&EPA
United States
Environmental Protection
Agency
Environmental Monitoring
Systems Laboratory
P.O. Box 15027
Las Vegas NV 89114
EPA-600/3-79-123
December 1979
Research and Development
Trophic Classification
of Selected Illinois
Water Bodies:
Lake Classification
Through Amalgamation
of LANDSAT Multispectral
Scanner and Contact-
Sensed Data
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development, U.S. Environmental
Protection Ageney, have been grouped into nine series. These nine broad categories
were established to facilitate further development and application of environmental
technology. Elimination of traditional grouping was consciously planned to foster
technology transfer and 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 ECOLOGICAL RESEARCH series. This series
describes research on the effects of pollution on humans.plant and animal species, and
materials. Problems are assessed for their long-and short-term influences. Investiga-
tions include formations, transport, and pathway studies to determine the fate of
pollutants and their effects. This work provided the technical basis for setting standards
to minimize undesirable changes in living organisms in the aquatic, terrestrial, and
atmospheric environments.
This document is available to the public through the National Technical Information
Service, Springfield, Virginia 22161
-------
EPA-600/3-79-123
December 1979
TROPHIC CLASSIFICATION OF SELECTED
ILLINOIS WATER BODIES
Lake Classification Through
Amalgamation of LANDSAT Multispectral
Scanner and Contact-Sensed Data
by
D. H. P. Boland
Advanced Monitoring Systems Division
Environmental Monitoring Systems Laboratory
Las Vegas, Nevada 89114
David J. Schaeffer and Donna F. Sefton
Division of Water Pollution Control
Illinois Environmental Protection Agency
Springfield, Illinois 62706
Robert P. Clarke
Environmental Programs
Illinois Environmental Protection Agency
Springfield, Illinois 62706
Richard J. Blackwell
Earth Resources Applications Group
Jet Propulsion Laboratory
Pasadena, California 91103
ENVIRONMENTAL MONITORING SYSTEMS LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
LAS VEGAS, NEVADA 89114
-------
DISCLAIMER
This report has been reviewed by the Environmental Monitoring Systems
Laboratory-Las Vegas, U.S. Environmental Protection Agency, and approved for
publication. Mention of trade names or commercial products does not con-
stitute endorsement or recommendation for use.
n
-------
FOREWORD
Protection of the environment requires effective regulatory actions that
are based on sound technical and scientific information. This information
must include the quantitative description and linking of pollutant sources,
transport mechanisms, interactions, and resulting effects on man and his
environment. Because of the complexities involved^ assessment of specific
pollutants in the environment requires a total systems approach that
transcends the media of air, water, and land. The Environmental Monitoring
and Support Laboratory-Las Vegas contributes to the formation and enhancement
of a sound monitoring data base for exposure assessment through programs
designed to:
• develop and optimize systems and strategies for monitoring
pollutants and their impact on the environment
• demonstrate new monitoring systems and technologies by
applying them to fulfill special monitoring needs of the
Agency's operating programs
This report describes in detail a project utilizing a combination of
LANDSAT-1 multispectral scanner data and contact-sensed data for the
classification of selected lakes and artificial reservoirs in Illinois. It
provides trophic rankings and classifications for 145 Illinois water bodies
based on trophic indicator and multivariate trophic index estimates derived,
in part, from satellite-acquired data. The information will be used by the
Illinois Environmental Protection Agency in one segment of its program to
meet the Federal mandate requiring the State to identify and classify,
according to trophic condition, all publicly-owned freshwater lakes (Public
Law 92-500, Section 314). The report will also provide useful information to
other governmental agencies and organizations in the private sector that are
considering the use or are already using LANDSAT in a lacustrine trophic
classification program. Further information on this subject can be obtained
from the Advanced Monitoring Systems Division.
George B. Morgan
Director
Environmental Monitoring Systems Laboratory
Las Vegas
m
-------
SUMMARY
This report describes the specific techniques utilized to generate a
satellite-based classification of 145 Illinois lakes and interprets and
disseminates the resultant products and ancillary information. The report
represents one segment of an ongoing effort of the Illinois Environmental
Protection Agency (IEPA) to meet the mandates of Public Law 92-500, the
Federal Water Pollution Control Act Amendments of 1972 and Public Law 95-217,
the Clean Water Act of 1977. Under Section 314-a of both laws, the
multifaceted legislation requires that each State prepare or establish, and
submit to the EPA Administrator for his approval:
"(1) an identification and classification according to eutrophic
condition of all publicly-owned fresh water (sic) lakes in such State-;
(2) procedures, processes, and methods (including land use
requirements), to control sources of pollution of such lakes; and
(3) methods and procedures in conjunction with appropriate Federal
Agencies to restore the quality of such lakes."
The State of Illinois has over 80,000 impoundments, of which 2,700 have
surface areas of 2.4 ha or more and are classified as lakes by the Illinois
Department of Conservation. In light of the Federal mandate and the
magnitude of the task (Illinois has about 775 publicly owned lakes), the IEPA
elected to investigate new approaches to the lake classification problem. To
receive serious consideration, an approach had to have the potential of being
cost effective and rapid and of yielding results of practical value.
The approach described in this report employs a combination of
satellite-acquired and contact-sensed data along with multivariate
statistical techniques to classify a group of Illinois lakes. The remote
sensor under consideration is the multispectral scanner (MSS) on board NASA's
LANDSAT-1.
LANDSAT MSS data acquired October 14-16, 1973, for 145 Illinois water
bodies were extracted from computer-compatible tapes (CCT's) using a digital
image-processing system at NASA's Jet Propulsion Laboratory (JPL). Counts of
picture elements (pixels), the MSS's basic unit of spatial resolution, were
transformed to lake surface area estimates. LANDSAT MSS digital number (DN)
mean values for each spectral band were adjusted to a common date using
regression analysis. The date-adjusted MSS data were then examined for the
existence of natural groups or clusters by applying a complete linkage
clustering algorithm to the four LANDSAT-derived spectral measurements made
on each lake.
IV
-------
Two multivariate trophic indices were developed through principal
component analyses of five trophic indicators (chlorophyll a_, inverse of
Secchi depth, total phosphorus, conductivity, and total organic nitrogen)
measured in 1973 by the U.S. Environmental Protection Agency's (EPA) National
Eutrophication Survey (NES). Next, three sets of regression models were
developed from the contact-sensed and remotely sensed data for 22 NES-sampled
lakes. Four trophic indicators (chlorophyll a^ inverse of Secchi depth,
total organic nitrogen, and total phosphorus) and the two multivariate
indices were treated as dependent variables; the four LANDSAT MSS bands
(including standardized and transformed versions) were utilized as
independent variables. The three sets of regression models were then
extended to the remaining 123 lakes in the study group. Next^ the 145 lakes
were ranked by each of the four trophic indicators, a composite trophic rank
parameter, and each of the two multivariate trophic indices using the
LANDSAT-derived regression model estimates. In addition, complete linkage
clustering algorithms were employed to delineate lake groups or clusters
using the estimated values of four trophic indicators (Secchi depth,
chlorophyll j*, total organic nitrogen, and total phosphorus) for the 145
lakes. Interpretation and validation of the classification results were
augmented by a lake water quality data base acquired by the State of Illinois
in its 208 planning effort in 1977 and ancillary data from other sources.
The analyses of LANDSAT multispectral scanner (MSS) and near-concurrent
contact-sensed data and ancillary information for 145 Illinois water bodies
indicate that lake clusters can be derived from LANDSAT MSS raw data and
MSS-estimated trophic indicator values. Each cluster is distinctive and
identifiable in terms of general water quality, use impairment, and lake
characteristics.
MSS data can also be used with contact-sensed data to develop regression
models to provide relative estimates of Secchi depth, chlorophyll a_, total
organic nitrogen, total phosphorus, and two multivariate trophic state
indices. Although less accurate and precise than contact-sensed trophic
indicator values, the LANDSAT-derived parameter estimates can be used to
develop generalized rankings of Illinois lakes. Regression models developed
from MSS and contact-sensed data for lakes with large parameter value ranges
and minimal contact-sensed data were more effective for the Illinois lakes
studied when spectral ranks rather than normalized spectral data were used.
Parameter estimation models developed from raw MSS data were least reliable.
Lacustrine water quality and use impairment in Illinois are significantly
impacted by suspended particulate matter. Lake morphology and hydraulic
factors affect the suspensoid load and general water characteristics. The
best quality lakes studied were generally deep, for Illinois, with long
retention periods (one year or more). Water quality and general use
potential decreased for lakes with shorter retention times and shallower
depths; generally the upper portions of reservoirs (areas of major stream
inflow) exhibit increased use impairment. Although all of the study lakes
are affected by high levels of phosphorus, overall water quality is basically
influenced by suspended particle impacts on water transparency.
-------
In other geographic areas, LANDSAT-derived lake surface area estimates
have exhibited excellent correspondence with area estimates derived from
concurrently acquired aerial photography; such imagery was not available for
use in this project. The LANDSAT MSS provided surface area estimates
generally within 10 percent of values derived for the study lakes from State
of Illinois data files.
The LANDSAT MSS is an economically viable tool for the acquisition of
data from all Illinois lakes and artificial reservoirs of significant areal
extent (i.e., four or more hectares). The extraction and processing of lake
MSS data is effectively accomplished, both costwise and timewise, through the
use of an image-processing system recently developed by the JPL for lake
classification purposes.
By virtue of its repetitive coverage, synoptic overview, and ability to
generate permanent records amenable to automated image-processing techniques,
the LANDSAT MSS is attractive for purposes of environmental assessment and
monitoring. In this study LANDSAT provided a view of the past; it also
provides a monitoring strategy that is objective, uniform, frequent, resource
tolerant, and cost-effective for the future. However, LANDSAT is not a
panacea; there are limitations and problems associated with it. A number of
problems were encountered during the study (e.g., a very limited selection of
cloud-free imagery, missing MSS internal scene calibration data, atmospheric
effects). Most of the problems encountered during this study were
successfully addressed. From the perspective of the State, this project was
highly successful, both in the insights that were gained while addressing
these problems and in the development of a high quality, comprehensive,
objective, short-term data base on Illinois lakes.
The following recommendations are made in light of the study results and
are consistent with reports on the assessment of lake problems and "clean
lakes" strategy completed under the Statewide 208 Water Management Planning
Program. A routine lake monitoring and assessment program should be
developed by the IEPA and coordinated with various State, Federal, and local
agencies. The program could be designed to incorporate LANDSAT MSS data.
Completion of the multiyear data baseline for key Illinois lakes is
desirable. Data collection could be coordinated with LANDSAT flyover. Water
transparency and suspended particulates measurements should be emphasized.
A statistically significant number of lakes should be sampled during
May-June and August-September to assess the range of lake quality and user
impairments throughout the recreational season. Efforts should be made to
extrapolate the contact-sensed data collected from these lakes to other
significant Illinois lakes using LANDSAT technology. Water samples collected
from a given lake and used in conjunction with LANDSAT data should be
integrated from the surface down to the maximum depth sensed by the MSS
(i.e., Secchi depth). The location and number of sampling sites should be
governed, in part, by the spectral and spatial characteristics of the
scanner. Summer is generally the best sampling period for Illinois lakes
that are to serve as LANDSAT benchmark or reference water bodies. The lakes
VI
-------
selected for calibration or modeling purposes should be representative of the
range of lakes to be monitored and classified. This may require more lakes,
and more types of lakes, than were used in this study. The Illinois EPA
should consider classifying the larger Illinois reservoirs in a mapping or
spatial context. Many of the larger water bodies are not homogeneous, and
they exhibit substantial differences in water quality, both in the vertical
and horizontal dimensions. The identification and location of such water
types are of importance to lake management programs. On large water bodies,
the MSS data should be calibrated against specific sample site information
acquired through contact sensing. The use of LANDSAT MSS data as a means of
identifying land cover and land use practices, and the relationship of land
use to lake characteristics, should be examined.
When processing the LANDSAT MSS data, forward overlap as well as side
overlap lake data should be extracted; this will provide better quality
control. In addition, certain areas of the State should be defined as
control points for the specific purpose of removing atmospheric effects from
the MSS data.
The data acquired during a future recreational high-use season, the
summer of 1977, the 1973 three-season National Euthrophication Survey (NES),
and the fall of 1973 by LANDSAT should be evaluated, along with other
information, to determine seasonal and long-term stability of the lake
characteristics, to define, to identify, and to map the spatial-temporal
distributions of .lake quality, and to determine representative sample
parameters, locations of sampling stations, and times for routine monitoring
purposes.
Based upon analyses of the lake data base, the Illinois EPA should
evaluate the lake assessment and the "clean lakes" strategy developed under
the 208 program. Methods to control the sources of degradation causing
fertility and sedimentation problems or management procedures to minimize
adverse impacts should be determined for the lakes. Short-term remedial
measures and long-range policies and programs should be assessed to maximize
lake life span and usability by the public.
The applicability of using different multivariate approaches to ordinate
and classify Illinois lentic water bodies should be further examined. For
example, it may be inappropriate to reduce the number of contact-sensed
parameters (as was done in this study) prior to the implementation of a
segregation procedure designed to separate those lakes with sediment-related
turbidity problems from those with turbidity problems related primarily to
the presence of algae.
This project has successfully demonstrated that LANDSAT MSS data can be
used to classify the lakes and artificial reservoirs of Illinois. It has
served as the vehicle through which both Federal and State scientists,
resource planners, and managers have developed a better understanding of the
LANDSAT MSS's capabilities and limitations in the area of lacustrine trophic
state assessment. The information relating to the technical aspects of the
vn
-------
project (e.g., digital image-processing techniques, scanner specifications,
multivariate statistical techniques) will largely appeal to the scientist.
Resource managers and planners will find the cluster diagrams, lake rankings,
and tabular lake data (including ancillary information) to be of practical
use. Overall, the report should be of value in assisting the State of
Illinois to meet its obligations under Public Law 92-500.
viii
-------
CONTENTS
Foreword i i i
Summary iv
Figures x
Tabl es xi i i
Abbreviations and Symbols xiy
Acknowledgments xvi i
1. Introduction 1
Background 1
Objectives 2
2. Conclusions 4
3. Recommendations 6
4. Geography and Lakes of Illinois 8
Geography 8
Lakes of Illinois 17
5. Methodological Overview 27
Optical properties of pure water and natural waters 28
LANDSAT multispectral scanner 32
Satellite sensing of Illinois lakes 38
6. Methods 49
Design overview 49
Data acquisition 49
Data processing 72
Data analysis approach 103
7. Results and Discussion 116
Water body surface area 116
Trophic indicator and index estimation 120
Cluster analysis 133
General discussion 164
References 172
Ap pendi x 183
IX
-------
FIGURES
Number
1 NASA'S LANDSAT space observatory 3
2 Physiographic provinces of Illinois 9
3 Generalized surficial geology of Illinois 12
4 Major soil orders in Illinois 14
5 Principal streams of Illinois 15
6 Hypothetical productivity growth-curve of a hydrosere 24
7 Reflection characteristics of filtered and unfiltered water
samples • 31
8 Schematic diagram of the LANDSAT-1 MSS scanning arrangement ... 33
9 Generalized spectral reflectance curve for a single picture
element (pixel) of a hypothetical lake 34
10 Generalized output of the LANDSAT MSS in response to the
spectral distribution illustrated in Figure 9 34
11 Some components and interactions of light with a hypothetical
lake and the atmosphere 36
12 IR2 image of LANDSAT scene 1448-16023 39
13 IR1 image of LANDSAT scene 1448-16023 40
14 RED image of LANDSAT scene 1448-16023 41
15 GRN image of LANDSAT scene 1448-16023 42
16 Characteristic "fingerprints" of clear or distilled water,
humic water, and algal-laden water 44
17 LANDSAT MSS spectral signatures and residual curves for
three types of 1akes 45
18 Range of MSS data for 145 Illinois lakes 47
19 LANDSAT coverage pattern for the State of Illinois 51
20 IR2 image of LANDSAT scene 1448-16023 52
21 Enlarged portion of LANDSAT IR2 print containing lakes and
reservoirs found in scene 1448-16023 53
22 IR2 image of LANDSAT scene 1448-16030 54
23 IR2 image of LANDSAT scene 1448-16032 55
24 IR2 image of LANDSAT scene 1448-16035 56
25 IR2 image of LANDSAT scene 1449-16082 ,57
26 IR2 image of LANDSAT scene 1449-16084 58
27 IR2 image of LANDSAT scene 1449-16091 59
28 IR2 image of LANDSAT scene 1449-16093 60
29 IR2 image of LANDSAT scene 1450-16140 61
30 IR2 image of LANDSAT scene 1450-16142 62
31 Geographic distribution of the Illinois water bodies 73
32 GRN, RED, IR1, and IR2 images of a LANDSAT scene 1448-16035
subscene ?5
-------
FIGURES (Continued)
Number
33 GRN, RED, IR1, and IR2 images of a LANDSAT scene 1448-16035
subscene after the application of the binary mask 76
34 Interactive user console with video-display device and
trackbal 1 78
35 Crab Orchard Lake with default 50-by-50 element box 80
36 Crab Orchard Lake contained within correctly positioned box ... 80
37 Crab Orchard Lake in binary form before editing of extraneous
water information 80
38 Crab Orchard Lake in final, edited form 80
39 Geometric interpretation of the water-detection algorithm 83
40 LANDSAT MSS histograms of four bands of Cedar Lake (serial
number 55) 84
41 IR2 concatenation of 33 Illinois water bodies extracted from
LANDSAT scene 1448-16023 85
42 IR2 concatenation of 12 Illinois water bodies extracted from
LANDSAT scene 1448-16032 86
43 IR2 concatenation of 20 Illinois water bodies extracted from
LANDSAT scene 1448-16032 87
44 IR2 concatenation of 27 Illinois water bodies extracted from
LANDSAT scene 1448-16035 88
45 IR2 concatenation of 20 Illinois water bodies extracted from
LANDSAT scene 1449-16082 89
46 IR2 concatenation of 35 Illinois water bodies extracted from
LANDSAT scene 1449-16084 90
47 IR2 concatenation of nine Illinois water bodies extracted from
LANDSAT scenes 1449-16093 and 1450-16140 91
48 IR2 concatenation of 13 Illinois water bodies extracted from
LANDSAT scene 1450-16142 92
49 Geometrical interpretation of the principal components 97
50 Sequence of procedures in numerical classification 107
51 Dichotomized scheme depicting the salient properties of
clusteri ng methods 109
52 Geometric aspects of Euclidian distance between two entities .. Ill
53 Geometric and computational aspects of Euclidian distance
between three entities 113
54 Sequence of procedures as applied to the numerical
classification of Illinois water bodies 115
55 LANDSAT-derived consecutive-day surface area estimates for
21 Illinois water bodies 118
56 Comparison of October 14 and 16 LANDSAT-derived surface area
estimates with IEPA file values for 20 Illinois water
bodi es 121
57 Comparison of October 15, 1973, LANDSAT-derived surface area
estimates with IEPA file values for 20 Illinois water
bodi es 122
58 Dendrogram of 145 Illinois water bodies based on complete
linkage clustering on four spectral attributes 140
XI
-------
FIGURES (Continued)
Number
59 Spectral index signature curves for six lake clusters 143
60 Visual correlation between contact-sensed data and scaled
MSS RED band data at the six-cluster level 146
61 Light extinction relationship between MSS RED band index
(percent) and Secchi depth measurements taken in 1977 147
62 Total suspended solids relationship with Secchi depth 148
63 Average annual runoff for Illinois in inches/square mile/
year 150
64 Distribution of precipitation in centimeters for Illinois
for the period October 11-14, 1973 152
-------
TABLES
Number
1 Description of lake types in Illinois 19
2 Optical properties of pure water 29
3 LANDSAT MSS scenes ordered for Illinois lake study 50
4 Serial list of Illinois water bodies 63
5 Area! statistics for example lakes 81
6 MSS DN statistics for example lakes 81
7 Regression models used to adjust multispectral scanner data ... 94
8 R-mode Pearson product-moment correlation matrix of five
trophic state indicators 98
9 Normalized eigenvectors and eigenvalues 99
10 R-mode Pearson product-moment correlation matrix of five
trophic state indicators 100
11 Normalized eigenvectors and eigenvalues 101
12 Trophic state indices and rankings for 31 NES-sampled
Illinois water bodies 102
13 Some SAHN clustering methods 108
14 Comparison of surface area estimates for 22 Illinois water 117
15 Three sets of regression models for the estimation of
trophic indicators and multivariate trophic state
i ndi ces 123
16 Trophic indicator and multivariate trophic index observed,
estimated, and residual values for the Set Three regression
model s 129
17 Comparison of the range of NES sample values with the 95 percent
confidence limits of predicted values 131
18 Summary statistics for Set Three (spectral rank) regression
model s 132
19 Rankings of 145 Illinois lakes based on Set Three models
and ordered by name 134
20 Composite ranking of 145 Illinois water bodies based on
Set Three models and ordered by increasing trophic state 138
21 Index signatures for various water body types in Illinois 142
22 Index signatures and general descriptions for six Illinois
1 ake clusters 142
23 Spectral data and contact-sensed data for the clusters and
subclusters developed using complete linkage clustering
on four attributes (GRN, RED, IR1, IR2) 144
24 Trophic classes developed from cluster analysis of trophic
indicator estimates from Set Three regression models 157
25 Interpretation of clusters developed from Set Three regression
model s 162
xm
-------
ABBREVIATIONS AND SYMBOLS
ABBREVIATIONS
A-space
C
CHLA
CCT
CDC 3300
COMNET
COND
DN
EPA
EROS
FARINA
GRN
GRNIR1
GRNIR2
GRNRED
GRNRK
ha
IBM 360
I DOC
I EPA
IFOV
IPL
IR1
IR1IR2
IR1RK
IR2
IR2RK
ISEC
JPL
attribute space
Celsius
chlorophyll ji
computer-compatible tape
Control Data Corporation's Model 3300 digital computer
Computer Network Corporation, 5185 MacArthur Blvd N.W.,
Washington, D.C. 20016
conductivity
digital number
U.S. Environmental Protection Agency
U.S. Geological Survey's Earth Resources Observation
System Data Center, Sioux Falls, South Dakota
image-processing software program designed to mask out
corresponding spectral channels or bands
green band, that portion of the spectrum between 500 and
600 nanometers, also known as band 4
ratio of the green and near infrared-one band values
ratio of the green and near infrared-two band values
ratio of the green to red band values
lake rank based on the LANDSAT MSS-measured green
band value
hectare (1 x 104 square meters)
International Business Machines Model 360 digital computer
Illinois Department of Conservation
Illinois Environmental Protection Agency
instantaneous field of view
Image Processing Laboratory
near infrared-one band, that portion of the spectrum
between 700 and 800 nanometers, also called band 6
ratio of the near infrared-one and near infrared-two band
values
lake rank based on the LANDSAT MSS-measured near infrared-
one band value
near infrared-two band, that portion of the spectrum
between 800 and 1,100 nanometers, also called LANDSAT
band 7
lake rank based on the LANDSAT MSS-measured near
infrared-two value
inverse of Secchi depth transparency
Jet Propulsion Laboratory, a NASA facility in Pasadena,
California
xiv
-------
LAKELOG
LANDSAT
LN
MSS
NASA
NES
nm
PCI
PC1F5
PC1Y5
pixel
r
R
R2
r-matrix
RAT100
RED
REDIR1
REDIR2
REDRK
RTSGRN
SAHN
SAS
SCS
SEC
S6RN
SIPS
Jet Propulsion Laboratory's interactive computer system
for the extraction and analysis of water body images
from LANDSAT and aircraft-acquired multispectral scanner
computer-compatible tapes
Land satellite; e.g., LANDSAT-1, LANDSAT-3; satellites in
NASA's Earth Resources Technology Satellite Program,
formerly known as ERTS
natural logarithm; e.g., LNSEC is the natural logarithm
transform of SEC
LANDSAT multispectral scanner
National Aeronautics and Space Administration
U.S. Environmental Protection Agency's National
Eutrophication Survey Program
nanometer, 1 x 10"^ meters
generalized form of a multivariate trophic state index
developed through principal component analysis of trophic
state indicators
multivariate trophic state index derived through
principal component analysis of the fall sampling
round values of five indicators: Seechi depth,
chlorophyll <±, conductivity total phosphorus, and
total organic nitrogen
multivariate trophic state index derived through
principal component analysis of the mean values for
three sampling rounds of five indicators: Secchi
depth, chlorophyll a_, conductivity, total phosphorus,
and total organic nitrogen
picture element
Pearson product-moment correlation coefficient
multiple correlation coefficient
as used in regression, this is the square of the
multiple correlation coefficient, also called the
coefficient of multiple determination
product-moment correlation matrix whose elements are used
in principal components analysis
LANDSAT spectral ratio
red band, that portion of the spectrum between 600 and
700 nanometers
ratio of the red and near infrared-one band values
ratio of the red and near infrared-two band values
lake rank based on the LANDSAT MSS-measured red band value
square root transformation of standardized LANDSAT green
band values
sequential, agglomerative, hierarchic nonoverlapping
Statistical Analysis System; a set of computer programs
maintained by the SAS Institute, Inc., Post Office Box
10066, Raleigh, NC 27605
scene color standard
Secchi depth transparency
standardized form of the LANDSAT green band value
Statistical Interactive Programming System
xv
-------
SIR1
SIR2
SQRTS6RN
SRED
STORE!
TON
TSS
TPHOS
VERTSL06
VSS
standardized form of the LANDSAT near infrared-one band
value
standardized form of the LANDSAT near infrared-two band
value
square root of the LANDSAT standardized green band value
standardized form of the LANDSAT red band value
STOrage and RETrieval; EPA's computer-based water quality
data bank
total organic nitrogen
total suspended solids
total phosphorus
Jet Propulsion Laboratory's computer software program
that performs digital image data format conversion
and geometric conversion
volatile suspended solids
SYMBOLS
/\
*
**
Ajk
3-D
estimated value for a parameter; (e.g., TON: regression-
derived estimate of total organic nitrogen)
multiplication, as used in computer programming
exponentiation, as used in computer programming
delta(jk), Euclidian distance between lake j and
lake k
delta squared, squared Euclidian distance
three-dimensional space
xvi
-------
ACKNOWLEDGMENTS
Sincere thanks are extended to Dr. Robert E. Frenkel (Oregon State
University Department of Geography), Dr. Richard C. Swartz and
James A. Keniston (Newport Field Station, U.S. Environmental Protection
Agency), and William C. Tiffany (Corvallis Environmental Research Laboratory,
U.S. Environmental Protection Agency) for the logistical and instructional
support so vital to the clustering and regression analysis aspects of this
project.
It is a pleasure to acknowledge the assistance of Harold Williamson,
Robert Eisenhart, Donna Wallace, and William Sandage (State of Illinois
Environmental Protection Agency) for their efforts in compiling and verifying
maps and topographic data. LeVerne Hudson and Dan Goodwin provided Illinois
Environmental Protection Agency staff and administrative support. Dr. Robert
Haynes and Dr. Konanur G. Janardan (Sangamon State University) provided
technical advice.
The assistance of John D. Addington, Andree Y. Smith, and Albert L.
Mendoza (Jet Propulsion Laboratory Image Processing Laboratory) in processing
the LANDSAT computer-compatible tapes and developing the interactive image
processing system is gratefully acknowledged.
xvn
-------
SECTION 1
INTRODUCTION
BACKGROUND
Lakes and reservoirs have long been recognized as valuable resources
serving as focal points for many types of recreational activity such as
fishing, swimming, skiing, and boating. They also help to meet the water
demands of municipalities, thermal-electric plants, commercial navigation,
and irrigation projects. Many inland water bodies, particularly those in
pristine condition, are valued for their aesthetic qualities. However, they
also provide convenient places to dump the organic and inorganic wastes of
society. Some are being drained to provide additional farmland; still others
are encroached upon through landfill schemes. Many water bodies,
particularly those in agricultural areas and in or near population centers,
are exhibiting the secondary effects of man-induced eutrophication (e.g.,
nuisance algal blooms).
Eutrophication — the process of nutrient enrichment of water — occurs
both naturally and as a consequence of man's activities. Many of man's
approaches to the disposition of municipal sewage and industrial wastes,
along with his land-use practices, impose large nutrient loads on lakes and
reservoirs. This enrichment process may result in algal blooms and other
secondary effects of eutrophication and thereby make these water bodies less
attractive to users. In Illinois, significant impacts on lake water quality
and user preference are frequently associated with sediment pollution in the
water bodies.
The rational management of a State's or the Nation's inland lentic
resources requires, as an initial step, that an inventory be made that
focuses on geographic, morphometric, biotic, and physiochemical factors
characterizing these waters. Thus, in 1972, the U.S. Congress, responding to
citizens and organizations concerned with the decline in the quality of the
Nation's water resources, passed Public Law 92-500, the Federal Water
Pollution Control Act Amendments of 1972. The legislation requires that each
State identify and classify, according to trophic condition, all publicly
owned freshwater lakes under its jurisdiction (Sections 106 and 314).
For States with large numbers of lakes and artificial reservoirs, the
mandated task is one of major proportions. The State of Illinois, for
example, has over 80,000 impoundments (2,700 of which have surface areas of
2.4 hectares or more and are classified as lakes by the Illinois Department
of Conservation), and collecting the data necessary for identification and
classification entails a project of sizable logistical and monetary
proportions. The fielding of boat-equipped crews is hindered by budgetary,
-------
manpower, and time constraints. The use of helicopter-borne field crews,
although attractive, is beyond the financial budget of the State.
The aforementioned circumstances dictate that consideration be extended
to other than traditional approaches when attempting to trophically classify
large numbers of water bodies. Any such approach should be cost effective
and rapid and should yield results having practical value. One such approach
employs data collected by satellite-borne sensors to develop trophic
classifications of water bodies.
On July 23, 1972, the National Aeronautics and Space Administration
(NASA) inserted a satellite called LANDSAT-1 (Figure 1) into a
sun-synchronous near-polar orbit to monitor the earth's resources. A
companion satellite, LANDSAT-2, was successfully launched January 22, 1975.
The tandem satellites provide repetitive coverage of nearly every point on
the earth's surface on a 9-day basis. A third satellite, LANDSAT-C and now
known as LANDSAT-3, was successfully launched March 5, 1978. The operation
of the LANDSAT-1 multispectral scanner ceased on January 6, 1978. Several
types of instrumentation are carried by the satellites including an imaging
multispectral scanner (MSS). Numerous investigations (e.g., Blackwell and
Boland 1979, Yarger and McCauley 1975, Moore and Haertel 1975, Strong 1973,
Fisher and Scarpace 1975, Boland and Blackwell 1975, Boland and Blackwell
1978, Boland 1976, Rogers et al. 1976, Rogers 1977) have demonstrated the
potential capabilities of the LANDSAT MSS for lake monitoring and
classification when used in combination with contact-sensed data.
OBJECTIVES
The basic objective of this project is to classify Illinois lakes and
reservoirs as to their trophic status and other characteristics, using an
approach combining contact-sensed and remotely sensed (i.e., LANDSAT MSS)
data. Classification, in a restricted sense, is the procedure of placing n
objects or p attributes into groups as defined by certain decision criteria.
A broader definition of classification includes the process of ordination and
its resultant products. In this report, classification is used in its less
restrictive sense.
Specific project objectives are to:
1. Develop rankings for some 150 Illinois lakes and
artificial reservoirs based on estimated magnitudes of
the following trophic indicators:
a. Secchi disc transparency (SEC);
b. Chlorophyll i (CHLA);
c. Total organic nitrogen (TON); and
d. Total phosphorus (TPHOS).
2. Develop lake rankings based on multivariate trophic indices.
3. Compile lake surface area estimates.
4. Identify lake trophic classes and their quality.
-------
SOLAR ARRAY
ORBIT ADJUST TANK
DATA COLLECTION ANTENNA
ATTITUDE CONTROL SUBSYSTEM
RETURN BEAM
VIDICON CAMERAS (3)
WIDEBAND
RECORDER
ELECTRONICS
WIDEBAND ANTENNA
ATTITUDE MEASUREMENT SENSOR
MULTISPECTRAL SCANNER
S BAND ANTENNA
Figure 1. NASA's LANDSAT space observatory (NASA 1972).
Develop procedures to classify Illinois water bodies into groups
with similar qualities and to rank the groups in order of
quality.
Identify the parameters that most significantly affect lake
quality and use.
-------
SECTION 2
CONCLUSIONS
The analyses of LANDSAT multispectral scanner (MSS) and near-concurrent
contact-sensed data, and ancillary information, for 145 Illinois water bodies
indicate that:
1. Lake clusters can be derived from LANDSAT MSS raw data and MSS-
estimated trophic indicator values; the clusters compare well with
near-concurrent contact-sensed data and ancillary information. The
application of an unsupervised clustering algorithm to raw MSS data
resulted in the establishment of distinctive lake clusters and
subclusters, eaclTidentifiable in terms of general water quality,
use impairment, and lake characteristics.
2. LANDSAT MSS data can be used with contact-sensed data to develop
regression models that provide relative estimates of Secchi depth,
chlorophyll £, total organic nitrogen, total phosphorus, and two
multivariate trophic state indices. Though less accurate and
precise than contact-sensed trophic indicator values, the LANDSAT-
deriyed parameter estimates can be used to develop generalized
rankings of Illinois lakes.
3. Regression models developed from MSS and contact-sensed data for
lakes with large parameter value ranges and minimal contact-sensed
data were more effective for the Illinois lakes studied when
spectral ranks rather than normalized spectral data were used.
