS_^£.%
NECROSIS %
Y-AXIS COUNTS AT 767nm.
-------
-174-
450 500 550 600 650 700 750
WAVELENGTH (nm)
REC»RDJilH__
SPECTRUM ay
S02 CONC
FOLIAR INJURY:
Y-AXIS COUNTS AT 767nm
RECORD Uo|H
ttg/m3
TOTAL
CHLOROSIS
NECROSIS
U^'b^
h
%
%
%
450 500 550 600 650 700
WAVELENGTH (nm)
RECORQjk
750
SPECTRUM Soybeans
902 CONC /otto
FOLIAR INJURY:
Y-AXIS COUNTS AT
Mq/m3
L_Li£_h
TOTAL 30 %
CHLOROSIS_t»_%
N£CROSIS_i0_%
450 500 550 600 650 700 750
WAVELENGTH (nm)
SPECTRUM Soybeans
S02 CONC.
RECORD
FOLIAR INJURY:
Y-AXIS COUNTS AT
TOTAL.
CHLOROSIS.
NECROSIS_
10
^> 1, t ' • '—
450 500 550 600 650 TOCT
WAVELENGTH (nm)
750
SPECTRUM Soybeans
S02 CONC _ /QV.QO
FOLIAR INJURY:
RECORD
M q/m3 ,'.
TOTAL 20
NECROSIS
Y-AXIS COUNTS AT 767nm
-------
-175-
450 500 550 600 650 700 750
WAVELENGTH (nm)
SPECTRUM Soybeans RECORD.. lip/1*
S02 CONC
FOLIAR INJURY:
Y-AXIS COUNTS AT
TOTAL __J2£_%
CHLOROSIS _££_%
NECROS6 %
450 500 550 600 650 700 750
WAVELENGTH (nm)
SPECTRUM 6r*ti fond RECORD L(3>
S02 CONC ng/m3 h
FOLIAR INJURY:
TOTAL,
CHLOROSIS.
NECROSIS _
Y-AXIS COUNTS AT 767nm
450 500 550 600 650 700
WAVELENGTH (nm)
750
SPECTR UM Soybeans
S02 CONC /S72.0
FOLIAR INJURY:
Y-AXIS COUNTS AT
RECORDj^lS.
TOTAL
10 %
CHLOROSIS
NECROSIS
'o
.%
450 500 550 600 650 700
WAVELENGTH (nm)
SPECTRUM Soybeans RECORD, JMJ1
750
902 CONC
ug/m3
FOLIAR INJURY:
TOTAL 30 %
CHLOROSIS_J£_%
NECROSIS %
Y-AXIS COUNTS AT 767nm
450 500 550 600 650 700 750
WAVELENGTH (nm)
SPECTRUM joybeans
SOoCONC I67Z.O
FOLIAR INJURY:
RECORDJ
ng/m3
TOTAL
US
/
. 5TO
IfO
h
%
CHLOROSIS
NECROSIS
Y-AXIS COUNTS AT 767nm 3
450
550 600 650 700
WAVELENGTH {nm)
750
SPECTRUM Soybeans
S02 CONC >5-l 2.0
FOLIAR INJURY:
Y-AXIS COUNTS AT
RECORD Jk
Hg/m3 t
TOTAL I
CHLOROSIS.
NECROSIS _
-------
-176-
450 500 550 600 650 700 750
WAVELENGTH (rim)
RECORD _11«/IL__
Mg/n3 _ h
S02 CONG
FOLIAR INJURY:
TOTAL
CHLOROSIS.
NECROSIS_
Y-AXIS COUNTS AT 767nm
_ Vo
_%
_%
450 500 550 600 650 700 750
WAVELENGTH (nm)
SPECTRUM Soybeans RECORD
S02 CONC /r?*-" Mg/m3__L5£_h
FOLIAR INJURY: TOTAL __!£__%
CHLOROSIS_££_%
NECROSIS %
Y-AXIS COUNTS AT 767nm,
450 500 550 600 650 700
WAVELENGTH (nm)
RECORD_ik
750
SPECTRUM r
S02 CONC r
FOLIAR INJURY:
TOTAL
CHLOROSIS.
NECROSIS _
.
_%
450
500 560 600 650 700
WAVELENGTH (nm)
SPECTRUM ^Soybeans RECORQjklUl.
750
Y-AXIS COUNTS AT 767nm
450 500 550 600 650 700
WAVELENGTH (nm)
750
SPECTRUM Soybeans
S02 CONC___LL2±£
FOLIAR INJURY:
RECORD /L/3
TOTAL
CHLOROSIS
NECROSIS
Y-AXIS COUNTS AT 767nm
450
550 600 650 700
WAVELENGTH (nm)
750
SPECTRUM £
S02 CONC L
FOLIAR INJURY:
RECORD JkJLi
TOTAL
5 o o/o
Y-AXIS COUNTS AT 767nm
CHLOROSIS go %
NECROSIS %
Y-AXIS COUNTS AT 767nm a^Q 00-&
-------
-177-
450 500 550 600 650 700
WAVELENGTH (nm)
750
SPECTRUM Sovbeons
S02 CONC I
FOLIAR INJURY:
RECORD
, 50 h
TOTAI lO %
CHLOROSIS _lO_%
_%
NECROSIS
Y-AXIS COUNTS AT 767nm
450 500 550 600 650 700 750
WAVELENGTH (nm)
SPECTRUM
S02 CONC
RECORD II
[3.°
FOLIAR INJURY:
TOTAL
%
NECROSIS.
Y-AXIS COUNTS AT 767nm (nil?)0!
450 500 550 600 650 700 750
WAVELENGTH (nm)
SPE CTR UM Soybeans
S02 CONC
FOLIAR INJURY:
RECORD IUI
. SO
TOTAL
NECROSIS
Y-AXIS COUNTS AT 767nm
450
500 550 600 650 700 750
WAVELENGTH (nm)
SPECTRUM Soybeans
802 CONC _
FOLIAR INJURY:
RFmpn
TOTAL
CHLOROSIS
NECROSIS
. ro
.%
.%
Y-AXIS COUNTS AT 767nm
450 500 550 600 650 700 750
WAVELENGTH (nm)
SPECTRUM Soybeans RECORD III 31
S02 CONC
FOLIAR INJURY:
TOTAL
CHLOROSIS
NECROSIS
Y-AXIS COUNTS AT 767nm
500 550 600 650 700 750
WAVELENGTH (nm)
SPECTRUM Soybeans
SOg CONC / 17 2.0
FOLIAR INJURY:
RECORD IV. )^
Mg/m3 .50 h
TOTAL 5 %
CHLOROSIS 5" o/n
NECROSIS %
Y-AXIS COUNTS AT 767nm
-------
-178-
450 500 550 600 650 700 750
WAVELENGTH(nm)
SPECTRUM gray/Wl RECORD-JJ^iL—
S02 CONC Hg^i3 h
FOLIAR INJURY: TOTAL
CHLOROSIS.
NECROSB_
Y-AXIS COUNTS AT 767nm_liSo6
. fo
.%
.%
450 500 550 600 650 700
WAVELENGTH (run)
750
SPECTRUM
S02 CONC
RECORD \\a\Z\t
;. z. z. f
FOLIAR INJURY:
TOTAL
CHLOROSIS.
NECROSIS _
Y-AXIS COUNTS AT 767nm
%
450 500 550 600 650
WAVELENGTH (nm)
700 750
SPECTRUM (Ztyt,
S0£ CONC
FOLIAR INJURY:
RECORD.
TOTAL
CHLOROSIS.
NECROSIS _
Y-AXIS COUNTS AT 767nm
450 500 550 600 650 700 750
WAVELENGTH (nm)
SPECTRUM Soybeans RFCORD ]vi|a5
SOgCONC lot to Mg/m3__
FOLIAR INJURY: TOTAL _
CHLOROSIS_/2_%
Y-AXIS COUNTS AT 767nm
450 500
550 600 650 700
WAVELENGTH (nm)
750
SPECTRUM Soybeans
S02 CONC
FOLIAR INJURY:
RECORD llalYI
nq/m3 I. iff h
TOTAL /Q %
Y-AXIS COUNTS AT 767nm
450
600 650 700
WAVELENGTH (nm)
750
SPECTRUM Soybeans
S02 CONC
FOLIAR INJURY:
RECORD.
Hg/nv^ /,,
TOTAL __Jii2_%
CHLOROSIS_^2_%
NECROSIS %
Y-AXIS COUNTS AT 767nm
-------
-179-
450 500 550 600 650 700 750
WAVELENGTH(nm)
SPECTRUM Soybeans
S02 CONC /ft^/fro
FOLIAR INJURY: TOTAL
10
NECROSIS.
Y-AXIS COUNTS AT 767nm
450 500 550 600 650 700 750
WAVELENGTH {nm)
SPECTRUM 6r^ P*H, I RECORD
S02 CONC , .- Mg/m^
FOLIAR INJURY:
TOTAL.
CHLOROSIS.
NECROSIS_
Y-AXIS COUNTS AT 767nm
_%
_%
450 500 550 600 650
WAVELENGTH (nm)
700 750
SPECTRUM Soybeans
S02 CONC
FOLIAR INJURY:
RECORD
0
TOTAL
CHLOROSIS IfO %
NECROSIS _ %
Y-AXIS COUNTS AT 767nm
450 500 550 600 650 700
WAVELENGTH (nm)
SPECTRUM Soybeans
SO2 CONC _
FOLIAR INJURY:
750 .
TOTAL * 0 %
CHLOROSIS_3«_%
NECROSIS %
Y-AXIS COUNTS AT 767nm
450 500 550 600 650 700 750
WAVELENGTH (nm)
SPECTRUM Soybeans RECORD lulM
S02 CONC_
FOLIAR INJURY: TOTAL.
2£_%
CHLOROSIS.
NECROSIS_
Y-AXIS COUNTS AT 767nm_r2ifiai
450
600 650 700
WAVELENGTH (nm)
750
SPECTRUM Soybeans
S0£ CONC
FOLIAR INJURY:
RECORD
Hg/m3
TOTAL
CHLOROSIS 5o %
NECROSIS %
Y-AXIS COUNTS AT 767nm
-------
-180-
450 500 550 600 650 700 750
WAVELENGTH(nm)
SPECTRUM £V^ _P*y,ei RECORDji<|3k__
S02 CONC Hg/^n3 h
FOLIAR INJURY: TOTAL %
CHLOROSIS %
NECROSIS %
Y-AXIS COUNTS AT 767nm blWb
450 500 550 600 650 700 750
WAVELENGTH (nm)
SPECTRUM Soybeans
SOg CONC /oygO
FOLIAR INJURY:
RECORD .Haft ft
Mq/m3
TOTAL
40 %
CHLOROSIS HO %
NECROSIS %
Y-AXIS COUNTS AT 767nm.
450 500 550 600 650
WAVELENGTH (nm)
700 750
SPECTRUM &TAU
S02 CONC
FOLIAR INJURY:
RECORDJL]
ng/m3
TOTAL
CHLOROSIS.
NECROSIS _
Y-AXIS COUNTS AT 767nm
450 500 550 600 650 700 750
WAVELENGTH (nm)
SPECTRUM Soybeans RECORQis4ll
S02CQNC
FOLIAR INJURY:
TOTAL 5 %
CHLOROSIS_JL_%
NECROSIS %
Y-AXIS COUNTS AT 767nm
450 500 550 600 650 700 750
WAVELENGTH (nm)
SPECTRUM Soybeans RECORD.
S02 CONC_
FOLIAR INJURY: TOTAL.
3 a %
CHLOROSIS.
NECROSIS_
J£_%
%
Y-AXIS COUNTS AT 767nm .
450 500 550 600 650 TOO"
WAVELENGTH (nm)
SPECTRUM SoybaanE
S02 CONC /o«KO
FOLIAR INJURY:
75O
RECORD.
Hg/rn3 /.tzf
TOTAI 5
CHL6ROSIS.
NECROSIS _
Y-AXIS COUNTS AT 767nm JH3Jil_
-------
-181-
450 500 550 600 650 700
WAVELENGTH(nm)
SPECTRUM Sovbaons RECORD.
SOo CONC lO^fiQ MgAi3_
FOLIAR INJURY; TOTAL_
750
CHLOROSIS J0_%
NECROSIS %
Y-AXIS COUNTS AT 767nm
450 500 550 600 650 700
WAVELENGTH (nm)
SPECTRUM^.*-."/
S02 CONC
FOLIAR INJURY:
RECORD.
Mg/m3_
TOTAL _
CHLOROSIS.
NECROSIS_
Y-AXIS COUNTS AT 767nm__T^£lL
450 500 550 600 650 700
WAVELENGTH (nm)
SPECTRUM Soybeans
S02 CONC lot80
FOLIAR INJURY:
RECORD.
rig/m'
TOTAL i
CHLOROSIS.
NECROSIS _
750
%
750
Y-AXIS COUNTS AT 767nm
450 500 550 600 650 700 750
WAVELENGTH (nm)
SPECTRUM Soybeans RFr.ORn I^HS
S02 CONC 'OVSO HQ/m3 /. ^ t
-------
-182-
450 500 550 600 650 700 750
WAVELENGTH(nm)
SPECTRUM &r
S02CONC
450 500 550 600 650 700 750
WAVELENGTH (nm)
RECOROJkiiS
450 500 550 600 650 700 750
WAVELENGTH (nm)
SPECTRUM Soybeans
S02 CONC
FOLIAR INJURY:
Y-AXIS COUNTS AT 767nm
450 500 550 600 650
WAVELENGTH (nm)
700 750
SPECTRUM _£liij_ p-sne (.
S02 CONC _
FOLIAR INJURY:
RECORD [
ng/m3
TOTAL _
CHLOROSIS.
NECROSIS _
_ /o
_%
SPECTRUM Soybeans
902 CONC _££±&£_
FOLIAR INJURY:
TOTAL 10 %
CHLOROSIS_22_%
NECROSIS %
Y-AXIS COUNTS AT 767nm
450 500 550 600 650 700 750
WAVELENGTH (nm)
SPECTRUM Sovbaans
SQoCONC
RECORD Itg
FOLIAR INJURY:
TOTAL
CHLOROSS_2£_%
NECROSIS %
Y-AXIS COUNTS AT 767nm_
450 500 550 600 650 700
WAVELENGTH (nm)
750
SPECTRUM Soybeans
S02 CONC.
FOLIAR INJURY:
RECORD.
Hg/m3
TOTAL_
Y-AXIS COUNTS AT 767nm
CHLOROSIS.
NECROSIS _
Y-AXIS COUNTS AT 767nm.
-------
-183-
450 500 550 600 650
WAVELENGTH (nm)
700 750
450 500 550 600 650
WAVELENGTH (nm)
700 750
SPECTRUM Sovbeons
S02 CONC /ot&O
FOLIAR INJURY:
RECORD.
Hg/m3_
TOTAL_
CHLOROSIS _i£_%
NECROSIS %
Y-AXIS COUNTS AT 767nm
450 500 550 600 650 700 750
WAVELENGTH (nm)
SPECTRUM
S02 CONC _
RECORD
FOLIAR INJURY:
TOTAL
CHLOROSIS.
NECROSIS _
Y-AXIS COUNTS AT 767nm
450 500 550 600 650
WAVELENGTH (nm)
700 750
SPECTRUM.,
S02 CONC
FOLIAR INJURY:
RECORD
rig/m'
TOTAL
CHLOROSIS.
NECROSIS _
°/
1%
.%
Y-AXIS COUNTS AT 767nm
SPECTRUM Soybeans
902 CONC /^6°
FOLIAR INJURY:
RECORD_lk£i__
Mq/m3 1.2-iS- h
TOTAL 10 %
CHLOROSIS_1£_%
%
NECROSIS
Y-AXIS COUNTS AT 767nm
450 500 550 600 650 700 750
WAVELENGTH (nm)
SPECTRUM^
S02CONC
RECORD.
FOLIAR INJURY:
TOTAL
CHLOROSIS.
NECROSIS_
Y-AXIS COUNTS AT 767nm
_%
450
550 600 650 700~
WAVELENGTH (nm)
750
SPECTRUM .
S02 CONC
FOLIAR INJURY:
RECORD.
Mg/m3
TOTAL _
CHLOROSIS.
NECROSIS _
Y-AXIS COUNTS AT 767nm
-------
350
400
450
500 550 600
WAVELENGTH
Normalized Individual Curves - 0.0 yg/m3 (Control) Plants
Curve i
Curve 2
Curve 17
... Curve 19
Curve 25
Curve 26
I
I—1
oo
Figure B
-2. Normalized Individual Curves of Visible Spectra - Winter Wheat - 0.0 Hg/m3 (Control)
-------
35
30
25
j20
U
Z
<
t—
10
0
350
400 450 500 550 600
WAVELENGTH (NM)
650
700
750
Normalized Individual Curves
3930 yg/m3 (1.5 ppm)/3 h
00
Curve 5
Curve 8
Curve 12
Curve 23
Figure B-3. Normalized Individual Curves of Visible Spectra - Winter Wheat - 3930 M8/m3 (1-5 ppm)/3h
-------
35
30
25
j20
U
LUIS
10
/ I
350 400 450 500 550 600
WAVELENGTH (NM)
650
700
750
Normalized Individual Curves
5240 ug/m3 (2.0 ppm)/3 h
Curve 11
Curve 18
00
CTv
Figure B-4. Normalized Individual Curves of Visible Spectra - Winter Wheat - 5240 |jg/m3 (2.0 ppm)/3h
-------
35
30
25
uj20
U
Z
ml5
10
U350
400
450
J •
TTT*
v
~-—•>»
500 550 600
WAVELENGTH (NM)
650
7OO
750
Normalized Individual Curves
6550 Ug/m3 (2.5 ppm)/3 h _
Curve 3
Curve 10
Curve 21
Curve 22
00
~>J
I
Figure B-5. Normalized Individual Curves of Visible Spectra - Winter Wheat - 6550 M8/m3 (2.5 ppm)/3h
-------
30
uj20
U
ml5
10
^
00
CO
350 400 450 500 550 600
WAVELENGTH (NM)
650
700
750
Normalized Individual Curves
7860 yg/m3 (3.0 ppm)/3 h
Curve 7
Curve 9
Curve 24
Figure B-6. Normalized Individual Curves of Visible Spectra - Winter Wheat - 7860 |Jg/m3 (3.0 ppm)/3h
-------
35
30
25
S?
uj20
U
Z
<
10
co
350 400 450 500 550 600
WAVELENGTH (NM)
650
700
750
Normalized Individual Curves
9170 UE/m3 (3.5 ppm)/3 h
Curve 15
Curve 20
Figure B-7. Normalized Individual Curves of Visible Spectra - Winter Wheat - 9170 |Jg/m3 (3.5 ppm)/3h
-------
APPENDIX C
ANALYSIS OF VARIANCE, LABORATORY EXPERIMENT
-------
-191-
EXPLANATION OF PROCEDURE FOR ESTIMATING
FOLIAR EFFECTS OF S02 ON CROP SPECIES
L = Percentage of leaves affected on an average affected plant
A = Percentage of leaf area affected on an average affected
leaf
P = Percentage of plants affected in field
L x A = Percentage of total leaf area affected on an average
affected plant
LxAxP=T= Average of total leaf area affected in a given field
Light foliar injury = up to 10 percent affected.
Moderate = 11-25 percent.
Severe = Greater than 25 percent.
Symptoms of foliar injury:
Chlorosis = visible yellowing of foliage due to bleaching of chloro-
phyll. Veins on leaves usually remain green if the
chlorosis is caused by S02-
Necrosis = death of tissue. Appears as ivory, gray, or white inter-
veinal markings which eventually drop out.
-------
-192-
TABLE C-l. RESULTS OF ONE-WAY ANALYSIS OF VARIANCE FOR SOYBEANS
Chlorosis
Reflectance
Band
Blue
Green
Red
IR
IR/R
Class
0
1
2
3
4
0
1
2
3
4
0
1
2
3
4
0
1
2
3
4
0
1
2
3
4
X
7269
7660
7762
7039
7691
50966
54020
55141
53521
59767
54650
58607
58658
53267
61690
272670
271735
276752
280978
273767
5.48
5.56
5.81
6.36
5,12
Significant
s df F (a = .05)
1938 4, 135 0.39 No
2709
2883
2439
2215
11029 4, 135 0.04 No
15642
15936
15434
11554
19687 4, 135 0.38 No
26888
29359
27498
26241
23899 4, 135 0.50 No
29836
27771
33952
27122
1.79 4, 135 1.08 No
2.30
2.59
2.52
1.90
Measurement for x and s in counts, except IR/R ratio.
-------
-193-
TABLE C-2. RESULTS OF ONE-WAY ANALYSIS OF VARIANCE FOR SOYBEANS
Necrosis
Reflectance
Band
Blue
Green
Green
Red
Red
Infrared
(IR)
Infrared
(IR)
IR/R
IR/R
Class
0
1+2+3
0
1+2+3
0
2+3
0
1+2+3
0
2+3
0
1+2+3
0
2+3
0
1+2+3
0
2+3
7
8
5
5
6
5
8
5
9
2
2
2
2
X
.269xl03
.595xl03
.098xl04
.098xl04
.647xl04
.465xl04
.554xl04
.465xl04
.133xl04
.727xl05
.6l7xl05
.727xl05
.66lxl05
5.48
3.66
5.48
3.21
Significant
s df F (a = .05)
1.
2.
1.
1.
1.
1.
2.
1.
2.
2.
8.
2.
1.
794xl03 1, 25 0.83
I4lxl03
021xl04 1, 25 1.80
021xl04 1, 10 4.94
173xl04
824xl04 1, 25 1.93
599xl04
824xl04 1, 10 7.34
456xl04
213xl04 1, 25 4.93
362xl03
213xl04 1, 10 0.30
234xl04
1.66 1, 25 0.49
1.27
1.66 1, 10 5.96
1.11
No
No
Yes
No
Yes
Yes
No
No
No
x and s in counts, except IR/R ratio.
-------
-194-
TABLE C-3. RESULTS OF ONE-WAY ANALYSIS OF VARIANCE FOR WINTER WHEAT
Reflectance
Band
Blue %
Green %
Red %
Blue %
Green %
Red %
Class
1
2
3
4
1
2
3
4
1
2
3
4
1
2
3
1
2
3
1
2
3
X
6.4
7.3
7.9
8.7
16.2
17.6
21.1
23.7
10.3
12.4
21.6
24.3
6.4
7.7
8.4
16.2
20.2
21.9
10.3
19.0
20.3
Significant
s df F (a = .05)
Necrosis
.44
.17 3, 17 13.92 Yes
.37
.32
4.14
2.89 3, 17 16.78 Yes
6.78
.25
3.86
9.62 3, 17 36.44 Yes
9.68
1.84
Chlorosis
.44
.36 2, 18 10.45 Yes
.78
4.14
10.22 2, 18 6.41 Yes
7.20
3.86
33.57 2, 18 6.54 Yes
32.76
x and s data in percent reflectance
-------
APPENDIX D
OBSERVATIONS OF FOLIAR INJURY
-------
-196-
TABLE D-l. OBSERVATIONS OF S02-EXPOSED SOYBEAN FIELDS
NEAR SHAWNEE STEAM PLANT, 1978
Field
Designation
3
4
1107
1
2
7
1119
1117
1122
1121
1120
1118
13
15
16
17
Point
1
2
3
4
5
6
7
1
2
3
4
1
2
3
No point data
1
2
3
1
2
3
1
1
1
1
1
1
1
1
2
3
1
2
3
4
5
1
2
3
LxAxP
(%)
2.0
0
0
0
0.4
1.2
1.6
3.0
1.2
1.4
1.6
14.4
10.5
12.6
12.0
16.2
13.5
9.0
7.5
6.0
6.0
0.0
0.0
0.0
3.2
4.0
9.0
1.8
4.8
6.0
3.2
4.0
6.0
7.2
10.0
2.4
2.0
2.7
Mean
Injury (%)
0.7
2.0
12.6
13.5
7.5
6.0
0.0
0.0
0.0
3.2
4.0
9.0
4.8
6.0
2.4
(continued)
-------
-197-
TABLE D-l (continued)
Field
Designation
18
19
20
21
22
23
25
26
14
8
1143
1145
(continued)
Point
1
2
3
1
2
3
1
2
3
1
2
3
1
2
3
1
2
3
1
2
3
4
5
1
2
3
4
5
1
1
1
2
3
1
LxAxP
(%)
1.2
1.8
1.8
9.6
2.0
12.8
12.8
8.4
12.0
9.0
16.0
22.5
25.0
24.0
13.5
0.4
0.8
0.4
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.6
0.0
12.0
28.0
21.6
0.0
Mean
Injury (%)
1.8
11.2
12.8
16.0
18.0
0.8
0.0
.- - -- -
0.0
0.6
0.0
19.2
0.0
-------
-198-
TABLE D-l (continued)
Field
Designation
1137
1114
1113
1112
1110
1109
1108
Point
1
2
1
2
3
4
5
6
7
1
2
3
4
5
1
2
3
4
1
2
3
4
5
1
2
3
1
2
3
LxAxP Mean
(%) Injury (%)
35.0 35.0
40.0
3.2 9.0
12.0
4.8
8.4
16.8
9.0
19.2
4.8 4.8
1.2
7.5
4.8
4.5
2.4 2.4
4.8
1.2
3.2
10.0 12.0
4.8
18.0
4.5
16.2
0.0 4.8
6.0
6.0
6.0 4.5
3.6
7.2
-------
-199-
oo
i—
o
O X
W
W
H (»(
CO
ss
h^ O
O CO
CJ
«-• «r-t-
-------
-200-
<
9*^ 'tot 'O •& w o -*f^^ n n <^ * • 41 M o> o> r- r-.