Parameter estimation models obtained from raw MSS data were least
reliable.
4. LANDSAT spectral index signatures developed from raw MSS data
provide an acceptable characterization of general lake quality and
use impairment and can be used to establish an objective assessment
baseline for lake ordination and inventory purposes.
5. The LANDSAT MSS provided surface area estimates generally within 10
percent of values derived from State of Illinois data files
for the study lakes. In other geographic areas, LANDSAT-derived
lake surface area estimates have exhibited excellent correspondence
with area estimates derived from concurrently acquired aerial
photography; such imagery was not available for use in this project.
6. The LANDSAT MSS is an economically viable tool for the acquisition,
within three days and at a spatial resolution of about 0.64
4
-------
hectares, of data from all Illinois lakes and artificial reservoirs
of significant areal extent (i.e., four or more hectares).
7. The extraction and processing of lake MSS data is effectively
accomplished, with respect to both cost and time, through the
use of LAKELOC, a recently developed image-processing system.
8. Lacustrine water quality and use impairment in Illinois is
significantly impacted by suspended particulate matter. Lake
morphology and hydraulic factors affect the suspensoid load and
general water characteristics.
9. The best quality lakes studied were generally deep, for Illinois,
with long retention periods (one year or more). Water quality and
general use potential decreased for shorter retention-time lakes
with shallower depths; generally the upper portions of reservoirs
(areas of major stream inflow) exhibit increased use impairment.
10. Although all of the study lakes are affected by high levels of
phosphorus, overall water quality is basically influenced by
the effect of suspended particles on water transparency.
-------
SECTION 3
RECOMMENDATIONS
Most of the problems encountered during this study were successfully
addressed. From the perspective of the State, this project was highly
successful, both in the insights that were gained while addressing these
problems and in the development of a high quality, comprehensive, objective,
short-term data base on Illinois lakes. The following recommendations are
made in light of the study results and are consistent with reports on the
assessment of lake problems and "clean lakes" strategy completed under the
Statewide 208 Water Management Planning Program (IEPA 1978a, IEPA 1978b).
1. A routine lake monitoring and assessment program should be developed
by the Illinois Environmental Protection Agency (IEPA) and
coordinated with various State, Federal, and local agencies. The
program could be designed to incorporate LANDSAT multispectral
scanner (MSS) data.
2. Completion of the multiyear data baseline for key Illinois lakes
is desirable. A significant number of lakes should be sampled
during May-June and August-September to assess the range of lake
quality and user impairments throughout the recreational season.
Data collection could be coordinated with LANDSAT flyover. Water
transparency and suspended particulates measurements should be
emphasized.
3. Efforts should be made to extrapolate the contact-sensed data
described in the second recommendation to significant Illinois
lakes using LANDSAT technology.
4. The data acquired during a future recreational high-use season, the
summer of 1977, the 1973 three-season National Eutrophication Survey
(NES), and the fall of 1973 by LANDSAT should be evaluated, along
with other information, to determine seasonal and long-term
stability of the lake characteristics; to define, identify, and map
the spatial-temporal distributions of lake quality, and to
determine representative sample parameters, locations of sampling
stations, and times for routine monitoring purposes.
5. Based upon analyses of the lake data base, the Illinois EPA should
evaluate the lake assessment and the "clean lakes" strategy
developed under the 208 program. Methods to control the source(s)
of degradation causing fertility and sedimentation problems or
-------
management procedures to minimize adverse impacts should be
determined for the lakes. Short-term remedial measures and
long-range policies and programs should be assessed to maximize
lake lifespan and usability by the public.
6. Water samples collected from a given lake and used in support
of LANDSAT should be integrated from the surface down to the maximum
depth sensed by the MSS (i.e., Secchi depth). The location and
number of sampling sites should be governed, in part, by the
spectral and spatial characteristics of the scanner. Summer is
generally the best sampling period for Illinois lakes that are to
serve as LANDSAT benchmark or reference water bodies.
7. The lakes selected for calibration or modeling puposes should be
representative of the range of lakes to be monitored and classified.
This may require more lakes, and additional types of lakes, than
were used in this study.
8. The Illinois EPA should consider classifying the larger Illinois
reservoirs in a mapping or spatial context. Many of the larger
water bodies are not homogeneous, and they exhibit substantial
differences in water quality, both in the vertical and horizontal
dimensions. The identification and location of such water types is
of importance to lake management programs. On large water bodies,
the MSS data should be calibrated against specific sample site
information acquired through contact sensing.
9. The use of LANDSAT MSS data as a means of identifying land cover and
land-use practices and their relationship to lake characteristics
should be examined.
10. When processing the LANDSAT MSS data, forward overlap as well as
side overlap lake data should be extracted; this will provide better
quality control.
11. Certain areas of the State should be defined as control points for
the specific purpose of removing atmospheric effects from the MSS
data.
12. The applicability of using different multivariate approaches to
ordinate and classify Illinois lentic water bodies should be further
examined. For example, it may be inappropriate to reduce the
number of contact-sensed parameters (as was done in this study)
prior to the implementation of a segregation procedure designed
to separate those lakes with sediment-related turbidity problems
from those with turbidity problems related primarily to the
presence of algae.
-------
SECTION 4
GEOGRAPHY AND LAKES OF ILLINOIS
GEOGRAPHY
An overview of Illinois physical geography is provided in this
subsection. More detailed descriptions can be found elsewhere (e.g.,
Strahler 1969, Fenneman 1938, Thornbury 1965). The information presented in
the following subsections draws heavily from the work of the State of
Illinois Environmental Protection Agency (IEPA 1976).
Physiography and Geology
Illinois is largely a low prairie plain with an average elevation of
approximately 193 meters above sea level (Kofoid 1903 as cited by Gunning
1966). With an elevation of 377 meters, Charles Mound, located in the
northwestern corner of the State near the Illinois-Wisconsin border, is the
highest topographic point (Rand McNallv 1971). [Gunning (1966) reports an
elevation of 383 meters. Bethel (1969) records 378 meters as the elevation
of Charles Mound.] The lowest point in the State is at Cairo where the
low-water mark of the Ohio River has been measured as 82 meters (Gunning
1966). Local relief, the difference in elevation between lowest and highest
topographic points in adjacent locations, is typically less than 60 meters
(IEPA 1976). Fenneman (1928) has divided the State into four physiographic
provinces (Figure 2). Leighton et al. (1948) have further subdivided
Illinois into eight physiographic sections using the following criteria:
bedrock topography, extent of the several glaciations, glacial morphology
differences, age differences of the uppermost drift, height of the glacial
plain above the main lines of drainage, glaciofluvial aggradation of basin
areas, and glaciolacustrine action.
More than 90 percent of Illinois is located in the Central Lowland
province. With the exception of the Wisconsin Driftless section located in
extreme northwestern Illinois (Jo Daviess and Carroll Counties), all of the
province has been glaciated. The Ozark Plateaus, Interior Low Plateaus, and
Coastal Plain province account for the remainder of Illinois.
Three sections constitute the bulk of the Central Lowland province in
Illinois: the Great Lakes section, the Wisconsin Driftless section, and the
Till Plains section. The Great Lakes section, located in northeastern
Illinois, is largely dominated by an area called the Wheaton morainal
country. Wedged between the morainic area and Lake Michigan is the Chicago
lake plain. The lake plain is characterized by a flat surface underlain
largely by till (a heterogeneous mixture of rock fragments ranging in size
8
-------
Till Plains Section
CWttconsirf/T
>.Driftlesi/
^ \SectionV .
^^ V IL —Rock River
Great Lake
£ 1 Section
Morainal"TCnicago
V^ountrv
Water body not included in
L rankings and dendrograms because
of faulty or missing MSS data
NES - sampled water body used in
| regression model development and
included in rankings and dendrogram*
» Water body included in rankings and
dendrograms
Ozark Plateaus Province
Interior Low Plateaus Province
Coiiitl Pl»in Province
^ Central Lowland Prowmce
Figure 2. Physiographic provinces of Illinois. The water bodies marked on
the map were proposed for inclusion in this project. Figure
adapted from Illinois State Geological Survey map (IEPA 1976).
-------
from clay to boulders and associated with glaciers) and having a gentle slope
toward Lake Michigan. Postglacial erosion has been of a relatively small
magnitude and is most evident in the valleys along the Fox and Des Plaines
Rivers (IEPA 1976).
The topography of the Wisconsin Driftless section (Figure 2) reflects the
physical-chemical character of the underlying rock and is a consequence of
the apparent lack of glacial activity in the area. Dendritic drainage
systems, tributary to the Mississippi River, dissect the section. A
substantial amount of drainage occurs through small caves and solution
channels. However, features associated with karst topography are not readily
evident.
Approximately 80 percent of the State is classified as belonging to the
Till Plains section, an area dominated by broad till plains with poorly
developed drainage systems. The western and southern limits of the
Bloomington ridged plain effectively define the boundary separating the older
Illinoian deposits from those Wisconsin!"an in age. Physiographic contrasts
(e.g., degree of drainage integration, extent of soil development, erosional
modification of the topography) have developed because of these age
differences.
A fourth section of the Central Lowland province, the Dissected Till
Plains section, is found in three western Illinois counties, Hancock, Adams,
and Pike. In this section glaciation occurred in an earlier stage, the
Kansan, and this accounts for the submaturely dissected landscape (Fenneman
1938). The longer time element has allowed the Dissected Till Plains section
to reach a more advanced point in the postglacial erosion cycle than the Till
Plains section to the east (Fenneman 1938).
The Ozark Plateaus province is represented in Illinois by two sections,
the Lincoln Hills section and the Salem Plateau section. Both sections are
found along the southwestern boundry of the State (Figure 2). The sole
representative of the Interior Low Plateaus province in Illinois is the
Shawnee Hills section (also known as the Shawnee section) found in the
southern portion of the State. The southern tip of Illinois lies in the
Mississippi Alluvial Plain section (commonly called the Mississippi
Embayment) of the Coastal Plain province. Detailed descriptions of these
relatively minor components of Illinois physiography can be found in Fenneman
(1938) and Thornbury (1965).
The geologic units of Illinois are classified into four major divisions:
1) Precambrian basement rock; 2) consolidated sedimentary bedrock; 3) a
variably thick blanket of glacial drift; and 4) a mantle of loess. The
Precambrian basement rock lies hidden, covered by some 600 to more than
3,960 meters of sedimentary strata. Borings, both in Illinois and
immediately outside the State's border, have yielded plutonic and volcanic
rocks of granitic or closely related composition (Bradbury and Atherton
1965).
Consolidated sedimentary rocks, ranging in age from Cambrian through
Tertiary, are found throughout the State. Extensive exposures of these rocks
10
-------
are limited to extreme northwestern Illinois and to the southwestern and
southern part of the State in areas along the Mississippi and Ohio Rivers
falling outside the glacial margin. Scattered bedrock outcrops, usually
limited to areas of high bedrock or the walls of the more deeply cut valleys,
are found within the glaciated region. The Ordovician and Silurian rocks are
largely dolomite. The Mississippian rocks are dominated by limestone, and
the Pennsylvanian rocks by shale and sandstone. Illinois is noted for its
large deposits of mineable coal, which are found in the Pennsylvanian strata.
Approximately 90 percent of Illinois is blanketed by glacial deposits
consisting of both stratified drift and till deposited during each of the
four major glacial stages (Figure 3). Drift of Nebraskan age, the first
glacial stage, is known in Illinois only from subsurface deposits in the
western part of the State. The Kansan glaciation covered most of the State,
but subsequent glacial stages have restricted exposed Kansan tills to a small
area in western Illinois, the Dissected Till Plains section. The third
glaciaticn, the Illinoian, marked the southern most incursion of the
continental ice sheet in North America. The ice extended to within
33 kilometers of the Mississippi Embayment in southernmost Illinois, laying
down an extensive blanket of drift over much of the State. The last
glaciation, the Wisconsinian, is recognized as a major factor in molding the
character of the modern landscape (Flemal 1972). The Wisconsinian deposits,
largely confined to northeastern Illinois, have not been exposed to
weathering as long as the older Illinoian deposits and, therefore, exhibit
marked soil differences when compared to II linoian-derived soils.
Willman and Frye (1970) report that as much as half of the material
comprising the tills of Illinois has been transported less than
160 kilometers from its bedrock source. The glacial drift ranges in
thickness from zero to about 180 meters, with the average being about
30 meters. Relatively thin layers of drift are found in western and southern
Illinois. The northeastern quarter of the State is mantled by a thick drift
deposit. As might be expected, extremely thick drifts blanket the major
bedrock valleys.
Much of Illinois is covered by loess, a silty windblown deposit that
consists chiefly of dust from dry glacial river valley floors. It is
thickest east of the valleys because of the prevailing westerly winds. It
becomes progressively thinner as the distance from the source increases.
Loess, the most extensive parent material of Illinois soils, varies in
average thickness from about 2 to 6 meters in western Illinois to 0.5 to
1.5 meters in the eastern part of the State. The Army Corps of Engineers
(1969) reports it as being the major constituent of the suspended sediment
transported in Illinois streams. The State's loess-covered plains are widely
recognized for their agricultural productivity.
Soils
The major parent materials forming Illinois soils are loess, outwash,
till, and alluvium. Bedrock and accumulations of organic material (e.g.,
peat) are of relatively minor importance. The following is a reiteration of
lEPA's (1976) description of Illinois soils, which in turn is based largely
on the work of Fehrenbacher et al. (1967).
11
-------
SURFICIAL
GEOLOGY
/Loess thickness (fi).
-,\* Hochures point m
r direction of thinning
Loess-Corertft Areas
] Giociol drift (Wisconsmon), kvgely
till, coicoreous to top except
where loess is less thon 4 ft
thick ond leoched
Gtociol drift (Illtnoion), momly till, upper 5 to
^~~~^ 10 ft generally noncolcoreous
S GkKiol drift (Kanson). largely till, upper 15 to
30 ft generally noncolcareous
F!'ffi''j' j Sand,gravel, or silt (Cretaceous-Tertiary)
! I Bedrock (Paleozoic), various consolidated rock!
Loess Absent or Very Thin
-• ] Alluvium, river and stream deposits — silt, sat
(Cahokia Alluvium)
| Lake deposits-silt, cloy, sand (Equality Formation)
j Sand dunes-wind-deposited sand (Parkland Sand)
I Glacial outwash - major outwosh plains and terraces of
sond ond grovel (Henry Formation)
Figure 3. Generalized surficial geology of Illinois. In uneroded areas the
loess thins from 25 to 100 feet thick in the bluffs of the major
valleys to 1 to 2 feet in areas farthest from the valleys.
(Figure courtesy of the Illinois State Geological Survey.)
-------
Of the 10 major soil orders comprising the 7th Approximation (Soil Survey
Staff 1960) -- Aridisols, Vertisols, Spodosols, Ultisols, Oxisols, Histosols,
Inceptisols, Mollisols, Alfisols, Entisols -- the latter three dominate
Illinois (Figure 4).
Histosols occur mainly in extreme northeastern Illinois. They are
organic soils (peats and mucks) that are distinguishable from those of the
other nine orders by gross composition differences, most notably their high
organic content of 20 percent or better (Simonson 1962). Histosols are wet
soils, generally requiring drainage when used for most agricultural purposes.
Alfisols predominate in southern Illinois. A light color is generally
associated with the surface layer; if the color is dark, the surface layer is
less than 25 centimeters thick. In either case the surface layer has an
organic content of less than one percent throughout its total thickness.
Another characteristic of Alfisols is a recognizable B horizon of clay
accumulation.
Mollisols cover about 49 percent of the State with most occurring in
central and northern Illinois. These are mineral soils possessing what is
called a mollic epipedon. This darkened surface layer has an organic content
of more than one percent, is high in base saturation, and is friable.
Inceptisols are usually found in small or narrow areas and are included
with Mollisols in the bottom lands and with the Alfisols in the uplands.
Although lacking the mollic epipedon of the Mollisols and the B horizon of
clay accumulation associated with Alfisols, Inceptisols do have recognizable
horizons or show evidence of the beginning of horizon development.
Entisols are mineral soils whose profiles contain few and faint horizons.
They cover an estimated 1.5 percent of Illinois, occurring along streams and
in very sandy areas. Entisols are known to form in regoliths consisting of
highly resistant minerals, in areas where the accumulation of surface
materials keeps pace with horizon differentiation, and in localities where
the erosion rate keeps up with horizon differentiation (Simonson 1962).
Climate
According to the Koppen-Geiger system of climate classification, the
northern third of Illinois is classified as humid continental; the remainder
of the State is placed in the humid subtropical category (Strahler 1969).
The State's geographic location results in great variations in temperature
and precipitation over the course of a gjven year. The mean annual
temperature is about 8 degrees Celsius ( C) in the north, increasing to
15 C in the south. Normally, January is theocoldest month with a mean
temperature of about -6 C in the north and 2 C in the south. July,
usually the hottest month, exhibits a similar north to south temperature
trend with the mean values being about 23 C and 27 C, respectively. Mean
annual precipitation also displays a north-to-south trend -- 80 centimeters
and 120 centimeters, respectively. Except for a small area in the southern
part of Illinois, the period of maximum precipitation occurs during the
13
-------
MOLLISOL ORDER
ALFISOl ORDER
ENTISOL ORDER
Figure 4. Major soil orders in Illinois (from Fehrenbacher et al. 1967).
Of the 10 soil orders in the 7th Approximation, five —
Mollisols, Alfisols, Entisols, Histosols, and Inceptisols — are
important in Illinois. The Inceptisols and Histosols cover rela-
tively small areas and are not shown in the above figure.
-------
growing season. In general, northern Illinois receives a larger percentage
of its total precipitation during the growing season than does the southern
part. This is important because it tends to make the lesser precipitation in
the north as effective for growing crops as the greater annual quantities in
southern Illinois (Page 1949).
Drainage
Three major rivers border Illinois: the Mississippi on the west, the
Ohio on the south, and the Wabash on the southeast (Figure 5). Large streams
that provide the internal drainage of the State include the Illinois, Rock,
Kaskaskia, Big Muddy, Embarras, and Little Wabash Rivers. The Illinois
River, the largest stream within Illinois, drains the State in a generally
southwesterly direction. The Rock River, the
State, drains the northwestern portion of
adjacent southern Wisconsin. Three major
second largest stream in the
Illinois as well as parts of
sub basins -- the Pecatonica,
Kishwaukee, and Green River Basins -- comprise the Rock River Basin.
Figure 5. Principal streams of Illinois (from O'Donnell 1935),
15
-------
The Kaskaskia and Big Muddy Rivers join the Mississippi River in
southwestern Illinois. The Kaskaskia, the larger of the two, originates in
east-central Illinois. The Big Muddy River is confined to the southern
portion of the State. Four Illinois streams, the Embarras, Little Wabash,
Vermilion-Wabash, and Saline, all flow in generally southeasterly courses and
feed the Ohio and Wabash Rivers.
Approximately 81 percent of Illinois is drained directly to the
Mississippi River by means of southwest flowing streams (e.g., Illinois and
Rock Rivers). Almost 19 percent is drained indirectly to the Mississippi
River by way of the Ohio and Wabash Rivers. Less than 0.1 .percent of the
State's area is drained into the Great Lakes system (McCarthy 1972).
Land Use
Illinois covers an area of approximately 146,076 square kilometers
(14,607,600 hectares), not including the 3,950 square kilometers of Lake
Michigan. The actual land area has been calculated as 144,872 square
kilometers (Bethel 1969). Eighty-three percent of Illinois human population
is classified as urban with about 80 percent being clustered into 10 Standard
Metropolitan Statistical Areas (SMSA).
Agriculture is the dominant land use in Illinois, with tilled crops
dominating. Cropland distribution is more dense in the central and northern
parts of the State than in the western and southern parts. The Illinois
Cooperative Crop Reporting Service (1974) found the major crop to be
soybeans, followed in order by corn, wheat, and hay. IEPA (1976) report that
about 74 percent of the inventoried land is devoted to tilled crops; cropland
increased from 96,620 square kilometers in the 1958 inventory to 98,745
square kilometers in 1967. Inventoried land is defined as the total surface
area of the State minus Federally owned land, water areas, and land used for
urban purposes. The inventory was conducted by the Illinois Inventory of
Soil and Water Conservation Needs to determine the use and condition of
privately owned rural land in Illinois. Inventory participants included the
United States Department of Agriculture and various State agencies.
Inventoried woodlands and forests cover approximately 11 percent of
Illinois, with most being classified commercial. The noncommercial forests
serve very important recreational and conservation functions. Forest lands
have declined from 1958 to 1967 as indicated by the area estimates of 15,783
and 14,569 square kilometers, respectively.
The third largest use of rural land is grazing. Approximately 10 percent
of the inventoried land is used to pasture livestock. A slight expansion in
pasturage has been noted, with 13,395 square kilometers reported for 1958 and
13,557 square kilometers for 1967. Grazing lands, generally confined to the
more hilly regions, are most dense in northwestern, western, and southern
Illinois.
Another land-use category, "other land" (rural land that is not used as
cropland, forest land, or pastureland), has declined from 7,284 square
kilometers in 1958 to 6,070 square kilometers in 1967. Some "other land" is
16
-------
used for recreation and wildlife. In some localities the quarrying of sand
and gravel and the mining of coal are of major importance. The extraction of
sand and gravel occurs mainly near major urban centers (e.g., Chicago).
Illinois, the fourth-ranked State in coal production (IEPA 1976), has large
coal deposits in its bedrock of Pennsylvanian age. Coal strip mines are
found along the outcrop margins of Pennsylvanian strata. Elsewhere, the
depth of the coal from the surface necessitates the use of underground
mi ni ng.
LAKES OF ILLINOIS
There is far from universal agreement as to what constitutes a lake.
Veatch and Humphrys (1966) suggested that to give the word "lake" a precise,
limited meaning would probably be an exercise in futility because the word
has been in use for a long time in a variety of applications. The word is
used as a synonym for pond, reservoir, and sea. It has been applied to
bodies of fresh water and saline water, to standing water and widenings in
rivers, to bodies of water measuring less than a hectare and to those gauged
in hundreds of thousands of hectares, to naturally occurring water bodies and
manmade reservoirs, to water-filled or partially filled basins, and to basins
void of water. "Lake" is generally more prestigious than other common names
(e.g., pond, slough, reservoir) and is preferred by promoters of water-based
tourist and recreational businesses and commercial developers of shoreline
property (Veatch and Humphrys 1966). Nevertheless, numerous attempts have
been made to define and delimit the members of lentic series (i.e., lake,
pond, marsh, and their intergrades).
Fore! defined a lake as a body of standing water occupying a basin and
lacking continuity with the sea, and a pond as a lake of slight depth (Welch
1952). Welch defined a lake as a "... body of standing water completely
isolated from the sea and having an area of open, relatively deep water
sufficiently large to produce somewhere on its periphery a barren, wave-swept
shore." He employed the term "pond" "... for that class of very small, very
shallow bodies of standing water in which quiet water and extensive occupancy
by higher aquatic plants are common characteristics ..." and suggested that
all larger bodies of standing water be referred to as lakes. Zumberge (1952)
defined a lake as an inland basin filled with water. Harding (1942)
described lakes as "... bodies of water filling depressions in the earth's
surface." The Illinois Department of Conservation (IDOC) uses a size
criterion (2.4 hectare or larger in surface area) to define lake. In this
report no deliberate effort will be made to carefully distinguish a lake from
another lentic body on the basis of a definition. We are using the term lake
in its broadest sense. For example an artificial reservoir may be called a
lake or, at times, a water body.
The Illinois Surface Water Inventory prepared by the State of Illinois
Department of Conservation (1972) reported that the State had 2,706 lakes
(2.4 ha or larger in surface area) and 73,595 ponds (less than 2.4 ha in
surface area) in 1972. Of the 2,706 lakes, only 709 were natural -- either
glacial lakes in the north or flood plain or oxbow lakes along the major
rivers. The remaining 1,997 were ma made (IDOC 1972, Gunning 1966, and
17
-------
Bennett 1960). An inventory of Illinois lakes (IEPA 1975) reveals a total of
773 publicly owned freshwater lakes, with 273 lakes having surface areas in
excess of 8 hectares. Excluding the portion of Lake Michigan within the
Illinois boundary, the total area of publicly-owned freshwater lakes is
58,818 hectares (IEPA 1975).
Illinois lacks pristine lakes such as those found in northern Minnesota
and Wisconsin. Illinois lakes, however, exhibit a range of morphologic,
water quality and usability characteristics. Table 1 lists brief comments
regarding lake types in Illinois.
Lake Succession
Lakes, although giving the impression of permanence when measured on the
scale of the human lifespan, are transitory features of the earth's surface.
Most lakes, regardless of their origin, pass through the process of
ecological succession that ultimately results in a terrestrial environment.
The ephemeral nature of natural lakes is a consequence of two fundamental
processes — the downcutting of the outlet and, more importantly, the
deposition of allochthonous and autochthonous materials in the basin. While
downcutting of the outlet is of little consequence to Illinois artificial
lakes and reservoirs, loss of depth and capacity occurs because of sediments
delivered by streams and from bank erosion.
Many lakes (presumably) commence the successional process as bodies
possessing relatively low concentrations of nutrients and, generally, low
levels of productivity. Edmondson (1974) suggested that the idea that all
lakes are born oligotrophic and gradually become eutrophic as they age is an
old misconception. The importation and deposition of materials (e.g.,
sediment) from the shoreline and the surrounding watershed gradually
decreases the lake depth. The addition of allochthonous materials normally
enriches the water and thereby stimulates the production of autochthonous
organic materials. Autochthonous materials increase the sedimentation rate
and accelerate succession. Marked floral and faunal changes occur. Algal
blooms become more common along with submergent and, eventually, emergent
aquatic macrophytes. Desirable game fish may be replaced by less desirable
species, the so-called rough fish. A lake eventually becomes a marsh or
swamp that, in turn, terminates as dry land.
Lindeman (1942) stressed the productivity aspects in relation to lake
succession. Figure 6 represents the probable successional productivity
relationships for a hypothetical hydrosere developing from a moderately deep
lake located in a fertile humid continental region. Productivity is
initially low, a consequence of low nutrient levels, but increases rapidly as
nutrients become more available. The length of time required for completion
of the successional process is a function of several factors including lake
basin morphology, climate, and the rate of influx and nutrient value of
allochthonous materials. It is readily apparent that allochthonous nutrients
can drastically increase lake productivity and thereby shorten the lifespan
of a lake.
18
-------
TABLE 1. DESCRIPTION OF LAKE TYPES IN ILLINOIS
Parameter
Natural Lakes
Glacial
Backwater
Normal
Flowage
Location Limited to northeastern
Illinois including Cook,
Lake, and McHenry Counties.
Physical/ Small (most are less than
morpholog- 80 ha). Long retention
ical char- time. Location of maxi-
acteristic mum depth is highly
variable. Shoreline tends
to be regular while the
basin is cone shaped.
Very long life expectancy,
a consequence of low
nutrient levels and the
inorganic nature of lake
bottom.
Water Start out oligotrophic
quality (low nutrient levels) and
eventually become eutro-
phic (nutrient rich).
Impacted by culturally
induced eutrophication.
Northeastern II1inoi s,
particularly the Fox-
Chain of Lakes.
Varying in surface
area from about 40 ha
to several thousand
ha. Very short
retention time with
large flow through.
See normal glacial
lakes (except that
water quality is
significantly affected
by sediment-related
turbidity).
Along major streams including
the Illinois, Mississippi,
and Ohio Rivers.
Wide range of sizes. Typically
(very shallow with a flat
'bottom profile and short
retention time.
Water quality is a reflection
of river water quality. High
sediment-related turbidity,
generated by rough fish, wind
action, and runoff, is a major
limiting factor. Generally high
in nutrients with large amounts of
sediment deposited when the river
floods.
(continued)
-------
TABLE 1. (continued)
Parameter
Natural Lakes
Glacial
Backwater
Normal
FTowage
PO
o
Water Exhibit classical symptoms
quality associated with eutrophica-
tion (e.g., algal blooms,
aquatic macrophytes, deple-
tion of dissolved oxygen).
Water transparency varies.
Lakes tend to have a
homogeneous water quality.
Alkalinity, hardness, and
conductivity are generally
high. Ground water con-
tributes to recharge.
Watershed/ Watersheds tend to be
shoreline small, with the drainage
uses area/lake capacity ratio
also being small. Water-
shed is generally urban
or suburban. Typically,
having much development
around the lakeshore.
Lake usage Very heavy recreational
use, primarily boating,
fishing, and swimming.
Very large watersheds,
with drainage area/
lake capacity ratios
also extremely large.
Otherwise, they are
similar to normal
glacial lakes.
(See normal glacial)
Very large watershed. Usually
having little shoreline
development.
Used primarily for flood pro-
tection, hunting, and fishing.
Waterfowl habitat.
(continued)
-------
TABLE 1. (continued)
Parameter
Natural Lakes
Glacial
Backwater
Normal
Flowage
Comments
Very difficult to manage and
limit pollution inputs because
of the riverine influence.
Parameter
Artificial
Reservoirs
Impoundments
North
Central
South
Other
(Borrow Pits, Strip Mines,
and Quarries)
Location
Physical/
morpholog-
ical char-
acteristics
Varying in surface
area but with
moderate to long
retention. Max-
imum depth at the
dam. Irregular
shoreline.
Varying in surface
area from very
small to more than
4,000 ha. Short
retention time
with maximum depth
at the dam. The
lakebed is rich
agricultural
soil , high in
nutrients and or-
ganic materials.
The shoreline is
irregular.
Surficial recharge.
Varying in surface
area with short to
long retention time.
Maximum depth at
the dam. Very
irregular shoreline.
Lakebed consists
primarily of claypan
soils that are
lower in nutrients
and organic content
than the central
reservoirs.
Surficial recharge.
Borrow pits are con-
centrated in or near
urban centers and along
highways. Generally, have
long retention periods and
are regulated by ground
water level.
(continued)
-------
Table 1. (continued)
Parameter
Artificial Impoundments
Reservoirs
North
Central
South
Other
(Borrow Pits, Strip Mines,
and Quarries)
Water quality
High alkalinity,
conductivity,
and hardness.
ro
Watershed/
shoreline
uses
Medium to high
alkalinity, con-
uctivity, and
hardness. Gen-
erally eutrophic
from the day res-
ervoir started to
fill. Extremely
high nitrogen and
phosphorus levels.
Turbidity and
siltation are major
problems. May be
light limited and,
therefore, do not
exhibit some of
the secondary
effects associated
with classical
eutrophication.
Watersheds vary
in size. Drain-
age area/lake
capacity ratios
vary. Watershed
is primarily in
row crops. Non-
point pollution
is a major prob-
lem. Some urban
Low alkalinity, con-
ductivity, and hard-
ness. Generally
clearer than
central Illinois
impoundments
except when there
is heavy runoff.
Some have algae and
macrophyte problems.
Clear unless there is
heavy runoff. Wide
range in water quality.
Watersheds vary in
size but are gen-
erally smaller than
those of central
Illinois. Watersheds
are primarily forest-
ed or not under as
intensive row crop
cultivation as those
of central Illinois.
Generally fed by
ground water.
(continued)
-------
Table 1. (continued)
Parameter
Artificial Impoundments
Reservoirs
North
Central
South
Other
(Borrow Pits, Strip Mines,
and Quarries)
Watershed/
shoreline
Lake usage
PO
CO
Recreation is the
primary use. A
few serve as
public water
suppl ies.
influence. There
may be a moderate
degree of develop-
ment around the
shorel ine
Many serve as
public water
supplies. Rec-
reational usage
is primarily
fishing and
boating.
Little urban in-
fluence except in
some cases where
developed along the
shorel ine.
Many serve as
public water sup-
plies. Recreational
usage includes swim-
ming, fishing, and
boating.
-------
EUTROPHY
ro
O
Q
O
CC
Q.
CLIMAX
VEGETATION
BOG FOREST
OLIGOTROPHY
MAT
SENESCENCE
TIME
Figure 6. Hypothetical productivity growth-curve of a hydrosere (adapted from Lindeman 1942).
Lindeman describes the curve as representing a hydrosere "developing from a
moderately deep lake in a fertile cold temperate climatic condition." It must be
kept in mind that this is a generalized curve and that not all lakes will follow
it in total. For example, lakes that are light limited because of suspended
inorganic materials may never experience the initial dramatic increase in productivity.
-------
In Illinois lakes, sedimentation is considered to be more significant
than lake production in terms of affecting their usable lifespan. Most
Illinois lakes, whether recently impounded or old, have high productivity
potential.
Eutrophication
The word eutrophication is often used to denote the process whereby a
pristine water body (e.g., lake) is transformed into one characterized by
dense algal scums, obnoxious odors, and thick beds of aquatic macrophytes.