«-«OOOOOOOOOOOOOOOOOOOO—iO OCO-^OOOOOOOOOOOOOOOOOOOOOOOOOO
0*0000000000 o ooooooooooo ooooo ooooooooooooooo ooo oooooo o
OJ
^ ^^«*->ooooooooo ooooooooooo oooooooooooooooooooooooooooooo
ooo ooooo ooooooooooooooo ooc>oo ooooo oooooooooooooooooooo
o
CJ
. OOO'OOOOOO-OO OOOOOOOOOCDOO OOOOO OOOOOOOOOO OOCOOOOOOOOOOO O
ooooooooooooooooooooooo oooooooooooooooooooooooooooooo
OOOOOOOOOOO OOOOOO-OOCDO^^O OO^O^OC3CiOOOOO OOOOO^OOOOOOOOOO^OO O
tfioonomifYOm Om oomom omoom oo ncnomH
mmmomiAOinmomOOiriiAOiniAOmoino omomomomotnmovtmooiAOomomoomoinom
X
_ m m lf\ M\m A iO ^1 ^ J5 <0 ^1 ^> ^ iO fr— r-- V- f*~ •*. V^ r-- «^- »*- •*. **» o» Ok «• m n •*» «n •« in nb **k o* r»i a
O
-------
-201-
OOOOOOOOOOOOOOOO OOOO—(OOOOOOOOOOOOOOOOOOOOOOGOOOOOOOGO
u| oooooooooooooooo o oo oo o o o o oo o o oo o oo o o o ooo -3 oooo o o oo ooo o
/—s
(U
oooooooooooooooo ooooooooooooooooooooooooooooooooooooo
o [[[
4-> ce oooooooooooooooo oooooo o ooo oooo •> ooooo o ooo o ooo oo oooo oo o
a
o
u
oooooooooooooooo o o o o o o o o o o o o o o o o o o o o o o o o o o 'o o o o o o o o o o o
CM 1C OOOOOOOOOOOOOOOO O O O OOO O OO OO OOOO O OO O O O OOO O OOOO OO OO O OO O
Q
OOOOOOOOOOOOOOOO OOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO
E-i
Vim o-Mvin OOOmnoomoonOmomoir\inomc\oinoir\ooinno«\in!rkOOoooooirkiftv\inmn«tOo
x ctv« wvm '
-------
-202-
«o o-« a^r^
<-«O — «— * O OC C4
ooooooooooooooooo oooo ooooo oooooo
•*-•.,••...-.........-..... ......
000*00000 oooo oooooo oooo oooo o ooooo
a
•H
4J
o
U
CSJ
I
Q
PQ
<
H
-
OOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO
ooooooooooooooooo ooooooooooooooo
.-*« r* o» tn m » ^» o w*jr- om Nrsio^ i^* <-• in •«• trt m ^
mrxiNcsjrarucvimmrnmrMmmmcnMrvtncw^mtn^
OOOOOOOOOOOOOOOOO OOOOOOOOOOOOOOO
ooooooooooooooooo ooooooooo ooooo o
omom o w\ o u*. o ar\ o m o otntcvoommo ooo
-------
-203-
TABLE D-3. OBSERVATIONS OF FOLIAR INJURY TO WINTER WHEAT
EXPERIMENTAL PLOT
Subplot
Cl
C2
C3
C4
C5
C6
C7
C8
C9
CIO
Cll
C12
Row
1
2
3
1
2
3
1
2
3
1
2
3
1
2
3
1
2
3
1
2
3
1
2
3
1
2
3
1
2
3
1
2
3
1
2
3
Necrosis (%)
9
9
9
8.5
8,5
r»
XD
7
7
7
8
8
8
10.5
10.5
10.5
11.5
11.5
11.5
11
11
11
11.5
11.5
11.5
6
6
6
6.5
X
X
6.5
6.5
6.5
6.5
6.5
6.5
Chlorosis (%)
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
S02 Conc.a
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
(continued)
-------
-204-
TABLE D-3 (continued)
Subplot
CIS
C14
i
!
Al
A2
\
A3
A4
A5
A6
A7
A8
A9
A10
Row
1
2
3
1
2
3
1
2
3
1
2
3
1
2
3
1
2
3
1
2
3
1
2
3
1
2
3
1
2
3
1
2
3
1
2
3
Necrosis (%)
8
8
8
8.5
8.5
8.5
90
90
90
10
10
10
92.5
92.5
92.5
89.5
89.5
89.5
52.5
52.5
52.5
58
58
58
86.5
86.5
86.5
29
29
29
51.5
51.5
51.5
38.5
38.5
X
Chlorosis (%)
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
S02 Cone.3
0
0
0
0
0
0
4
4
4
1
1
1
4
4
4
3
3
3
2
2
2
3
3
3
2
2
2
1
1
I
3
3
3
2
2
2
(continued)
-------
-205-
TABLE D-3 (continued)
Subplot Row
All 1
2
3
A12 1
2
3
A13 1
2
3
A14 1
2
3
A15 1
2
3
A16 1
2
3
Necrosis (%)
19.5
19.5
19.5
55
55
55
55
55
55
53.5
53.5
53.5
69
69
69
19
19
19
Chlorosis (%)
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
S02 Conc.a
1
1
1
4
4
4
2
2
2
3
3
3
4
4
4
1
1
1
o
Concentration
Number 0 = 0 |Jg/ra"
1 = 2620
2 = 5240
3 = 7860
4 = 10480
J for 2-h avg.
X = not observed
-------
TABLE D-4.
-206-
S02 EXPOSURE, OBSERVATION, AND SCANNING DATES-
WINTER WHEAT EXPERIMENTAL PLOT
S02
Subplot
Al
A2
A3
A4
A5
A6
A7
A8
A9
A10
All
A12
A13
A14
A15
A16
EXPOSURE
Month
5
5
5
5
5
5
5
5
4
4
4
4
4
4
4
4
DATES
Day Year
1 1979
1
7
7
1
1
7
7
30
30
30
30
27
27
27
27
OBSERVATION DATES
All observations done on 5-14-79
SCANNING DATES
Subplots
Month
Day Year
C1,C2,C3,C4,A1,A2,A3,A4
All others
5
5
10
14
1979
1979
-------
TABLE D-5.
-207-
REFLECTANCE, FOLIAR INJURY, AND YIELD OF S02-AFFECTED AND
UNAFFECTED SOYBEANS GROWN IN EXPERIMENTAL PLOT
Reflectance (%)
Subplot
Row
Green
Red
IR
IR/Red -
Injury
Chlorosis
(%)3
Necrosis
Yield
(kg/ha)
Control
Cl
C2
C4
Mean for
Control
Subplots
1
2
3
1
2
3
1
2
10.
10.
10.
10.
10.
9.
9.
10.
10.
25
79
25
79
61
89
89
25
34
9.17
9.71
9.35
9.71
9.53
9.17
9.17
9.53
9.42
53
56
53
52
51
49
47
46
51
.42
.47
.59
.88
.80
.46
.12
.58
.42
5
5
5
5
5
5
5
4
5
.82
.81
.73
.44
.43
.39
.14
.89
.46
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
3047
3320
1824
2730
S02-Affected
Al
A2
A3
A4
Mean for
Affected
Subplots
1
2
3
1
2
3
1
2
3
2
3
10
10
10
12
10
10
10
9
10
10
10
10
.25
.43
.07
.05
.61
.61
.07
.89
.97
.79
.43
.56
9.35
9.35
9.17
12.95
10.97
9.71
9.35
9.17
10.07
10.79
10.43
10.12
51
52
50
41
39
48
45
44
48
36
32
44
.26
.34
.72
.19
.93
.92
.32
.78
.74
.51
.55
.75
5.
5.
5.
3.
3.
5.
4.
4.
4.
3.
3.
4.
48
60
53
18
64
04
85
88
84
38
12
50
25
25
25
6
16
21
25
6
10
15
8
16.5
0
0
0
80
54
1
0
0
0
30
30
17.7
2596
2234
2759
2352
2485
Injury calculations by LxA method.
-------
-208-
TABLE D-6. S02 EXPOSURE, OBSERVATION, AND SCANNING DATES-
SOYBEAN EXPERIMENTAL PLOT
S02 EXPOSURE DATES
Subplot Month Day Year
All 8 16 1979
SCANNING DATES
Subplot Month Day Year
All 8 21 1979
OBSERVATION DATES
Subplot Month Day Year
All 8 21 1979
-------
TABLE D-7. REFLECTANCE, NECROSIS, S02 CONCENTRATION, AND YIELD FOR S02-EXPOSED
AND UNEXPOSED WINTER WHEAT GROWN IN EXPERIMENTAL PLOT
Subplot
Cl
C2
C3
C4
C5
C6
C7
C8
C9
CIO
Cll
C12
CIS
C14
Mean ± S.D. of
Unexposed (control)
Subplots
S02 Cone.
(2-h avg)
0 pg/m3
0
0
0
0
0
0
0
0
0
0
0
0
0
Mean
Green
10.54
9.80
13.83
11.38
10.07
11.90
10.54
10.43
10.97
11.65
9.77
8.71
9.41
12.64
10.8311.31
Reflectance (%)
Red
9.21
8.65
11.53
10.83
9.12
11.13
9.73
9.96
10.00
12.21
9.93
9.62
10.04
13.45
10.39±1.26
IR
32.55
29.55
43.06
30.23
37.03
40.23
36.22
34.46
34.44
30.97
25.18
21.64
23.88
30.83
32.1615.83
IR/Red
3.53
3.42
3.73
2.79
4.06
3.61
3.72
3.46
3.44
2.54
2.54
2.25
2.38
2.29
3 . 13±0 . 36
Mean
Necrosis (%)a
9.0
8.5
7.0
8.0
10.5
11.5
11.0
11.5
6.5
6.0
6.5
6.5
8.0
8.5
8.511.9
Yield
(kg/ha)
3318
2661
2466
2306
3297
2919
3250
3357
3406
3832
2482
3217
3301
3499
30941436
N5
O
(continued)
-------
TABLE D-7 (continued)
Subplot
Al
A2
A3
A4
A5
A6
A7
A8
A9
A10
All
A12
A13
A14
A15
A16
Mean ± S.D. of
Exposed Subplots
S02 Cone.
(2-h avg) Green
10480 |Jg/m3
2620
10480
7860
5240
7860
5240
2620
7860
5240
2620
10480
5240
7860
10480
2620
Mean Reflectance (%)
Red
10.67
9.42
12.90
11.67
10.29
10.72
9.30
10.59
10.09
10.41
9.78
8.40
9.03
9.93
11.26
12.41
10.4311.16
IR
12.09
8.36
14.44
14.06
11.47
10.77
12.03
9.64
10.61
11.15
11.13
9.62
10.43
10.95
12.43
13.17
11.4011.57
IR/Red
23.35
28.02
32.12
27.75
29.14
33.02
24.06
36.46
46.13
28.02
23.02
17.01
19.41
21.38
23.62
29.26
27.6116.90
1.93
3.35
2.22
1.97
2.54
3.07
2.00
3.78
4.35
2.51
2.07
1.77
1.86
1.95
1.90
2.22
2.4710
Mean
Necrosis (%)
90.0
10.0
92.5
89.5
52.5
58.0
86.5
29.0
51.5
38.5
19.5
55.0
55.0
53.5
69.0
19.0
.75 54.3125.7
Yield
(kg/ha)
1512
3710
948
1261
2524
2670
2576
3648
2562
2365
2723
2105
2450
2469
3010
3219
24851739
I
to
alnjury calculations by LxA method (Appendix D). Light necrosis is classified as £10 percent; moderate, 11-25 percent;
and severe, >25 percent. Necrosis in control is associated with natural senescence that occurred after scanning but
before observations were made.
-------
TABLE D-8. REFLECTANCE, FOLIAR INJURY, S02 CONCENTRATION, AND YIELD FOR S02-EXPOSED AND UNEXPOSED SOYBEANS
GROWN IN EXPERIMENTAL PLOT
Subplot
Cl
C2
C4
Mean ± S.D.,
Control
Subplots
Al
A2
A3
A4
Mean ± S.D. ,
S02-Exposed
Subplots
S02 Cone.
(2-h avg)
0 pg/m3
0
0
2620 |Jg/m3
5240
7860
10480
Mean Reflectance (%)
Green
10.43
10.43
10.07
10.3110.17
10.25
11.09
10.31
10.61
10. 57±0.33
Red
9.41
9.47
9.35
9.4110.05
9.29
11.21
9.53
10.61
10.1610.78
IR
54.49
51.38
46.85
50.9113.14
51.44
43.35
46.28
34.53
43.9016.14
IR/Red
5.79
5.42
5.02
5.4110.31
5.54
3.95
4.86
3.25
4.4010.87
Injury
Chlorosis
0
0
0
0
25.00
14.33
13.67
11.50
16.1315.23
(%)a
Necrosis
0
0
0
0
0
45.00
0
30.00
18.75119.49
Yield
(kg/ha)
3047
3320
1824
27301650
2596
2234
2759
2352
24851205
Ni
I— '
I—"
1
Injury calculations by LxA method (Appendix D). Light effects are classified as £10 percent; moderate effects,
11-25 percent; and severe effects, >25 percent.
-------
APPENDIX E
SPECTRAL CURVES, EXPERIMENTAL PLOTS
-------
PLOT Cl
ROW I
-213-
450 500 550 600 650 TOO 750
WAVE LENGTH (nm)
RECORD 7-2
SPECTRUM I SOYBEANS; 2 GRAY PANEL
S02 CONC. __0__A9/n>s . 2h
NECROSIS O % CHLOROSIS Q 96
Y-AXIS COUNTS. 750 nm: SOYBEANS 2-97
PLOT Cl
ROW 3
_l l-
_l L.
JL
WAVE LENGTH (nm)
RECORD _Z=£_
SPECTRUM I SOYBEANS; 2 GRAY PANEL
S02 CONC 0 /<(f]/m> , 2 h
NECROSIS . Q 3& CHLOROSIS Q %
Y-AXIS COUNTS. 750 nm: SOYBEANS 2.98
PLOT A I
ROW 2
WAVE LENGTH (nm)
RECORD 7-10
SPECTRUM I SOYBEANS; 2 GRAY PANEL
S02 CONC._2_6JLQ_Ag/ms , 2 h
NECROSIS Q % CHLOROSIS __2J
Y-AXIS COUNTS, 750 nm: SOYBEANS 2.91
PLOT CI
ROW 2
. i ' ' 1 1 1 1—
450 500 550 600 650 700 750
WAVE LENGTH (nm)
RECORD
7-4
SPECTRUM I SOYBEANS; 2 GRAY PANEL
SOj CONC.
. 2 h
NECROSIS 0 % CHLOROSIS __P__%
Y-AXIS COUNTS. 750 nm: SOYBEANS 3.1 4
PLOT A I
ROW I
WAVE LENGTH (nm)
RECORD _LZ£__
SPECTRUM I SOYBEANS? 2 GRAY PANEL
S0a CONC. J62C__Ag/rT|S f 2 h
NECROSIS Q ?& CHLOROSIS 25 %
Y-AXIS COUNTS. 750 nms SOYBEANS 2.65"
PLOT A I
ROW 3
Figure E-l. Individual Spectral Curves - Soybeans
WAVE LENGTH (nm)
RECORD 7-12
SPECTRUM I SOYBEANS; 2 GRAY PANEL
SO, CONC. _2JL20_/ig/m» , 2 h
NECROSIS 0 % CHLOROSIS 25 %
Y-AXIS COUNTS, 750 nm: SOYBEANS _2_JL?_
-------
-214-
PLOT 02
ROW I
450 500 550 600 650 700 750
WAVE LENGTH (nm)
RECORD 7~14
SPECTRUM I SOYBEANS; 2 GRAY PANEL
SO, CONC __P__^g/m' . 2 h
NECROSIS _2 3& CHLOROSIS _J°__%
Y-AXIS COUNTS. 750 nm: SOYBEANS 2.94
PLOT 02
ROW 3
(Record destroyed
PLOT 02
ROW 2
450 500 550 600 650 700 750
WAVE LENGTH (nm)
RECORD 7-1 6
SPECTRUM I SOYBEANS; 2 GRAY PANEL
SO, CONC __P___/tg/m> . 2 h
NECROSIS _0__2& CHLOROSIS __Q__96
Y-AXIS COUNTS. 750 nm: SOYBEANS 2.88
PLOT A2
ROW I
J L
J L
WAVE LENGTH (nm)
RECORD 7-1 8
SPECTRUM I SOYBEANS; 2 GRAY PANE.I
SOi CONC _ Q ./tg/mi . 2 h
NECROSIS _ Q_% CHLOROSIS 0 _%
Y-AXIS COUNTS, 750 nm: SOYBEANS 2.75
WAVE LENGTH (nm)
RECORD 7-20
SPECTRUM I SOYBEANS; 2 GRAY PANEL
SO, CONC 5240 ^g/m» . 2 h
NECROSIS 8Q'% CHLOROSIS 6 %
Y-AXIS COUNTS. 750 nm: SOYBEANS 2.29
PLOT A2
ROW 2
WAVE LENGTH (nm)
RECORD 7-22
SPECTRUM I SOYBEANS; ? GRAY PANLl
or. rr+jr 5240 , .
^u, CONG ./tg/m5 , 2 h
NECROSIS .54..S CHLOROSIS 16 <«
r'-AXIS COUNTS. 750 nm: SOYBEANS 2.22_
PLOT A2
ROW 3
WAVE LENGTH (nm)
RECORD 7-24
SPECTRUM I SOYBEANS; 2 GRAY PANEL
SO, CONC 5240 ^g/m> , 2 h
NECROSIS
CHLOROSIS
Y-AXIS COUNTS. 750 nm: SOYBEANS _ 2-72
-------
-215-
PLOT ca
ROW
450 500 550 600 650 700 750
RECORD
WAVE LENGTH (nm)
7-14
SPECTRUM I SOYBEANS; 2 GRAY PANEL
S02 CONC ° /tg/m» . 2 h
NECROSIS _2 % CHLOROSIS __°__%
Y-AXIS COUNTS, 750 nm: SOYBEANS 2.94
PLOT C2
ROW 3
(Record des-Weyec/ rftei
J
L
WAVE LENGTH (nm)
RECORD 7- I 8
SPECTRUM
SOj CONC
SOYBEANS;2 GRAY PANEL
0. ./tg/m' . 2 h
NECROSIS Q_9& CHLOROSIS 0 %
Y-AXIS COUNTS, 750 nm: SOYBEANS 2.75
PLOT A2
ROW 2
JL
_l_
WAVE LENGTH (nm)
RECORD 7r22 _
SPECTRUM I SOYBEANS; 2 GRAY PANEL
K9An
S02 CONC ° ^3/m» , 2 h
NECROSIS 54 % CHLOROSIS 16
450 500 550 600 650 700 750
WAVE LENGTH (nm)
RECORD ...7-16
SPECTRUM I SOYBEANS; 2 GRAY PANEL
S02 CONC. ___0___^g/m5 , 2 h
NECROSIS 0 % CHLOROSIS P_ %
Y-AXIS COUNTS, 750 nm: SOYBEANS 2.88
WAVE LENGTH (nm)
RECORD -7-20
SPECTRUM I SOYBEANS; 2 GRAY PANEL
SOj CONC. _5_£40__A9/™' , 2 h
NECROSIS ^8O'% CHLOROSIS _S %
Y-AXIS COUNTS, 750 nm: SOYBEANS —2.29
WAVE LENGTH (nm)
7-24
Y-AXIS COUNTS, 750 nm: SOYBEANS 2.22
RECORD
SPECTRUM I SOYBEANS; 2 GRAY PANEL
S02 CONC. _5240_Ag/rr,3 , 2 h
NECROSIS L2fc CHLOROSIS 2J_%
Y-AXIS COUNTS, 750 nm: SOYBEANS 2.72
-------
-216-
PLOT C4
ROW J
-I 1 1 L
WAVE LENGTH (nm)
RFCDRD 7-26
SPECTRUM I SOYBEANS; Z GRAY PANEL
S02 CONC Q___/tg/m3 t 2 h
NECROSIS O % CHLOROSIS 0 %
Y-AXIS COUNTS, 750 nm: SOYBEANS 2.62
PLOT A4
ROW 2
450 50*0 biJo et)o bio /oo 75*0
WAVE LENGTH (nm)
RECORD _ 7-40
SPECTRUM I SOYBEANS; 2 GRAY PANEL
S02 CONC. IQ480iA9/m3 t 2 h
NECROSIS 3Q % CHLOROSIS _15_96
Y-AXIS COUNTS, 750 nm: SOYBEANS 2.03
WAVE LENGTH (nm)
RECORD 7-32
SPECTRUM I SOYBEANS; 2 GRAY PANEL
SOj CONC 7860 ^g/n.5 , 2 h
NECROSIS 0 % CHLOROSIS 25 %
Y-AXIS COUNTS, 750 nm: SOYBEANS 2.52
PLOT C4
ROW 2
o Curve
*" r ofter-
^— av&r-o-a i OQ
J_
_L
_L
_L
_L
WAVE LENGTH (nm)
RECORD 7-28
SPECTRUM I SOYBEANS; 2 GRAY PANEL
S02 CONC 0 ,/tg/m* , 2 h
NECROSIS Q % CHLOROSIS O %
Y-AXIS COUNTS, 750 nm: SOYBEANS 2.59
_L
_L
_L
_L
_L
WAVE LENGTH (nm)
RECORD 7-42
SPECTRUM I SOYBEANS; 2 GRAY PANEL
SOj CONC 10480 ^0/m» , 2 h
NECROSIS 30 % CHLOROSIS 8 %
Y-AXIS COUNTS, 750 nm: SOYBEANS 1.81
PLOT A3
ROW 2
RECORD 7-3
WAVE LENGTH (nm)
SPECTRUM I SOYBEANS; 2 GRAY PANEL
S02 CONC._Z860_Ag/m3 >2h
NECROSIS 0 % CHLOROSIS 6 %
Y-AXIS COUNTS, 750 nm: SOYBEANS 2.49
-------
-217-
PLOT A3
ROW 3
PLOT A4
ROW 3
PLOT A3
ROW
WAVE LENGTH (nm)
RECORD 7~3g
SPECTRUM I SOYBEANS; 2 GRAY PANEL
SOj CONC. 7860 /[g/m* , 2 h
NECROSIS 0 % CHLOROSIS '0 %
Y-AXIS COUNTS, 750 nm: SOYBEANS 2-7 '
-------
-218-
PLOT A-//
450 500 550 600 650 700 750
WAVELENGTH (nm)
RECORD 3(/-$3
SPECTRUM I WHEAT; 2 GRAY PANEL
S02 CONC. 3 fe 5 O mq/n& £ h
FOUAR NECROSIS / ¥00387.
'GRAY
-------
-219-
PLOT C-/2
450 500 550 600 650 700 750
WAVELENGTH (nm)
RECORD /ti-gq
SPECTRUM I WHEAT; 2 GRAY PANEL
SOg CQNC. O mg/n? _ h
FOL/AR NECROSIS .^ %
Y-AX/S COUNTS TSOnm'WHEAT
PLOT C-/3
450 500 550 600 650 TOO 7
WAVELENGTH (nm)
RECORD 49-go
SPECTRUM » WHEAT; 2 GRAY PANEL
S02 CONC. O my/m3
FOL/AR NECROS/S JL£_%
Y-AXIS COUNTS 750nm'WHEAT A
' GRAY PANEL It* 33/0
•GRAY PANEL Oto/St.
PLOT A-14
450
500 550 600 650 700 750
WAVELENGTH (nm)
RECORDS - 3 a h
FOL/AR NECROS/S_Al£_%
Y-AX/S COUNTS 750nm: WHEAT/.