However, the word has been applied differently, according to the respective
interests of its users. Weber (1907) used the German adjectival form of
eutrophication, "nahrstoffreichere" (eutrophe), to describe the high
concentration of elements requisite for initiating the floral sequence in
German peat bogs (Hutchinson 1973). The leaching of nutrients from the
developing bog resulted in a condition of "mittelreiche" (mesotrophe) and
eventually "nahrstoffearme" (oligotrophe). Naumann (1919) applied the words
oligotrophic (underfed), mesotrophic, and eutrophic (well-fed) to describe
the nutrient levels (calcium, phosphorus, combined nitrogen) of water
contained in springs, streams, lakes, and bogs (Hutchinson 1973). Naumann
(1931) defined eutrophication as the increase of nutritive substances,
especially phosphorus and nitrogen, in a lake. Hasler (1947) broadly
interpreted eutrophication as the "enrichment of water, be it intentional or
unintentional ...." Fruh et al. (1966) defined the word as the
"... enhancement of nutrients in natural water ...." while Edmondson (1974)
suggested that many limnologists seem to use the term to describe "... an
increase in the rate of nutrient input ...." Hasler and Ingersoll (1968)
suggested that eutrophication is the "... process of enrichment and aging
undergone by bodies of fresh water ...." Vollenweider (1968) summarized the
eutrophication of waters as meaning "... their enrichment in nutrients and
the ensuing progressive deterioration of their quality, especially lakes, due
to the luxuriant growth of plants with its repercussions on the overall
metabolism of the water affected ...." A search of the literature on
eutrophication indicates that the meaning of the term, originally limited to
the concept of changing nutrient levels, has been gradually expanded to
include the consequences of nutrient enrichment.
Eutrophication occurs both naturally and as a result of rrvan1 s activities
(cultural or anthropogenic eutrophication). Many of man's practices relating
to the disposition of municipal sewage and industrial wastes and to land use
impose relatively large nutrient loadings on many lakes and rivers. In many
cases, the enrichment results in algal blooms and other symptoms of
eutrophication. The consequences of man-induced eutrophication often make
the water body less attractive to potential users. More importantly, at
least when a long-range viewpoint is adopted, eutrophication accelerates lake
succession and shortens the time period before a lake loses its identity.
A comment regarding eutrophication is in order. In the popular press and
the mind of the layman, the term is equated with a "bad" or highly
undesirable situation. Certainly when the enrichment levels reach extremes
and undesirable manifestations occur (e.g., algal blooms, fish kills,
obnoxious odors), the water body loses much of its value as a natural
25
-------
resource. However, enrichment of natural waters can result in increased
primary productivity, leading to a larger biomass of consumers. Eutrophic
water bodies often provide excellent warm-water fisheries.
The lakes of Illinois are undergoing eutrophication and the successional
process as previously described. However, many of the lakes do not exhibit
the secondary effects of eutrophication (e.g., algal blooms) because the
silt-related turbidity of these waters greatly reduces light penetration,
resulting in light-limited conditions. Water quality in Illinois streams and
lakes is appreciably affected by dissolved and suspended matter carried by
runoff from the land surface. IEPA (1976) provides an overview of the
situation in Illinois; the following paragraph draws heavily from their
reports.
Agricultural runoff and soil erosion are two nonpoint sources that affect
the water quality of Illinois lakes and streams. Other major nonpoint
sources (of a more localized nature) that affect water quality include active
and abandoned coal-mining areas, intensive livestock and specialized
agricultural operations, and storm drainage from urbanized areas and
construction sites. Agricultural runoff and runoff from ordinary
precipitation events contain many contaminants (e.g., organic materials that
are oxygen demanding, minerals derived from the soil or applied by man, fecal
coliform bacteria, pesticides, herbicides, fertilizers, and other chemicals)
from ground surface and ground cover that have accumulated through natural
processes and nonintensive land husbandry. When rainfall of sufficient
intensity occurs, soil erosion results. The severity and frequency of soil
erosion is a function of many factors including intensity of immediate
rainfall, prior climatic conditions, soil cover, soil texture, topography,
and antecedent human activities. The eroded soil contributes both dissolved
and suspended matter to the flowing waters. The suspended matter may impair
the recreational use of the body of water as well as such vital biological
functions as photosynthesis, respiration, reproduction, feeding, and growth.
The suspended materials contributed by agricultural runoff and erosion may
also be deposited in streambeds and lake beds. The deposited soil can bury
aquatic life, create an oxygen demand, and release nutrients and chemicals to
the flowing stream or overlying lake water. The influx of nutrients to a
lake, assuming they are not deposited on the bottom and overlain by other
materials, tend to make the water body more eutrophic. The accumulation of
materials on the lake or reservoir bottom decreases the water depth and moves
the water body closer to the time when its identity as a lake or reservoir is
lost.
26
-------
SECTION 5
METHODOLOGICAL OVERVIEW
Most attempts to classify or ordinate lakes employ contact-sensing
techniques coupled with the observations of the field crew to document the
characteristics of the water bodies. The major constraints of most
classification systems are the neccessity for elaborate field data,
difficulties in obtaining data for all lakes within a comparable time period
or comparable physical circumstances, and lack of sufficient or appropriate
sample locations to characterize the entire lake.
A good historical data base for most lakes in Illinois is either not
available or not suited to the development of an overall lake classification
system. Several attempts to characterize these lakes according to sample
data and field observations have had limited usefulness since the data were
not intensively collected within a short time period and since they relied,
in part, upon subjective observations of field personnel. It appears that
satellite-borne sensors such as the multispectral scanner carried by LANDSAT
are capable of collecting data of value for lake classification and
monitoring activities. The LANDSAT space observatories are attractive
because they provide repetitive coverage, a synoptic view, and a permanent
record. The LANDSAT capabilities offer a unique opportunity to obtain a data
base that could group the lakes into categories according to their spectral
responses and also provide the opportunity to study relationships between
certain trophic indicators and the spectral data with an eye toward the
development of predictive models. LANDSAT provides what may be an
economically viable technique for collecting data for the entire surface area
of each lake within a reasonable time period. In about 25 seconds the
LANDSAT multispectral scanner (MSS) collects data in four bands of the
spectrum for an area of the earth covering about 34,225 square kilometers.
In regions of the earth where lakes are very abundant, a typical LANDSAT
scene may contain several hundred to more than a thousand inland water
bodies. With two satellites in operation and assuming cloud-free conditions,
repetitive coverage is provided on a 9-day basis. Clearly, the satellite
offers certain advantages over conventional contact-sensing techniques.
Before discussing in more detail the characteristics, capabilities, and
limitations of LANDSAT in the area of lake classification, it is necessary
that we examine the optical properties of water.
27
-------
OPTICAL PROPERTIES OF PURE WATER AND NATURAL WATERS
It is readily apparent, even to the casual observer, particularly if he
or she is looking downward from an aircraft, that lakes differ in color and
brightness. Many investigations have been undertaken to develop an
understanding of the processes that result in the observed phenomena.
Although a detailed description of the interaction between electromagnetic
energy and the components of the hydrosphere and atmosphere is outside the
scope of this report, a brief discussion is essential to gain some
understanding of the principles that both permit and yet constrain the use of
remote-sensing techniques in lake classificatory work.
The interaction between electromagnetic energy and chemically pure water
has been studied by numerous investigators (e.g., Ewan 1894, Sawyer 1931,
Collins 1925, James and Birge 1938, Hulburt 1945, Raman 1922, Dawson and
Hulburt 1937). The transmission of electromagnetic energy through a material
medium is always accompanied by the loss of some radiant energy by
absorption. Some of the energy is transformed into other forms (e.g.,
chemical) or to some longer wavelength of radiation (e.g., thermal infrared)
(James and Birge 1938). Pure water is very transparent to violet, blue, and
green light. In the infrared region, the extinction coefficient is high with
a complementary low degree of transmission (Table 2). The absorption
spectral characteristics of pure water can be modified greatly through the
addition of dissolved and particulate materials.
The absorption spectra of natural waters (e.g., lake and ocean) have been
studied in detail by Jerlov (1968), Duntley (1963), Atkins and Poole (1952),
Birge and Juday (1929, 1930, 1931, 1932), and Juday and Birge (1933), to
mention a few. Hutchinson (1957) has summarized the more important attempts
to elucidate the interactions of light with natural waters, particularly with
regard to lakes.
An electromagnetic wave impinging on the surface of a lake decomposes
into two waves, one of which is refracted and proceeds into the aquatic
medium and the other of which is reflected back to the atmosphere (Jerlov
1968). The wave entering the water is refracted as it passes through the
air-water interface according to Snell's Law, which may be expressed as
n = sin(i)/sin(r)
where
(i) = angle of incidence
(r) = angle of refraction
n = refractive index, which for water is approximately 1.33.
Most of the electromagnetic energy entering a lake is attenuated through
the process of absorption. Although only a small percentage (less than 3
percent) (Davis 1941) of the incident energy is backscattered from the lake
water volume, this light (volume reflectance) is the focus of interest in the
28
-------
TABLE 2. OPTICAL PROPERTIES OF PURE WATER (ROOM TEMPERATURE)1
Wavelength
(nanometers)
820 (near infrared)
800
780
760
740
720
700
680 (red)
660
640
620 (orange)
600
580 (yellow)
560
540
520 (green)
500
480
460 (blue)
440
420
400 (violet)
380 (ultraviolet)
Extinction
Coefficient
2.42
2.24
2.31
2.45
2.16
1.04
0.598
0.455
0.370
0.310
0.273
0.210
0.078
0.040
0.030
0.016
0.0075
0.0050
0.0054
0.0078
0.0088
0.0134
0.0255
Percentile
Absorption
91.1
89.4
90.1
91.4
88.5
64.5
45.0
36.6
31.0
26.6
23.5
19.0
7.0
3.9
3.0
1.6
0.77
0.52
0.52
0.70
0.92
1.63
2.10
Refractive
Index
1.329
1.331
1.333
1.338
1.343
*Adapted from Hutchinson (1957)
remote-sensing aspect of water quality investigations. Its spectral
characteristics have been shaped by the materials found in the lake's waters
(dissolved and suspended materials, plankton, aquatic macrophytes, and air
bubbles).
The attenuation of electromagnetic radiation in lake waters is a
consequence of the relatively unselective effect of suspended particulate
materials and the highly selective effect of dissolved coloring matter,
usually of organic origin, on the electromagnetic spectrum. The dissolved
matter absorbs strongly in the violet and blue wavelengths, moderately in the
middle wavelengths (e.g., green), and only weakly at longer wavelengths
(Hutchinson 1957). When the dissolved materials are present in small
quantities, the water will be most transmissive in the green wavelengths.
Lake waters with large amounts of dissolved substances are more transmissive
in the orange and red wavelengths than in the shorter wavelengths. However,
the transmission of red and orange light is still greater in pure water than
29
-------
in water containing participate or dissolved materials. As water
transparency diminishes, the detectable electromagnetic energy will be of
progressively longer wavelength at increasingly shallower depths (Hutchinson
1957).
The color of a lake is the color of the electromagnetic energy back-
scattered from the lake body and reflected from the lake surface to the
sensor. Lake color need not be, and is usually not, the same as the color of
the lake's water. Lake color ranges from the blue of pure water through
greenish blue, bluish green, pure green, yellowish green, greenish yellow,
yellow, yellow brown, and clear brown (Hutchinson 1957). Welch (1952)
defines water color as "... those hues which are inherent within the water
itself, resulting from colloidal substances or substances in solution" (i.e.,
true color). Lakes that are blue in color lack appreciable quantities of
humic materials and colored materials in suspension (e.g., phytoplankton).
The bluer the lake color, the smaller the amount of free-floating organisms
contained in the water (Ruttner 1963). Waters with a high plankton content
possess a characteristic yellow-green to yellow color. The characteristic
color may not be apparent because of masking by other materials (e.g.,
suspended sediments). Ruttner (1963) suggested that:
"A lake with very transparent and dark blue, blue-green
or green water is always oligotrophic. On the other
hand, eutrophic lakes always have a relatively low
transparency and are yellow-green to yellow-brown in
color; but the determination of these optical properties
alone will not establish the productivity type, for the
turbidity can be of inorganic origin, and the color can
come from humic substances."
Seston color (color that is attributable to the reflection spectra of
suspensoids of microscopic or submicroscopic size) is often observed in
highly productive lakes. Lakes containing large quantities of suspended
inorganic matter (e.g., silt) may acquire a characteristic seston color, but
in most cases the color is related to large concentrations of phytoplanktonic
organisms (Hutchinson 1957).
Scherz et al. (1969) have investigated the total reflectance (surface
reflectance plus volume reflectance) curves of pure water and natural waters
under laboratory conditions using a spectrophotometer. They reported that
the addition of dissolved oxygen, nitrogen gases, and salts (e.g., NaCl,
Na2$04, Na3P04*H20) had no apparent effect on the reflection curve
However, water from lakes in the Madison, Wisconsin, area had reflectance
curves that differed both from the distilled water curve and from each other.
They attribute these differences to the presence of different algal
organisms, since filtration of the lake waters produced similar reflectance
curves for all lakes even though the curves differed from those of pure water
(Figure 7).
The color of natural waters is the end result of optical processes that
are both numerous and complex. It is relatively easy to detect differences
30
-------
10
o
LJ
_J
u.
Ill
or I
uj 0.4
o
o:
0.0
FILTERED
UNFILTERED
200 400 600 800 1,000 1,200
WAVELENGTH (NANOMETERS)
u.
UJ
a:
10
f: 4
o
UI
U
O
a:
ui
a.
0.4
0.0
— UNFILTEREO
LAKE
KE60NSA
\
200 400 600 800 1,000 1,200
WAVELENGTH (NANOMETERS)
Figure 7. Reflection characteristics of filtered and unfiltered water
samples from two Wisconsin lakes in the area of Madison
(adapted from Scherz et al. 1969).
31
-------
in color within a lake and also among a population of lakes. It is, however,
more difficult to attach physical, chemical, or biological significance to
the color, particularly when quantitative estimates are desired. The
difficulty is compounded in waters having more than one class of particulates
present, which is normally the case in natural water (McCluney 1976), and by
seasonal differences in color within a lake. The degree of success in
sensing and interpreting the significance of color is partially a function of
the sensor type employed for the collection of spectral data.
LANDSAT MULTISPECTRAL SCANNER
Scanner Characteristics
The LANDSAT MSS is a line-scanning radiometer that collects data by
creating images of the earth's surface in four spectral bands: Green (GRN)
500 to 600 nanometers (nm); Red (RED) 600 to 700 nm; near infrared-one (IR1)
700 to 800 nm; near infrared-two (IR2) 800 to 1,100 nm.
The MSS scans crosstrack swaths 185 km in width, simultaneously imaging
six scan lines for each of the four bands (Figure 8). The resultant analogue
signals are sampled, digitized, arranged into a serial digit data stream, and
transmitted to ground stations either in real time or by delayed
transmission. LANDSAT data enter the public domain through the U.S.
Geological Survey's EROS (Earth Resources Observation System) Data Center
near Sioux Falls, South Dakota. The data are available in the form of
photographic products (e.g., black-and-white prints) and computer-compatible
tapes (CCT's). The CCT's contain the data in digital form.
The MSS, as found on LANDSAT-1, -2, and -3, is a low resolution device,
both spatially and spectrally speaking. Three of the bands (GRN, RED, and
IR1) are 100 nm in width while the IR2 band covers 300 nm. Figure 9
illustrates a generalized spectral reflectance curve for a single picture
element (the MSS spatial resolution unit, also called a pixel) of a
hypothetical lake. The width of the MSS bands disallows the recording of the
fine details in the curve. The MSS output more closely resembles Figure 10.
Responses are given as values for the various wavelengths bands (e.g., 500 to
600 nm, GRN) instead of specific values for the entire spectral range. This
procedure crudely defines an entire range of wavelength responses as four
single readings.
The nominal MSS pixel measures 57 by 79 meters, and covers an area of
0.3933 hectares (ha). Through the use of resampling techniques it is
possible to adjust the pixel size (e.g., an 80-meter by 80-meter pixel
corresponding to 0.64 ha was employed in this study). It must be kept in mind
that the MSS gathers energy over the area of its nominal pixel. Many
measurements made using contact techniques are of the point type, a direct
contrast to those acquired by the LANDSAT MSS. It is commonly recognized
that some LANDSAT MSS pixels contain a mixture of water and land features.
This normally occurs along the water-land interface or in situations where
the water body.is much smaller than the pixel or, conversely, where an island
is much smaller than the pixel. The pixel size also tends to give small
water bodies or those with very irregular shorelines a "blocky appearance.
32
-------
SCANNER
•OPTICS
6 DETECTORS
PER BAND:
24 TOTAL
SCAN MIRROR
(OSCILLATES
NOMINALLY
± 2.89°)
NOTE: ACTIVE SCAN IS
WEST TO EAST
FIELD OF VIEW
= 11.56 DEGREES
ACTIVE
"'' SCAN
NORTH
WEST EAST
PATH OF
SPACECRAFT
TRAVEL
6 LINES/SCAN/BAND
Figure 8. Schematic diagram of the LANDSAT-1 MSS scanning arrangement
(adapted from the Data Users Handbook (NASA 1972)).
33
-------
GRN
500 600 700 800 900 1000 1100
Wavelength (nm)
Figure 9. Generalized spectral reflectance curve for a single picture
element (pixel) of a hypothetical lake.
V)
V)
u»
GRN RED IR1
IR2
I
500 600 700 800 900 1000 1100
Wavelength (nm)
Figure 10. Generalized output of the LANDSAT MSS in response to the
spectral distribution illustrated in Figure 9.
34
-------
A visual examination of imagery generated from the LANDSAT MSS will usually
detect a pattern of stripes running nearly orthogonal to the satellite's
path. This is a consequence of an imbalance among the MSS's 24 detectors.
The problem is particularly noticeable when working in the digital domain
(i.e., with the CCT's).
It should be kept in mind that, although the LANDSAT MSS was designed
with the earth's resources in mind, it was not developed specifically for
water. Most of the incident solar energy entering a water body is attenuated
through absorption. The volume reflectance of a water body is generally less
than three percent of the incident light. Thus, the energy reaching the MSS
from water bodies is relatively small in magnitude compared to that received
from land features. While it is possible to increase the MSS's gain in the
GRN and RED bands, this is not normally done.
Peripheral Effects
The character of the electromagnetic energy impinging on the remote
sensor, the LANDSAT-1 MSS in this case, has been shaped through interactions
with numerous environmental phenomena (Figure 11). Some of the interactions
are highly desirable because they mold the character of the light, which may
then be interpreted in terms of some parameter of interest (e.g., Secchi
depth). Other interactions of light energy with the environment (e.g.,
atmospheric scatter) may be detrimental to a particular study. What may be a
vitally important interaction in one study may be devastating in another.
The earth's atmosphere has a pronounced effect on the solar spectrum and
on lake water color as sensed from aircraft and satellite altitudes.
Atmospheric conditions (e.g., degree of cloudiness; presence of fog, smoke,
and dust; amount of water vapor) affect the degree of insolation attenuation.
Weather conditions strongly affect the distribution of energy between
sunlight and skylight (Piech and Walker 1971), contributing a degree of
uncertainty in water quality assessment through remotely sensed color
measurements. Hulstrom (1973) has pointed out the adverse impact that cloud
bright spots can have on remote-sensing techniques that utilize reflected
energy.
The degree of scattering and absorption imposed on the return signal from
water bodies is related to atmospheric transmittance and can result in
changes in lake color when sensed at aircraft-high flight and satellite
altitudes. The attenuated return signal is also contaminated by
electromagnetic radiation from the air column (path radiance). Rogers and
Peacock (1973) have reported that solar and atmospheric parameters have a
serious adverse impact on the radiometric fidelity of LANDSAT-1 data. Path
radiance was found to account for 50 percent or more of the signal received
by the MSS when viewing water and some land masses. The magnitude of the
adverse atmospheric effects can be reduced, though not completely eliminated,
by using imagery or digital data collected on clear, cloudless days. This is
the approach used in this investigation.
The LANDSAT-1 spacecraft passes over the same point on the earth at
essentially the same local time every 18 days. However, even though the time
35
-------
Sun
01
LANDS AT
Skylight
LJULT
i
Backscatter
Volume
Reflectance
Surface
Effects Bottom
Effects
L
Cloud
Shadow
Figure 11. Some components and interactions of light with a hypothetical lake and the atmosphere,
-------
of flyover will remain essentially the same throughout the year, solar-
elevation angle changes cause variations in the lighting conditions under
which the MSS data are obtained. The changes are primarily the result of the
north or south seasonal motion of the sun (NASA 1972). Changes in solar-
elevation angle produce changes in the average scene irradiance as seen by
the sensor from space. The change in irradiance is influenced both by the
change in the intrinsic reflectance of the ground scene and by the change in
atmospheric backscatter (path radiance). The actual effect of a changing
solar-elevation angle on a given scene is very dependent on the scene itself
(NASA 1972). For example, the intrinsic reflectance of sand is significantly
more sensitive to a changing solar-elevation angle than are most types of
vegetation (NASA 1972). The effects of a changing solar-elevation angle are
of particular importance when comparing scenes taken under significantly
different angles. The use of color ratios in lieu of raw data values may be
of value in reducing the magnitude of the solar angle-induced effects by
normalizing the brightness components. The approach is given some
consideration in this study.
A portion of the radiation impinging on the lake surface will be
reflected. The percentage of surface-reflected energy is a strong function
of the angle of incidence. The light reflected from the water-atmosphere
interface is composed of diffuse light from the sky (skylight) and specularly
reflected sunlight. Specular reflection areas contained in a scene are of
little value in most water studies, the possible exception being the
determination of surface roughness. The specularly reflected radiation
exceeds, by several orders of magnitude, the reflected energy emanating from
beneath the water surface (Curran 1972). Surface-reflected skylight,
containing no water-quality color information, can comprise from 10 percent
of the return signal on a clear day to 50 percent on a cloudy day (Piech and
Walker 1971). The surface-reflected skylight not only increases the apparent
reflectance from the water body (volume reflectance) but also affects the
shape of the reflectance curve. Surface roughness is known to have an effect
on the percentages of light reflected and refracted at the interface (Jerlov
1968). However, the effect of surface roughness is negligible in estimating
total radiation entering a water body when the solar-elevation angle is
greater than 15 degrees (Hutchinson 1957).
The lake bottom characteristics (color and composition) will also affect
the intensity and the spectrum of the volume reflectance in settings where
water transparency permits the reflection of a significant amount of
radiation from the bottom materials. In studies involving the estimation of
water depth or the mapping of bottom features, it is essential that the lake
bottom be "seen" directly or indirectly by the sensor. Bottom effects are
capitalized upon and put to a beneficial use. However, in this
investigation, bottom effects are considered to be an undesirable peripheral
effect. A sensor with the capabilities of the LANDSAT MSS is not able to
"see" much deeper into a lake than the Secchi depth. The Secchi transparency
of Illinois lakes is, in most cases, relatively small (e.g., less than one
meter) when cqmpared to the mean depth of each lake. The assumption is made,
as a first approximation, that the bottom effect is relatively insignificant
when considering each of the selected lakes as an entity.
37
-------
It is evident that many factors influence the intensity and spectral
characteristics of the electromagnetic radiation that is collected by the
sensor. Absolute quantification of remotely sensed phenomena requires that
all of the adverse effects be accounted for in the return signal. Failure to
account for all of the variation introduced by the detrimental effects might
be criticized as being simplistic or naive. However, given the present
"state of the art" along with manpower, time, and monetary constraints, and
the ex post facto nature of the project, a complete accounting is not
possible.
SATELLITE SENSING OF ILLINOIS LAKES
A visual examination of LANDSAT MSS imagery indicates that gray-tone
differences can be detected in the study population of Illinois lakes.
Figures 12 through 15 represent, respectively, the IR2, IR1, RED, and GRN
gray-tone images of LANDSAT scene 1448-16023. The IR2 image (Figure 12)
clearly demonstrates a great contrast between water bodies and terrestrial
features. Water is an excellent absorber of radiation wavelengths comprising
the IR2 band, so water bodies appear black. Figure 13, the scene's IR1
counterpart, exhibits a similar contrast between water and land. A careful
examination of the water bodies suggests surface or near-surface phenomena in
some lakes. Gray-tone differences both within specific water bodies and
among members of the lake population are most pronounced in the RED image
(Figure 14). In this band, lakes with extremely turbid water often meld with
the terrain features, a consequence of similar gray-tone values. A vivid
example is presented in Boland (1976). Though less obvious to the eye,
gray-tone differences are also noted among water bodies in the GRN-band image
(Figure 15).
When viewing LANDSAT scenes such as the black-and-white standard
photographs produced by the EROS Data Center, it should be kept in mind that
no special effort has been made to enhance water bodies and related
phenomena. Indeed, a loss of spectral information occurs when the MSS
digital data are transformed into photographic products. Specifically, the
products have a relatively small density range compared to the sensitivity
range of the MSS. This results in a scale compression when the MSS data are
transformed into a film image on an electron beam recorder. In addition, the
range of energy returns from water bodies is small and concentrated at the
lower end of the MSS intensity scale. Scale compression coupled with the
small range of digital number (DN) values adds to the difficulty of
determining trophic state index and indicator values through visual and
densitometric evaluation of "standard" EROS black-and-white photographs.
As can be seen from Figures 12, 13, 14, and 15, it is possible to detect
spectral differences for Illinois lakes using LANDSAT imagery coupled with
photointerpretive techniques. The real problem is one of relating the
spectral variations to chemical, biological, and physical phenomena
measurable through contact-sensing techniques or acquired through
ground-level observation.
As indicated earlier, the quantity and spectral composition of radiation
directed upward across the water-atmosphere interface is, in part, a function
38
-------
Figure 12.
IR2 image of LANDSAT scene 1448-16023 (October 14, 1973). The
IR2 band is excellent for separating water bodies from land.
39
-------
Figure 13. IR1 image of LANDSAT scene 1448-16023 (October 14, 1973). The
band is excellent for discriminating water from terrain.
Surface or near-surface phenomena are evident in some of the
water bodies.
40
-------
Figure 14. RED image of LANDSAT scene 1448-16023 (October 14, 1973).
Variations in water body gray tones suggest differences in
water qua! ity.
-------
Figure 15. 6RN image of LANDSAT scene 1448-16023 (October 14, 1973).
While lacking the contrast of the RED, IR1, and IR2 images
of the scene, gray scale differences are still evident among
the water bodies. Compare with Figures 12, 13, and 14.
42
-------
of the dissolved substances and particulate materials in the water. While
water itself is capable of scattering and absorbing light, the major portion
of the scattering is caused by materials in the water. Scattering as a
result of dissolved color is highly selective, while suspended solids tend to
affect volume reflectance in a rather nonselective fashion. It then follows
that increases in suspended particulate materials in lake water will tend to
increase the reflectance in the LANDSAJ bands.
It should be noted that some natural waters will, at least for a portion
of the spectrum, exhibit a lower volume reflectance than pure water. Humic
waters have this characteristic as demonstrated by Rogers (1977) and shown in
Figure 16. Humic or "brown water" lakes are relatively common in the
northern portions of Minnesota, Wisconsin, and Michigan. They are much less
common in Illinois, and none of the lakes included in this project are of the
humic type.
Rogers (1977) and Scherz et al. (1975) have demonstrated a simple and
practical technique to determine the spectral signatures of lakes. The term
"satellite residual fingerprint" is used to identify the isolated spectral
signature. Their approach is as follows.
If a very deep, clear lake is found in a LANDSAT scene, and assuming that
bottom signals are not present or at least are insignificant, it follows that
the electromagnetic signal received from the lake by the MSS is attributable
to lake surface signals, plus a very small amount of backscatter from the
water molecules, and atmospheric effects. If another lake containing
dissolved and or suspended solids that interact with light is present in the
scene, then subtracting the spectral band values for the clear lake from the
corresponding band values of the second lake will result in a residual
spectral curve that is related to the impurities present. Computation of the
difference for one or several MSS bands can result in the aforementioned
satellite residual fingerprint. A graphic example of the technique is seen
in Figure 17. The upper portion of the figure illustrates the raw spectral
curves (i.e., spectral signatures) for three Wisconsin water bodies: Yellow
Lake (algal-laden), Moose Lake (humic), and Grindstone Lake (clear). The
lower set of curves (i.e., satellite residual fingerprints) were obtained by
subtracting the clear lake MSS DN value for each band from its counterpart in
each of the other lakes.
Though lacking the elegance often associated with attempts to remove
atmospheric effects through mathematical modeling, the satellite residual
fingerprint technique has much appeal. It is, however, dependent on the
presence of deep, clear lakes. With the exception of Lake Michigan, Illinois
lacks such lakes.
Piech and Walker (1971) have demonstrated a method called the scene color
standard (SCS) technique* for the removal of peripheral effects in water
quality surveys. The technique is attractive because no ground truth is
required for removing the peripheral effects. The SCS approach employs a
Patent pending.
43
-------
1.5
1.0
O
i
i
3
s
-0.5"
-1.0
-— Algal- Ladened Water
— Clear or Distilled Water
A—Humic Water
-Range of LANDS AT MSS Bands-
IR2
CRN I RED I IR1 I
1
400 500 600 700 80O 900
Wavelength (nm)
1000
1100
Figure 16. Characteristic "fingerprints" of clear or distilled water, humic
water, and algal-laden water. The ordinate is scaled in
percentage of laboratory difference (D^ = (pvi - Pv])/Ppi
where D»i and P..I are the volume reflectances for the wat
er
where Pvi an pvi
in question and distilled water, respectively, and ppi_ is the
"fingerprint" obtained under laboratory conditions (adapted from
Rogers (1977).
44
-------
O)
40
30-
B-
M,
Q 10
z
Q ior
Q>
O
2
(0 -10 •
c
0)
• Yellow Lake (Algal-Laden)
A Grindstone Lake (Clear)
• Moose Lake (Humic)
MSS Residual Signals For Three Types Of Lakes
< Spectral Range Of LANDS AT MSS Bands
400 500 600 700 800 900 1000 1100 1200
Wavelength (nm)
Figure 17. LANDSAT MSS spectral signatures and residual curves for three
types of lakes (modified from Rogers (1977)).
45
-------
combination of known reflectances from natural objects and certain
characteristic shadow areas within the scene. Unfortunately, shadows of a
size compatible with the spatial resolution of the LANDSAT MSS are either
lacking in the Illinois scenes or confined to a geographically restricted
area. In addition, the reflectances of sunlit natural objects in
juxtaposition with the shadows (in this case cloud shadows) are not known.
Therefore, the SCS technique cannot be employed to remove the peripheral
effects from Illinois MSS lake data available for this project.
It has been well documented that the MSS is incapable of directly
detecting substances such as nutrients (e.g., phosphorus) in water. This
does not mean, however, that it is impossible to get some estimate of such
substances. Phosphorus, for example, is known to be a key element in primary
productivity, stimulating the production of biomass. Differences in nutrient
levels are often directly related to the magnitude of the manifestations of
eutrophication (e.g., turbidity, chlorophyll a^, algal blooms). Such
phenomena are sensible to the MSS. Again, it should be kept in mind that the
energy return from natural water bodies is generally low compared to that
from land features. Thus, all of the water quality related information is
contained in a relatively small range of DN levels for each band for the
Illinois lakes (Figure 18). This precludes developing trophic indicator
estimates that have the accuracies and precisions of the contact-sensed data.
This project is based on the premise that the volume reflectances of
water bodies represent distinct characteristics of their optical properties,
which are then interpretable in terms of parameters considered important in
assessing the trophic state. This concept assumes:
1. Waters with similar optical properties will yield similar
spectral responses.
2. Under identical light conditions, the volume reflectance as
measured in all LANDSAT bands will generally be lowest for
clean water lakes. The inverse is also assumed.
3. Detritus, phytoplankton, suspended solids, and most other natural
large particulates are Mie scatters and, therefore, scatter
approximately uniformly over the spectrum sensed by the MSS.
As the quantity of scattering materials increases, there is a
relatively uniform increase in the reflectance curve (Piech
and Walker 1971). In other words, the reflectance curve
will become higher and flatter as the water becomes more
turbid.
4. Substances (e.g., phosphorus) that are not sensible to the MSS,
can be sensed indirectly through their effects on parameters
that are sensible to the MSS.
5. Shifts in dominant-color reflectance from the blue range toward
the red-brown range reflect increases in lake productivity or are
associated increases in dissolved color or inorganic turbidity.
46
-------
60-
Reflectance (DN)
50-
40-
30-
20-
10-
\
GRN
V
\
s.
V*
X
RED
\
\
V,
X
\
IR1
^•^
\
IR2
500 600 700 800 900 1,000 1,100
LANDSAT MSS Bands (nm)
Figure la. Range of MSS data for 145 Illinois lakes. The data, averages
for each lake, were acquired October 14, 15, and 16, 1973, and
adjusted to a common date (October 15) using regression analysis.
47
-------
It should be recognized that the data used for calibration purposes in
this project were collected with no thought of their being used in a
satellite-related project. Thus, some highly desirable parameters (e.g.,
total suspended solids, organic particulates, inorganic particulates) were
not measured during the time of satellite flyover. In some cases the
location of the contact-sensed data stations was less than nominal when
viewed through the "eye" of the satellite.
48
-------
SECTION 6
METHODS
DESIGN OVERVIEW
LANDSAT MSS data can be used to classify lakes and reservoirs with little
or no a priori knowledge about the water bodies. However, past experience
has shown that the full capabilities inherent in the LANDSAT MSS can only be
utilized if the MSS. data are used in conjunction with water quality data
obtained through contact sensing concurrent or nearly concurrent with
satellite flyover. The inclusion of contact-sensed data in the study
provides the opportunity to examine the satellite MSS data for statistical
relationships (correlations) with specific water quality parameters. If
correlations exist, it may then be possible to develop models of practical
value for the estimation of trophic indicators and index values.
At this time no generalized model exists that incorporates LANDSAT MSS
and contact-sensed data for the estimation of a trophic indicator or index
for all inland water bodies at different times of the year; it is necessary
to develop regression models specific to a date of LANDSAT coverage. This is
accomplished by elucidating the relationships between the MSS data and
specific trophic indicators and indices for a relatively small group of
benchmark lakes. The resulting regression models are then employed to
estimate the magnitudes of indicators and indices for other lakes in the
LANDSAT scenes. The need for some contact-sensed data requires that either
field crews be dispatched to sample a relatively small number of lakes in the
State, in this case Illinois, or that the requisite data be drawn from
existing data banks. The second option is attractive for both economic and
logistic reasons and was selected for this project.
DATA ACQUISITION
In 1973, the U.S. Envirormental Protection Agency's National
Eutrophication Survey (NES) sampled 31 Illinois lakes and reservoirs during
three periods -- May 7-12, August 7-10, and October 16-19. Details on the
sampling procedures and analytical techniques are found in U.S. EPA (1975).
The data are stored in the U.S. EPA's STORET system.
Over 100 Illinois lakes were sampled by IEPA during late spring and
simmer of 1977 (June 15 through August 21). Lake selection criteria
(generally adhered to) included a minimal surface area of 40 hectares and
public access. The water bodies, well dispersed geographically, were sampled
by means of boats. Each was visited once and generally sampled at three
sites. Parameters measured included temperature, dissolved oxygen, Seccni
49
-------
depth, alkalinity, conductivity, pH, total suspended solids, volatile
suspended solids, total phosphorus, ammonia-nitrogen, and nitrite-nitrate
nitrogen. Field observations were recorded and phytoplankton were identified
and enumerated. In addition, a problem assessment was conducted for over 350
Illinois lakes under the Section 208 Water Quality Management Program.
Existing lake data were collected from various sources, and qualitative
evaluations of lake quality and problems were made by persons familiar with
the lakes. The details of the 1977 IEPA lake sampling program and the
Section 208 lake problems assessment have been published in a separate report
(IEPA 1978a).
LANDSAT-1 and LANDSAT-2 cover more than 95 percent of the State of
Illinois in three consecutive passes (Figure 19). A search was initiated
through the EROS Data Center to determine the availability of LANDSAT MSS
scenes for Illinois that were concurrent or nearly concurrent (within a few
days) with the NES sampling dates for the 31 Illinois lakes that were to
serve as benchmark lakes from which regression model development would be
attempted. Past experience (Boland 1976, Rogers 1977) indicates that lakes
in the north-central part of the United States are best characterized as to
trophic status during the latter part of summer (August to September). Cloud
coverage prevented the use of LANDSAT data from the spring and summer NES
sampling periods. For the most part, complete LANDSAT coverage was available
for Illinois for October 14 to 16, 1973; the NES sampled the lakes from
October 16 to 19. Photographic prints and CCT's for 10 LANDSAT scenes were
ordered through the EROS Data Center (Table 3). The IR2 image of each scene
is displayed as a black-and-white print (Figures 20-30). Each edge of a
scene represents a distance of 185 km on the earth's surface; a scene
typically covers 34,200 square km. The study lakes are identified by serial
number callouts. The names corresponding to the serial numbers are in
Table 4.
TABLE 3. LANDSAT MSS SCENES ORDERED FOR ILLINOIS LAKE STUDY
Path Number Date Scene Number
24 10-14-73 1448-16023
1448-16030*
1448-16032
1448-16035
25 10-15-73 1449-16082
1449-16084
1449-16091*
1449-16093
26 10-16-73 1450-16140
1450-16142
*NASA-Goddard did not produce the CCT.
50
-------
N
Foreward
Overlap "~\
LANDSAT
Coverage
Of
Illinois
October
14-16, 1973
Figure 19. LANDSAT coverage pattern for the State of Illinois
with scenes for October 14-16, 1973.
51
-------
Figure 20. lR2 image of LANDSAT scene 1448-16023 (October 14, 1973). The
massive dark object dominating the scene is the southern end
of Lake Michigan. Study lakes and .reservoirs are identified by
serial number call outs. See Table 4 for their names.
52
-------
Figure 21. Enlarged portion of LANDSAT IR2 print containing lakes
and reservoirs found in scene 1448-16023.
53
-------
IU088-30
Ha
-Ticy ; •
'£" •/.^*^W.-.--i'f',% -g(h
-
Figure 22. IR2 image of LANDSAT scene 1448-16030 (October 14, 1973).
The scene's CCT's were not available from NASA-Goddard.
-------
U888-38I
ness-eel
U089
I40CT73 C N38-59/WB88-20 N
IN838-38
Figure 23. IR2 image of LANDSAT scene 1448-16032 (October 14, 1973)
5E
-------
.;-." *- r-
-
U689-B8I W866-
I«OCT73 C N37-3-VHB88-49 N N37-38/U888-42 tBS SUN EL36 H21«8 I98-B2«-N-1-N-D-IL
uese-ee
Figure 24. IR2 image of LANDSAT scene 1448-16035 (October 14, 1973).
Several Illinois lakes are partially or wholly obscured by
clouds.
56
-------
Figure 25. IR2 image of LANDSAT scene 1449-16082 (October 15, 1973)
57
-------
W089-30I
I50CT73 C N«e-29- !6B&a-7 81
Figure 26.
IR2 image of LANDSAT scene 1449-16084 (October 15, 1973).
The CCT's for this scene lacked the necessary
internal calibration data. Most of the water bodies were
picked up because they appear in forward or side overlap
areas of adjacent scenes; seven were dropped.
58
-------
I50CT73 C N40-2
IU090-38
.ABB9-I9 N N08-2B/W089-I I tlSS 7 D SUN EL36 FKI50 ISB-e^Nnj^L^^gS^WS
14090-001 U089-30I
ISOCT73 C NSe-B
:
9B-6260-N-1-NH3ML NflSf .-1^9-16091-701
Figure 27. IR2 image of LANDSAT scene 1449-16091 (October 15, 1973).
The scene's CCT's were not available from NASA-Goddard.
Most of the water bodies were picked up because they appear
in forward or side overlap areas of adjacent scenes.
59
-------
I50CT73 C N37-32/U096-I7 N N37-ZB-T-BS*
N037-80I IU091-B
HZ148 I98-6268-N-I-N-D-IL NRSR ERTS E-l
r~
9-16833-7 81
4-
U096-38I
Figure 28. IR2 image of LANDSAT scene 1449-16093 (October 15, 1973).
60
-------
160L
.-
IU89I-38
UB9I-00I
Figure 29. IR2 image of LANDSAT scene 1450-16140 (October 16, 1973),
61
-------
Figure 30. IR2 image of LANDSAT scene 1450-16142 (October 16, 1973).
-------
TABLE 4. SERIAL LIST OF ILLINOIS WATER BODIES PROPOSED FOR INCLUSION IN TROPHIC CLASSIFICATION PROJECT
O>
CO
Serial
Number
001
003
003
004
005
006
008
009
010
Oil
012
013
014
015
016
017
018
019
020
Name of STORET
Water Body Number County
Horseshoe
Greenville
New City
(Governor
Bond)
DePue 1752
Spr i ng
Fuller, Taylor,
Bundy
Swan
Spring
Sanganois Con-
servation Area
Me red os i a
Sangchris 1753
Taylorville
Lincoln Trail
State Park
Carlyle 1706
Pa rad i se
Charleston 1708
Wolf
Calumet
Bakers
McGinnis Slough
(Or! and)
Alexander
Bond
Bureau
Bureau
Calhoun
Calhoun
Carroll
Cass, Morgan
Cass, Morgan
Christian
Christian
Clark
Clinton,
Bond, Fayette
Coles
Coles
Cook
Cook
Cook
Cook
Nearby Town
01 ive Branch
Greenville
DePue
Bureau
Graf ton
Grafton
Savanna
Beardstown
Me red os i a
Kincaid
Taylorville
Marshall
Carlyle
Mattoon
Charleston
Chicago
Chicago
Barrington
Or land Park
Latitude
(North)
37-38-32
38-55-45
41-18-38
41-18-05
--
38-57-52
42-03-06
40-06-00
39-53-00
39-38-40
39-31-00
39-20-33
38-43-43
39-25-00
39-27-30
41-40-05
41-41-08
42-08-08
41-38-03
Longitude
(West)
89-21-30
89-33-40
89-19-09
89-21-02
--
90-33-13
90-08-00
90-20-00
90-33-00
89-28-50
89-15-30
87-43-03
89-16-14
88-26-15
88-08-25
87-31-02
87-35-13
88-06-42
87-51-26
Surface Area
(Hectares)
765
314
212
106
61
949
1437
993
685
876
465
59
10522
71
145
170
648
53
127
(Acres)
1890
775
524
262
150
2345
3550
2451
1692
2165
1148
146
26000
176
359
419
1600
130
313
(continued)
-------
TABLE 4. (continued)
Serial
Number
021
022
023
024
025
026
027
028
029
030
031
032
033
034
035
036
037
038
039
040
041
042
043
044
Name of
Water Body
Saganashkee
Skokie Lagoons
Tampier
Mat toon
Paris Twin
Sara
Vandal i a City
Moses
Rend
West Frankfort
Old
West Frankfort
New
Old Ben Mine
We-Ma-Tuk
Anderson
Rice
Canton
Murphysboro
Carbondal e
Cedar
Kinkaid
Sam Parr State
Eagle
Fowl er
Gilbert
STORET
Number County
Cook
Cook
Cook
Cumberland
Edgar
Effingham
1764 Fayette
Franklin
1735 Franklin,
Jefferson
Franklin
Franklin
1765 Franklin
1761 Fulton
Fulton
Fulton
Ful ton
Jackson
Jackson
Jackson
Jackson
Jasper
Jersey
Jersey
Jersey
Nearby Town
Willow Springs
Winnetka
Or land Park
Neoga
Paris
Effingham
Vandal ia
Benton
Benton
West Frankfort
West Frankfort
Sesser
Fiatt
Marbletown
Banner
Canton
Murphysboro
Carbondale
Carbondal e
Grimsby
Newton
Graf ton
Graf ton
Graf ton
Latitude
(North)
41-41-28
42-07-13
41-38-53
39-22-00
39-38-27
39-07-50
39-00-35
38-01-13
38-02-35
37-53-37
37-54-18
38-06-15
40-31-56
40-12-00
40-27-30
40-43-53
37-46-52
37-41-49
37-39-52
37-47-47
29-01-10
38-59-45
39-01-50
38-57-30
Longitude
(West)
87-53-04
87-46-42
87-54-22
88-27.-50
87-41-30
88-37-45
89-07-15
88-52-00
88-57-05
88-48-44
88-47-56
89-00-45
90-10-14
90-11-25
89-56-55
89-58-25
89-23-02
89-13-61
89-16-58
89-25-53
88-07-15
90-33-40
90-34-10
90-31-00
Surface
(Hectares)
132
76
66
310
89
237
267
69
7650
59
87
43
60
552
560
101
58
55
728
1113
73
40+
94
122
Area
(Acres)
325
190
163
765
220
586
660
170
18900
146
214
106
149
1364
1383
250
144
136
1800
2750
180
100+
231
300
(continued)
-------
TABLE 4. (continued)
en
Serial
Number
045
049
050
051
052
053
054
055
056
057
058
059
060
061
062
063
064
065
066
067
068
069
070
071
072
Name of
Water Body
Flat, Brushy,
Deep, Long
Dutchman
Storey
Holiday
Long
East Loon
SI oc urn
Cedar
Bangs
Diamond
Catherine
Channel
Fox
Grass
Marie
Nippers ink
Petite
Pistakee
Round
Spring
Crystal
Wonder
Dawson
Bloomington
Evergreen
STORE?
Number
1751
1754
1725
1757
1758
1759
1755
1756
1727
1733
1750
1703
County
Jersey
Johnson
Knox
LaSalle
Lake
Lake
Lake
Lake
Lake
Lake
Lake
Lake
Lake
Lake
Lake
Lake
Lake
Lake,
Me Henry
Lake
McDonough
Me Henry
Me Henry
McLean
McLean
McLean,
Woodford
Nearby Town
Grafton
Buncombe
Galesburg
Somonauk
Long Lake
Antioch
Williams Park
(Island Lake)
Lake Villa
Wauconda
Mundelein
Antioch
Antioch
Fox Lake
Spring Grove
Antioch
Fox Lake
Lake Villa
Fox Lake
Round Lake
Macomb
Crystal Lake
Wonder Lake
LeRoy
Bloomington
Bloomington
Latitude
(North)
39-01-00
37-28-41
40-59-20
41-36-35
42-22-33
42-27-18
42-15-34
42-25-17
42-16-13
42-15-00
42-29-08
42-29-03
42-25-03
42-25-58
42-27-58
42-24-08
42-25-47
42-23-18
42-21-41
40-30-30
42-14-04
42-24-00
40-24-30
40-39-43
40-38-36
Longitude
(West)
90-04-15
88-56-02
90-24-30
88-39-30
88-08-10
88-04-25
88-11-20
88-05-22
88-07-46
88-00-30
88-07-04
88-08-17
88-08-28
88-09-50
88-08-18
88-10-43
88-07-39
88-12-22
88-04-33
90-43-25
88-21-27
88-20-40
88-43-30
88-56-20
89-02-30
Surface
(Hectares)
283+
48
54
133
136
66
87
115
120
60
59
129
692
598
209
240
67
829
87
112
92
295
61
257
283
Area
(Acres)
700+
118
133
328
335
163
215
284
297
149
146
318
1709
1478
516
592
165
2048
215
277
228
729
150
635
700
-------
TABLE 4. (continued)
Serial
Number
073
074
075
076
077
078
079
080
081
082
083
085
086
087
088
089
090
091
092
Name of STORE!
Water Body Number
Decatur 1714
Carlinville
Gillespie New
Otter
Highland (Silver) 1740
Stephen A. Forbes
Central 1 a
Raccoon 1762
Marshall County
Public Hunting
& Fishing Area
(Babb, Sawyer,
Wightman)
Goose
Chain, Ingram,
Sangamon, Staf-
ford, Stewart,
Snicarte
Crane
Clear
Chautauqua
Liverpool
Matanzas
Quiver
Mermet Conserva-
tion Area
Keithsburg
National Wild-
life Refuge
County
Macon
Macoupin
Macoupin
Macoupin
Madison
Marion
Marion
Marion
Marshall
Marshall
Mason
Mason
Mason
Mason
Mason
Mason
Mason
Massac
Mercer
Nearby Town
Decatur
Carl invi lie
Gillespie
Girard
Highland
Omega
Central i a
Central i a
Lacon
Sparland
Snicarte
Liverpool
Havana
Liverpool
Havana
Havana
Mermet
Keithsburg
Latitude
(North)
39-49-30
39-14-30
39-08-20
39-24-12
38-46-05
38-43-13
38-33-24
38-32-40
41-00-53
41-15-45
—
40-07-15
40-25-00
40-22-30
40-22-10
40-15-00
40-20-08
37-15-28
—
Longitude
(West)
88-57-11
89-52-00
89-05-20
89-54-30
89-41-50
88-45-12
89-00-12
89-06-15
89-25-35
89-14-45
—
90-16-40
89-57-00
90-01-00
90-02-25
90-06-00
90-02-30
88-50-47
—
Surface
(Hectares)
1252
68
84
310
223
213
182
374
1035
526
1458+
306
592
1442
63
146
165
183
72
Area
(Acres)
3093
168
207
765
550
525
450
925
2557
1300
3600+
756
1463
3562
155
361
407
452
178
(continued)
-------
TABLE 4. (continued)
Sft
Serial
Number
093
094
095
096
097
098
099
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
Name of STORET
Water Body Number
Swan
Coffeen 1711
Lou Yaeger 1726
Jacksonville
Mauvaise Terre
DuQuoin
Pinckneyville
New Pittsfield
Goose
Turner
Sawmill
Senachwine
Baldwin 1763
Olney East Fork
01 ney New
George
Frank Hoi ten
State Park Pond
Three (Grand
Marais)
Glen 0. Jones
Harrisburg
Springfield 1742
Pana
Shelbyville 1739
County
Mercer
Montgomery
Montgomery
Morgan
Morgan
Perry
Perry
Pike
Putnam,
Bureau
Putnam
Putnam
Putnam
Randolph,
St. Clair
Richland
Richland
Rock Island
St. Clair
Saline
Saline
Sangamon
Shelby,
Christian
Shelby,
Moultrie
Latitude Longitude
Nearby Town (North) (West)
New Boston
Coffeen
Litchfield
Jacksonville
Jacksonville
DuQuoin
Pinckneyville
Pittsfield
Henry
Granville
Henry
Baldwin
Olney
Olney
Andalusia
East St. Louis
Equal ity
Raleigh
Springfield
Pana
Shelbyville
41-14-35
39-02-15
39-11-15
39-40-15
39-42-35
38-04-00
38-06-00
--
--
41-18-47
__
41-10-00
38-12-25
38-45-08
38-47-04
41-25-11
38-34-48
37-41-00
38-50-45
39-41-14
39-21-00
39-24-30
91-03-35
89-23-45
89-35-58
90-12-45
90-12-45
89-13-30
89-24-10
--
--
89-15-00
_-
89-21-00
89-51-55
88-04-15
88-03-49
90-49-50
90-05-13
88-23-00
88-35-18
89-38-58
89-01-25
88-46-30
Surface Area
(Hectares)
49
420
514
193
70
99
67
98
1143
122
255
1346
796
379
56
68
54
43
85
1630
89
4452
(Acres)
120
1038
1269
477
172
244
165
241
2823
300
630
3324
1967
935
138
167
133
105
209
4025
220
11000
(continued)
-------
TABLE 4. (continued)
01
oo
Serial
Number
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
134
135
136
137
Name of STORE!
Water Body Number
Lake of the
Woods
Pekin
Worl ey
Spring
Little Grassy
Lyerla-Autumnal
Flooding
LaRue-P1ne
Hills Ecological
Area
Vermilion 1748
Washington
County
Sam Dale State
Cattail
Sunfish
Crab Orchard 1712
Devil's Kitchen
Marion
Pierce
Horseshoe 1766
Lake of Egypt
Big
Mud, Sand
Lily
Commonwealth
Edison-Dresden
Nuclear
County
Tazewel 1
Tazewel 1
Tazewel 1
Tazewell
Williamson,
Jackson
Union
Union
Vermil 1on
Washington
Wayne
Whiteside
Whiteside
Williamson
Williamson
Williamson
Winnebago
Madison
Johnson,
Williamson
Brown
Calhoun
Cass
Grundy, Will
Nearby Town
Pekin
Pekin
Pekin
Manito
Makanda
Reynoldsville
LaRue
Danville
Nashville
Johnsonville
Fulton
East Clinton
Carterville
Marion
Marion
Rockford
Granite City
Goreville
Versailles
Gilead
Beardstown
Morris
Latitude
(North)
40-35-11
40-35-00
40-35-49
40-30-59
37-38-12
--
—
40-09-24
38-16-20
38-32-29
41-51-55
--
37-43-50
37-38-06
37-40-49
42-20-55
38-41-01
37-37-15
39-58-15
39-08-12
--
41-21-30
Longitude
(West)
89-38-52
89-35-15
89-38-05
89-48-30
89-07-45
__
__
87-39-03
89-21-30
88-35-00
90-08-30
--
89-08-30
89-06-18
88-57-26
88-58-50
90-06-48
88-56-43
90-31-00
90-41-07
--
88-15-00
Surface
(Hectares)
44
43
105
520
405
105
382
246
119
79
47
72
2819
328
89
66
853
931
106
110
115
526
Area
(Acres)
108
105
259
1285
1000
259
943
608
295
194
115
178
6965
810
220
163
2107
2300
262
271
285
1300
(continued)
-------
TABLE 4. (continued)
O)
10
Serial
Number
138
139
140
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
Name of STORE!
Water Body Number
Coal City
Recreation Club
Goose( Village
Club)
South Wilmington
Fireman's Beach
and Park Club
Snyder's Hunting
Club
Spring Arbor
Apple Canyon
Bracken
St. Mary's
Sand
Countryside
Crocked
Deep
Fourth
Gages
Highland
(Old Taylor's)
Zurich
Third
West Loon
Argyl e
Griswold
McCullom
Sunset
County
Grundy
Grundy
Grundy
Jackson
Jackson
Jo Daviess
Knox
Lake
Lake
Lake
Lake
Lake
Lake
Lake
Lake
Lake
Lake
Lake
McDo rough
McHenry
Me Henry
Macoupin
Nearby Town
Coal City
Morris
E, Brooklyn
Elkville
Carbondale
Apple River
Galesburg
Mundelein
Lindenhurst
Mundelein
Lake Villa
Lake Villa
Lake Villa
Graysl ake
Grayslake
Lake Zurich
Grayslake
Antioch
Colchester
Island Lake
McHenry
Girard
Latitude
(North)
__
--
—
37-54-17
37-38-48
42-26-00
40-51-30
42-16-57
42-24-33
42-15-20
42-25-22
42-55-22
42-23-27
42-21-03
42-31-47
42-11-45
42-22-31
42-27-13
40-27-15
42-17-17
42-21-42
39-26-12
Longitude
(West)
M •»
—
—
89-10-50
89-10-09
90-10-00
90-21-00
87-59-44
88-02-31
88-30-15
88-02-31
88^.04-01
88-01-29
87-59-52
88-03-52
88-06-27
88-30-47
88-05-03
90-47-30
88-13-16
88-17-32
89-51-20
Surface
(Hectares)
129
109
41
81
41
194
70
41
47
57
53
81
126
56
45
92
64
66
38
57
99
59
Area
(Acres)
318
268
101
200
100
480
172
101
115
141
130
200
310
139
110
228
157
163
95
141
245
146
(continued)
-------
TABLE 4. (continued)
Serial
Number
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
Name of
Water Body
Holiday Shores
Wild wood
Mound
Moscow
Jack, Swan,
Grass
Bath
Otter
Ki rineman
Petersburg
Fyre
Moredock
Swan
Thunderbird
Open (Marshy)
Sahara Coal
Company
Big
Long
Sugar Creek
(Curry)
Yorky
Upper Smith
(Atkinson)
Lower Smith
Powerton Cooling
Vermilion
Fishing Club
STORE!
Number County
Madison
Marshall
Mason
Mason
Mason
Mason
Mason
Mas sac
Menard
Mercer
Monroe
Putnam
Putnam
Saline
Saline
Schuyler
Schuyler
Schuyler
Schuyl er
Scott, Morgan
Scott
Tazewel 1
Vermilion
Nearby Town
Edwardsville
Vorna
Snicarte
Bath
Bath
Bath
Snicarte
Unionville
Petersburg
Sherrard
Valmeyer
Henry
Putnam
Stonefort
Carrier Mills
Frederick
Browning
Frederick
Beardstown
Naples
Naples
Pekin
Oak wood
Latitude
(North)
38-55-05
41-04-16
40-07-00
40-10-35
40-11-30
40-07-00
40-04-00
37-04-56
39-59-15
41-18-00
38-19-19
41-10-00
41-12-21
—
—
40-04-30
40-08-30
40-05-40
--
39-47-00
39-46-15
—
_-
Longitude
(West)
89-56-19
89-16-25
90-22-00
90-09-45
90-10-45
90-09-45
90-16-45
88-32-27
89-50-55
90-19-30
90-18-40
89-14-00
89-26-53
—
--
90-24-45
90-20-00
90-23-50
--
90-35-30
90-35-30
—
--
Surface
(Hectares)
174
89
140
104
673
56
117
42
77
67
55
115
46
222
47
44
45
49
188
112
51
577
43
Area
(Acres)
430
220
345
258
1662
138
289
103
191
165
135
285
113
548
115
108
111
121
465
277
125
1426
105
(continued)
-------
TABLE 4. (continued)
Serial
Number
184
185
186
Name of
Water Body
Mesa
Little Swan
Summerset
STORE!
Number County
Wabash
Warren
Winnebago
Nearby Town
Lancaster
Avon
Durand
Latitude
(North)
38-31-43
40-40-00
42-27-10
Longitude
(West)
87-51-37
90-32-00
89-23-40
Surface
(Hectares)
41
101
115
Area
(Acres)
102
250
285
-------
Figure 31 depicts the geographic distribution of the water bodies in a county
framework.
NASA-Goddard experienced difficulty in generating the CCT's and was
ultimately unable to provide the CCT's for two scenes (1448-16030 and
1449-16091). In addition, the CCT's for scene 1449-16084 arrived without
internal calibration data; this scene was eventually dropped from
consideration.
DATA PROCESSING
Multispectral Data Processing
LANDSAT MSS data are available from EROS in photographic and digital
(i.e., CCT) formats. Data relating to lacustrine trophic state can be
extracted from both products. However, in water-related studies the use of
digital data is preferred to avoid the uncertainties introduced when digital
data are coded into a photographic product and then requantified through
microdensitometry. The digital approach, selected for this project, permits
the rapid determination of picture element (pixel) counts and descriptive
statistics, and the application of a multitude of digital image enhancement,
processing, and classification techniques.
The LANDSAT CCT's were processed in the Image Processing Laboratory (IPL)
at NASA's Jet Propulsion Laboratory (JPL) using an IBM 360/65 and associated
software and peripherals. The system was operated in two modes, batch and
interactive.
MSS Data Preprocessing-
Prior to attempting classifications of any sort, certain multispectral
data processing procedures must be implemented. Preprocessing functions,
applied to MSS data, are employed to make corrective changes for both
cosmetic purposes and geometric reasons. The cosmetic processing corrects
for line dropouts, slipped or missing lines, and other obvious defects in the
imagery. In terms of geometric corrections, the LANDSAT computer-compatible
tapes (CCT's) are not in a format compatible to the processing approaches
used in the IPL. The CCT's, as received from the EROS Data Center, have the
data for the four MSS bands interleaved. The IPL software program, VERTSLOG,
separates the interleaved data and creates a separate image for each band.
Next, various geometric corrections are made to compensate for mirror
velocity changes and panorama. The data are resampled to create an
instantaneous field of view (IFOV) approximating 80 meters. In addition, the
MSS data are expanded from seven bits of precision in the green (GRN), red
(RED), and near infrared-one (IR1) bands, and six bits in the near
infrared-two (IR2) band, to eight bits of precision resulting in 256 digital
number (DN) levels (0-255).
Lake Extraction Methodology—
The primary thrust of this task is water quality monitoring and lake
classification. The project is not concerned with land use or land-use
practices as they relate to water quality at this time. The image-processing
72
-------
LEGEND
0 Kilometers 50
Water body not included in
rankings and dendrograms because
of faulty or missing MSS data.
NES - sampled water body used in
regression model development and
included in rankings and dendrograms.
Water body included in rankings and
dendrograms.
Figure 31.
ILLINOIS WATER BODIES
Geographic distribution of the Illinois water bodies in
a county framework.
-------
techniques used early in this project were designed to extract and manipulate
MSS pixels, representing surface water, in a batch mode of operation. The
extraction procedure, explained in detail by Blackwell and Boland (1975), is
outlined below.
Upon completion of the previously described preprocessing functions, a
hard-copy image is generated from the rescaled band 7 (IR2) data (e.g.,
Figure 24). A candidate lake is selected from the scene, and a polygon is
constructed around it. The polygon's coordinates are input to the computer
system, and four new images, each an MSS band rendition of the subsection of
the LANDSAT scene, are generated depicting water body, surrounding terrain,
and a histogram of DN values for all of the pixels comprising the subsection
(Figure 32).
Through inspection, and after comparative testing, it has been determined
that an IR2 DN value of 2"8 provides good segregation of water and land
features. A binary mask is developed from the IR2-extracted lake image by
setting IR2 data values between 0 and 28 equal to 1 and all other IR2 DN
values (29 to 255) equal to 0. The binary mask, in which water pixel values
equal 1 and nonwater pixel values equal 0, is then used to eliminate all but
water-related features in the subscene. Multiplication of each MSS band
subsection image [4 (GRN), 5 (RED), 6 (IR1), 7 (IR2)] pixel by its IR2
binary-mask counterpart produces an image for each band. If processed
correctly, the images will represent only pixels containing water-related
information. Figure 33 is an example of the image produced by masking each
subscene image with its counterpart IR2 binary mask.
Some final editing is required to eliminate rivers, streams, and other
water-related features not considered to be part of the lake proper. Once
editing is completed, listings are generated of pixel counts, DN histograms,
and mean DN values for each band for the entire water body, along with their
associated standard deviations. The water body's mean DN values (Appendix
Table A-l) for each of the four LANDSAT MSS spectral bands are used for model
development and classification purposes.
Interactive Lake Extraction Methodology--
The lake extraction methodology previously described was originally
designed to handle a relatively small number of lakes in a batch processing
computer mode. The method, though accurate, was inherently slow since the
image-processing analyst necessarily had to wait for products before
continuing with the next phase of processing. However, during the course of
this project, the capability for interactive image processing at JPL's Image
Processing Lab was developed. The interactive system enabled analysts
associated with this task to develop and utilize a series of three programs
that effectively reduced the time to isolate a lake, increased the accuracy
of the water-detection scheme, and output a statistical and surface area
listing for any given lake. The overall system is called LAKELOC. A
detailed description of the hardware, program operations, water-detection
algorithm, and associated outputs follows.
Hardware--The host computer is currently an IBM 360/65. The display
controller used is a Ramtek G100B, a versatile video-display device that can
74
-------
1448;1S03! 1 BED
tor-l» U088-49 HDG190
PPC121 SCBLE 79.35M/
CP«B ORCHHRC L*E • 12
1448-1603? 6 IR1 140CT7J
1448-1603? 7 IRS 140CT73 ILLIHD
LHt'ES
»i EL.38 OZ14S
F*Clil
CPWB ORCHnfi'D LHKE ' 127 >
flPlTPIC IP r8^-03^16^094OT7| m.H'*LM
.IPL iMaot p»aet::i« LABOROTDRY
M£flN-,'7.0e (T-
IIPU PIC ID78/03 16--084416I "I M i
JPL IMflGE PROCESSING LBBWOTORY
IIPL PIC It.?.f
JPL IH«QE PCDCE'i^INIj LuEDPOTDRV
Figure 32- GRN RED, IR1, and IR2 images of a LANDSAT scene 1448-16035 subscene. The histograms depict
the DN distributions for the subscene including both the water bodies and the land cover.
The large body is Crab Orchard Lake (serial number 127) in Williamson County.
-------
"Hvlil
CRO! V-.tttl •
14W
. •
I
M=39.15 0=2.291
MEAN=24.32 o=2.67|
|MEAN=17.36 0=5.331
|MEAN=6.76 o=6.33l
•
Figure 33. GRN, RED, IR1, and IR2 images of a LANDSAT scene 1448-16035 subscene after the application
of the binary mask. The histograms depict the DN distributions for Crab Orchard Lake
(serial number 127).
-------
be used to display grey-level images and graphics data. The Ramtek is a
solid-state refresh memory system with a display format of 512 lines by 640
elements. Readback from the refresh memory is available under software
control. It is possible to display six-bit gray-level images along with two
graphics planes, and the user may selectively write or erase the displays
point by point. Manipulation of the graphics data can be accomplished with
the aid of a trackball cursor. Figure 34 illustrates the configuration of
the interactive hardware as it is arranged for the operation of the LAKELOC
program.
Operation of LAKELOC—For the purposes of illustration, the operation of
LAKELOC as It would be applied to a scene in southern Illinois is described
below. Although any number of lakes could be extracted from a scene, this
description will be limited to one lake, in this case Crab Orchard Lake,
located in scene 1448-16035.
For a given digital data scene, such as LANDSAT, the user may selectively
display 512-by-640 element subsections until he locates the water body of
interest. Automatic linear contrast stretching of the displayed scene can be
performed during this operation to aid in the location of the lake.
Once a lake has been located, the trackball cursor is set on the desired
lake and a default 50-by-50 element box is drawn on the graphics plane about
the cursor position. Figure 35 illustrates the default box drawn about the
cursor positioned on Crab Orchard Lake. Since only the area within the box
will be acted on by the water detector, the user must correct the size and
the position of the box relative to the lake so that the lake is contained
within the box boundaries. The size is changed by a simple command to the
program that allows the manipulation of the trackball cursor to control the
box dimensions. The position of the box is also controlled in the same
manner by the trackball. Figure 36 illustrates Crab Orchard Lake completely
enclosed by the box after manipulation of the cursor.
Once the box has been satisfactorily positioned about the lake
boundaries, the user is able to invoke the water detector to isolate the
water body in a binary form. A detailed description of the water detector
follows in the subsection entitled Water Detection Algorithm. In the binary
form, the water bodies appear as white, and nonwater features as black. At
this time the user can magnify the area within the box boundaries by issuing
a "zoom" command with the appropriate magnification factor. The zoom command
redisplays a magnified picture of the boxed area directly over the existing
image. At the conclusion of the edit session, the magnified image is erased
from the screen leaving the original image. This allows the user to continue
uninterrupted with other lakes contained in the existing scene.
Magnification of the image allows the user to easily determine the exact
boundaries of the lake as opposed to any extraneous water information that
may also be displayed in the scene. Figure 37 illustrates Crab Orchard Lake.
The detached white areas represent extraneous water information that is not
associated with the lake. The task of editing out extraneous pixels has been
in the past the most time-consuming chore in the water analysis project.
Aided only infrequently by a map, the user must decide what constitutes the
77
-------
..
Figure 34. Interactive user console with video-display device and trackball. The system is located
in the Jet Propulsion Laboratory's Image Processing Laboratory, Pasadena, California.