••GRAY PANEL
PLOT A-/6
7&0
"450 500 550 600 650 TOO
WAVELENGTH (nm)
RECORD 39-
-------
ho
K>
O
350 400 450 500 550 600 650 700 750
WAVELENGTH (NM)
Figure 3. NASA GRAY TARGET NO. 4 SPECTRAL REFLECTANCE
-------
APPENDIX F
ANALYSIS OF VARIANCE, EXPERIMENTAL PLOT
-------
-222-
TABLE F-l. ANALYSIS OF VARIANCE FOR GREEN BAND BY DOSE SET - SOYBEANS
Dose Set
1
2
Treatments
Error
Total
Cone. X S
(iJg/m3) No. Obs. (counts) (counts)
Control 8 0.58 0.02
2620] 11 0.59 0.03
5240 ,. ,
-,a/-n — combined
/obu
10480_
ANOVA Data
SS df MS
6.98 x 10"4 1 6.98 x 10"4
0.01 17 8.01 x 10"4
0.01 18
Sum
(counts)
4.60
6.46
F
0.87
F = 0.87 does not exceed F „,. 17 = 4.45.
* UD j x j I /
Null hypothesis is not rejected.
Means of green reflectance of the two dose sets are not significantly
different.
-------
-223-
TABLE F-2. ANALYSIS OF VARIANCE FOR RED BAND BY DOSE SET - SOYBEANS
Dose
1
2
Cone. X
Set (|Jg/m3) No. Obs. (counts)
Control 8 0.52
2620] 11 0.56
5240 , . .
-•off. — combined
7obO
10480.
ANOVA Data
SS df
Treatments 0.01 1
Error 0.04 17 2
Total 0.05 18
S Sum
(counts) (counts)
0.01 4.18
0.06 6.19
MS F
0.01 3.07
.45 x 10"3
F = 3
Null
Means
.07 does not exceed F ,.,. 1 17 = 4.45.
hypothesis is not rejected.
of red reflectance of the two dose sets
are not significantly
different.
-------
-224-
TABLE F-3. ANALYSIS OF VARIANCE FOR IR BAND BY DOSE SET - SOYBEANS
Dose Set
1
2
Treatments
Error
Total
Cone. X
(|Jg/ra3) No. Obs. (counts)
Control 8 2.86
2620] 11 7.55
5240 . . ,
-.ot-r, — combined
/ooO
10480,
ANOVA Data
SS df
0.45 1
1.68 17
2.13 18
S Sum
(counts) (counts)
0.19 22.87
0.38 28.01
MS F
0.45 4.58
0.10
F = 4.58 does exceed F _,. 1 17 = 4.45.
Null hypothesis is rejected.
Means of IR reflectance of the two dose sets are significantly different.
-------
-225-
TABLE F-4. ANALYSIS OF VARIANCE FOR IR/RED RATIO BY DOSE SET - SOYBEANS
Dose Set
1
2
Treatments
Error
Total
Cone , X
(|Jg/m3) No. Obs. (counts)
Control 8 5 . 46
2620"] 11 4.51
5240 , . ,
?860_ combined
10480^
ANOVA Data
SS df
4.23 1
10.35 17
14.58 18
S Sum
(counts) (counts)
0.33 43.69
0.98 49.56
MS F
4.23 6.95
0.61
F = 6.95 does exceed F
= 4.45.
1 17
j J. j J. /
Null hypothesis is rejected.
Means of IR/red ratios of the two dose sets are significantly different.
NOTE: STATISTICS FOR CHLOROTIC VERSUS UNAFFECTED (CONTROL) SOYBEANS ARE
IDENTICAL TO THE PRECEDING STATISTICS FOR DOSE SETS 1 AND 2.
-------
-226-
TABLE F-5. ANALYSIS OF VARIANCE FOR GREEN BAND BY NECROSIS SET
SOYBEANS
Necrosis Set
Unaffected
Necrotic
No. Obs.
5
8
X
(counts)
0.58
0.60
S
(counts)
0.02
0.04
Sum
(counts)
4.60
3.02
ANOVA Data
SS df MS
Treatments
Error
Total
2.59 x 10"3
0.01
0.01
1
11
12
2.59 x 10"3
7.56 x 10"4
3.42
F = 3.42 does not exceed F nr- 17 = 4.45.
• \J J • J. • -L /
Null hypothesis is not rejected.
Means of green reflectance of the two necrosis sets are not significantly
different.
-------
-227-
TABLE F-6. ANALYSIS OF VARIANCE FOR RED BAND BY NECROSIS SET
SOYBEANS
Necrosis Set
Unaffected
Necrotic
Treatments
Error
Total
X S
No. Obs. (counts) (counts)
5 0.52 0.01
8 0.61 0.07
ANOVA Data
SS df MS
0.02 1 0.02
0.02 11 1.74 x 10"3
0.04 12
Sum
(counts)
4.19
3.05
F
13.12
F = 13.12 does exceed F „,- , = 4.84.
• U J } X y JL A
Null hypothesis is rejected.
Means of red reflectance of the two necrosis sets are significantly
different.
-------
-228-
TABLE F-7. ANALYSIS OF VARIANCE FOR IR BAND BY NECROSIS SET
SOYBEANS
X
Necrosis Set No. Obs. (counts)
Unaffected 8 2.86
Necrotic 5 2.21
ANOVA Data
SS df
Treatments 1 . 28 1
Error 0.71 11
Total 1.99 12
S Sum
(counts) (counts)
0.19 22.87
0.34 11.07
MS F
1.28 19.73
0.06
F = 19.73 does exceed F =4.84.
• UD y j. j x J.
Null hypothesis is rejected.
Means of IR reflectance of the two necrosis sets
are significantly
different.
-------
-229-
TABLE F-8. ANALYSIS OF VARIANCE FOR IR/RED RATIO BY NECROSIS SET
SOYBEANS
Necrosis Set No. Obs. X
Unaffected 8 5.46
Necrotic 5 3.67
ANOVA Data
SS df
Treatments 9.80 1
Error 3.26 11
Total 13.06 12
S Sum
0.33 43.65
0.79 18.36
MS F
9.80 33.02
0.30
F = 33.02 does exceed F .,. 1 . = 4.84.
• UO j X j 1 J.
Null hypothesis is rejected.
Means of IR/red ratio of the two necrosis sets
are significantly
different.
-------
TABLE F-9.
-230-
ANALYSIS OF VARIANCE FOR CROP YIELD BY CHLOROSIS SET
SOYBEANS
Chlorosis Set
Unaffected
Chlorotic
No. Obs.
i
8
5
X
(kg/ha)
2844
2497
S
(kg/ha)
642
220
Sum
(kg/ha)
22749
27471
ANOVA Data
SS df
MS
Treatments
Error
Total
555312
3369044
3924356
1
17
18
555312
198179
2.80
F = 2.80 does not exceed F „ 17 = 4.45.
• v/O * J. * x /
Null hypothesis is not rejected.
Means of yield of the two chlorosis sets are not significantly different.
-------
-231-
TABLE F-10.
ANALYSIS OF VARIANCE FOR CROP YIELD BY NECROSIS SET
SOYBEANS
Necrosis Set
Unaffected
Necrosis
Treatments
Error
Total
No. Obs.
8
5
ANOVA
SS
973298
2900863
2844
X S
(kg/ha) (kg/ha)
2844 642
2281 65
Data
df MS
1 973298
11 263715
12
Sum
(kg/ha)
22749
11406
F
3.69
F = 3.69 does not exceed F 05 1 11 = 4'8^'
Null hypothesis is not rejected.
Means of yield of the two necrosis sets are not significantly different.
-------
-232-
TABLE F-ll. ANALYSIS OF VARIANCE FOR RED BAND BY DOSE SET
WINTER WHEAT
Dose Set
I
2
3
4
5
Treatments
Error
Total
Cone.
(H8/m3)
Control
2620
5240
7860
10480
X
No. Obs. (counts)
37 0.57
12 0.59
11 0.63
12 0.64
11 0.68
ANOVA Data
SS df
0.13 4
0.67 78
0.80 82
S Sum
(counts) (counts)
0.08 21.21
0.11 7.05
0.05 6.91
0.10 7.73
0.13 7.51
MS F
0.03 3.82
0.01
F = 3.82 does exceed F nc , _0 = 2.50.
.UD,4,/o
Null hypothesis is rejected.
Means of red reflectance of the five dose sets are significantly
different.
-------
-233-
TABLE F-12. ANALYSIS OF VARIANCE FOR GREEN BAND BY DOSE SET
WINTER WHEAT
Cone.
Dose Set (pg/m3)
1 Control
2 2620
3 5240
4 7860
5 10480
Treatments
Error
Total
X S
No. Obs. (counts) (counts)
37 0.60 0.08
12 0.58 0.06
11 0.54 0.04
12 0.59 0.07
11 0.59 0.09
ANOVA Data
SS df MS
0.04 4 0.01
0.46 78 0.01
0.49 82
Sum
(counts)
22.33
6.34
5.95
7.07
6.49
F
1.37
F = 1.37 does not exceed F Q5 , -g = 2.50.
Null hypothesis is not rejected.
Means of green reflectance of the five dose sets are not significantly
different.
-------
-234-
TABLE F-13. ANALYSIS OF VARIANCE FOR IR BAND BY DOSE SET
WINTER WHEAT
Dose
1
2
3
4
5
Cone.
Set (|Jg/m3)
Control
2620
5240
7860
10480
Treatments
Error
Total
X
No. Obs. (counts)
37 1.81
12 1.62
11 1.38
12 1.52
11 1.26
ANOVA Data
SS df
3.52 4
7.41 78
10.93 82
S Sum
(counts) (counts)
0.35 66.98
0.29 19.48
0.25 15.22
0.28 18.24
0.23 13.83
MS F
0.88 9.25
0.10
F = 9
Null
Means
.25 does exceed F
05,4,78 = 2'50-
hypothesis is rejected.
of IR reflectance
of the five dose sets
are significantly
different.
-------
-235-
TABLE F-14. ANALYSIS OF VARIANCE FOR IR/RED RATIO BY DOSE SET
WINTER WHEAT
Cone.
Dose Set (pg/m3) No
1 Control
2 2620
3 5240
4 7860
5 10480
. Obs. X
37 3.21
12 2.86
11 2.21
12 2.39
11 1.85
ANOVA Data
SS df MS
Treatments 21.
Error 28.
Total 49.
49 4 5.37
48 78 0.37
97 82
S Sum
0.70 118.59
0.77 34.31
0.35 24.26
0.53 28.70
0.14 20.34
F
14.71
F = 14.71 does exceed F n,-
• "O ,
Null hypothesis is rejected
4,78 = 2'5°-
•
Means of the IR/red ratio for the five dose sets are
significantly
different.
-------
-236-
TABLE F-15. ANALYSIS OF VARIANCE FOR GREEN BAND BY NECROSIS SET
WINTER WHEAT
Necrosis Set
1
2
3
4
Range of
injury (%)
0-10
11-25
26-50
51-100
No. Obs.
28
18
5
32
X
(counts)
0.60
0.60
0.59
0.57
S
(counts)
0.09
0.07
0.01
0.07
Sum
(counts)
16.73
10.88
2.94
18.35
Treatments
Error
Total
SS
.01
.47
.48
ANOVA Data
df MS
3 4.71 x 10"3
79 0.01
82
F
0.80
F = 0.80 does not exceed F Q5 3 7g = 2.73.
Null hypothesis is not rejected.
Means of green reflectance for the four necrosis sets are not significantly
different.
-------
-237-
TABLE F-16. ANALYSIS OF VARIANCE FOR RED BAND BY NECROSIS SET
WINTER WHEAT
Necrosis Set
1
2
3
4
Range of
injury (%)
0-10
11-25
26-50
51-100
No. Obs.
28
18
5
32
X
(counts)
0.57
0.60
0.57
0.65
S
(counts)
0.09
0.09
0.05
0.10
Sum
(counts)
15.93
10.72
2.85
20.91
Treatments
Error
Total
SS
0.12
0.68
0.80
ANOVA Data
df
3
79
82
MS
0.04
0.01
F
4.58
F = 4.58 does exceed F Q5 3 79 = 2'73'
Null hypothesis is rejected.
Means of red reflectance for the four necrosis sets are significantly
different.