-------
boundaries of the lake in question. Previously, the user relied on pixel
listings and hard-copy photographs to locate the lake boundaries. In the
case of a large lake, one was often hampered by cumbersome pixel listings
that had to be carefully pieced together to re-create a lake image. With
LAKELOC, the magnification factor, used in conjunction with the easily
manipulated trackball, allows the user to perform the editing task in a
matter of minutes as opposed to a duration of several days.
The removal of water bodies not associated with the lake of interest can
be performed in two ways. In the first method the trackball controlled
cursor is set point by point on the areas to be removed. The default size of
the area removed is one pixel; however, the user can specify the number of
surrounding pixels to be removed for each erase operation. This method is
most useful when working in close proximity to the boundaries of the lake of
interest, where it is imperative not to remove too large a section of pixels
close to the lake. The second method utilizes continuous erasures as the
trackball cursor is moved across the screen. The size of the area about the
cursor position to be erased can also be controlled by the user in this mode.
Figure 38 depicts Crab Orchard Lake after all extraneous information has been
removed during the editing phase.
Once the user is satisfied that he has isolated the lake of interest, he
assigns the lake a name and commands the program to save the binary image of
the lake on a disk data set. The lake's position in the disk data set is
exactly the same as it is in the original LANDSAT scene. LAKELOC returns to
the user the exact position and size of the extracted lake image as it
appears on the disk data set; a parameter data set that contains this
positional information is also created. This information is necessary to the
operation of the follow-on programs for LAKELOC. At this time the user is
able to continue processing any number of lakes or, if finished, to fetch the
follow-on programs that will process the statistical data.
Follow-on programs--The output from LAKELOC consists of a binary mask
disk data set containing the extracted lakes and a parameter data set
containing the positional information and lake names. The output size of the
binary disk data set is exactly the same as the size of the original LANDSAT
image used as input to LAKELOC. In the next step, the binary data set is
used by the program FARINA to mask out of each corresponding spectral channel
in the original LANDSAT frame the water features that have been processed
through LAKELOC. The output is four data sets containing the original DN
values for each lake in each of the spectral bands. This output can in turn
be used as input to the program STATUS and as input to follow-on MSS
classification programs.
STATUS, utilizing the parameter data set from LAKELOC or punched
parameter cards, produces a statistical analysis in the form of a hard-copy
listing of lake statistics in all four spectral channels. The lakes are
listed by name and ranked according to size. Two tables are printed. The
first (Table 5) consists of lake statistics such as pixel count, surface area
calculations, and shoreline perimeter calculations. The second, Table 6,
lists lake MSS statistics, such as the mean DN level for each lake in each
spectral channel, and the corresponding standard deviations.
79
-------
Figure 35. Crab Orchard Lake with
default 50-by-50 element box. This
is a subsection of scene 1448-16035
(October 14, 1973).
Figure 36. Crab Orchard Lake
contained within correctly
positioned box.
Figure 37. Crab Orchard Lake in
binary form (with magnification
factor of 2) before editing of
extraneous water information.
Figure 38. Crab Orchard Lake in
final, edited form.
-------
TABLE 5. AREAL STATISTICS FOR EXAMPLE LAKES*
** LAKE NAME **
*** LAKE STATISTICS ***
** TOTAL PIXELS ** ** SURFACE AREA **
SQUARE FEET ACRES HECTARES
** SHORELINE **
FEET METERS
Plnckneyville 99
OuQuoin 98
Washington Co 123
Devil's Kitchen 128
Little Grassy 119
Cedar 39
Egypt 132
Klnkald 40
Crab Orchard 127
Rend 29
59
74
103
338
547
855
1121
1267
4027
10694
4064451.0
5097786.0
7095567.0
23284432.0
37682283,0
58900095.0
77224569.0
87282363.0
277416003.0
736698966.0
93.3
117.0
162.9
534.5
865.1
1352.2
1772.8
2003.7
6368.6
16912.2
37.8
47.4
65.9
216.3
350.1
547.2
717.4
810.9
2577.3
6844.1
11650.1
17379.6
31260.7
84010.2
101302.8
161456.4
240158.9
239724.2
363440.0
583720.1
3550.8
5297.0
9527.8
25605.1
39875.6
49209.5
73196.9
73064.4
110771.1
177909.3
•Computer-generated table.
TABLE 6. MSS DN STATISTICS FOR EXAMPLE LAKES*
LAKE NAME
WEEN
*** LAKE MSS STATISTICS ***
** MEAN ** * ** STANDARD DEVIATION **
RED IR1 IR2 * GREEN RED IR1 IR2
PlnckneyvUle 99
DuQuoin 98
Washington Co 123
Devil's Kitchen 128
Little Grassy 119
Cedar 39
Egypt 132
Kinkaid 40
Crab Orchard 127
Rend 29
45.03
35.04
40.63
33.12
36.83
34.68
38.27
35.89
38.84
40.50
28.71
18.28
27.75
16.12
18.60
19.15
22.81
18.94
23.69
26.73
23.71
19.22
25.39
17.38
16.98
17.80
20.06
17.58
20.82
19.87
11.36
9.88
13.42
9.89
9.08
9.16
10.92
9.68
8.13
7.03
2.75
2.33
5.93
1.78
2.54
2.51
3.68
3.00
3.11
5.46
2.23
2.03
8.53
1.86
2.52
3.00
3.50
3.65
3.47
6.51
5.86
5.92
8.98
6.80
6.74
6.38
7.83
7.30
6.37
7.49
6.65
6.44
7.81
7.89
7.86
7.54
9.10
7.77
6.40
6.11
•Computer-generated table.
81
-------
Water detection algorithm -- In the past, the detection of pixels whose
instantaneous field of view (IFOV) is that of water was done in the
straightforward manner of thresholding band 7 (IR2) as described previously.
The low reflectance of water in this spectral range conveniently produced a
bimodal distribution of DN's — one peak for water, another peak for
nonwater. This technique works quite well except in the case where the IFOV
of the scanner is viewing a combination of water and nonwater areas such as
the shoreline of a lake or where cloud shadows straddle the water-land
interface. In this situation, the problem becomes one of trying to estimate
the proportion of each material in the IFOV.
Horowitz et al. (1971) and Work and Gilmer (1976) have investigated the
proportion-estimation problem with encouraging results. Work and Gilmer have
estimated the proportions of water, bare soils, and green vegetation using
LANDSAT bands 5 and 7. This technique requires an estimate of the spectral
signature for pure water, pure bare soil, and pure vegetation. While the
spectral signature of water is fairly easy to estimate, that for soil and
vegetation becomes more difficult. The many variables involved, such as
different soil types, vegetative cover types, and thickness of the vegetative
cover, cause considerable error when estimation is attempted by a completely
automatic processor.
An alternate approach, and the one chosen for implementation, considers
the mixture classes to be only water and nonwater. Bands 5 (RED) and 7 (IR2)
are used in the detection process; bands 4 (GRN) and 6 (IR1) offer little
additional information. The estimation of the spectral signature for water
and nonwater is made over a region within, and immediately surrounding, the
water body.
The spectral signatures (mean DN values) are estimated by an iterative
procedure. First, the two-dimensional space (band 5 vs. band 7) is
partitioned into two regions in which the populations of water and nonwater
typically cluster. The mean is then recomputed for those DN's that fall
within a neighborhood of the initial mean. This process is continued until a
convergent mean has been found for each region.
The proportion estimation that was implemented uses a technique proposed
by McCloy (1977). In Figure 39, W is the mean for water, U is the mean for
nonwater, and P is the DN for any given pixel. P' is the projection of P
onto the line segment WU. If /WU/ is the length of the line segment WU, and
/WP1/ is the length of line segment WP1, then the proportion estimate q for
water is:
where
0 4q 41
If P1 does not fall between W and U, it is given the position of the closest
point, W or U. A decision threshold is set for q at which the pixel is
defined to be water or nonwater.
82
-------
CO
00
P1
Band 5
Figure 39. Geometric interpretation of the water-detection algorithm.
Both the batch and interactive modes of lake extraction provide the user
with MSS data related to the specific water bodies under consideration. For
this project, the band averages for each lake were used, a consequence of
several factors including time and cost.
Average spectral responses for each band for all pixels in a lake do not
account for the variability of responses for a specific portion of a lake and
its associated optical properties. Thus, lake characterization using average
spectral responses for each band demonstrates the average lake response and
not the variation actually measured. An examination of LANDSAT DN-level
histograms for all study lakes supports the idea that lakes, as viewed by the
MSS, are heterogeneous bodies. For example, Figure 40 illustrates the
nonhomogeneous nature of Cedar Lake (serial number 55). Although this aspect
of lakes is recognized, data extraction and subsequent analyses largely
utilized a "whole lake" concept by using band values averaged over all the
lake pixels. This technique provides a general spectral response for a given
lake but does not differentiate between extreme readings within a wavelength
band, nor does it demonstrate precise variations in spectral composition.
The data for each of the project lakes in a LANDSAT scene were treated in
this manner. After final editing, the IR2 images of lakes from a particular
scene were concatenated into one or two photographs (Figures 41-48). In
general, side overlap water bodies, those found in LANDSAT scenes of two
consecutive dates (e.g., October 14 and 15), appear in only one
concatenation. Forward overlap water bodies, those found in the 10 percent
forward overlap area of two scenes of the same date (e.g., Sawmill Lake
(serial number 103) in scenes 1449-16082 and 1449-16084), were only extracted
from one scene and appear in only one concatenation.
83
-------
6O-
50-
w
•® 40-
X
£
!ao-
0)
.0
E
3
Z 20-
10-
0-
1-1
r
i
GRN Band
n
J
. r1
RED Band
J
/I
S n
IR1 Band
•I
^i™. t
1
VJKJL
r
IR2 Band
\rUlNK
•60
-50
-40 .2
X
£
H-
-30 0
k
1
-20 J
-10
-0
30 35 40 45 50 55 10 15 20 25 30 35 40 5 10 15 20 25 30 35
DN Level
DN Level
DN Level
51 5 10 15 20 25 30
DN Level
Figure 40. LANDSAT MSS histograms of four bands of Cedar Lake (serial number 55). The data were
extracted from scene 1448-16023.
-------
CHTHEPINE
CHHHMEL
MnPIE
GRASS
PETITE
EH:T LOCH!
UEiT LQQH
CEIUP
FIX:
NIPPER ;ir«
Pi:Tf*.EE
LUNG
POUND
BEEP
SAMB
GMGE:
HIGH-HHP
» * ^ » **
153 146 154 57
COUNTRY; IK
ZURICH
DIMHOND
BANGS
iLOCUM
it DUE
MCGINNIS
THMPIER
CROOKED
FCR.STH
THIRD
20 23 149
Figure 41. IR2 concatenation of 33 Illinois water bodies extracted from
LANDSAT scene 1448-16023 (October 14, 1973).
85
-------
I I I I I 1 ( I I ( I I I I I ( I ( I I I t I 1 I > > I I I ( I I 1 1 I I 1 I I I I I I I I t I I I I I I 1 I I I 1 I t I 1 I I I I i I I I I
13
15
134
24
107
/*"
124
25 PARIS EAST
16 CHARLESTON
15 PARADISE
c'4 MATTOON
13 LINCOLN TRAIL STATE PARt-
26 SARA
41 SAM PARR STATE
107 OLMEY ICW
10e DLMEY E. FORK
134 MESA
7S STEPHEN A. FORBES
1£4 SAM DALE STATE
LANBSAT FRAME 1448-1603E
ILLINOIS LAKE PROJECT
JPL EPA
12 LAKES EXTRACTED
;TRETCH
[PL PIC D 77/05XS1/13EQ34I
JPL IMAGE PROCESSING LABORATORY
Figure 42. IR2 concatenation of 12 Illinois water bodies extracted from
LANDSAT scene 1448-16032 (October 14, 1973). See Figure 18
for 20 more lakes extracted from the scene.
86
-------
CH&PLEITDH
PAP^BILE
MATTDDM
LIMCDLH TPt,IL ITATE
Figure 43. IR2 concatenation of 20 Illinois water bodies extracted from
LANDSAT scene 1448-16032 (October 14, 1973). See Figure 17.
87
-------
142
-
99 98 £8 31 30 111
V. * * $
37 36 29
-
V
127
32 49
121
110
29
99
98
28
31
30
111
142
175
40
37
38
39
143
119
127
128
129
132
32
121
49
168
110
120
frENB
PINCKrCYVILLE
IlUOUDIn
MO:E1
UECT FRWOTDPT tCU
UE1T FP»*T[PT DLL
': HUNTING CLUE
< COM. COMPANY
MUPPHYIKPD
LITTLE GP^OIY
CPAE DPCH<*D
DEVIL: KITCHEN
I*
-------
:UMMERSET
PIERCE
WONBEP
MCCULLOM
GPi:UDLB
CRYSTAL
HOLIDAY
MCGINNIi
-* r r
COMMONWEALTH EDISON-DRESDEN
NUCLEAR
LAKE VILLAGE CLUE
CDAL CITY RECREATION CLUE
S. WILMINGTON FIREMAN: BEACH
AND CLUE
IiEPUE
TURNER
GDQ:E
:AU MILL
:UAN
THUriliEREIPIl
:ENACHWINE
Figure 45. IR2 concatenation of 20 Illinois water bodies extracted
from LANDSAT scene 1449-16082 (October 15, 1973).
89
-------
UILDUQOD
GOO IE
UOPLE .-'
WKE or THE uani
PEKDI
FDUEPTOH ODOLIHG
llj life ISi
LIE-MiH-TlJK
LIVEPPDOL
PICE
1.FFIHG
MIDEP10N
MOUMD
OUIVEP
MOICDU
MIGwP CPEEK
EIO
DTTEP
Dull' DM
PETEP1EUPG
EVEPGPEED
ELQQnltiGTDtl
Figure 46. IR2 concatenation of 35 Illinois water bodies extracted from LAND-
SAT scene 1449-16084 (October 15, 1973). The CCT's for this scene
were supplied without the necessary internal calibration data.
-------
1£1 LA RUE-PINE HILLS
ECOLOGICAL
HORSESHOE
121
FRAME 1449-16093
8 SPRING
144 APPLE CANYON
CATTAIL
SUNFISH
108 LAKE GEORGE
170 FYRE
93 SMAN
186 108 170 93
FRAME 1450-16140
Figure 47. IR2 concatenation of nine Illinois water bodies extracted
from LANDSAT scenes 1449-16093 (October 15, 1973) and
1450-16140 (October 16, 1973).
91
-------
KETHSBURG NUR
BRACKEN
LITTLE SHAH
SPRING
APGYLE
BIG
tCU PITTlFIELIi
UPRER SMITH
LDUEP SMITH
LILY
SfcMGAMDIS CONSEPVATIDr
Figure 48. IR2 concatenation of 13 Illinois water bodies extracted
from LANDSAT scene 1450-16142 (October 16, 1973).
92
-------
Milltispectral Data Adjustment
Complete multispectral scanner data coverage for Illinois requires
overlapping the data from three passes of the satellite, as shown earlier in
Figure 19. Thus, the raw data set consists of three subsets that are defined
by the orbit during which the information was gathered. This stratification
in the raw data set required that certain transformations be effected to
create a single, unified, date-independent base that could be used for
further analyses.
To accomplish this restructuring, use is made of the fact that lakes on
the eastern and western edges of the scenes (i.e., side overlap lakes) appear
on successive passes. Thus, within the larger data set, there are two
subsets — the October 14 and 15 overlap and the October 15 and 16 overlap.
It appears reasonable, from an examination of the operation of the
satellite, the multispectral scanner, and the resultant data, to expect that
the raw data pairs for the side overlap lakes represent the sum of small
random and systematic effects. If the effects are almost entirely random,
they will be reflected in any statistical adjustments as increases in the
errors of estimation. Although systematic effects can contribute both
additively and multipl icatively, multiplicative effects should be minimized
by calibration of the MSS. Thus, it appeared reasonable, both from
consideration of the processes involved in generating the data and from
plotting side overlap data pairs, that simple linear relationships could be
established for each of the four bands and for each of the two pairs of
dates.
The models that were developed are presented in Table 7. As expected,
the slopes are close to 1.00, the relationship being better for the 14th to
15th conversion than for the 16th to 15th. This may be largely the result of
the greater number of degrees of freedom available for the former estimates.
Data from scene 1449-16084 could not be used in developing these models
because there were no internal calibration data found with it. This was
unfortunate since a substantial number of lakes that might have been included
in developing the models could not be used because their overlap on the 15th
was on this unusable scene. Attempts to develop an internal calibration for
this scene were unsuccessful. Unfortunately, we learned this after extensive
efforts to develop meaningful clusters based on spectral data, and attempts
to develop models relating spectral and chemical data, failed.
Using these models, a data base consisting of 14th and 16th data adjusted
to the 15th, and raw 15th data, was prepared. Duplicates in the data base
arising from the overlap lakes were removed, retaining original 15th data in
preference to calculated values, except for scene 1449-16084 where only
calculated values were used. For this latter reason, seven lakes
(Bloomington, Springfield, Sanghris, Dawson, Evergreen, Goose (Sparland), and
Wildwood) appearing in scene 1449-16084 could not be incorporated into the
final data set. The raw and final sets are presented in Appendix Table A-l.
93
-------
TABLE 7. REGRESSION MODELS USED TO ADJUST MULTISPECTRAL SCANNER DATA FOR OCTOBER 14
AND 16 TO OCTOBER 15, 1973
Dependent
Variable
Models for
GRN15
RED15
IR115
IR215
Models for
GRN15
RED15
IR115
IR215
Regression
Constant
Adjusting the
0.5580
-1.1481
0.9823
1.6842
Adjusting the
12.4990
6.3859
7.6123
3.5378
Regression
Coefficient
October 14th
0.9683
1.0161
0.9473
0.8822
October 16th
0.7291
0.7781
0.6159
0.6340
Independent
Variable
Data
GRN14
RED14
IR114
IR214
Data
GRN16
RED16
IR116
IR216
R2
0.8226
0.9279
0.9221
0.9210
0.9445
0.9918
0.9464
0.8652
Standard
Error of
Regression
2.0807
1.2853
1.0811
0.8931
0.6952
0.4725
0.8927
0.7306
Degrees of
Freedom:
Regression,
Residual
1,13
1,13
1,13
1,13
1,5
1.5
1,5
1,3
-------
Trophic Indices Development
A multiplicity of classificatory schemes has evolved to group and rank
lakes. Examples of some approaches to lake typology are found in Lueschow
et al. (1970), Rawson (1956, 1960), Margalef (1958), Hansen (1962),
Jarnefelt (1958), Larkin and Northcote (1958), Moyle (1945, 1946), Pennak
(1958), Round (1958), Whipple (1898), Winner (1972), Zafar (1959), Beeton
(1965), Donaldson (1969), Uttormark and Wall (1975), Gerd (1957), and Taylor
et al . (in press). Hutchinson (1957, 1967) has reviewed many of the attempts
to arrange lakes into orderly systems. The term "classification" is often
used in the restricted sense of placing entities into distinct groups,
thereby excluding arrangements showing no distinct division (e.g.,
ordination). The term is used here in the broader context suggested by
Sneath and Sokal (1973) and includes ordination.
Lacustrine trophic state is a multidimensional concept and amenable to
analysis by multivariate statistical techniques (e.g., cluster analysis,
principal component analysis). Multivariate techniques minimize the personal
bias often present when data are examined for groups and rankings are
developed. They are of particular value in situations where large numbers of
objects or parameters are to be classified. Principal components analysis
can be used to develop trophic state indices.
Principal components analysis, an ordination technique, may be used to
reduce the dimensionality of a multivariate system by representing the
original attributes as functions of themselves. The main object is to
summarize most of the variance in the system with a lesser number of
"artificial" va Mates (i.e., principal components).
The computation of principal components can be undertaken using either a
covariance matrix (S) or a p x p matrix of Pearson product-moment correlation
coefficients (r). Use of the r-matrix is indicated when the variates are
measured in different units (e.g., grams and meters). Computation of the
r-matrix principal components involves the extraction of eigenvalues
(characteristic or latent roots) and eigenvectors (characteristic or latent
vectors). The eigenvalues are a set of p nonzero, positive, scalar
quantities. The sum of the eigenvalues of the r-matrix is the trace of the
matrix, which is equal to the number of dimensions in the original system
(i.e., the number of variates, p). The rank of the matrix is equal to p.
Normalized eigenvectors give the attribute-space (A-space) coordinates
of an orthogonal set of axes known as the principal axes. The normalized
eigenvectors are comonly designated as principal components. The first
principal component of the observations of the p-variates Xi,...,Xp is
the linear compound
Yj = alj xl + ••• + apj x
j lj l ••• pj p
whose coefficients are the elements of the eigenvector associated with the
jth largest eigenvalue extracted from the r-matrix. The jth eigenvalue is a
measure of the variance of the jth principal component.
95
-------
The proportion of the total sample variance in the cloud of dimensionless
standard scores attributable to any component is found by dividing its
eigenvalue by p. The first principal component has the innate property of
explaining the greatest proportion of the sample variance with each
successive component explaining progressively smaller amounts of the total
sample variance. Frequently, a consequence of the decreasing order of the
variance is that K < p dimensions will adequately summarize the variability
of the original variates, Xi,...,Xp. The first three components
generally account for most of the variation, thereby permitting the
ordination of the subjects in one-, two-, and three-dimensional (3-D) space.
All of the dispersion in the data can be accounted for by using p dimensions,
but this negates the analysis objective, which is the reduction of
dimensionality or, as Seal (1964) stated, the "... parsimonious summarization
of a mass of observations."
The principal components of N p-variate observations are defined
geometrically as "... the new variates specified by the axes of a rigid
rotation of the original response coordinate system into an orientation
corresponding to the directions of maximum variance in the sample scatter
configuration" (Morrison 1967). The normalized eigenvectors give the
directions of the new orthogonal axes, and the eigenvalues determine the
lengths (i.e., variance) of their respective axes. The coordinate system is
expressed in standard units (zero mean, unit variance) when the components
are extracted from the r-matrix. Figure 49 is a hypothetical bivariate
example of the geometrical meaning of principal components. Detailed
descriptions of the theoretical and computational aspects of principal
components are found in Hotelling (1933a, 1933b, 1936), Anderson (1958), and
Morrison (1967).
Principal components analysis was used to develop two trophic indices for
the 31 Illinois water bodies sampled by the National Eutrophication Survey.
The first index (PC1F5) uses the fall sampling values (Appendix Table A-2)
for five trophic indicators: CHLA, ISEC, COND, TPHOS, and TON. The second
index (PC1Y5) is generated using sampling-year mean values (Appendix Table
A-3) for the same trophic indicators. The same methodology was employed in
the development of both indices. The methodology is briefly explained below
using the fall trophic index generation for illustrative purposes. The raw
trophic indicator data (Appendix Tables A-2 and A-3) for the 31 NES-sampled
lakes are skewed and were, therefore, natural log (LN) -transformed to give a
distribution more closely approximating a normal one. The transformed
indicator data are identified as LNCHLA, LNISEC, LNCOND, LNTPHOS, and LNTON.
The data matrix was further standardized by attributes using the relationship
where
z = standardized value for attribute i of observation j (i.e., lake)
*ij - LN-transformed value of attribute i of observation j
96
-------
FIRST PRINCIPAL AXIS
CORRESPONDING TO THE
FIRST PRINCIPAL
COMPONENT S
SECOND PRINCIPAL AXIS
CORRESPONDING TO THE
SECOND COMPONENT S
CHLOROPHYLL a (X,J
Figure 49. Geometrical interpretation of the principal components
for a hypothetical bivariate system.
Principal components may be interpreted geometrically as the variates
corresponding to the orthogonal principal axes of observation scatter in
A-space. The elements of the first normalized eigenvector (i.e.,
coefficients of the first principal component) define the axis that passes
through the direction of maximum variance in the scatter of observations.
The associated eigenvalue corresponds to the length of the first principal
axis and estimates the dispersion along it. The second principal component
corresponds to the second principal axis, the length of which represents the
maximum variance in that direction. In our example, the first component
accounts for most of the dispersion in the data swarm, and the original
2-dimensional system could be summarized in one dimension with little loss of
information. The new variate value (PCI) for each lake is obtained by
evaluating the first component
Y! = aXj + bX2
The PCI for each lake in one-dimensional A-space is its coordinate on the
first component axis, which is shown diagrammatically by projecting each
observation to the principal axis (modified from Brezonik and Shannon 1971).
97
-------
x-j = the mean of attribute i
s.j = standard deviation of attribute i.
Eigenvectors and eigenvalues were then extracted from the associated
correlation matrix (Table 8) and displayed in Table 9. Next, the first
normalized eigenvector was evaluated for. each of the 31 lakes, resulting in
31 trophic index (PC1F5) values, one for each water body. The PC1F5 value
defines a water body's position on the first principal axis. The correlation
coefficients, eigenvectors, and eigenvalues for the sampling-year trophic
index (PC1Y5) are found in Tables 10 and 11, respectively. The resultant
trophic index values for the 31 NES water bodies are found in Table 12.
TABLE 8. R-MODE PEARSON PRODUCT-MOMENT CORRELATION MATRIX OF FIVE TROPHIC
STATE INDICATORS*
CHLA
(LNCHLA)
CHLA 1.000
(LNCHLA) (1.000)
COND
(LNCOND)
ISEC
(LNISEC)
TPHOS
(LNTPHOS)
TON
(LNTON)
COND
(LNCOND)
0.219
(0.405)
1.000
(1.000)
ISEC
(LNISEC)
0.242
(0.479)
0.142
(0.124)
1.000
(1.000)
TPHOS
(UVTPHOS)
0.649
(0.593)
0.596
(0.459)
0.444
(0.641)
1.000
(1.000)
TON
(LNTON)
0.827
(0.758)
0.327
(0.400)
0.370
(0.407)
0.836
(0.771)
1.000
(1.000)
Correlations computed using mean data values of the fall sampling period for
31 NES-sampled water bodies. Numeric values enclosed by parentheses are
correlation coefficients for natural log-transformed data.
Index values for 145 Illinois lakes were calculated from regression
models developed from 22 of the NES lakes (LANDSAT data were available for
only 22 NES lakes). In these models the trophic indices, derived from
principal components analyses, were taken as the dependent variables and the
LANDSAT MSS bands (or some variation thereof) were the independent variables.
The data for the 22 NES lakes were used to develop the models. These models,
found in "Trophic Indicator and Index Estimation" were then used to estimate
trophic state index values for the entire set of 145 lakes.
98
-------
TABLE 9. NORMALIZED EIGENVECTORS AND EIGENVALUES*
ID
Eigenvector
Number
1
2
3
4
5
LNCHLA
0.480
0.029
0.470
0.642
-0.368
LNCOND
0.325
0.764
-0.504
0.181
0.156
LNISEC
0.385
-0.634
-0.505
0.297
0.327
LNTPHOS
0.513
-0.099
-0.208
-0.570
-0.599
LNTON
0.502
0.065
0.476
-0.377
0.612
Eigenvalue
3.085
0.887
0.557
0.363
0.108
57000
Variance
(%)
61.70
17.74
11-. 14
7.26
2.16
Cumulative
Variance
(%)
61.70
79.44
90.58
97.84
100.00
*Eigenvalues and eigenvectors calculated from correlation matrix values based on trophic indicator
data collected during the fall sampling period from 31 water bodies.
-------
TABLE 10. R-MODE PEARSON PRODUCT-MOMENT CORRELATION MATRIX OF FIVE TROPHIC
STATE INDICATORS*
CHLA
(LNCHLA)
CHLA 1.000
(LNCHLA) (1.000)
COND
(LNCOND)
ISEC
(LNISEC)
TPHOS
(LNTPHOS)
TON
(LNTON)
COND
(LNCOND)
0.382
(0.439)
1.000
(1.000)
ISEC
(LNISEC)
0.292
(0.412)
0.028
(-0.033)
1.000
(1.000)
TPHOS
(LNTPHOS)
0.559
(0.622)
0.422
(0.254)
0.443
(0.699)
1.000
(1.000)
TON
(LNTON)
0.937
(0.807)
0.360
(0.399)
0.275
(0.312)
0.652
(0.699)
1.000
(1.000)
Correlations computed using sampling year-mean values for 31 NES-sampled
water bodies. Numeric values in parentheses are correlation coefficients
for the natural log-transformed data.
All of the computational aspects of trophic state index development were
executed on a Control Data Corporation digital computer (CDC 3300) at Oregon
State University using the Statistical Interactive Programming System (SIPS).
A detailed explanation of SIPS and its operation is found in Guthrie et al.
(1973).
Surface Area Estimation
The nominal size of a LANDSAT-1 MSS pixel is 57 meters by 79 meters,
resulting in an areal coverage of 0.4503 hectares per pixel. As indicated
earlier in this study, the surface area of a water body is defined by pixels
having IR2 levels of 28 or less. The total surface area of a water body is
calculated by summing the number of pixels having DN values within the above
range (Appendix Table A-l) and then multiplying by the appropriate conversion
factor. During the CCT preprocessing phase, the MSS data were resampled,
resulting in pixels having nominal edge measurements of 80 m and an area of
0.6400 ha. The value of 0.6400 was used as the multiplication factor to
convert water body IR2 pixel summations to surface area in hectares (Appendix
Table A-l).
Side overlap coverage (October 14 and 15, October 15 and 16) was
available for 22 Illinois water bodies. This self-pairing situation
permitted the comparison of surface area estimates derived from LANDSAT data
collected on consecutive days.
100
-------
TABLE 11. NORMALIZED EIGENVECTORS AND EIGENVALUES*
o
Eigenvector
Number
1
2
3
4
5
LNCHLA
0.519
0.182
-0.278
-0.630
-0.473
LNCOND
0.269
0.703
0.639
0.002
0.160
LNISEC
0.375
-0.617
0.434
-0.362
0.399
LNTPHOS
0.514
-0.258
0.166
0.639
-0.483
LNTON
0.504
0.160
-0.546
0.253
0.598
Eigenvalue
2.926
1.127
0.562
0.265
0.120
5.000
Variance
(%)
58.52
22.54
11.24
5.30
2.40
Cumulative
Variance
(%)
58.52
81.06
97.30
97.60
100.00
*Eigenvalues and eigenvectors calculated form correlation matrix values based on sampling-year
trophic indicator mean values for 31 water bodies.
-------
TABLE 12. TROPHIC STATE INDICES AND RANKINGS FOR 31 NES-SAMPLED ILLINOIS
WATER BODIES
Name of Water
Body
Baldwin
Bloomington
Carlyle*
Cedar*
Charleston*
Coffeen
Crab Orchard*
Decatur*
DePue*
East Loon*
Fox*
Grass*
Highland-Silver
Holiday*
Horseshoe
Long*
Lou Yaeger
Mari e*
Old Ben Mine*
Pistakee*
Raccoon*
Rend*
Sangchris
Shelbyville*
S locum*
Springfield
Storey*
Vandal ia*
Vermilion
We-Ma-Tuk*
Wonder*
Serial
Number
105
71
14
55
16
94
127
73
3
53
60
61
77
51
131
52
95
62
32
65
80
29
11
114
54
112
50
27
122
33
69
PCIF5
-1.74
-1.00
-0.78
-2.67
0.78
-2.52
-0.22
-0.51
2.61
-0.55
0.67
1.71
-1.76
0.86
1.31
2.59
-1.38
1.43
2.38
1.52
-1.01
-1.19
-1.89
-1.35
4.56
-1.21
-0.47
-1.13
-0.09
-1.33
2.33
(Rank)
( 5)
(13)
(14)
( 1)
(21)
( 2)
(18)
(16)
(30)
(15)
(20)
(26)
( 4)
(22)
(23)
(29)
( 6)
(24)
(28)
(25)
(12)
(10)
( 3)
( 7)
(31)
( 9)
(17)
(11)
(19)
( 8)
(27)
PCIY5
-2.00
-1.30
-0.90
-2.89
0.02
-2.55
0.14
-0.17
2.41
-0.77
1.68
1.84
-0.77
1.22
2.43
1.98
-0.77
0.33
1.55
1.63
-0.48
-1.13
-1.48
-1.62
4.31
-1.09
-1.07
-1.32
-0.16
-1.35
2.29
(Rank)
( 3)
( 8)
(12)
( 1)
(19)
( 2)
(20)
(17)
(29)
(15)
(25)
(26)
(13)
(22)
(30)
(27)
(14)
(21)
(23)
(24)
(16)
( 9
( 5)
( 4
(31)
(10)
(11)
( 7)
(18)
1 6)
(28)
NES
504
296
345
528
225
454
347
201
139
399
212
244
229
247
313
195
241
303
240
253
330
442
369
339
210
283
333
323
227
367
183
(Rank)
( 2)
(16)
( 9)
( 1)
(25)
( 3)
( 8)
(28)
(31)
( 5)
(26)
(20)
(23)
(19)
(14)
(29)
(21)
(15)
(22)
(18)
(12)
( 4)
( 6)
(10)
(27)
(17)
(11)
(13)
(24)
\ /
( 7)
\ /
(30)
*Used in the development of regression models for the estimation of trophic
state using LANDSAT MSS bands as the independent variables.
102
-------
DATA ANALYSIS APPROACH
Regression and cluster analysis are two techniques that were employed to
explore the relationships suspected to exist between and within the remotely
sensed and contact-sensed data sets.
Regression Analysis
MSS data presented as false-color or gray-scale imagery provide immediate
visual recognition, not only of structural features such as water bodies,
land surfaces, and vegetation, but also of gradations within and between such
features. Expressed as spectral density histograms, information is derived
about the evenness of the distribution of reflected radiation, without regard
to the spatial distribution of the features giving rise to the pattern.
Using clustering techniques, which will be described later, items such as
lakes may be conveniently grouped together in physically significant ways
solely on the basis of their reflectance patterns. All of these techniques
need to utilize only the information inherent in the MSS data themselves to
produce significant results.
In environmental monitoring studies such as this on lake trophic state,
MSS data by themselves are of limited use. Rather, the scientist needs to
use this readily available data to estimate physical, chemical, and
biological parameters in lakes. To make these estimates, regressions are
developed from a set of lakes for which concurrent contact-sensed and MSS
data are available.
Regression is a statistical term describing the relationship of a
dependent variable (Y) to one or more independent variables (X). Regression
has many uses. In converting MSS data from the 14th or 16th of October to
the 15th, for example, regression was used to effect a change of scale.
Regression was also used to develop predictors or estimators of Y, for given
values of X, as described below.
While several statistical methods for developing regression models exist,
least-squares analysis is most commonly employed. According to Snedecor and
Cochran (1967), three assumptions are made about the relation between Y and X
in univariate models. First, for each value of X there is a normal
distribution of Y from which a sample value of Y is taken at random. Second,
the population of Y's corresponding to the selected value of X has a mean p,
which lies on the straight line
u.. .. = a + BX
where x = X - X
the value of the population mean corresponding to x = 0
a
3
the slope of the line which is the incremental change in Y per
unit change in X.
103
-------
Third, the standard deviation, ffy.x, of Y about the mean is the same for
all values of Y. Then
Y = a + px + «
where « = a random variable drawn from normal population with mean 0 and
standard deviation
-------
Next, the actual models were developed by stepwise multiple regression
using the date-corrected MSS bands as independent variables and trophic
indicators and trophic indices as dependent variables. (A large number of
"new" independent and dependent variables, "created" through a variety of
techniques (e.g., log transformation, ratios, standardization), were examined
during the modeling effort.) The regression was usually carried about two
stages above that of maximum change in the multiple correlation coefficient
(R) or its square (R2). At this point the regression was backstepped until
the most parsimonious model was reached. Selection of a particular model was
made after an examination of the multiple correlation coefficient, mean
residual square, significance of the regression and individual regression
coefficients, number of independent variables, and scatter in the residuals
for several models. A thorough discussion of the analysis and selection of
variables in multiple regression is available in Hocking (1976).
Regressions were developed using the Statistical Interactive Programming
Systan (SIPS) on a CDC 3300 digital computer located at Oregon State
University (Rowe and Barnes 1976). With the particular regression subsystem
used, it was discovered that the last-added independent variable was not
necessarily the first removed in the backstepping mode (Hocking 1976). The
subsystem was consistent in generating a particular model in this fashion,
and the statistics for the chosen model could be confirmed by independent
calculation. The three sets of models in Table 15 were developed in this
manner. These models were used to estimate the values of trophic indicators
and trophic indices for 145 Illinois water bodies. The estimated values were
then used to rank the 145 water bodies.
Cluster Analysis
Boesch (1977) has recently written an excellent review of the application
of numerical classification in ecological investigations of water pollution.
Excerpts from his description of numerical classification are given below:
"In simplest terms, classification is the ordering of entities
into groups or sets on the basis of the relationships of their
attributes. Classification is an important biological process
which must predate man, but the science of classification has
had a fairly recent and parallel development in several dis-
ciplines (Sokal 1974)....
"Numerical classification ojr cluster analysis encompasses a
wide variety of techniques for ordering entities into groups
on the basis of certain formal pre-established criteria rather
than on subjective and undefined conceptions. Numerical classi-
fications have certain advantages over subjective classifica-
tions, notably: (1) they can be based on a much larger number
of attributes than is allowed by human mental capacity; and ('2)
once the classificatory criteria are set, their results are
repeatable by any investigator studying the same data set.
105
-------
"It is important to distinguish classification from several
other processes and analyses. First, the process of "identi-
fication," involving the allocation of additional unidentified
entities to the most appropriate class, once such classes have
been established (Dagnelie 1971, Sneath and Sokal 1973, Sokal
1974), is excluded from classification. ... The optimal
splitting of a continuous into a discontinuous series
(Clifford and Stephenson 1975), is here considered a case of
classification. Secondly, various multivariate analyses other
than numerical classification may be applied to ecological
data. These include, in addition to various regression and
correlation approaches, a broad group of techniques known as
ordination. In ordination the relationships among entities
are expressed in a simplified spatial model of few dimensions,
with no attempt to group or draw boundaries between classes
(Pielou 1969, Whittaker 1967, Whittaker and Gaich 1973, Sneath
and Sokal 1973, Orloci 1975). Ordination includes such
techniques as principal components analysis, factor analysis,
principal coordinates analysis, correspondence analysis, and
multidimensional scaling.
"To orient the reader ..., a brief description of the chain of
procedures in numerical classification is in order. Numerical
classifications are generally directed by a set of algebra-
ically expressed criteria (an algorithm). This chain of opera-
tions begins with the original data, in one or more forms which
may be further transformed to conform to certain preconditions.
In ecological applications the original data are generally in
the form of a matrix of some measure of abundance of each
species in a series of collections (See Figure 50).
"From the original or transformed data matrix most numerical
classifications then require the computation of a resemblance
measure between all pairs of entities being classified. This
is a numerical expression of the degree of similarity, or,
conversely, dissimilarity, between the entities on the basis
of their attributes. In ecology, the entities being classi-
fied may be collections (representing sites, stations, or
temporal intervals) with species content as the attributes.
This may be referred to as a normal classification as opposed
to an inverse classification of species as entities with
their presence or abundance in the collections as attributes
(Williams and Lambert 1961). 'Normal' or 'inverse' are syn-
onymous with the widely used terms 'Q analysis' and 'R
analysis', respectively. However, the Q/R distinction has
been confused in the past (Ivimey-Cook, Proctor, and Wigston
1969) and the normal/inverse inverse terminology is fast
becoming standard in ecology....
106
-------
.£ 3
o
Collections
B C 0 E
1
2
3
S
6
5
3
7
0
0
1
10
12
1
1
0
0
2
7
9
2
3
0
1
2
6
5
0
1
0
3
1
to
2
4
|
Species
II
w
M
1-
0.45
1
0.31
0.23
1
0.32
0.10
0.22
1
i
0.14
0.11
0.18
0.26
1
m o n OB >
Collections
|1 2 3 4 S 6
F
1
1
1
1
i
0.36
1
0.21
0.28
1
0.11
0.18
0.14
1
/
0.09
0.16
0.13
0.17
1
0.08
0.1 S
0.10
0.21
0.18
1
•••••
1
2
3
4
5
6
1
1
r
i
I' r
3 1 I
1
I
B 1
i
: 1
1
_L
i i
j
5n
k*^.
o
5-
in
oj
; i
i
0)
' B
[ e
.
»
1 i
2
I -
I
!
|
i C
Collections
* W«»BI «•««••««••« . • WWW VIWW
Species
Q Analysis or normal
Classification J
Analysis or Inverse
Classification j
Interpretation of Classifications
Figure 50. Sequence of procedures in numerical classification (modified
from Boesch 1977). (The Illinois water bodies were classified
using LANDSAT MSS bands or LANDSAT estimated trophic indicators
as "species", i.e., attributes).
107
-------
"Matrices of inter-entity resemblance measures are usually re-
quired to perform normal or inverse analyses. These matrices
are symmetrical in that one corner is the mirror image of the
other across the 'self-match' diagonal and thus it is
necessary to display only half of the matrix, ... as the
excluded portion is repetitous.... From the resemblance
matrix one can go further and seek to group entities into
groups on the basis of their patterns of resemblance....This
is the essence of clustering.
Many clustering concepts and methods are described in the literature (Sneath
and Sokal 1973, Anderberg 1973). Although algorithms have been developed for
many clustering methods, the investigator is often forced to select from a
very limited number because he lacks access to computers with the necessary
operational software. In biological studies, the most commonly employed
strategies for finding clusters are those that can be described by the
acronym SAHN (sequential, agglomerative, hierarchic, nonoverlapping) (Sneath
and Sokal 1973). Several SAHN clustering methods are found in Table 13.
TABLE 13. SOME SAHN CLUSTERING METHODS*
Cluster Method Synonyms
Single linkage Nearest neighbor
Minimum method
Complete linkage Furthest neighbor
Maximum method
Average linkage
Arithmetic average
Unweighted (UPGMA) Group average
Weighted (WPGMA)
Centroid
Unweighted centroid Centroid
Weighted centroid Median
*Adapted from Sneath and Sokal 1973
The complete linkage (also called maximum or furthest neighbor) method of
clustering was selected for this project, primarily on the basis of
availability. The complete linkage method may be described as an exclusive,
intrinsic, hierachicals agglomerative, combinatorial approach to clustering
(Figure 51).
108
-------
I
NONEXCLUSIVE EXCLUSIVE
I
I
EXTRINSIC INTRINSIC
I
HIERARCHICAL NONHIERARCHICAL
I
DIVISIVE AGGLOMERATIVE SERIAL SIMULTANEOUS
| I OPTIMIZATION OPTIMIZATION
MONOTHETIC POLYTHETIC J
COMBINATORIAL NON-COMBINATORIAL
Figure 51. Dichotomized scheme depicting the salient properties of clustering methods (modified
from Boesch 1977 and Williams 1971). Complete linkage properties are interconnected
by the arrow-defined pathway.
-------
Although this report will not detail the above description of the
complete linkage method (see Sneath and Sokal 1973, Boesch 1977, Anderberg
1973), it bears mentioning that the process is one of progressive fusions.
Clustering proceeds "...by forming one cluster for each observation in the
analysis. The two closest clusters are combined into one cluster, then the
two closest of the new set of clusters are combined into a cluster, and so
forth.... The distance between two clusters is defined to be the maximum
distance between an observation in one cluster and an observation in the
other cluster" (Barr et al. 1976).
Squared Euclidian distance (A2) was selected as the resemblance measure
(dissimilarity coefficient) between lakes. As with the clustering method,
selection was made on the basis of availability. When working in an
attribute space (A-space) of p orthogonal dimensions, Euclidian distance is
the linear distance between any pair of entities (e.g., lakes) in that space.
The distance between two entities is computed as the square root of the sum
of the squared differences of the entity-paired attribute values
where Ajk = Euclidian distance between lake J and K
X-jj = the value of the ith attribute for lake J
Xjk = the value of the ith attribute for lake K
p = the number of attributes.
If two entities are identical in terms of their attributes, they will occupy
the same position in p-dimensional A-space and the Euclidian distance between
them will be zero. As the distance between entities increases, the disparity
between them increases (Sneath and Sokal 1973). Distance is the complement
of similarity.
The use of Euclidian distance (or its square) as a measure of resemblence
has much intuitive appeal. It is relatively easy to grasp through the use of
algebraic and geometric techniques (Figure 52). Recalling from geometry that
the Pythagorean theorem for right triangles states that the square of the
length of the hypotenuse is equal to the sum of the squares of the lengths of
the two remaining sides, or (using the notation in the upper part of Figure
52)
C2 = a2 + b2
it follows that the length of the hypotenuse (distance between Point A and
Point B) may be defined as
110
-------
(8
-------
Using different notation (lower portion of Figure 52), the Pythagorean
theorem is extended to two lakes, A and B, whose locations in two-dimensional
attribute space (A-space) are defined by the coordinate pairs Ui,a,
X2 a) and (Xi K» X2 b)» respectively. The squared Euclidian distance
between the tikes mly be described as
*ab2 ' Ixl,a-Xl,bl2+ Ix2,a-X2,bl2
where | | refers to "absolute magnitude" and is employed to eliminate the
use of negative distances. The squaring of differences also ultimately
eliminates the negative distance aspect and thus the equation may be
rewritten as:
*ab2 - (xl,a - xl,b>2 + a - X2jb)2
The Euclidian distance (Aab) between the two lakes is described by
and will either have or lack units depending upon which preclustering
procedures are used. If, for example, the attribute data in the data matrix
were standardized (mean of zero, unit variance), the Euclidian distance would
be dimensionless. On the other hand, using the LANDSAT MSS bands as the
attributes and foregoing transformations and standardizations (i.e., using
MSS band raw data), the Euclidian distance would be measured in digital
number (DN) levels.
Figure 53 illustrates the geometric aspects of three lakes (A, B, and C)
in three-dimensional A-space. In this case squared Euclidian distance
between Lake A and Lake B is computed as
and Aab = U^)
It is difficult to visualize the geometry of four-dimensional A-space and
it can not be depicted graphically because of the orthogonal axes
requirement. However, it is possible to extend most geometric theorems of
three-dimensional spa^e to p dimensions in Euclidian hyperspace; this can be
demonstrated algebraically (Sneath and Sokal 1973). Thus, the Euclidian
distance dissimilarity measure is not limited to three dimensions by
theoretical constraints.
112
-------
Lake Coordinates
/arX2,a, X3/a)
b X2/b, X3(C)
(X1rC/ X2/c, X3/c)
2 _
(X2,a -
(X3.a -
= (Aab2)V2
Figure 53. Geometric and computational aspects of Euclidian distance between
three entities (i.e., lakes) defined by coordinates in three-
dimensional attribute space (A-space). [In this example, the
attributes are three LANDSAT bands (GRN, RED, IR1). The
attributes could be three trophic indicators (e.g., chlorophyll
a^, Seech i depth, total phosphorus) (adapted from Boesch 1977 and
Ineath and Sokal 1973).]
113
-------
As indicated earlier, the complete linkage clustering method was employed
using squared Euclidian distance as the measure of resemblance. The U-k2)
sequence of procedures is illustrated in Figure 54. While many
clustering "runs" were made, only the results for two, made on 145 Illinois
water bodies in A-space, will be interpreted in the discussion in Section 7
of this report. These runs are:
1. Four-band LANDSAT clustering -- A-space of four dimensions defined
by LANDSAT GRN, RED, IR1, and IR2 bands. Date-adjusted MSS data
were used without standardization or transformation.
2. LANDSAT-estimated four trophic indicator clustering — A-space of
four dimensions defined by the trophic indicators CHLA, TPHOS, TON,
and SEC. Trophic indicator values were estimated from regression
models developed from LANDSAT MSS-trophic indicator relationships
elucidated for 22 NES-sampled lakes. The values for each trophic
indicator were standardized (mean of zero, unit variance) prior
to the computation of the resemblance matrix.
The actual clustering was accomplished using two different computer
programs. The first program, CLUSTER, was developed by Barr et al.
(1976) as part of the SAS 76 statistics program package. It uses the
clustering scheme described by Johnson (1967). The cluster "map" generated
by the program was found to be physically bulky, difficult to visually
interpret, and thus unsuitable for inclusion in this report. Fortunately,
this program also generates a listing of the clusters and their respective
members. A second program (also called CLUSTER) rewritten by Keniston from
programs by Keniston, Faruqui, and Carkin (Keniston 1978) was used to
generate products (e.g., dendrograms or phenograms) of a more desirable
nature. This program was run on a Control Data Corporation (CDC) CYBER 73
digital computer at Oregon State University. The dendrograms were then
generated on a Gerber Model 1000 flatbed plotter, an "off-line" device, using
the tape output from the CYBER 73.
The outputs of a clustering program represent an attempt to simplify
.complex data sets. Numerical classification per se does not provide an
ecological interpretation of the products. Post-clustering analyses aimed at
the interpretation of the dendrograms are presented in Section 7 of this
report.
114
-------
LAKES
I 6*i ,
I iiJS
(Q "^ (Q W ~Q
(0 ^f tf% ^L £
°t|g!
M""*
L
8 2*< •§
S 2 s SiS
s g * =5 c •;
« £ 3 §i
« 8 S «3 1-
J S o -gSl
o -28 • 8-i
v E C 3 ? i
•5 o « o- 5 c
1
S
n
3
k
5
L
1
C
t
5
Mir
L
A2
0
ror
AKE
• • •
A2
A2
O
f
S
A2
A2
A2
0
N
L
t
t
t
L
i2 1
,2 •<
k« . 5
i2 •'
I
o
1
Q.
|
I
a.
a
S "
few
S "o
•n c
c <
«
Q
CLUSTERING )
ss-_______/
/ \_
CLUSTER
STATISTICS
(0
111
I
V
I—I—I—I—I
0.0 A max
Euclidian Distance
1
INTERPRETATION
Figure 54. Sequence of procedures as applied to the numerical
classification of Illinois water bodies.
115
-------
SECTION 7
RESULTS AND DISCUSSION
As Indicated in Section 1, the basic objective of this project was to
classify, in the broad sense of the term, Illinois lakes and reservoirs as to
trophic state. It was the original intention to include all Illinois lentic
bodies with surface areas equal to or greater than 40 ha. This would have
included the 31 NES-sampled lakes. A combination of circumstances (e.g.,
cloud cover, missing or defective MSS data) resulted in dropping
approximately 40 of the proposed 179 lakes. The decrease in the number was
partially offset by the inclusion of several water bodies with surface areas
under 40 ha. The end resglt was the use of 22 NES-sampled water bodies in
the development of the regression models. Rankings and classes were
generated for 145 Illinois water bodies, including the 22 sampled by NES.
Subsequent subsections of this section will discuss the results.
WATER BODY SURFACE AREA
Surface area estimates were made for 153 Illinois water bodies (Appendix
Table A-l) including those found in scene 1449-16084, the scene without
internal calibration data. As evidenced by Appendix Table A-l, side overlap
lakes have two LANDSAT-derived area estimates, one for each consecutive day
of satellite coverage. Two questions immediately come to mind: How precise
are the consecutive day area estimates? How accurate are the estimated area
values? Table 14 was prepared in response to these questions. The lakes
included in this table are those for which side overlap coverage was
available, excluding those from scene 1449-16084. Lake Senachwine was not
included in the statistical calculations because its area estimate for
October 16, 1973, is less than optimal, a consequence of a portion of the
lake falling outside scene 1450-16140.
An indication of the precision of LANDSAT MSS-derived area estimates can
be gained by examining the self-pairing estimates made for 21 water bodies.
Ideally, the consecutive-day estimates for a particular water body would be
the same, and indeed this is the case for three water bodies: Commonwealth
Edison-Dresden Nuclear Lake (serial number 137), Lake Pinckneyville (serial
number 99), and Snyders Hunting Club Lake (serial number 142). The
difference between the self-paired estimates was computed as (A-B) where A is
the area estimate of'October 14 or 16 and B is the estimate of October 15.
The mean difference for the water bodies is 0 ha with a sample standard
deviation of 13 ha. The correspondence between the estimates A and B is very
good, as demonstrated by Figure 55 and the area ratios (A/B) in Table 14.
116
-------
TABLE 14. COMPARISON OF SURFACE AREA ESTIMATES FOR 22 ILLINOIS WATER BODIES
HAVING LANDSAT SIDE OVERLAP COVERAGE (OCTOBER 14 AND 15 OR
OCTOBER 15 AND 16, 1973)
Area Estimates (ha)
LANDSAT Date File*
Name
Carbondale
Cedar**
Commonwealth
DeQuoln
Goose t
Goose (Village)
Kinkaid
McGinnis
Murphy sboro
Pierce
Pinckneyville
Saganashkee
Senachwine f
Skokie Lagoons
Snyders Hunting
South Wilmington
Spring
Spring Arbor
Summerset
Tampier
Thunderbi rd
Sum
Mean
Standard Error
of Mean
Maximum Value
Minimum Value
Range
Sample Standard
Deviation
Serial
Number
38
39
137
98
101
139
40
20
37
130
99
21
1 t\A
11)4
22
142
140
4
143
186
23
173
14th or
(A)
29
72
460
55
741
74
855
114
47
60
49
146
1 Of\Q
ItU?
80
17
99
137
15
94
90
38
3510
167
51
855
15
840
16th 15th
(B)
27
77
460
52
696
75
876
102
54
58
49
167
i cm
loUo
84
17
101
135
17
92
94
42
3505
167
50
876
17
859
(C)
55
728
526
99
1143
109
1093
127
58
66
67
132
1 1Ad
U*tO
76
81
41
106
41
115
66
45
4258
213
73
1143
41
1102
Area
(A-B)
+ 2
- 5
0
+ 3
+45
- 1
-21
+12
- 7
+ 2
0
-21
- 4
0
- 2
+ 2
- 2
2
- 4
- 4
0
3
45
-21
66
13
Differences
(A-C)
- 26
- 66
- 44
-402
- 35
-238
- 13
- 11
- 6
- 18
+ 14
+ 4
- 64
+ 58
+ 31
- 26
- 21
+ 24
- 7
- 40
23
58
-402
460
104
(B-C)
- 28
- 66
- 47
-447
- 34
-217
- 25
- 4
- 8
- 18
+ 35
+ 8
- 64
+ 60
+ 29
- 24
- 23
+ 28
- 3
- 42
25
60
-447
507
111
Area Ratios
(A/B)
1.07
0.94
1.00
1.06
1.07
0.99
0.98
1.12
0.87
1.03
1.00
0.87
0.95
1.00
0.98
1.02
0.88
1.02
0.96
0.91
0.99
0.02
1.12
0.87
0.25
0.07
(A/C)
0.53
0.88
0.56
0.65
0.68
0.78
0.90
0.81
0.91
0.73
1.11
1.05
0.21
2.42
1.29
0.37
0.82
1.36
0.84
0.90
0.10
2.41
0.21
2.21
0.46
(B/C)
0.49
0.88
0.52
0.61
0.69
0.80
0.80
0.93
0.88
0.73
1.26
1.11
0.21
2.46
1.27
0.42
0.80
1.42
0.93
0.92
0.10
2.46
0.21
2.25
0.47
* Surface area values taken from I EPA data files.
** Cedar Lake was just filling in 1973. The file value is for 1977. The
lake is not included in the calculations.
t Goose Lake (101) area difference (A-B) exceeds the sample standard
deviation by more than three times.
t Senachwine (104) surface area data not included in calculations because
a portion falls outside scene 1450-16140.
117
-------
900
800
U
300
200
100
• 40
Line of 1:1 Correspondence
Serial Number and Name
3 Depue
4 Spring
20 McGinnis
21 Saganashkee
22 Skokie Lagoons
23 Tampier
37 Murphysboro
38 Carbondale
39 Cedar
40 Kinkaid
98 Duquoin
99 Pinckneyville
101 Goose
130 Pierce
137 Commonwealth
139 Goose (Village Club)
140 South Wilmington
142 Snyder Hunting
143 Spring Arbor
173 Thunderbird
186 Summerset
100
200
300
400
500
600
700
800
900
Surface Area Estimate of October 15 (ha)
Figure 55. LANDSAT-derived consecutive-day surface area estimates for 21 Illinois water bodies,
-------
The line in Figure 55 is defined by the ideal area ratio of 1:1; it is not a
regression line generated from area data for the 21 water bodies. If the two
area estimates for a particular water body are the same, the point
representing the values would fall somewhere on the line. With a few
exceptions, the points fall on or very close to the line. The area ratios
(A/B) for several lakes are a source of some concern; these include McGinnis
(1.12), Murphysboro (0.87), and Saganashkee (0.87). Assuming that the areal
discrepancies are not a consequence of heavy rains or drawdown, several
factors may be responsible for the variation in the estimates.
One such factor, probably the prime one, is atmospheric variability.
Atmospheric path radiance can change markedly from one day to the next. For
example, the intrinsic (independent of atmosphere) radiance may be the same
for a lake for two consecutive LANDSAT flyovers. However, as viewed through
the atmosphere by the satellite, the lake may exhibit an apparent change in
its intrinsic radiance, a consequence of a change in path radiance. As path
radiance increases, the apparent intrinsic radiance of most lakes will
increase and the converse is also true. Thus, the DN values for the pixels
representing a lake and surrounding terrain may change from day to day. As
given previously, an IR2 level of 28 was used as the threshold between land
and water features. If a pixel has an IR2 DN value of 0 to 28, inclusive, it
was classified as water; if 29 to 255, it was labeled as nonwater. The IR2
DN threshold of 28 was treated as a rigid boundary. Increases in path
radiance could result in an apparent emigration of pixels across the
boundary, effectively reducing the number of pixels in the IR2 0 to 28 DN
group defined as water. Decreases in path radiance could result in the
immigration of more pixels to the IR2 0 to 28 DN group, leading to a larger
surface area estimate.
In situations where water bodies are in juxtaposition and have poorly
defined natural boundaries (e.g., flowages and backwaters), the judgment of
the computer operator becomes the critical factor in separating the water
bodies. In this case, the operator defines the location of the common
boundary. He is not always able to duplicate the boundary's location in
successive imagery. The situation is further complicated by maps that give
conflicting locations for the boundary or, and this is more common, only name
the bodies, leaving the map reader the task of defining boundaries. The
common boundary problem can lead to different surface area estimates.
Illinois reservoirs tend to be long and narrow, with highly developed
shoreline configurations (numerous fingers and bays). With so much
land-water interface (Lake Sangchris, for example, has more than 160 km of
shoreline), a substantial number of pixels will contain both land and water
features. In many cases the land portion of the pixel may contribute enough
energy to shift the pixel DN value above the water threshold. This could
lead to an underestimation of lake surface area.
The accuracy of LANDSAT-derived area estimates is another point of prime
concern. This concern can be best addressed by examining the LANDSAT area
values in light of estimates derived from other time-proven techniques (e.g.,
aerial photography, surveying) employed concurrently with the LANDSAT
flyover. The concurrent aspect of the contact-sensed data is important
119
-------
because lentic bodies are dynamic and can experience substantial changes in
surface area. Prolonged dry spells, drawdown, large quantities of
precipitation, outlet downcutting, flow control structures, and landfill
projects can produce changes in surface area. Even in the case of concurrent
acquisition of surface area data, the problem of defining the water-land
interface exists. It would appear that defining the boundary of a lentic
body on an aerial photograph or through survey is an easy task, and in some
cases it is. However, more often the task of locating the interface is
complicated by the presence of marshes or swamps that grade into a lake over
distance. Where to "draw the line" becomes a problem that is left to the
judgment of the photo interpreter, cartographer, or surveyor. Once this
determination has been made and the requisite computations performed, an area
estimate is entered into a data file. It is important to recognize the
limitation of surface area figures taken from files.
Given the historical nature of this project — NES sampled the water
bodies prior to the formulation of the project protocol -- surface-area data
were not readily available. This severely handicapped the project in making
a detailed analysis of the accuracy of LANDSAT-derived surface estimates.
However, realizing its limitations, a comparison was made using surface area
figures taken from IEPA files (Table 14). For reservoirs, in most cases, the
file values are estimates of the surface area at spillway level at the time
of construction. However, in many cases the surface area at the spillway
level has probably changed over time because of shoreline erosion,
sedimentation, and other factors. Pool level elevations are not available
for October 1973. An examination of Table 14 and accompanying Figures
56 and 57 suggests that the LANDSAT-derived values tend, on the average, to
underestimate the IEPA values. A lot of scatter is apparent about the 1:1
line of correspondence. Hence, data of sufficient quality are not available
to determine how well LANDSAT data, as processed, estimate the surface area
of Illinois lakes.
It may well be that the LANDSAT area values are underestimates. Boland
and Blackwell (1978), reporting on a study involving several Colorado lakes,
indicate that LANDSAT area estimates were, on the average, about 1.2 percent
under those estimated from concurrently acquired aerial photographs (N = 8).
The Colorado data do not, however, prove that LANDSAT underestimates the
surface area of Illinois water bodies.
The selection of the IR2 DN level of 28 as the water-land threshold value
was made based on past experience and its intuitive appeal; it is a simple
approach. As suggested by Blackwell and Boland (1978), it may be more
appropriate to use the technique (or some variation) proposed by McCloy
(1977). This technique, discussed in Section 6, may result in more precise
and accurate estimates of surface area.
TROPHIC INDICATOR AND INDEX ESTIMATION
In order to obtain estimates of four trophic indicators (CHLA, SEC, TON,
and TPHOS) and two multivariate trophic indices (PC1F5, PC1Y5) using LANDSAT
MSS spectral data, three sets of multiple regression models were developed
(Table 15). Initially, regressions were developed for the indicators and
120
-------
900 r
800
I
<0 700
r- 600
w
flJ
.O
o
W 500
"5
5
CD 400
rv>
w
UJ
8
•g
3
0)
300
200
100
40
101
•
137
Line of 1:1 Correspondence
Serial Number and Name
3 Depue
4 Spring
20 McGinnis
21 Saganashkee
22 Skokie Lagoons
23 Tampier
37 Murphysboro
38 Carbondale
40 Kinkaid
98 Duquoin
99 Pinckneyville
101 Goose
130 Pierce
137 Commonwealth
139 Goose (Village Club)
140 South Wilmington
142 Snyder Hunting
143 Spring Arbor
173 Thunderbird
186 Summerset
100 200 300 400 500 600 700 800 900
Surface Area Estimate From IEPA Files (ha)
1000 1100 1200
Figure 56. Comparison of October 14 and 16 LANDSAT-derived surface area estimates with IEPA file
values for 20 Illinois water bodies.
-------
ro
ro
900
800
2 700
600
500
400
300
200
100
140 ~ A •'••
130 " *1"
v»i»A,r
X**3"*1."
A
101
Line of 1:1 Correspondence
Number and Name
Depue
Spring
McGinnis
Saganashkee
Skokie Lagoons
Tampier
Murphysboro
Carbondale
Kinkaid
Duquoin
Pinckneyville
Goose
Pierce
Commonwealth
Goose (Village Club)
South Wilmington
Snyder Hunting
Spring Arbor
Thunderbird
Summerset
100 200 300 400 500 600 700 800 900
Surface Area Estimate From IEPA Files (ha)
1000
1100
1200
Figure 57. Comparison of October 15, 1973, LANDSAT-derived surface area estimates with IEPA file
values for 20 Illinois water bodies.
-------
TABLE 15. THREE SETS OF REGRESSION MODELS FOR THE ESTIMATION OF TROPHIC INDICATORS AND MULTIVARIATE
TROPHIC STATE INDICES
Model
Degrees
of
Freedom
Calculated
F-Value
100
Standard
Error of
Estimate
Set One: Models based on LANDSAT MSS spectral bands and band ratios
9.55 74.90
GJ
SEC = -26.206 - 0.610 * GRN + 0.939 * RED + 17.246 * 5,16
GRNRED + 1.276 * REDIR1 - 0.827 * IR1IR2
CHLA = EXP (37.337 + 0.476 * GRN - 0.922 * RED 5,15
-15.327 * GRNRED -8.079 * REDIR1 + 1.922 *
REDIR2)
TON = [(X * X-l.) **2./ (4. * X * X)] where X is: 4,17
-2.023 - 0.223 * GRN + 0.325 * IR1 + 3.117 *
GRNRED + 0.826 * REDIR2
TPHOS = |EXP[EXP (6.202 - 0.076 * IR1 -2.293 * GRNIR1 + 3,18
0.244 * GRNIR2)](/100.
PC1F5 = (12.279 - 0.177 * IR1 - 4.631 * GRNIR1 + 0.572 * 3,18
GRNIR2) **2.-3.
PC1Y5 = 12.782 - 0.193 * IR1 - 5.017 * GRNIR1 + 0.701 * 3,18
GRNIR2) **2. -3.
22.30
25.16
18.74
30.46
39.32
88.14
85.55
75.75
83.54
86.76
0.31
0.31
0.71
(continued)
-------
TABLE 15. (continued)
Model
Degrees
of Calculated
Freedom F-Value
R2X
100
Standard
Error of
Estimate
Set Two: Models based on normalized LANDSAT spectral bands
SEC =
CHLA =
TON =
TPHOS =
PC1F5 =
PC1Y5 =
EXP[EXP (1.396 - 0.234 * SIR1 - 1.198 * LNSRED
+ 0.227 * SGRN + 0.200 * SIR2)] /100.
[-34.677 - 12.735 * SIR1 - 16.209 * LNSIR2 -
5.304 * RED + 94.014 * LNSIR1]**2.
|EXP[EXP(8.796 + 0.087 * SIR1 - 4.522 * RTSGRN
+ 0.647 * SGRN - 0.042 * SIR2)]|/100.
[1.173 + 1.407 * LNSIR1 - 1.682 * LNSGRN +
5.371 * RTSIR2 - 5.176 * LNSIR2]**5.
79.188 + 19.827 * LNSIR1 - 48.892 * SQRTSGRN +
5.399 * SGRN - 7.459 * LNSIR2
-10.922 - 5.041 * SIR1 - 9.384 * LNSIR2 - 2.276
4,17
4,17
4,17
4,17
4,17
4,17
72.6
80.2
70.4
73.7
81.2
87.6
0.33
28.91
0.34
0.16
0.80
0.61
* SGRN + 44.592 * LNSIR1
(continued)
-------
TABLE 15. (continued)
IX)
en
Degrees Standard
Model of Calculated &K Error of
Freedom F-Value 100 Estimate
SEC =
+ 0.006
CHLA =
TON =
TPHOS =
PC1F5 =
PC1Y5 =
Set Three: Models based on lake LANDSAT spectral
JEXP[EXP(1.552 - 0.006 * REDRK + 0.003 * IR2RK 4,17
* GRNRK - 0.005 * IR1RK)]|/100.
[2.147 - 0.011 * GRNRK - 0.010 * IR2RK + 0.008 4,17
* RAT100 + 0.017 **IR1RK] **5.