-------
-238-
TABLE F-17. ANALYSIS OF VARIANCE FOR IR BAND BY NECROSIS SET
WINTER WHEAT
Necrosis Set
1
2
3
4
Treatments
Error
Total
Range of X S Sum
injury (%) No. Obs. (counts) (counts) (counts)
0-10 28 1.68 0.34 46.97
11-25 18 1.74 0.53 31.40
26-50 5 1.84 0.27 9.21
51-100 32 1.37 0.27 43.83
ANOVA Data
SS df MS F
2.51 3 0.84 6.37
10.38 79 0.13
12.89 82
F = 6.37 does exceed F 05 3 79 = 2'73'
Null hypothesis is rejected.
Means of IR reflectance for the four necrosis sets are significantly
different.
-------
-239-
TABLE F-18. ANALYSIS OF VARIANCE FOR IR/RED RATIO BY NECROSIS SET
WINTER WHEAT
Necrosis Set
1
2
3
4
Treatments
Error
Total
Range of
injury (%) No. Obs. X
0-10 28 2.99
11-25 18 3.17
26-50 5 3.28
51-100 32 2.11
ANOVA Data
SS df
19.05 3
31.11 79
50.16 82
S Sum
0.61 83.63
0.88 53.93
0.73 16.42
0.46 67.43
MS F
6.35 15.92
0.40
F = 15.92 does
Null hypothesis
exceed F 05j3)79 = 2.73.
is rejected.
Means of the ratios for the four necrosis sets
are significantly
different.
-------
-240-
TABLE F-19. ANALYSIS OF VARIANCE FOR CROP YIELD BY NECROSIS SET
WINTER WHEAT
Necrosis Set
1
2
3
4
Treatments
Error
Total
Range of X S
injury (%) No. Obs. (kg/ha) (kg/ha)
0-10 28 3044 498
11-25 18 3128 235
26-50 5 3135 703
51-100 32 2229 606
ANOVA Data
SS df MS
14419609 3 4806536
20985118 79 265634
35404727 82
Sum
(kg/ha)
85220
56295
15674
71313
F
18.09
F = 18.09 does exceed F n, „ 7Q = 2.73.
• UD j j j / y
Null hypothesis is rejected.
Means of crop yield for the four necrosis sets are significantly
different.
-------
APPENDIX G
PHOTOMETRIC CALIBRATION
-------
-242-
Photometric Analyses Related to Soy Bean Crop
Yield and SCL Stress
Final Report Calspan Number 6258-M-l
Contract No. TV-48080A
30 May 1978
Prepared For:
The Tennessee Valley Authority
Division of Environmental Planning
Muscle Shoals, Alabama 37401
Mr. C. Daniel Sapp
Prepared By:
John E. Walker
Calspan Corporation
Advanced Technology Center
Post Office Box 400
Buffalo, New York 14225
-------
-243-
PHOTOMETRIC CALIBRATION
The atmospheric and illumination contributions to the total exposure
of an image on film are, by definition,1
(1) a multiplicative attenuation of the energy for every object in the
scene (a) and
(2) an additive contribution of energy from the optical path between the
objects and the sensor (|3).
To measure these two parameters (a and p), one must first derive the
relationship between image density (D), which is measured from the images,
and the exposing energy (E) that caused the image density. This is accom-
plished by using a step wedge (series of known exposures) that has been
imaged on the film before it is processed. TVA furnished one set of EPA
aerial photographs with a step wedge (Colbert) and two sets of TVA photo-
graphs without wedges (Johnsonville). The densities of the steps in the
Colbert wedge were measured and plotted against their known relative log
exposure values. The resultant curves are shown in Figures G-l, G-2, and
G-3 for the infrared energy (red filter), red energy (green filter), and
green energy (blue filter). The Eastman Kodak Handbook curve for CIR film
was also plotted. The comparison of these curves showed a significant
difference in the low-exposure end, with severe density compression pre-
sent in the EPA film. Examination of both the wedge and scenes under SOX
magnification revealed a microscopic pattern generally related to process-
ing to correct for underexposure during data collection. This processing
problem could manifest itself in several ways in photometric analyses if
the analyst were not aware of its presence.
The first manifestation could be in the photometric calibration pro-
cess to derive the additive contribution to image exposure (p). This task
is accomplished by performing a regression analysis of the exposures of
dark and light objects in the scene illuminated by skylight (shadow) and
also by sunlight plus skylight. Because of density compression due to pro-
cessing, the exposure range between objects illuminated by skylight and
those illuminated by skylight plus sunlight is reduced. Thus, the plot-
ting scale used in the computerized regression program to derive p must
be expanded to obtain meaningful values for 6. This actually occurred in
the first attempt to derive p for the Colbert data, with negative p's
being output by the regression program. Expanding the plotting scale,
however, did allow the derivation of meaningful p's. The expanded scale
and plot allowed the operator to recognize erroneous data points and elim-
inate them from the analysis, an important interactive step in calibration.
Another manifestation that could occur is that the density of healthy
soybean fields located in the extreme corners of the film format, especially
in the red information band (green filter-chlorophyll absorption band),
could reach the maximum density limit imposed by the processing problem;
1. Piech, R., and J. R. Schott. Atmospheric Corrections for Satellite
Water Quality Studies. Proceedings of the SPIE, 51:84-89, 1974.
-------
-244-
3.0i
o HANDBOOK
• EPA
RED FILTER
2.0
H
EQ
1.0
O O
I
I
0.0
1.0 2.0
RELATIVE LOG EXPOSURE
Figure G-l. Comparison of D-log E curves—red filter.
-------
-245-
O HANDBOOK
• EPA FILM
GREEN FILTER
3.0
o
2.0
1.0
o"
5
j
0.0 1.0 2.0
RELATIVE LOG EXPOSURE
Figure G-2. Comparison of D-log E curves—green filter.
-------
-246-
3.0
H
2.0
1.0
o HANDBOOK
• EPA FILM
BLUE FILTER
0.0
•
O
1.0
RELATIVE LOG EXPOSURE
2.0
Figure G-3. Comparison of D-log E curves—blue filter.
-------
-247-
even with the correction for lens fall off, an incorrect value of red
reflectance could result. This did not occur in the case of Colbert data,
but the analyst should be aware that processing problems can result in
maximum density compression.
A third manifestation to be aware of, because of this type of process-
ing problem, is the effect of the microscopic pattern on densitometry. If
a large-aperture (1 mm) densitometer is used, it integrates the variations
in density within the microscopic pattern anomaly with little or no effect
on the densitometry. However, if a densitometer with a small sample aper-
ture (50 |Jm) is used, the placement of the aperture is critical. The pro-
cessing pattern consists of circular areas, under 100 pm in diameter, over
the entire film format. In the center of each circular area of normal dye
concentration is an abnormal "snowflake" pattern that is generally of higher
density than the normal dye concentration surrounding the "snowflake."
Therefore, the analyst must be sure he places the small aperture in the
normal dye area and not on a "snowflake" when making an image density
measurement.
Fortunately, the Colbert photographs were acquired in late August,
when most soybean crops had reached full cover condition. Therefore, it
was possible to use the 1-mm aperture and have the effect of the process-
ing problem averaged into the noise of the overall measurement process.
The a's and P's derived for the Colbert photographs are shown in
Table G-l.
TABLE G-l. a'S AND p'S FOR COLBERT PHOTOGRAPHS OF AUGUST 29, 1977
Parameter Infrared band Red band Green band
Of 206.92
p 6.72
157.14
3.45
174.70
10.37
For the Johnsonville photographs it was necessary to assume charac-
teristic curves to obtain a's and (3's. The Johnsonville films were exam-
ined under 30X magnification to determine whether the microscopic pattern
present in the Colbert wedge and scenes was still present. The pattern
was not present in either the July 13, 1977, films furnished by EPA to TVA
or in the August 2, 1977, TVA films. The maximum densities of these films
were checked (i.e., zero exposure in borders), with the expectation that
they would be significantly higher. However, because the maximum densi-
ties were not found to be significantly higher, we decided to use the
original wedge data for the Colbert scene to derive a's and P's for
Johnsonville. The a's and (3's for the two sets of Johnsonville films are
given in Table G-2.
-------
-248-
TABLE G-2. a'S AND p'S FOR JOHNSONVILLE PHOTOGRAPHS
Parameter Infrared band Red band Green band
July 13, 1977
a 322.588 130.486 138.611
P 13.39 3.21 4.27
August 2, 1977
a
P
329.139
7.15
214.169
3.28
219.411
3.45
The assumptions necessary to obtain a photometric calibration of
these films were pointed out to the TVA representative. The necessity for
the exposure of a 21-step wedge on the original aerial films and a dupli-
cate of this wedge for subsequent duplicates of scenes to be analyzed in
any future field experiment was also emphasized. This requirement was the
first experimental design criterion resulting from this effort.
Because of the problems of sensitometric and densitometric control
for the aerial photographs, absolute spectral reflectances and reflectance
ratios could not be obtained from photometric interpretation; however,
important relative reflectance information within any one scene could still
be obtained, depending on the level of reflectance. The film covering the
Colbert site (frames 5353 through 5355, August 29, 1977) was selected for
analysis because it did have a wedge. Thus, these reflectances would be
closest to absolute values.
The TVA representative measured the spectral reflectance and reflect-
ance ratio properties of soybean plants in five fields at the Colbert site
(fields A through E). The number of measurements per field was limited by
the field of view of the densitometer used. This was a 1-mm-diam aperture,
representing a 12-m-diam area on the ground. A 0.6-m-diam area could have
been used, but this was not considered necessary for these fields because
they were mature fields with very full cover. The 12-m-diam area still
allowed about 50 samples to be measured per field in less than 15 min.
This is considered a very important point relative to the comparison
of any potential aerial photometric method of soybean stress measurement
to a ground survey method. It would be very difficult to obtain a sample
rate of 200 samples/h per field when using a ground survey method, whether
they are spectral reflectances from a ground-based TSR or L x A x P (see
Appendix D) estimates. This is a very practical advantage of an aerial
photographic approach. Furthermore, the photograph would provide a valua-
ble permanent record of the crop condition.
The TVA representative compared the mean IR/R reflectance ratios for
these fields and found the most injured (4 percent by the L x A x P method)
plants had a ratio of 3.010.49, whereas the lease injured (0 percent)
-------
-249-
plants had a ratio of 5.7±0.34. In previous photometric interpretation
analyses of vegetation stress (chlorosis-necrosis) by Calspan, the mean
IR/R reflectance ratio has always been high when stress was low and low
when stress was high. Thus, even with known processing problems, the
reflectance and ratio results obtained were consistent with previous
studies.
However, one inconsistency was noted. The standard deviation in the
red (chlorophyll absorption) band is usually extremely large in comparison
with that in the infrared band. These data showed the infrared band to
have the larger standard deviation. The presence of a sunspot image in
the Colbert scene could have caused these inconsistencies.
Field B, having 2 percent injury by the L x A x P method, appeared
on three frames (nos. 5353, 5354, and 5355) at different format positions
relative to the sunspot and center of the format. Spectral reflectance
gradient functions from the center of the sunspot image through the center
of each format to the edge of the film away from the sunspot were derived
from regressions of soybean reflectances against distance from the center
format. Deciduous tree reflectance was also regressed relative to distance
with almost identical results. This result would be expected since both
canopies are highly textured and contain multiple reflectances .
A second-order regression resulted in a functional relation with high
correlation coefficients (r2):
F = B2X2 + BiX + B0,
where X = the distance of an image from the center of the film format (0)
to (+) the sunspot and away from (-) the sunspot, in inches along the
sunspot, center format line.
The constants BO, BI, and 62 are shown in Table G-3, with the correla
tion coefficients for each spectral band.
TABLE G-3. SPECTRAL REFLECTANCE GRADIENT CONSTANTS
Constants
R
IR
R.
R
BO
Bl
BS
r"
0.0248
0.0139961
0.001159
0.728
0.0575
0.008437
0.0007353
0.879
0.0539
0.0076
0.0008
0.865
Eight to nine spectral reflectance measurements were made in field B
on each of the three frames on which the field was imaged. The results
of the measurements are shown in Table G-4.
-------
-250-
TABLE G-4. MEASUREMENTS OF REFLECTANCE--FIELD B
Frame
no.
5353
5354
5355
All data
P. rr/'o/'\a p n("l\^ P
IR au°J KR CTU; KG
29.26 ±7.5 05.65 ±3.4 06.14
28.89 ±3.0 05.76 ±6.0 05.83
24.88 ±5.0 04.51 ±4.0 04.38
27.5 ±9 05.14 ±5.3 05.38
a(%)a
±4.1
±7.0
±5.0
±15.8
RIR/RR
5.2
5.4
5.5
5.4
a(%)a
±8
±4
±2
±6
o
Standard
deviation, as percentage of mean reflectance.