)EXP[EXP(1.263 + 0.002 * IR1RK - 0.001 * GRNRK 3,18
- 0.001 * IR2RK) **2.]f/100.
(0.697 + 0.005 * IR1RK - 0.003 * GRNRK - 0.001 3,18
* IR2RK) **5.
0.264 + 0.076 * IR1RK -0.043 * GRNRK - 0.030 * 3,18
IR2RK
0.304 + 0.076 * IR1RK - 0.040 * GRNRK- 0.038 3,18
ranks
59.0
74.0
62.5 0.39
61.1 0.20
65.3 1.04
64.6 1.03
* IR2RK
-------
multivariate indices using contact-sensed and raw spectral data available
from 22 NES-sampled lakes. In developing these models (Set One, Table 15)
using stepwise regression, additional variates were created by taking the
ratios of bands (i.e., green band to red band, GRNRED). The use of such
ratios has precedent in similar studies (e.g., Boland 1976). While the
models gave every appearance of having practical utility (e.g., high RS
low standard error of estimate, good estimates of indicator and index
values), when they were applied to the full set of 145 lakes major
inconsistencies appeared. The inconsistencies were particularly obvious when
the estimated values of the four trophic indicators for the 145 lakes were
aggregated into a combined or composite rank for each lake and a list was
generated displaying the lakes in ascending order of eutrophication.
To create this list, the estimated value for each trophic indicator
(CHLA, TON, SEC, TPHOS) was replaced by its rank value. The rank values for
the four indicators were summed to give a grand value for each of the 145
lakes, which were then ranked in ascending order according to the magnitude
of the grand value. An examination of this list disclosed several seriously
misclassified lakes. For example, Lake Calumet, a highly polluted water
body, appeared near the top (best) of the list. These reversals were
suspected to be the result of at least four causes.
First, although efforts were made to normalize the dependent (trophic
indicator or multivariate index) parameter in developing the regression
models, no effort was made to do this for the spectral data. This was not
deemed important since, with the exception of the IR2, the bands showed an
approximately normal distribution of data, with skew and kurtosis of about
0.0 and 3.0, respectively.
Second, although the IR2 band has the poorest discrimination, and the
lowest information content of any of the bands, it weighed consistently and
heavily in all of the models. When this band was excluded, the resulting
models were statistically unsatisfactory. Thus, we are faced with the
paradox of a poorly resolved factor contributing significantly to the
predictive power of the regressions. Since, in at least some of the cases
(e.g., Calumet), the IR2 intensity did not appear to vary to the extent or
direction of the other bands, that is, did not increase as water quality
deteriorated, the highly significant IR2 values would contribute nonlinearly
to the estimates.
Third, it was recognized that the ratios used in developing the models
contributed nonlinear components to the regressions. This effect, coupled
with the nonlinearity contributed by the IR2, resulted in "linear" models
with nonlinear components. Alternatively, the effect may be viewed as that
of a p-dimensional plane operating in a nonorthogonal or bivalued
p-dimensional hyperspace; that is, the model is a linear hyperplane cutting
both curved and linear spaces.
Fourth, the spectral data for the subset of 22 water bodies used to
develop the models cover a narrow range in relation to that of the full set
of 145. Thus, while the models appear to operate well for lakes within this
narrow band, they have very poor discrimination outside these limits.
126
-------
Whether the failure to afford reasonable extrapolations is inherent in
the data, as discussed above, or in the models, as suggested here, is moot
since no simple corrections for these effects are apparent. Consequently,
the project is constrained by these models to estimate trophic indicator and
multivariate trophic index values only for lakes with LANDSAT MSS data that
fall within the range of those for the 22-lake subset or to develop other
models as will now be discussed.
In developing new models, it was imperative that the information about
the full range of the whole set be included in the subset used for developing
the regressions. To accomplish this, the spectral data for the subset were
standardized by dividing each observation by the standard deviation for that
band for the entire set of 145. The models (Set Two, Table 15) that were
developed using standardized spectral data gave better estimates of trophic
indicator and index values for the subset of 22 than did those constructed
from the raw spectral data. Further, the estimates for the full set showed
few reversals or non-linearities.
At least two problems were apparent with some of the individual trophic
indicator and index estimates made from Set Two models. First, these models
generated negative values for lakes with LANDSAT spectral data either much
higher or much lower than the regression set, for Secchi and chlorophyll a,
and total organic nitrogen and total phosphorus respectively. Unreasonably
high estimates of Secchi depth were obtained from very low (or for the other
indicators, for very high) spectral data values.
Trophic indicator estimates made for the 145 water bodies from the second
set of models, when converted to ranks, gave ordered lists which were in
reasonable agreement with expectation. Clusters developed from the estimated
trophic indicator values appeared to have physical significance in that the
clusters could be sequenced in a rough hierachy of trophic condition.
The third set of models was developed using data ranks in place of raw or
normalized spectral data. To create the regression models (Set Three, Table
15), the normalized spectral data for the 145 lakes were sequenced in
ascending order, and ranks were then assigned to each value for each of the
four bands. Next the spectral ranks for the 22 NES lakes were extracted from
the set of 145 and used to develop the third set of models defining
statistical relationships between the remotely sensed data and ground truth.
For Secchi depth, total organic nitrogen, and total phosphorus, highly
significant correlations were developed using combinations of the four sets
of ranked spectral data. Chlorophyll a^ presented a problem, however, in that
this technique afforded statistically significant, but 'physically
unimportant, models with poor predictive power. Careful examination of the
results from the other models, clustering results from spectral data, and
consideration of the interactions occurring between chlorophyll a_ and the
high sediment loads present in many Illinois lakes suggested that some
spectral ratio of the LANDSAT IR1 or IR2 and the red and green bands might
make a statistically significant contribution to the model for chlorophyll a..
For these reasons two ratios were developed, one between the RED and IR1
bands and the other, subsequently named RAT100, between the GRN and IR2
bands. The former was statistically insignificant when included as a
127
-------
regression variable, while the latter was highly significant. The use of the
variable RAT100, defined as (GRNRK + IR2RK) / (GRNRK - IR2RK -1.), where
GRNRK and IR2RK are a lake's rank for the set of 145, caused mathematical
problems for Diamond Lake (serial number 57) and potential problems for Grass
Lake (serial number 61), Swan Lake (serial number 93), and Pierce Lake
(serial number 130). For the first lake the denominator of the ratio was
zero, while for the other three it was either plus or minus 1, indicating
that GRNRK = IR2RK. For these four lakes RAT100 was assigned a value of
zero.
Trophic indicator and multivariate trophic index estimates generated from
the Set Three models are in general agreement with the contact-sensed data
for the 22 NES lakes (Table 16). The descriptive statistics for the actual
and predicted values are in (subjectively) close agreement, thereby
indicating that the models are capable of reproducing population, as well as
point, information. The 95 percent confidence limits of the predicted values
for the 22 NES lakes used in developing the Set Three regression models are
within the range of the sample values (Table 17). These confidence limits
are those obtained for each estimated point. They are not the same, being
broader, as the standard error about the mean of the dependent variable.
The trophic indicator and index estimates for the full set of 145
Illinois lakes developed from the Set Three models are found in Appendix
Table A-4. The population statistics for the full set are (subjectively) in
agreement with those of the subset of NES lakes (Table 18). This is
important for two reasons. First, it demonstrates that the NES lakes
comprise a random subset; that is, they adequately reflect the variability of
the population. Second, unlike the previous models (Sets One and Two), the
predictive equations are essentially linear over the data range. Therefore,
meaningful extrapolations can be made from the Set Three models.
While it is possible to gauge the accuracy of the trophic indicator and
multivariate index estimates for the 22 NES lakes through an examination of
the models' residuals, the necessary concurrent contact-sensed data are not
available for the remaining 123 lakes, thereby preventing a similar
comparison. However, the estimates for the full set appear to be
"reasonable" for Illinois lakes.
IEPA sampled 72 of the 145 lakes during the summer of 1977. It is not
appropriate to gauge the accuracy of the 1973 parameter estimates in terms of
the 1977 field measurements. The 1973 data were collected in the middle of
October. By contrast, the 1977 data were collected during the summer when
the Secchi depth would be expected to be lower because of plankton biomass
and recreational uses that suspend and resuspend particulate matter and
intense agricultural activities that contribute substantially to the
particulate load. Recognizing the limitations imposed by the dynamic nature
of inland water bodies, a comparison was made for the 72 lakes using Secchi
depth, a parameter common to both data sets. The mean Secchi depth values (N
= 72) for 1973 and 1977 are 0.98 and 0.97 m, respectively. The
product-moment correlation coefficient, significant at the 0.01 level, is
0.505. However, because the data were collected or estimated for different
128
-------
TABLE 16. TROPHIC INDICATOR AND MULTIVARIATE TROPHIC INDEX OBSERVED, ESTIMATED, AND RESIDUAL VALUES
FOR THE SET THREE REGRESSION MODELS
ro
Name
Carlyle
Cedar
Charleston
Crab Orchard
Decatur
DePue
East Loon
Fox
Grass
Holiday
Long
Marie
Old Ben Mine
Pi stakee
Raccoon
Rend
Shelbyville
Si ocum
Storey
Vandal ia
We-Ma-Tuk
Wonder
Serial
Number
14
55
16
127
73
3
53
60
61
51
52
62
32
65
80
29
114
54
50
27
33
69
SEC
0.48
2.77
0.25
0.36
0.46
0.15
0.91
0.36
0.31
0.46
0.31
0.56
0.48
0.31
0.41
0.61
0.48
0.31
0.89
0.71
1.07
0.46
SEC
0.42
1.19
0.27
0.62
0.29
0.25
0.85
0.35
0.31
0.43
0.44
0.36
0.57
0.54
0.31
0.58
0.67
0.34
0.93
0.89
0.62
0.31
SEC-SEC
0.06
1.58
-0.02
-0.26
0.17
-0.10
0.06
0.01
0.00
0.03
-0.13
0.20
-0.09
-0.23
0.10
0.03
-0.19
-0.03
-0.04
-0.18
0.45
0.15
CHLA
19.9
5.6
18.0
46.7
21.4
42.4
26.8
37.4
46.1
67.0
61.2
70.7
24.6
66.5
10.6
15.6
12.8
241.4
30.0
13.5
8.3
198.0
CHLA
10.3
26.7
36.0
26.5
48.6
45.4
25.8
46.5
46.8
37.6
43.6
59.9
18.4
33.5
19.6
18.0
15.4
229.4
30.4
19.4
5.0
157.5
CHLA-CHLA
9.6
-21.1
-18.0
20.2
-27.2
- 3.0
1.0
- 9.1
- 0.7
29.4
17.6
10.8
6.2
33.0
- 9.0
- 2.4
- 2.6
12.0
- 0.4
- 5.9
3.3
40.5
TON
0.742
1.105
1.200
0.843
0.590
2.020
1.380
0.970
1.773
1.200
2.074
1.896
1.690
1.635
0.900
0.971
0.595
5.940
0.980
1.019
0.588
1.788
TON
0.598
1.160
1.308
1.010
1.410
1.492
1.225
1.252
1.391
1.477
1.326
1.612
1.307
1.142
0.856
0.854
0.828
5.118
0.898
0.853
0.820
2.068
TON-fON
0.144
-0.055
-0.108
-0.167
-0.820
0.528
0.155
-0.282
0.382
-0.277
0.754
0.284
0.383
0.493
0.044
0.117
-0.233
0.822
0.082
0.166
-0.232
-0.280
(continued)
-------
TABLE 16. (continued)
co
o
Name
Carlyle
Cedar
Charleston
Crab Orchard
Decatur
DePue
East Loon
Fox
Grass
Hoi id ay
Long
Marie
Old Ben Mine
Pistakee
Raccoon
Rend
Shelbyville
Sloe urn
Storey
Vandal i a
We-Ma-Tuk
Wonder
Serial
Number
14
55
16
127
73
3
53
60
61
51
52
62
32
65
80
29
114
54
50
27
33
69
TPHOS TPHOS TPHOS-fPHOS PC1Y5
0.091
0.053
0.207
0.114
0.143
0.499
0.087
0.212
0.347
0.173
0.785
0.286
0.930
0.216
0.145
0.065
0.063
1.330
0.125
0.175
0.103
0.423
0.056
0.135
0.349
0.138
0.272
0.413
0.176
0.200
0.226
0.238
0.187
0.326
0.328
0.156
0.155
0.083
0.083
1.425
0.174
0.122
0.171
0.468
0.035
-0.082
-0.142
-0.024
-0.129
0.086
-0.089
0.012
0.121
-0.065
0.598
-0.040
0.602
0.060
-0.010
-0.018
-0.020
-0.095
-0.049
0.053
-0.068
-0.045
-0.90
-2.89
0.02
0.14
-0.17
2.41
-0.77
1.68
1.84
1.22
1.98
0.33
1.55
1.63
-0.48
-1.13
-1.62
4.31
-1.07
-1.32
-1.35
2.29
PC1Y5 PC1Y5-PC1YE
-1.22
-0.36
1.03
-0.14
1.10
1.36
-0.17
0.88
0.75
0.63
0.46
1.47
0.11
0.12
-0.03
-0.96
-1.20
4.92
-0.69
-1.02
-1.15
1.96
-0.32
2.53
1.01
-0.28
1.27
-1.05
0.60
-0.80
-1.09
-0.59
-1.52
1.14
-1.44
-1.51
0.45
0.17
0.42
0.61
0.38
0.30
0.20
-0.33
> PC1F5
-0.78
-2.67
0.78
-0.22
-0.51
2.61
-0.55
0.67
1.71
0.86
2.59
1.43
2.38
1.52
-1.01
-1.19
-1.35
4.56
-0.47
-1.13
-1.33
2.33
£C1F$ PC1F5-PC1F5
-1.58
-0.26
1.27
-0.19
1.11
1.64
0.06
0.71
0.77
0.78
0.46
1.51
0.74
0.08
-0.11
-1.03
-1.18
5.15
-0.34
-0.80
-0.63
2.18
-0.80
2.41
0.49
0.03
1.62
-0.97
0.61
0.04
-0.94
-0.08
-2.13
0.08
-1.64
-1.44
0.90
0.16
0.17
0.59
0.13
0.33
0.70
-0.15
-------
TABLE 17. COMPARISON OF THE RANGE OF NES SAMPLE VALUES WITH THE 95 PERCENT CONFIDENCE LIMITS OF
PREDICTED VALUES
SEC
OBS* PRED**
CHLA
OBS* PRED**
TPHOS
OBS* PRED**
TON
OBS*
PRED**
Carlyle
Cedar
Charleston
Crab
Orchard
Decatur
DePue
E. Loon
Fox
Grass
Hoi id ay
Long
Marie
Old Ben
Mine
Pistakee
Racoon
Rend
Shelby-
ville
Sloe urn
Storey
Vandal i a
We-Ma-Tuk
Wonder
0.4-0.6
2.8
0.3
0.3-04
0.5
0.2
0.9
0.3-0.4
0.3
0.3-0.6
0.3
0.3-0.8
0.5
0.3
0.4-0.5
0.5-0.8
0.3-0.9
0.3
0.8-0.9
0.6-0.9
0.9-1.2
0.3-0.6
0.2-1.3
0.4-4.8
0.1-0.7
0.3-1.9
0.1-0.7
0.1-0.6
0.3-3.0
0.2-0.9
0.1-0.8
0.2-1.2
0.2-1.2
0.2-0.9
0.2-1.9
0.2-1.6
0.1-0.8
0.3-1.7
0.3-2.0
0.1-1.2
0.3-3.5
0.3-3.0
0.2-2.0
0.1-0.8
9.8-38.8
5.6
11.2-24.9
29.4-77.6
6.3-33.9
42.4
26.8
24.3-80.6
37.7-54.6
66.5-67.5
60.7-61.8
59.6-81.9
24.6
59.7-73.3
3.1-18.1
10.3-21.7
5.3-26.3
241.4
16.2-43.8
10.1-18.0
5.2-11.4
188.0-211.0
1.1-47.7
5.6-88.1
8.6-109.3
5.9-81.8
13.5-135.7
11.4-129.0
5.3-83.5
12.2-133.7
12.9-132.2
9.1-113.2
11.4-125.8
17.4-161.0
2.9-70.5
8.2-100.4
3.4-69.7
3.3-62.2
2.7-55.5
73.7-571.0
5.5-108.2
3.6-67.5
0.4-28.2
47.0-416.9
0.076-0.
0.029-0.
0.172-0.
0.071-0.
0.111-0.
0.499
0.085-0.
0.283-0.
0.152-0.
0.744-0.
0.262-0.
0.920-0.
0.201-0.
0.073-0.
0.047-0.
0.023-0.
1.330
0.053-0.
0.044-1.
0.066-0.
0.363-0.
119
093
229
257
173
089
438
212
828
318
940
232
299
082
128
321
020
177
477
0.004-0.
0.020-0.
0.080-1.
0.023-0.
0.061-0.
0.098-1.
0.031-0.
0.037-0.
0.047-0.
0.048-0.
0.035-0.
0.076-1.
0.065-1.
0.027-0.
0.023-0.
0.010-0.
0.010-0.
0.394-3.
0.030-0.
0.019-0.
0.025-0.
0.119-1.
318
534
093
507
868
234
638
703
758
799
659
000
115
565
604
357
355
947
638
476
681
270
0.88-1.24
0.98-1.31
0.69-1.50
0.56-1.83
0.36-0.94
2.02
1.25-1.84
1.27-2.29
0.97-1.36
1.76-2.24
1.61-2.18
1.54-1.84
1.54-1.73
0.74-1.06
0.76-143
0.39-1.02
5.94
0.86-1.48
0.75-1.89
0.45-0.51
0.97-2.25
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
1.
0.
0.
0.
0.
32-1.30
58-2.72
65-3.11
53-2.16
71-3.31
72-3.58
61-2.85
63-2.93
70-3.26
72-3.58
66-3.11
79-3.93
62-3.31
59-2.57
45-1.89
46-1.81
45-1.74
70-21.27
47-2.23
46-1.81
43-1.86
95-5.38
*OBS = Observed sample range
**PRED = 95% limits for predicted value (point estimate)
-------
TABLE 18. SUMMARY STATISTICS FOR SET THREE (SPECTRAL RANK) REGRESSION
MODELS
Parameter
GRNRK
GRNRK
REDRK
REDRK
IR1RK
IR1RK
IR2RK
IR2RK
RAT100
RAT100
SEC
SEt
residuals
SEC
CHLA
CHLA
residuals
CHLA
TON
TON
residuals
fOff
TPHOS
TPHOS
resjduals
TPHOS
PC1F5
P~C1F5
residuals
PC1F5
PC1Y5
"PC1Y5"
residuals
PcTv?
N
22
145
22
145
22
145
22
145
22
145
22
22
22
145
22
22
22
145
22
22
22
145
22
22
22
145
22
22
22
145
22
22
22
145
Mean
64.68
73.00
69.82
73.00
55.45
73.00
40.86
73.00
5.75
2.71
0.60
0.52
-0.07
0.76
49.30
45.31
-3.98
30.60
1.44
1.36
-0.09
1.27
0.30
0.27
-0.03
0.30
0.46
0.47
0.00
0.48
0.35
0.35
0.00
0.13*
Median
64.00
73.00
63.50
73.00
52.00
73.00
31.00
73.00
2.29
1.42
0.46
0.44
-0.01
0.48
2"8.40
31.94
-0.50
24.67
1.15
1.23
-0.10
1.16
0.17
0.18
0.02
0.26
0.23
0.30
0.22
0.60
0.08
0.11
-0.17
0.20
Maximum
136.00
145.00
123.00
145.00
135.00
145.00
104.00
145.00
48.50
81.67
2.77
1.17
0.27
7.41
241.4
227.8
27.6
229.4
5.94
5.08
0.82
5.12
1.33
1.42
0.14
1.62
4.56
5.13
2.13
5.15
4.31
4.89
1.52
4.92
Minimum
8.00
1.00
11.00
1.00
6.00
1.00
4.00
1.00
-11.94
-71.75
0.15
0.25
-1.60
0.17
5.60
5.00
-40.48
0.91
0.59
0.59
-0.86
0.41
0.05
0.05
-0.60
0.02
-2.67
-1.60
-2.41
-3.61
-2.89
-1.24
-2.53
-3.70
Standard
Deviation
33.46
42.00
28.30
42.00
32.59
42.00
29.31
42.00
13.90
18.24
0.54
0.25
0.38
0.79
59.11
50.86
17.13
29.92
1.11
0.89
0.39
0.65
0.33
0.28
0.20
0.25
1.77
1.42
1.04
1.38
1.72
1.38
1.03
1.32
Skew
0.41
0.00
0.26
0.00
0.54
0.00
0.83
0.00
2.08
1.11
3.22
1.02
-3.16
4.67
2.35
2.69
-0.33
3.29
3.13
3.43
0.04
3.56
1.96
3.38
-2.27
2.90
0.41
1.48
-0.02
0.30
0.30
1.53
-0.37
0.22
Kurtosis
2.65
1.80
2.88
1.80
2.82
1.80
2.70
1.80
6.70
9.73
13.58
3.20
13.45
35.28
7.63
9.50
2.73
18.58
13.43
14.94
3.03
19.85
6.10
14.60
7.19
12.16
2.52
6.25
3.24
1.37
2.56
6.27
2.94
1.40
132
-------
seasons it would be inappropriate to suggest that, overall, Secchi
transparency has changed or, for that matter, remained the same.
Trophic indicator rankings for the 145 lakes are found in Table 19. A
composite rank value was determined for each lake by pooling its four trophic
indicator rank values and then once again ranking the 145 lakes (Table 20).
In general, the rank positions are in good agreement with the 1977
qualitative and quantitative information. As always, however, these
comparisons can only serve as a rough guide, since in making them it is
assumed that the lakes have been stable from 1973, rather than dynamic,
changing, and nonuniform, as is actually observed.
CLUSTER ANALYSIS
Cluster analysis ("numerical classification" in the words of Boesch 1977)
was used to order the 145 Illinois water bodies into groups. Separate
clusterings were accomplished using just the LANDSAT bands as attributes in
sets of three (6RN, RED, IR1) and four (GRN, RED, IR1, IR2). Additional
clusterings were made using, as attributes, the LANDSAT-estimated values for
four trophic indicators (CHLA, SEC, TON, and TPHOS).
Spectral Classification of the Lakes
In this subsection, the results of the cluster analysis of raw MSS data
are discussed. Natural lake groupings were established according to spectral
characteristics from the four attributes (i.e., GRN, RED, IR1, and IR2 bands)
for 145 Illinois water bodies. The objective of the classification process
was to demonstrate general lake characteristics associated with user
impairment. Interpretation of the dendrogram (Figure 58) established six
distinct water groups of dissimilar spectral composition and intensity. The
groups are comprised of water bodies having similar optical and physical
characteristics. The clustering algorithm provides a nonbiased perspective
of the spectral traits of the lakes. The spectral properties of each lake
provide an integrated characterization of water quality and relate to
water-use indices.
Cluster interpretation was accomplished by comparing the spectral
composition and uniqueness of each cluster with water quality data, field
evaluations, lake morphology, and watershed characteristics. The LANDbAT
flyover during October of 1973 was supported by chemical, physical, and
biological data collected by the NES sample program for 31 lakes. Since the
classification procedure utilized spectral responses, the NES data were of
limited value because many of the NES lake parameters have little or no
direct impact on the lake spectra. Field surveys were made during the summer
of 1977 for 72 of the MSS-sampled lakes. These surveys included physical,
biological, and chemical information useful for comparison of lake clusters.
Lake morphology, watershed information, and field evaluations of user
impairments were available for most lakes to supplement the ground
information base.
133
-------
TABLE 19. RANKINGS OF 145 ILLINOIS LAKES BASED ON SET THREE MODELS
AND ORDERED BY NAME
Serial
Lake Name Number
Anderson
Apple Canyon
Argyle
Bakers
Bangs
Bath
Big
Big
Bracken
Calumet
Canton
Carbondale
Carlyle
Catherine
Cattail
Cedar
Cedar
Central i a
Chain
Channel
Charleston
Chautauqua
Clear
Coal City
Recreation Club
Commonwealth Edison-
Dresden Nuclear
Countryside
Crab Orchard
Crane
Crooked
Crystal
Decatur
Deep
DePue
Devil's Kitchen
Di amond
DuQuoin
Dutchman
East Loon
Fourth Lake
34
144
157
19
56
166
134
176
145
18
36
38
14
58
125
39
55
79
83
59
16
87
86
138
137
148
127
85
149
68
73
150
3
128
57
98
49
53
151
SEC
127
29
40
110
32
108
70
117
21
142
73
134
79
25
86
76
30
31
116
50
121
140
124
10
92
51
57
137
65
23
111
9
130
6
37
67
11
47
115
CHLA
112
38
3
99
66
11
50
68
19
96
22
131
30
102
69
14
85
27
93
101
103
105
109
18
52
119
83
81
125
51
124
63
120
36
39
136
20
79
115
TON
86
45
5
104
35
70
32
57
36
88
10
82
4
80
127
29
73
52
90
81
94
95
102
3
14
91
49
78
115
19
105
30
118
47
13
140
39
85
89
TPHOS
84
39
22
132
16
109
55
74
50
119
25
127
3
36
134
63
34
54
107
44
95
111
117
11
15
67
35
108
125
13
75
7
118
33
5
124
52
48
80
PC1Y5
120
35
5
127
38
85
47
87
27
125
25
100
19
61
109
30
51
28
110
72
112
121
123
3
41
90
59
104
136
18
114
33
126
16
14
130
23
56
116
PC1F5
107
39
7
130
26
96
50
74
37
122
21
111
8
52
127
41
44
40
110
58
108
119
121
3
25
75
49
106
132
15
101
18
125
27
11
131
36
53
102
(continued)
134
-------
TABLE 19. (continued)
Lake Name
Fox
Fyre
Gages
George
Glen 0. Jones
Goose
Goose (Village)
Grass
Greenville New City
Griswold
Harrisburg
Highland
Holiday
Horseshoe
Jack, Swan, Grass
Keithsburg
Kinkaid
Kinneman
Lake of Egypt
Lake of the Woods
Larue-Pine
Lily
Lincoln Trail
Little Grassy
Little Swan
Liverpool
Long
Long
Lower Smith
Lyerla-Autumnal
Marie
Marion
Marshall
Matanzas
Mattoon
McCul 1 om
McGinnis
Meredosia
Mesa
Moscow
Moses
Mound
Serial
Number
60
170
152
108
110
101
139
61
2
158
111
153
51
1
165
92
40
168
132
115
121
136
13
119
185
88
52
177
181
120
62
129
81
89
24
159
20
10
184
164
28
163
SEC
94
53
28
18
2
145
99
106
80
66
55
35
78
85
123
75
1
101
19
109
107
135
3
5
89
119
77
131
88
59
91
48
114
138
62
22
133
141
16
87
20
90
CHLA
121
87
84
6
21
135
7
122
123
70
67
116
107
92
127
58
2
26
16
37
141
49
28
43
40.5
8
118
73
113
34
134
106
140
91
31
24
142
88
13
71
57
138
TON
87
120
54
12
16
133
37
103
101
117
99
100
114
132
121
112
1
44
6
113
143
72
42
38
21.5
60
96
69
15
76
126
134
108
84
11
28
144
83
20
62
111
122
TPHOS
59
93
24
18
14
138
87
65
76
100
78
43
69
135
112
130
1
104
4
131
145
123
32
20
56.5
113
53
105
58
110
89
126
121
91
19
28
144
115
17
82
86
101
PC1Y5
106
93
48
11
10
140
49
101
98
84
64
79
97
122
131
96
2
65
6
118
143
92
12
21
42.5
80
89
94
26
76
128
105
113
108
24
17
144
115
15
82
54
117
PC1F5
79
98
34
12
9
139
66
82
89
97
69
60
83
133
126
116
1
78
5
128
143
104
23
20
45.5
91
67
100
38
90
118
120
117
103
17
24
144
114
16
80
71
115
(continued)
135
-------
TABLE 19. (continued)
Lake Name
Murphysboro
New Pittsfield
Nippers ink
Old Ben Mine
Olney East Fork
Olney New
Open Pond
Otter
Pana
Paradise
Paris Twin
Pekin
Petite
Pierce
Pinckneyville
Pistakee
Powerton Cooling
Quiver
Raccoon
Rend
Rice
Round
Saganashkee
Sahara Coal Company
Sam Dale
Sam Parr
Sand
Sanganois
Sara
Sawmi 1 1
Senachwine
Shelbyville
Skokie Lagoons
SI ocum
Snyder's Hunting
South Wilmington
Spring
Spring
Spring
Spring
Spring Arbor
St Mary's
Serial
Number
37
100
63
32
106
107
174
167
113
15
25
116
64
130
99
65
182
90
80
29
35
66
21
175
124
41
147
9
26
103
104
114
22
54
142
140
4
8
67
118
143
146
SEC
14
129
102
64
13
39
74
60
49
83
42
126
84
52
72
68
71
100
104
63
103
33
96
118
43
41
15
122
4
144
143
56
95
98
69
7
120
139
93
82
27
17
CHLA
25
45
126
56
29
32
60
23
4
76
55
44
117
90
97
95
104
54
62
53
133
65
139
1
94
64
82
77
15
143
137
46
98
145
59
9
108
111
40.5
129
48
12
TON
43
31
106
93
9
74
119
53
8
59
109
61
97
66
79
71
68
41
25
24
125
34
135
48
110
128
51
56
26
142
124
18
137
145
123
2
136
98
21.5
116
107
27
TPHOS
41
66
70
90
6
81
133
79
31
61
92
98
71
51
62
42
45
64
40
12
97
10
122
114
68
129
23
102
29
142
103
9
141
143
116
2
139
96
56.5
85
99
60
PC1Y5
13
62
111
69
8
46
103
44
7
71
57
78
102
60
91
70
86
55
63
32
132
36
134
68
81
83
45
77
9
142
138
20
135
145
67
1
133
119
42.5
124
53
34
PC1F5
30
59
94
81
6
61
123
57
10
64
76
85
88
56
68
54
63
55
51
19
124
22
134
84
72
109
33
86
13
142
129
14
140
145
95
2
137
112
45.5
113
77
42
(continued)
136
-------
TABLE 19. (continued)
Lake Name
Stephen A. Forbes
Storey
Sugar Creek
Summerset
Sunfish
Swan
Swan
Tampier
Third
Thunderbird
Turner
Upper Smith
Vandal i a
We-Ma-Tuk
West Frankfort New
West Frankfort Old
West Loon
Wolf
Wonder
Worl ey
Yorkey
Zurick
Serial
Number
78
50
178
186
126
93
172
23
155
173
102
180
27
33
31
30
156
17
69
117
179
154
SEC
26
44
125
46
81
112
132
61
54
8
136
97
45
58
34
38
12
36
105
113
128
24
CHLA
47
89
114
74
33
17
132
35
128
5
130
72
61
10
78
110
80
42
144
86
75
100
TON
50
33
55
92
65
63
138
64
130
7
141
58
23
17
75
131
46
40
139
129
67
77
TPHOS
37
47
106
49
83
120
137
77
73
8
140
72
27
46
38
94
21
26
128
136
88
30
PC1Y5
37
40
74
73
66
75
139
52
107
4
141
88
31
22
50
99
39
29
137
129
95
58
PC1F5
35
43
87
62
70
93
138
65
99
4
141
73
28
32
47
105
31
29
136
135
92
48
Cursory examination of the remote-sensing data indicated that the natural
groupings of lakes developed by cluster analysis techniques discriminated
groups of water bodies, each with its own unique physical qualities. Data for
the four LANDSAT bands were rescaled in a manner suggested by Ruttner (1963).
Thus, the raw reflectance values were transformed into percentages that were
used as a four-element index of lake wavelength reflectance characteristics.
The maximum and minimum reflectance values recorded for each band were assumed
to represent a reasonable range to form the base index signatures.
Establishing the range from actual energy return signals for water bodies
minimized the path radiance interference from the base signal. Ground control
data for suspensoids and Secchi depths verified this assumption.
This characterization of lake waters by their optical spectrum relies upon
the relationships between the energy reflectance of the clearest Illinois lake
and waters containing greater concentrations of suspended or dissolved matter.
The procedure is similar to that used by Rogers (1977) except that no lake in
the Illinois MSS program is considered an exceptionally pure-water lake.