Next a preliminary gradient correction method
Calspan was applied to the spectral measurements.
ance data for the field are shown in Table G-5.
TABLE
under development at
The corrected reflect-
G-5. REFLECTANCES CORRECTED FOR ILLUMINATION
GRADIENT
Frame
no.
5353
5354
5355
All data
RIR CT(%) RR
-------
-251-
TABLE G-6. COMPARISON OF PHOTOMETRIC INTERPRETATION ACCURACY
WITH AND WITHOUT PRELIMINARY ILLUMINATION GRADIENT CORRECTION
Frame
no.
5353
5354
5355
Accuracy in *
With
0.3
2
1
f - RA
RIR(%)
V
Without With
6
5
10
(all data)
RA (all
5
4
9
- RA (field) „
data)
0
Without
10
12
12
100
With
2
3
4
RG(%)
Without
12
8
18
-------
-252-
SPECTRAL REFLECTANCES OF SOYBEAN FIELD COMPONENTS
Being acquired about four weeks earlier than the Colbert (8/29/77)
films, the Johnsonville (7/13/77) films had many immature, partially
covered fields and few mature, full cover fields. This necessitated the
use of the 0.6-m diam ground sample size densitometer because too much
soil and herbaceous cover were integrated into the 12-m diam samples.
Reflectance measurements of tilled soils, untilled soils, and weeds were
made as well as measurements of soybeans per se. These are other major
components of a less than full cover (immature) soybean field.
Table G-7 shows the reflectance and reflectance ratio data for five
tilled fields (visually white), as well as data for six untilled fields,
(visually cyan).
TABLE G-7. JOHNSONVILLE FIELD REFLECTANCE RESULTS
Field No.
RIR RR RG RIR/RR
RR/RG
RIR/RG
R
Visually White Fields (Tilled)
106.10+
110.02
201.08
115.08
110.01
Mean/
Std.Dev.
36.3
42.7
36.7
32.2
38.8
37.3
±3.8
30.9
37.1
23.7
25.7
33.0
20.9
24.7
18.6
14.5
22.5
,18
,52
,55
1.47
.50
.27
30.1 20.2
±5.4 ±3.9
1.25
1.18
1.3
±0.2
1.78
1.47
1.5
±0.2
1.74
1.73
1.97
2.22
1.72
1.9
±0.2
29.
34.8*
26.3**
24.1**
31.4
Visually Cyan Fields (Untilled)
100.07+
105.09+
107.01
107.02
108.01
108.02
Mean/
Std. Dev.
21.1
18.8
19.7
18.5
22.7
22.8
20.6
±1.9
16.8
18.3
15.5
14.7
17.1
20.3
17.1
+2.0
13.5
14.1
12.8
12.6
14.1
15.3
13.7
±1.0
1.26
1.03
1.27
1.26
1.33
1.12
1.2
±0.1
1.25
1.30
1.21
1.17
1.21
1.33
1.3
±0.1
1.57
1.34
1.54
1.47
1.61
1.49
1.5
±0.1
17.1**
17.1
16**
15.3
18**
19.5*
+A11 more than 5" from sunspot, all others less than 5" from sunspot.
*Brightest overall.
**Darker but with higher long to short wavelength ratio.
It is quite apparent that tilled and untilled soils are as easily
separable on the basis of these spectral reflectance properties, as they
are by visual color discrimination. Applying the illumination gradient
correction will only result in a slight increase in reflectances of fields
marked with the symbol (+) and slight decreases for those not so marked.
-------
-253-
This should not have a significant effect upon the differences between the
mean values shown in Table G-7.
These data all exhibit the normal characteristics of soil reflectance
ratios, i.e., long to short wavelengths, are greater than 1.0. It is possi-
ble to demonstrate a photometric method to assess moisture variations in
the fields based upon average brightness R and the RTR/Rp ratio of these
data. Field 110.02 is the brightest of the tilled fields. If moisture is
present it will cause a lower overall brightness, and an increase in the
infrared to green ratio. If the soil brightness decreases and the ratio
does also, then it is some other parameter that is causing the difference.
From the data of Table G-7 it appears that practically all of the variation
in tilled soil areas (3 out of 4 fields) is due to moisture whereas in the
untilled fields only 3 out of 5 appear to vary due to moisture. The soy-
beans in tilled fields are visually red whereas those in untilled fields
are visually pink. This alone suggests that tilling which reduces moisture
retention in fields may also reduce stress on soybean plants, or cause
stress if not practiced. The soil results suggest that in future TVA
efforts to assess injury to soybeans, an experiment should be conducted to
assess the effects of soil moisture levels in tilled and untilled fields
upon the soybean growth rate, reflectance properties, and yield. This is
the second experimental design criterion to be derived from this brief
photometric analysis of TVA's CIR films.
Seven fields (100, 101, 102, 103, 104, 105, and 106) were selected
for analysis at the Johnsonville site. These fields are all in close
proximity to each other and therefore relative reflectance data would be
valid without correction for the sunspot illumination gradient effect.
Data was acquired by the TVA representative from 11 other fields as well,
but most were located such that the gradient correction would have to be
applied for a comparison to the results from these fields. Field 100 and
101 are "double planted" fields. Field 106 is also about 12 weeks old
compared to 5 weeks for the other fields. Field 106 is a tilled field
whereas the rest are not tilled.
In reviewing the reflectance measurements acquired by the TVA repre-
sentative for Field 101, it was noted that five sets of readings were
identified as "weeds" and had been included in the computations for the
mean values of soybean reflectances and reflectance ratios in the field.
There were sufficient samples of both weeds and soybeans to derive a mean
value and standard deviation of reflectance and reflectance ratios for
each. These values are shown in Table G-8 along with the original values
for weeds plus soybeans and the values for the 11-week old Field 106.
TABLE G-8. SOYBEANS AND WEEDS IN FIELD 101
Vegetation
RIR
RR
RG
RIR/RR
Soybeans 23.214.5 6.510.9 7.511.3 3.610.73
Weeds 19.011.0 3.210.8 4.010.7 6.311.6
Weeds + Soybeans 22.2+4.3 5.4411.82 6.6811.55 4.55511.787
Mature Soybeans (106) 22.711.0 4.610.8 4.710.6 5.61.95
-------
-254-
The data of Table G-8 is interesting in several respects. First, a
visual check on the image of Field 101 by three interpreters results in
100 percent agreement that visually it was impossible to separate weeds
from soybeans by the visual assessment of color in this image. Therefore,
we suspect the representative "identified" weeds from a priori knowledge
of their location and presence from his field observations. However, there
is a definite signature for weeds if all facts are considered.
First, this is an untilled field, i.e., cyan-colored soil image.
Second, it is a field with immature soybean growth, i.e., row effects are
clearly defined in most of the field. Third, it is unlikely that under
these conditions, mature soybeans will be present anywhere in the field,
although the spectral properties of mature soybeans (106) are very similar
to the spectral properties of areas identified as weeds by the representa-
tive. Fourth, in view of the surficial drainage pattern of this field and
the lack of tilling, it would be feasible for weeds having a preference
for high moisture to develop to maturity long before soybeans which prefer
lower moisture levels. Fifth, weeds would not develop the stacked leaf
canopy of soybeans and would therefore have lower infrared reflectance
than soybeans at maturity although their visible light reflectance proper-
ties would be nearly identical. Sixth, the very high mean infrared-to-red
reflectance ratio of the identified weed areas (6.3±1.6) indicates a
significantly higher vegetal biomass is present at these locations that
at the identified soybean location where the mean ratio is 3.6±0.7.
Applying these facts to Field 101, and generating an infrared-to-red
ratio mask or sampling the reflectance properties of the vegetation in the
field on a grid basis, the weed areas can be mapped from the growing soy-
bean areas. This analysis illustrates the need for ancillary data to
develop an effective photometric measure of SC>2 stress on soybeans. Speci-
fically, for every field to be assessed, a preplanting, postpreparation-
for-planting photograph should be acquired after a period of rainfall to
assess the moisture pattern distribution. In view of the apparent impor-
tance of soil moisture retention related to soybean and weed conditions
it is recommended that an experimental effort be undertaken by TVA to
assess the full significance of moisture retention.
-------
APPENDIX H
NASA/ERL LETTER REPORT
-------
-256-
Lyndon B. Johnson Space Center ft |/\CH/\
Earth Resources Laboratory National
1010 Gause Boulevard Aeronautics arid
Slidell, Louisiana 70458 Space
Administration
ARTETPNYOTF° GC May 10, 1976
Mr. Herbert C. Jones
Supervisor, Air Quality Research
Tennessee Valley Authority
Muscle Shoals, Alabama 35660
Dear Mr. Jones:
We have analyzed the multispectral data taken in the vicinity of the
Joppa and Shawnee power plants near Paducah, Kentucky. The results of
this analysis indicate that spectral differences due to vegetation stress
associated with SOz plumes from power plants to be too subtle to be
resolved.
Examination of the excellent ground truth information supplied by your
organization revealed that the estimate of total leaf area affected in
a given field was low—usually less than five percent. When this is
considered, together with the percentage of ground cover usually around
75 percent, the percent of affected leaves in the field of view of the
scanner is very small. Thus, it is not surprising that areas of affected
leaves were not identifiable from the scanner data.
I do not believe that it would be accurate to extrapolate from this
particular set of data that the problem of detecting stressed vegetation
due to S02 from multispectral scanner data is generally untenable; however,
for this specific set of data this appears to be the case. For instance,
it would seem much more likely that stressed vegetation might be detected
in an area comprised of one crop and cultivated using similar farming
practices. This situation is quite different than existed here where there
were several crop types cultivated using diverse farming practices and at
different stages of growth. With this situation, there is much variability
in the scene causing the slight change in spectral quality of the data due
to S02 vegetation stress to essentially be in the "noise level" of the data.
If you have any questions in regard to the analysis of the data, please
contact me and we will be happy to furnish you with additional information.
E. L. TILTON, III
Assistant Director
-------
APPENDIX I
SUPERVISED CLASSIFICATION PROCEDURE
-------
-258-
SUPERVISED CLASSIFICATION OF MSS DATA
When preprocessing of the digital tapes is complete, the analyst may
implement the supervised classification procedure. The first step is
training sample selection and statistics calculation, second is divergence
analysis, and the third is the classification of the scene.
Training sample selection begins with an examination of ground truth
and collateral sources such as photography and maps to obtain information
on candidate sites for training samples. These samples are areas having
apparently homogeneous physical characteristics. They are outlined as
polygons on the display monitor using a trackball and cursor. Training
samples may be grouped as land cover classes. At least three training
samples per class are normally obtained. Statistics are computed to
determine the homogeneity of the training samples. Histograms are plotted
(Figure 1-1) to ascertain that the samples are normally distributed with
no biomodal tendencies or skips in the data. Coefficients of variation
are also checked.
Since the selection task is interactive, unacceptable samples may be
replaced with new ones and the appropriate data stored upon completion.
The divergence analysis step examines spectral separability between
classes. Inseparable classes may be grouped if this is considered advan-
tageous. When this step is completed, we are ready to classify the scene.
Supervised classification involves computation of n-dimensional (n =
number of channels) decision boundaries and assignment of pixels to classes
or categories. Hyperellipsoidal decision boundaries are computed from the
class statistics with a program called ELLIPSE. A second program, ASSIGN,
assigns pixels to classes or an unclassified category. The assigning pro-
cess is iterated for each pixel and scan line until the entire scene has
been classified. A classified data tape is the product. This tape can
serve as input to the display monitor for viewing and film recorder for
hard copy.
-------
CHW. 1 * « 8 PTS. 250
240
184
176
168
160
152
144
136
128
120
112
104
96
88
80
72
64
56
48
40
32
24
4 230
6
i
it
*
*
t
*
*
16 I *
81 11 *
aao
210
200
I9O
ISO
170
160
150
140
130
120
3 110
' 100
* »0
» BO
» 70
* 60
» 50
* 40
» 30
* 20
•2 10
375
264
253
242
231
220
209
1*8
187
176
165
154
143
132
121
110
99
88
77
66
55
44
33
22
1 11
336
384
30 B
294
280
266
252
238
224
210
1*6
182
168
154
140
126
1 12
98
84
70
56
42
D CHN. 5 ' * 5 PTS.
as i
14 I A
1101 2
105
100
95
90
85
80
75
70
65
60
55
50
45
40
35
30
25
20
15
*
*
*
»
*
3»
»*
3
*
*
*
*
10 I 2*
>
1
M
*
* 3
*4 *
3 43 83 3 43 S3 3 43 83 3 43 83 3 43 F3 123 163
FREQUEHCT
ct
84
80
'6
72
68
64
60
56
52
48
44
40
36
32
28
24
20
16
12
IN. 6 * * 4 PTS.
*
•
*
-
3 43 83 123
*
*
» 2
*
#*
**
**
**
+ *
**
**
**
**
**
**
1
150
144
138
132
126
120
1 14
108
102
96
90
84
78
72
66
60
54
1 48
* 42
» 36
* 30
* 24
• 11 18
>3 203
2 C
»
* 138
* 132
1 • 126
** 120
»» 114
<» 108
»« 102
»* 96
»* 90
** 84
»* 78
»« 72
«» 66
•* 60
•* 54
«» 48
»*3 42
»» 36
«» 30
• t 24
5*» 1 18
* ** * 12
3 ##* ** 6
3 43 83 123
W. B * = 6 PT5
A
*
3 43 P
CLASS NO. | SN
.
5
3 123
HISTOGRAMS FOR SN1042
FREQUENCY
Figure 1-1. Selected histograms describing homogeneity of a training sample used in supervised
classification of MSS data. This is unaffected soybeans.
-------
-260-
I ft (''• f: »,' f
i ^£ 3; ?l 3'' £*
L II ?^ t!- i? L
f " " -"• >' ' =-; ;? ; f* f-
* ? * i * 5 s 1 \
:.. •- I *, i •, . -, J 0 a
SYSTEMS AND SERVICES DIVISION
P. O. BOX 15027 D LAS VEGAS. NEVADA 89114 a TELEPHONE (AREA CODE 702) 736-2969, EXT. 271
April 25, 1978
Environmental Protection Agency
P. O. Box 15027
Las Vegas, Nevada 89114
Attention: G. A. Shelton
Subject; Supervised Classification of TVA Scenes, Project 7736
Dear Mr. Shelton:
This letter is in response to Dan Sapp's request. The paragraph
below is an explanation of the inability to perform supervised
classifications on TVA Tapes 319 and 320.
Training fields were selected when Mr. Sapp was here at EMSL-LV.
By error, interactive stats were run on the wrong data file and
output indicated that all classes were separable. This error was
discovered after Mr. Sapp's visit and stats were run on the correct
file. Results indicated that the classes were not separable.
Other training fields were selected by myself and again by Chuck
Tanner. However, in both cases, the stats run indicated non-
separability of classes.
Very truly yours,
David R. Williams
Environmental Analyst, Sr.
Remote Sensing Laboratory
DRW:mb
cc: J. R. Donaldson
H. V. Johnson
J. O. 11.01
A SUBSIDIARY OF LOCKHEED AIRCRAFT CORPORATION
-------
APPENDIX J
PSEUDOSUPERVISED CLASSIFICATION OF MSS DATA
-------
-262-
PSEUDOSUPERVISED CLASSIFICATION OF MSS DATA
GENERAL
The pseudosupervised classifier is an alternative to the time-consuming
supervised classifier. The pseudosupervised classifier implemented at
EMSL-LV is entitled MAXL4. It uses a fast clustering program in combination
with a limited amount of manual selection of training samples. The basis
of the procedure is a trainer called SEARCH.58
SEARCH
This program is an unsupervised trainer for a maximum likelihood
classifier that is unlike conventional cluster analysis. SEARCH evalu-
ates contiguous 6-scan line by 6-element blocks of 4-channel input data
through use of a covariance matrix that permits comparison of relative
variability about a different sized mean. Alternatively, a 3-scan line
by 3-element block may be evaluated. The output of SEARCH is means,
standard deviations, and covariances for the 4-channel data for each
training sample. These statistics are stored and the pair of statistics
with the smallest pairwise divergence is determined and merged reitera-
tively until the minimum divergence exceeds a specified value.
SEARCH can be run from either the background or foreground (interac-
tive) mode. The interactive mode permits the analyst to change program
parameters in real time. The interactive SEARCH program requires a sta-
tistics (STAT) file and a 4-channel data tape.
Classification Processor
A maximum likelihood scheme is used to classify the data with MAXL4.
This scheme assumes the samples within a given class (e.g., soybeans) are
distributed according to a normal multivariate probability density function.
The MAXL4 classifier begins with statistics obtained from the program
SEARCH. The mean vector M and the covariance matrix C are estimated.
MAXL4 classified each vector X, where X = (Xi, X2, Xs, X,j) as belonging
to one of NCL classes (NCL 5 63) or, under certain conditions, as "other."
Since MAXL4 can process only 4 channels of MSS data, the mathematics
involved is not complex. Baye's Rule is used to assign pixels to classes.
Baye's Rule as used by Pearson of NASA/ERL is as follows:
P(X i) • APR.
Pd x) = - P(XJ 1 (i)
-------
-263-
where
P(i X) - probability of class i given the occurrence of X.
P(X i) = probability of X given the occurrence of class i.
APR. = a_ priori probability of class i.
P(X) = Probability of X.
For the 4-channel case:
P(X i) =
- Mi)
t
- ^.
(27t)2 Dj
- Mi)
where
D. = determinant of matrix C.
_1
C. = inverse of C.
1 1
For a given vector X (P(X) = 1):
i X) =
,
Dj
n e
- Mi)
(27T)2
MAXL4 assigns vector X to class j if:
P(j X) > (P(i X)
(4)
where
l
-------
APPENDIX K
UNSUPERVISED CLASSIFICATION PROCEDURE
-------
-265-
UNSUPERVISED CLASSIFICATION PROCEDURE
PROGRAM PURPOSE
This program was developed by NASA/ERL to afford users the capability
to peruse training field areas as subsets of larger areas.59 The user
selects a general region of interest of which his training field is a sub-
set and the program classifies this region and determines homogeneity of
the training field.
DETAILED PROGRAM DESCRIPTION
Technique
The existing programs that produce channel composite images are
restricted by the arbitrariness in selection of bin edges in the grey
scale, the compression factor that results in a 36 to 1 count resolution
reduction for 255 count range aircraft data for a 2-channel gray scale,
and the loss of coherence information between channels. The unsupervised
sequential clustering technique program (UNSUP) was developed to alleviate
some of these restrictions and to facilitate users in selecting homogeneous
training fields. The major assumption in this program is that the data
has some homogeneous patterns. The data flow consists of the following:
(1) establishing the initial population, (2) developing tables for classes,
(3) determining if the sample belongs to established populations, and
(4) establishing additional populations.
Mathematics
The initial population is established in the following manner:
NVARBS
X(j,k) = NVARBS I l/XDATA(i,j)
(1)
where
l
-------
-266-
If this test holds for all channels (NCHAN) and variables (NVARBS),
it is accepted for ki initial population.
The program then generates tables for the NCHAN channels of the ki
population as follows:
LSTART = XBAR(J!,Ki) - (a(jik1)xPRBLTY) 1 < LSTART < 256
LEND = XBAR(ji,ki) - (a(j!,ki(XPRBLTY) 1 < LEND < 256
PTABLE (i,J!) = ki (3)
for LSTART
-------
-267-
XBARQi.lOjjpopj.
ZDATA(ji,m,n) (6)
\ / J.1J- Wi. J_l • J. /
where
NPOPL = the number of pixels in existing population statistics.
The PTABLE( ) describing population count ranges is updated as
described in eq. 3.
If the pixel does not fit any of the established populations since
the k of equation 4 equals zero, then this pixel is stored in XDATA(i,j)
array until i reaches the value of NVARBS. At this time it is evaluated
the same as the initial population was in equations 1 and 2 to test its
validity in establishing a new population.
-------
-268-
UPGRADIWG AND ENHANCEMENT OF UNSUP PROGRAM SOFTWARE FOR VARIAN V-75 COMPUTER
Purpose
Enhancements to UNSUP establish the relationship between classes,
select the optimal four channels for subsequent supervised data classifi-
cation, and reassign classes based upon a selected channel queueing. The
enhancements provided include: (1) feature selections, (2) class separa-
bility determination, and (3) reassignment of classes to match the
standardized color table (chip).
General Description of Upgrading and Enhancements
The existing UNSUP program is a multichannel sequential clustering
program. The channels 1 through 12 designate the channel bandwidth of the
Daedulus 1260 scanner. The UNSUP program determines the validity of an
initial population l>y the coefficient of variation. Upon establishing the
initial population, proceeding data elements are tested against a specified
Chi-Squared confidence. Any samples that satisfy this Chi-Squared confi-
dence interval are designated to be in that class. The elements that
exceed this confidence interval are stored in a buffer until the designated
number elements are accumulated for testing by coefficient of variation.
After complete processing of the input data set, an output is generated
consisting of the program-determined classes. These class statistics are
processed by a similarity criterion technique for selecting the four
optimum channels.
Average normalized distance separations for all four channel subsets
of the channels are computed. The four-channel subset with the largest
numerical average normalized distance separation is optimum. This proce-
dure compares each class with the other classes and also compares all class
combinations.
The upgrading also includes an automated color chip generation feature
for displaying classified outputs.
Processing
The upgrade to the UNSUP program is implemented through the following
equations:
(XBARi(K) - XBARj(K)
D(i,j) = 2
{STUDEVi(K)}{STUDEVj(K)}
where
i and j represent the classes being compared for similarity.
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
REPORT NO.
EPA-600/7-81-I14
2.
3. RECIPIENT'S ACCESSION NO.
TITLE AND SUBTITLE
Remote Sensing of Sulfur Dioxide Effects on Vegetation -
Final Report - Volume II - Data
5. REPORT DATE
July 1981
6. PERFORMING ORGANIZATION CODE
AUTHOR(S)
8. PERFORMING ORGANIZATION REPORT NO.
C. Daniel Sapp
TVA/ONR/ARP-81/6
. PERFORMING ORGANIZATION NAME AND ADDRESS
Office of Natural Resources
Tennessee Valley Authority
Norris, TN 37828
10. PROGRAM ELEMENT NO.
JLHE_625£L
11. CONTRACT/GRANT NO.
80 BDJ
2. SPONSORING AGENCY NAME AND ADDRESS
U.S. Environmental Protection Agency
Office of Research and Development
Office of Energy, Minerals, and Industry
Washington. D.C. 20460
13. TYPE OF REPORT AND PERIOD COVERED
Final 1976-1980
14. SPONSORING AGENCY CODE
5. SUPPLEMENTARY NOTES
This project is part of the EPA-planned and -coordinated Federal Interagency Energy/
Environmental R&D Program.
e. ABSTRACTThree techniques for detecting and mapping sulfur dioxide (S02) effects on
the foliage of sensitive crops and trees near large, coal-fired power plants were teste
and evaluated. These techniques were spectroradiometry, photometric analysis of aerial
photographs, and computer analysis of airborne multispectral scanner data.
Spectroradiometry is a useful, ground-based technique for measuring the changes in
reflectance that accompany exposure of sensitive crops to S02« Photometric analysis of
aerial color-infrared photographs has some practical advantages for measuring the
reflectances of forest species or for synoptic point-sampling of extensive areas; these
tasks cannot be done effectively by field crews. The relationships among reflectance,
foliar injury, and yield of crops are complex and are affected by many extraneous vari-
ables such as canopy density. The S0£ effects are easier to detect on winter wheat
than on soybeans, but in either case they cannot be consistently detected by airborne
remote sensors except under near-ideal conditions when the injury is moderate to severe
Airborne multispectral scanner data covering affected soybean fields were analyzed
using three computer-assisted procedures: unsupervised, supervised, and pseudosuper-
vised; the last method provided the best results. Landsat imagery was also investigated
but the foliar effects of S02 were too subtle to detect from orbit.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
Air pollution *
Electric power plants
Photointerpretation
Remote sensing *
Environmental surveys *
Infrared photography
Photometry
Reflectance
Sulfur dioxide *
Plant pathology
b.lDENTIFIERS/OPEN ENDED TERMS
Transport processes
Char., meas. & monit.
Crop & forest species
Digital image analysis
Multispectral scanning
Microdensitometry
Tennessee Valley
cos AT l Field/Group
18. DISTRIBUTION STATEMENT
Release to public
19. SECURITY CLASS (This Report}
Unclassified
21. NO. OF PAGES
280
20. SECURITY CLASS (This page)
Unclassified
22. PRICE
EPA Form 2220-1 (Rev. 4-77) PREVIOUS EDITION is OBSOLETE
------- |