Hence, Illinois lake index signatures of reflectance data represent actual
137
-------
TABLE 20. COMPOSITE RANKING OF 145 ILLINOIS WATER BODIES BASED ON SET
THREE MODELS AND ORDERED BY INCREASING TROPHIC STATE
Lake Name
Serial
Number RANKSUM
Kinkaid
South Wilmington
Thunderbird
Coal City Recrea-
tion Club
Lake of Egypt
Glen 0. Jones
George
Olney East Fork
Mesa
Argyl e
Sara
Pana
Diamond
McCullom
Lincoln Trail
Crystal
Little Grassy
Deep
Carlyle
St Mary's
Dutchman
Devil's Kitchen
Mattoon
Murphysboro
Bracken
Shelbyville
Canton
We-Ma-Tuk
Round
Wolf
Bangs
Apple Canyon
Rend
Vandal i a
West Loon
Stephen A. Forbes
Central i a
Sand
Commonwealth Edison-
Dresden Nuclear
40
140
173
138
132
110
108
106
184
157
26
113
57
159
13
68
119
150
14
146
49
128
24
37
145
114
36
33
66
17
56
144
29
27
156
78
79
147
137
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
17
17
18
20
20
22
22
24
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
Serial
Lake Name Number RANKSUM
Cedar
Gages
Big
Little Swan
Spring
Storey
Otter
Cedar
Crab Orchard
West Frankfort New
Olney New
Goose (Village)
Raccoon
Zurick
Tampier
Catherine
East Loon
Quiver
Pierce
Summerset
Sunfish
New Pittsfield
Moses
Lower Smith
Ki nneman
Channel
Pi stakee
Paradi se
Lye rl a- Autumnal
Spring Arbor
Sahara Coal Company
Powerton Cooling
Highland
Paris Twin
Bath
Harri sburg
Upper Smith
Liverpool
Moscow
Old Ben Mine
39
152
134
185
67
50
167
55
127
31
107
139
80
154
23
58
53
90
130
186
126
100
28
181
168
59
65
15
120
143
75
182
153
25
166
111
180
88
164
32
40
41
42
43
44
45
46
47
48
49
50
51
53
53
54
55
57
57
57
59
60
61
63
63
64
66
66
68
68
70
70
71
72
74
74
76
76
77
78
79
(continued)
138
-------
TABLE 20. (continued)
Lake Name
Serial
Number RANKS LM
Lake Name
Serial
Number RANKSUM
Pinckneyvil le
Swan
Sam Dale
Big
Countryside
Pekin
Long
Griswold
Fyre
Sanganois
Yorkey
Fox
Sam Parr
Snyder's Hunting
Hoi iday
Petite
West Frankfort
Old
Keithsburg
Long
Lily
Greenville New
City
Third
Open Pond
Lake of the Woods
Grass
Fourth Lake
Sugar Creek
Nippers ink
Crane
Matanzas
Chain
Anderson
99
93
124
176
148
116
52
158
170
9
179
60
41
142
51
64
30
92
177
136
2
155
174
115
61
151
178
63
85
89
83
34
80
81
82
83
84
85
86
88
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
108
108
108
110
111
Spring
Charleston
Marion
Decatur
Cattail
Me red os i a
Crooked
Marie
Horseshoe
Spring
Cal umet
Bakers
Chautauqua
Mound
Clear
Rice
Worl ey
DuQuoin
Skokie Lagoons
Carbondale
Marshall
Jack, Swan, Grass
DePue
Saganashkee
Spring
Senachwine
Wonder
Sloe urn
La rue-Pine
Swan
Turner
Goose
McGinnis
Sawmill
118
16
129
73
125
10
149
62
1
8
18
19
87
163
86
35
117
98
22
38
81
165
3
21
4
104
69
54
121
172
102
101
20
103
112
113
114
115
116
117
118
119
121
121
123
123
125
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
139
-------
i * *is
$«iji
o ! It*
^
-u-
Serial Number and Lake Name
Cluster Number
Figure 58. Dendrogram of 145 Illinois water bodies based on complete linkage clustering on four
spectral attributes (GRN, RED, IR1, IR2). The dissimilarity axis is in the distance
scale A. The attributes' values were not standardized prior to computing the A?
values for the resemblance matrix.
-------
optical qualities for these lakes and are not a comparison of these lakes to
"pure-water lakes" found elsewhere.
Generalized signatures for various types of Illinois water bodies are
found in Table 21. These signatures represent actual reflectance differences
in the LANDSAT MSS bands for lake waters integrated over Secchi depth. Since
most Illinois lakes exhibit a Secchi depth of less than one meter, and because
most solar radiation is absorbed within one meter (Ruttner 1963), the
procedure provides a reasonable demonstration of Illinois lake properties.
Use of the actual range of recorded spectral values to establish index
signatures provides an objective and comparable data base to establish a "real
world" interpretation of Illinois lake groups.
Cluster analysis resulted in'the delineation of 6 groups of 21 to 28 lakes
each, arranged in distinct subgroups (Figures 58 and 59). Index signatures
and general lake qualities for the six clusters are in shown Table 22.
Some semi quantitative characteristics of the lake clusters and their
subclusters at level six are readily apparent when compared to the ground
control data in Table 23. The six clusters can be ordered on the basis of
declining water quality as follows: 3-1-2-5-4-6. Based upon this order,
Figure 60 demonstrates the correlation between spectral signatures and 1977
Secchi depth and suspensoid levels for 72 of the 145 lakes. Suspensoid
contact-sensed data are available for 1977 for each subcluster of clusters 1
to 6 and represent 42 to 82 percent of the lakes in each group. The obvious
relationship between Secchi depth and proportionate concentrations of volatile
(organic) and total suspended solids suggest that the lake clusters can be
discriminated by Secchi depths, total suspended matter, and the proportions of
inorganic-organic suspensions.
Clusters 3 and 1 have similar organic suspensoid levels but are separable
by their inorganic suspensoid levels and associated changes in Secchi depths.
Clusters 5 and 4 demonstrate a similar pattern indicating that inorganic
turbidities account for the differences between Clusters 3 and 1 and also for
those between 5 and 4. Cluster 2 represents lakes exhibiting qualities
between these two sets. Algal concentrations for the 72 lakes sampled during
1977 closely follow the volatile suspended solids values indicating that
primary productivity increases from Cluster 3 to 5 and then decreases in
Cluster 4. The abrupt reduction in algal counts from Cluster 5 to 4
corresponds with a 100 percent increase in inorganic suspensions and a Secchi
depth reduction of 50 percent (to 0.28 m) in Cluster 4. Cluster 6 appears to
follow the ordered relationship according to actual spectral information;
contact-sensed data to demonstrate suspensoid levels and Secchi depth
relationships for this cluster are available for only four lakes. Generally,
lakes in Cluster 6 are considered poor quality lakes that were either
privately owned lakes or were too shallow to sample during the 1977 survey.
Figure 61 suggests that the 1973 NES data and the 1977 IEPA data for
Secchi depth qualitatively corresponds with the LANOSAT RED band. Figure 62
demonstrates the correspondence of suspensoid levels and light extinction
properties of Illinois lakes during 1977. The relationship between
suspensoids, Secchi depth, and spectral response is obvious. Generally, the
141
-------
TABLE 21. INDEX SIGNATURES FOR VARIOUS WATER BODY TYPES IN ILLINOIS
Type of Water Body
LANDSAT MSS Band Percentages
GRN RED IR1 IR2
Clean glacial lakes
Clear deep reservoirs
Low algal clear reservoirs
High algal reservoirs
High algal glacial lakes
Turbid reservoirs
River backwaters
Sloughs and harbors
5
5
20
30
25
40
55
85
10
5
20
30
25
45
55
90
5
15
15
25
20
30
50
90
5
25
10
10
5
25
30
90
TABLE 22. INDEX SIGNATURES AND GENERAL DESCRIPTIONS FOR SIX ILLINOIS LAKE
CLUSTERS
Cluster
Number
Signature
GRN-RED-IR1-IR2
Cluster Description
3
1
10-10-15-16
25-22-22-19
27-29-33-23
40-42-54-41
49-53-41-21
67-68-62-44
Deep clear lakes of excellent quality with
macrophytes
Deep somewhat clear lakes of good to very
good quality with slight sediment, algal,
and macrophyte problems
Moderate depth lakes of good quality with
moderate silt and algal problems, slight
macrophyte problems, and average Secchi
depths
Shallow reservoirs and backwaters of fair
quality with moderate sediment problems and
severe algal problems
Shallow lakes and backwaters of poor quality
with low Secchi depths and severe sediment-
related problems
Shallow lakes and backwaters of very poor
quality with sediment-related
turbidity problems
142
-------
X
LU
Q
801
70-
60H
50-
LU
O
z
<
5 40-
LU
30-
20-
10-
f-MSSGRN-Hf MSSRED-*- MSS IR1
I
MSS IR2
550 650 750 950
WAVELENGTH (Nanometers)
Figure 59. Spectral index signature curves for six lake clusters
based upon raw MSS data.
143
-------
TABLE 23. SPECTRAL DATA AND CONTACT-SENSED DATA FOR THE CLUSTERS AND SUBCLUSTERS DEVELOPED USINli
COMPLETE LINKAGE CLUSTERING ON FOUR ATTRIBUTES (GRN, RED, IR1, IR2)*
3d
b
a
c
3
Ib
c
a
d
1
2d
a
c
b
2
5a
b
e
c
d
5
Clusl
Nunb«
and
rer
•r
(Obs.) GRN
(10
( 3
( 8
( 7
{28
( 6
( 9
( 8
( 1
11
6
13
5
10
29
21
22
50
(24) 25
( 3
( 7
5
(6
11
29
29
30
(21) 27
( 5) 42
( 3) 34
( 3) 31
( 8
( 4
(23
44
38
40
Scaled Spectral
Signature
RED
9
2
13
11
10
23
22
19
19
22
17
32
26
34
29
46
38
27
47
41
42
IR1
6
13
18
25
15
24
17
27
17
22
35
26
37
35
33
41
46
72
55
58
54
IR2
5
33
17
30
16
21
10
29
15
19
46
10
34
19
23
36
37
52
32
61
41
IEPA Data
(1977)
TSS
6.4
1.3
8.8
8.7
7.0
6.6
10.3
9.3
1.0
9.7
3.0
14.5
15.3
39.3
21.8
13.7
27.0
25.0
31.0
65.0
31.7
VSS
5.4
1.0
6.4
6.9
5.9
4.4
9.0
3.7
0.0
5.8
2.0
9.0
10.6
18.3
11.9
8.7
16.0
21.0
16.0
40.0
21.5
SEC
1.32
3.31
1.14
0.95
1.43
1.40
1.13
1.04
1.97
1.27
1.65
1.51
0.70
0.39
0.88
1.37
0.33
0.34
0.30
0.14
0.59
Observed
Problems
4J
I .
5 s,
5
1.3
1.0
1.6
2.0
1.5
2.0
2.0
1.9
1.0
1.9
2.0
2.6
2.5
2.6
2.5
2.2
2.3
2.3
3.3
3.3
2.7
<
2.0
1.3
2.4
2.6
2.2
2.0
2.4
1.9
1.0
2.1
2.5
2.7
2.5
2.4
2.6
2.2
2.5
3.7
2.0
3.0
2.5
01
*
?
^f
2.7
1.7
2.7
2.7
2.6
2.4
1.8
2.7
2.0
2.3
3.0
1.9
3.0
1.8
2.3
1.8
2.3
3.0
1.7
2.0
2.0
NES Data
(1973)
SEC
2.77
0.91
1.83
0.70
0.43
0.49
0.40
0.49
0.46
0.43
0.18
0.31
0.24
CIILA
6
27
17
13
39
33
54
24
198
77
182
241
212
Morphology
I
3.72
7.53
3.84
3.57
4.12
4.57
3.63
4.63
8.75
4.42
2.65
2.23
2.80
2.23
2.44
4.60
1.07
0.76
1.49
0.61
1.83
*m
11.19
21.23
9.85
8.90
11.46
11.95
10.33
11.16
24.38
11.58
8.23
6.22
5.73
5.06
6.10
8.96
2.59
1.95
3.60
1.74
4.30
tershed
ulvalent
ntlmeters
533
90.9
61.2
39.4
22.1
53.B
33.0
43.4
25.9
56.9
34.8
31.2
18.8
23.4
19.8
23.1
13.5
16.0
6.9
30.0
13.2
Hydrology
c
o
iff f~
Vt O 4J
1 °^
4-» Of » 91
c § ** e
o» v 5 * •"
•ME f— r— t3
SI
4.20
1.81
1.46
0.71
2.25
1.30
2.09
1.00
1.72
1.46
0.97
0.57
0.94
0.95
0.90
0.54
0.78
0.34
1.07
0.69
££$
0.13
0.09
0.40
0.30
0.32
0.18
0.11
0.20
0.04
0.16
0.24
0.14
0.07
0.13
0.14
1.17
0.28
0.59
0.12
0.72
.c
-------
TABLE 23. (continued)
en
Cluster
Number
and
(Obs.
4c ( 5
d ( 6
a ( 8
) CRN
56
44
51
b ( 5) 46
4 (24) 49
6a (12) 59
b ( 4) 70
c ( 6
d ( 3
71
91
6 (25) 67
Scaled Spectral
Signature
RED
56
47
56
51
53
62
77
56
88
68
1R1
43
34
48
49
41
57
68
54
80
62
1R2
18
11
27
24
21
33
39
62
69
44
1EPA Data
(1977)
TSS
57.3
29.0
77.8
120.0
61.8
68.0
150.0
58.0
VSS
22.5
9.0
31.8
27.0
22.6
20.0
39.0
16.8
SEC
0.39
0.34
0.30
0.11
0.28
0.11
0.11
0.11
Observed
Problems
! %
5
3.0
3.0
3.4
3.3
3.2
3.2
3.2
2.5
4.0
3.1
5
2.4
2.0
2.0
1.7
2.1
1.5
2.2
1.3
1.0
1.6
I
v>"~ *
a> i- Z>
*J =J C
sss
19.6
3.8
5.8
49.0
11.7
23.1
15.0
0.3
13.2
Hydrology
c
o
ssx
O — «J
••- T> C
*> V « V
So, Is!
Jsl
0.88
0.16
0.27
2.22
0.61
1.14
0.53
0.56
££%
0.32
1.52
3.07
0.12
1.49
0.15
0.25
0.22
JC.
vt
s
n <•-
r-. c
91 >, 3
r* JO &
97.9
53.9
724.8
5.6
233.7
11.0
26.1
20.7
Biota
c
r— O
o> c
1-4 UJ
19.3
4.1
2.1
1.0
6.6
55.1
•Parameter means were constructed for 6 clusters and 25 subclusters.
Parameters, acronyms, and units are as follows:
Scaled spectral signature - LANDSAT bands were scaled (0-100) using
approach described 1n text.
1EPA Data (1977) - Data collected by Illinois Environmental
Protection Agency during summer of 1977. TSS - total suspended solids
(mg/Hter). VSS - volatile suspended solids (mg/llter). SEC -
Secchl depth (m).
Observed Problems - Problems noted through field observations and
adversely affecting use. Scaled 0-4 with 0=m1n1mal problem and
4=severe problem.
NES Data (1973) - Data collected by U.S. EPA's National Eutro-
phication Survey in October 1973. SEC - (m). CHLA - dig/liter)
Morphology - Data from 1EPA files.
maximum depth (m).
I - mean depth (m). Zm -
Hydrology - Data from 1EPA files. Watershed equivalent centimeters -
reservoir volume divided by watershed area (cnrVcm'). Detention
time (years). Volume loss related to sedimentation - annual loss of
lake volume expressed as a percentage. 1973 flush by rain runoff -
percent of lake volume replaced by runoff from heavy rains just prior
to dates of LANDSAT coverage.
Biota - Data from 1EPA files.
(cells x 1000/ml).
1977 algal enumeration
-------
A B
AD Algae (1000/ml)
B E3 Secchi Depth (cm)
I Total Suspended Solids (mg/l)
C EH Volatile Suspended Solids (mg/l)
D[ ] Red Index Reflectance %
C D
18-
12-
6-
•10
-50
-90
.130
60-
48-
36-
24-
12-
-60
-50
-40
-30
-20
-10
Figure 60.
2 5
Lake Cluster
Visual correlation between contact-sensed data and scaled
MSS RED band data at the six-cluster levels. The parameter
values are cluster means from the 1977 contact-sensed data.
146
-------
E
o
340-
300-
260-
220-
O 140-
o
100-
60-
20-
Figure 61.
• 1973 NES Secchi Depth
*1977 Secchi Depth for NES Lake
10 20
30 40 50
60 70
80 90
Red Band Index %
Light extinction relationship between MSS RED band index
(percent) and Secchi depth measurements taken in 1977.
The line is a trend line obtained by a visual fit, not a
regression.
147
-------
.*>
00
«••»
\
O)
V)
"o
•o
TJ
c
OJ
a
to
3
15
o
H-
200-i
100-
80-
60-
40-
20-
10-
5-
V-
*.•
%
*1 %
*
.7^
*'4 . .•
0 0
•
+• m . • •
• • ••• •
i i i i i i i i i *
40 80 120 160 200 240 260 280 320 400
Secchi Depth (cm)
Figure 62. Total suspended solids relationship with Secchi depth. Data are mean values for each lake
sampled during 1977.
-------
spectral reflectance values increase in cluster order 3-1-2-5-4-6 as shown in
Figure 59 and Table 23. Corresponding increases in total and volatile
suspended solids follow a similar pattern, while Secchi depths demonstrate a
progressive decrease in light penetration (Figure 60, Table 23). Algal data
for 1977 correlate with the general pattern found in the field observations of
use impairment and correspond to Secchi and volatile suspended solids data.
Similar trends for Secchi depths and chlorophyll a^ concentrations are evident
in the 1973 NES data, the extensive ground control data of 1977, the use
impairment evaluations, and the spectral index signatures.
Chlorophyll a^ levels and algal counts indicate primary productivity
increases from deep lakes to shallow water bodies until the inorganic
suspensoids decrease Secchi depths below one-half meter. The most turbid
water bodies are encountered in Clusters 4 and 6. These trends suggest that
user impairment is related to lake morphology and suspensoids. Algal
suspensoids selectively affect shallower water bodies unless sediment-related
turbidities are excessive. Suspended sediment appears to progressively impact
use impairment as water bodies become shallower and ultimately becomes the
single factor determining loss-of-use potential in the worst lakes.
Macrophyte use impairment is most significant in deep, clear waters and
gradually becomes less conspicuous as suspensoids increase and Secchi depths
decrease.
Thus, suspensoids are a significant factor in determining Illinois lake
qualities and impairments. Table 23 summarizes lake morphology and
hydrological information for lake clusters and subclusters to demonstrate the
obvious relationship of lake morphology to the use impairment factors
discussed above. The decrease in mean and maximum depths generally follow the
established lake cluster order (Table 23) as do the volatile and total
suspended solids relationships. Clusters 3 and 1 are deep, clear water bodies
with well-developed basins, while Clusters 5 and 4 consist of shallow turbid
lakes. Cluster 2 is between these.
Interpolation of hydrological data for these clusters provides insight as
to their behavior regarding suspensoid levels and Secchi depths. Analysis of
the equivalent flushing rate and normal detention time of lakes indicates that
the concentration of inorganic suspensoids is a function of inputs from
runoff. Calculations for detention times account for the average annual
runoff values (Figure 63) expected for each watershed. The detention period
for lake clusters decreases in the order 3-1-2-5-4-6, corresponding to the
inorganic suspensoid and MSS reflectance index signature increases in the same
order. This inverse relationship of detention time to lake quality can be
interpreted as a direct correlation of lake flushing to suspensoid levels and
overall qualities. Therefore, it follows that shallow reservoirs with large
watersheds will be more severely impaired than deep reservoirs with long
detention capacity.
Illinois lakes are significantly influenced by their flushing rates.
Short-detention lakes are of poorer quality, irrespective of appreciable
nutrient inputs, than are lakes with long hydraulic retention times.
Therefore, light inhibition may be more significant in determining Illinois
lake impairment than the classical nutrient trophic index procedures. It
149
-------
8
: 10
0 1 10 10
UWMfifTM
Jr\ . IMODU01I
15
Figure 63. Average annual runoff for Illinois in inches/square mile/year
(Upper Mississippi River Basin Study Commission 1970).
150
-------
appears that algal productivity is proportionate to flushing rates until the
inorganic suspensions mask the effects of nutrient flushing by restriction of
the euphotic zone. Obvious relationships of lake morphology and watershed
hydraulics to the overall quality of Illinois lakes are summarized in Table
23.
To further examine the relationships between lake quality characteristics
and lake morphology and hydraulic characteristics, equivalent flushing rates
were determined for the clusters from meteor!ogical conditions just prior to
satellite flyover (Figure 64 and Table 23). The equivalent flush occurring
just prior to the MSS sampling in 1973 does not account for the good-to-
excellent quality water bodies of Clusters 3, 1, and 2 since those were not
flushed by significant amounts of runoff prior to sampling. Clusters 4 and 5,
on the other hand, were flushed by significant amounts of runoff and are
comprised of lakes with poor water quality. The general flushing trend
increases in cluster order 3-1-2-5-4-6 and corresponds to higher suspensoid
loads than expected for annual watershed equivalent centimeter volumes and
detention times. The rainfall amounts that occurred during October, 1973,
were not unusual for this time of year.
Volume loss estimates given in Table 23 relate to morphology and hydrology
of the lake and its watershed. Watershed physiography, geology, and soil
types affect the sediment impact. Discussion of these factors and their
bearing on the lake clusters is given in the detailed lake cluster
descriptions. Volume loss from sedimentation can be misleading because this
figure represents a percent loss by lake volume and not a concentration in the
water column. In addition, lakes having larger watersheds generally have
proportionally lower sedimentation losses because less sediment reaches the
reservoir. Sediment-trap efficiency decreases as the reservoir capacity
decreases, and consequently an increased amount of sediment is passed through
the reservoir outlet. Thus, long-detention lakes may have greater trap
efficiency and higher volume losses than short-detention lakes, even though
their loadings are similar.
Lake Cluster Characteristics--
Table 23 lists representative data for the lake clusters. In addition,
values for subclusters were included to indicate the variability of lake types
within a cluster. Appendix A-5 gives individual data for each lake within the
clusters and subclusters. The lake classification developed by this effort
generally is based upon the physical properties and characteristics of lakes.
This approach provides an objective comparison of the water transparency as it
is affected by weed growth, nuisance algal blooms, impaired water quality,
sediment infilling, and turbidity. These problems are directly related to
impairment of use and do not consider the nutrient budgets as a means to
characterize lakes in Illinois.
The following discussion provides a characterization of the lake clusters
established by spectral responses. The agreement of ground truth information,
subjective observations of lake problems, watershed characteristics, and lake
morphology suggest the approach is valid. Watershed geology and soil
associations are included in the analysis.
151
-------
^ J., 0.-~, I s,,^,..,,! „.
Figure 64. Distribution of precipitation In centimeters for Illinois
for the period October 11-14, 1973. Data from Environmental
Data Service (1973).
152
-------
Cluster 3 characteristics--This cluster generally represents lakes in
Illinois having excellent characteristics and associated trophic qualities.
The average depth of visibility (Secchi depth) is 1.43 m. Volatile suspended
solids comprise over 80 percent of the total suspensoids by weight, indicating
that most of the light-extinction properties are of organic origin. The TSS
values are the lowest recorded for all lakes.
Field evaluations of lake conditions demonstrated that sediment-related
turbidities are minimal. Some problems occur as a result of algal blooms.
Macrophyte growths are moderate and may be associated with the relatively
clear waters in these lakes.
The cluster includes most of the good quality, natural glacial lakes of
less than 100 surface hectares in the Great Lakes section of the Central
Lowland province (Figure 2). Most other lakes in this cluster are small-to
medium-size south-central or southern Illinois impoundments in the Tills
Plains section and are associated with weathered Illinoian glacial drift
deposits (Figure 2). Their watershed soils are generally low in productivity,
with less than one percent organic matter, and are light-colored with high
clay content.
Cluster 3 lakes are generally deep (for Illinois) and well developed, with
the mean depth for the group averaging 4.12 m and the maximum depth averaging
11.46 m. The average lake volume, expressed as watershed equivalent
centimeters (reservoir volume divided by drainage basin area (cm3/cm2)),
is over 53.8. This high value represents lakes having average detention
periods of 2.25 years, generally ranging between 0.5 and 9 years. Rainfall
prior to satellite flyover was equivalent to an average volume displacement of
19 percent for all lakes in this group. This value represents the lowest
equivalent displacement of all clusters. The watersheds generally have less
than 1.25 m of loess deposits. The long detention periods and watershed soil
types result in an estimated annual lake capacity loss from sediments of 0.32
percent.
Cluster 1 characteristics—This cluster represents lakes of good quality
associated with moderate sediment, algal, and macrophyte problems. The
average Secchi depth is slightly less than Cluster 3 lakes, although the
organic suspensoids remain constant. The moderate increase in total suspended
solids, when compared to Cluster 3 lakes, indicates increased inorganic
turbidities and losses of light penetration. Field observations suggest that
the more sparse macrophyte growths are primarily the result of increased
sediment-related turbidity levels in this group.
This group of lakes has a mean depth averaging 4.42 m and a maximum depth
averaging 11.58 m. Although their morphology is similar to Cluster 3 lakes,
this group has larger watersheds. The average lake volume, expressed as
watershed equivalent centimeters, is less than 34.8. The average lake
detention period is 1.46 years and generally is from 0.3 to 6 years. Rainfall
prior to satellite flyover represented an average volume displacement of
nearly 30 percent, a significant increase over Cluster 3 lakes. Annual
average lake capacity loss from sediment is estimated to be approximately
0.16 percent.
153
-------
Most of these lakes have watersheds in glacial drift generally associated
with loess deposits of less than 1.25 m. All lakes are either natural glacial
lakes, gravel pits, or artificial reservoirs. The reservoirs vary from 19 to
2,953 ha and generally are associated with Illinoian glacial drift of the Till
Plain section. In general, sediment, algae, and macrophyte problems are
slight to moderate.
Cluster 2 characteristics—Cluster 2 lakes are similar to Cluster 3 and 1
lakes except that their water quality is lower. Moderate turbidity and algae
bloom problems occur. Macrophyte abundances are not excessive, a consequence
of the turbidity conditions. These lakes have almost twice the concentrations
of both total and volatile suspended solids and are associated with a
40-percent decrease in the depth of visibility when compared to Cluster 3
lakes. Sediment-related turbidity problems appear to contribute to the Secchi
depth average of less than one meter. The NES chlorophyll a_ values suggest
that algal problems are progressively worse from Cluster 3 to 1 to 2; changes
in volatile suspended solids values for these clusters qualitatively
correspond to changes in chlorophyll values.
This group of lakes has a mean depth averaging 2.4 m and a maximum depth
averaging 6.1 m. The average lake volume, expressed as watershed equivalent
centimeters, is about 23. The detention period ranges from 0.4 to 2 years and
averages 0.9 years. Thus, these lakes have significantly shorter detention
times than the previous clusters. Rainfall prior to satellite flyover
represented an average volume displacement of about 25 percent. Although this
average displacement value is less than the Cluster 1 value, the individual
lake values in Cluster 2 lakes varied less.
Lakes in Cluster 2 include natural glacial flowage lakes known to be
turbid and productive, shallow natural glacial lakes, several river
backwaters, and artificial reservoirs. Most of the reservoirs have watersheds
associated with Illinoian glacial drift with loess deposits having
light-colored soils with less than 1 percent organic matter. Average annual
lake capacity loss from sediment is estimated at 0.14 percent.
Cluster 5 characteristics—This cluster represents lakes having moderate
to severe sediment-related turbidities and algal blooms and moderate
macrophyte growths. In general, inorganic and organic suspensoids are high.
The volatile suspensoids account for nearly two-thirds of the total
suspensoids, which are nearly four times those found in the best quality lakes
(Cluster 3). Secchi depth averages about 0.6 m, as compared to 1.4 and 0.9 m
for Clusters 3 and 2, respectively.
This group of lakes has a mean depth averaging 1.8 m and a maximum depth
averaging 4.3 m. It includes natural backwater lakes in the bottom lands with
alluvium and no loess deposits, as well as sloughs and artificial reservoirs.
The reservoirs have an average lake volume of about 13, expressed in watershed
equivalent centimeters. Detention periods generally range from .007 to 2
years and average 0.69 years. Rainfall prior to satellite flyover represented
an average volume displacement of about 143 percent. This value represents a
significant displacement for reservoirs.
154
-------
Ground samples during 1977 indicated that the deeper reservoirs are
generally clear waters except after high runoff periods. The average annual
lake capacity loss from sediments is estimated to be 0.72 percent. The
backwater areas receive overflows from the adjacent rivers and have turbid
waters of poor to fair quality. Reservoirs deeper than 4 m are considered to
be fair to good quality lakes, while the shallower reservoirs are generally
of poor quality.
Cluster 4 characteristics—This cluster represents lakes having severe
sediment-related turbidities, moderate algal bloom problems, and minimal
macrophyte problems. Secchi depth is less than 0.3 m, or less than one-half
the depth observed in Cluster 5 lakes. The total suspended solids of Cluster
4 lakes is double the average for Cluster 5, although the volatile
suspensoids are at about the same level. The significant decrease in depth
of visibility corresponds with the increase in sediment.
This group of lakes has a mean depth averaging 2.4 m and a maximum depth
averaging 4.9 m. Most of the artificial reservoirs drain watersheds of
Wisconsin glacial drift with less than 1.25 rn of loess deposits. Their soils
are dark-colored and productive and have more than 1 percent organic matter.
The average lake volume, expressed in watershed equivalent centimeters, is
only 5.8. The detention period ranges from 0.002 to 0.546 years. Rainfall
prior to satellite flyover represented an average volume displacement of
nearly 360 percent. The annual lake capacity loss from sediment ranges
between 0.06 and 8.5 percent and averages about 1.8 percent. The quality of
the reservoirs is considered fair, while the natural backwaters are in poor
condition. These lakes do not have excessive macrophyte growth problems.
Field observations suggest this is primarily because of the sediment-related
turbidity.
Cluster 6 characteristics—Cluster 6 lakes are generally of poor quality,
with severe sediment-related turbidity problems. The group is composed of
many shallow-river backwater lakes as well as harbors or artificial
reservoirs experiencing significant sedimentation problems. Contact-sensed
data are limited to four lakes. Two reservoirs sampled in 1977 had low
suspensoids loads and good Secchi depth readings but are known to have
experienced severe suspended solids problems in the past. Both watersheds
have more than 1.25 m of loess deposits. The two backwater lakes sampled
during 1977 exhibited extremely high suspended solids values. The TSS value
for Meredosia was 150 mg/liter, the highest value recorded for any lake
sampled during 1977. Most of the natural backwater lakes in this group
occupy the glacial outwash area adjacent to the Illinois River. These
lateral levee lakes have been severely impacted by drainage improvements,
levee construction, river dam construction, navigation, and silt deposition.
Lack of water clarity is the result of suspended sediment from flowage and
resuspension of flocculent bottom materials. Prior to their deterioration,
these bottom land lakes supported luxuriant macrophyte populations and fish
and invertebrate fauna.
155
-------
LANDSAT-Estlmated Trophic Indicator Classification of the Lakes
In the preceeding subsection the dendrogram generated from the
date-adjusted LANDSAT spectral data for the 145 water bodies was interpreted.
In this subsection, the results of a parallel effort that employed
LANDSAT-derived trophic indicator estimates will be interpreted. Clustering,
using trophic indicator estimates derived through conventional lake sampling
techniques as attributes, is a well-accepted approach. In this case
clustering is based on attributes (i.e., trophic indicators) estimated from
the LANDSAT data set.
Four of the trophic indicators (SEC, TPHOS, TON, and CHLA) estimated by
the three sets of regression models were selected for cluster analysis.
Separate clustering runs were made for each data set using two programs
(CLUSTER, a SAS program on COMNET, and CLUSTER, a program on Oregon State
University's Cyber 73). While both programs are of the complete linkage
type, their hard-copy products and cluster results are similar though not
identical.
The regression results indicate that the Set One models are least
satisfactory and that the Set Three models provide the best overall
estimates. The following interpretation is limited to the Set Three
12-cluster table (Table 24) generated by the SAS program on COMNET.
Limnologists found the 12-cluster classification to be more satisfactory than
the other groupings examined (e.g., 6- and 9-cluster classifications).
Lake Cluster Character! stics--
In the following subsections an interpretation is made of the 12-cluster
SAS program output (Table 24). The cluster characteristics, as evidenced by
the 1973 and 1977 contact-sensed data, are given in Table 25. Appendix
Tables A-2, A-3, and A-5 list the contact-sensed data for each lake within
the clusters. The interpretation is made by comparing: 1) trophic indicator
estimates with 1973 NES data; 2) trophic indicator estimates with the 1977
lake water quality data; and 3) the trophic status of the clusters with each
other. Clusters are described in order of increasing eutrophy as determined
by estimated values for SEC, TON, TPHOS, and CHLA. Terse descriptions for
each group are based on cluster means established from estimated values for
SEC, TON, TPHOS, and CHLA.
Trophic class number 1 (Cluster 8)—Terse description: (very high SEC,
low TON, TPHOS, and CHLA). Lake Kinkaid (serial number 40), a southern
Illinois reservoir and the only lake in this class, has excellent water
quality (at least for Illinois lakes) with minimal sediment-related turbidity
and algal blooms. The aquatic macrophyte problem is rated as slight.
Results of IEPA summer 1977 sampling for TSS, VSS, and SEC averaged
1.0 mg/liter, 0.0 mg/liter, and 1.97 m, respectively. Lake Kinkaid has a
mean depth of 8.7 m a"nd a maximum depth of 24.4 m; water retention time is
1.72 years.
156
-------
TABLE 24. TROPHIC CLASSES DEVELOPED FROM CLUSTER ANALYSIS OF TROPHIC
INDICATOR ESTIMATES FROM SET THREE REGRESSION MODELS
Name
Class 1 (Cluster 8)
Kinkaid
Class 2 (Cluster 7)
Lincoln Trail
Sara
Little Grassy
Devil 's Kitchen
Glen 0. Jones
Coal City Recreation
South Wilmington
Thunderbi rd
Murphysboro
Dutchman
Bracken
St. Mary's
Olney East Fork
Lake of Egypt
George
Mesa
Sand
West Loon
Deep
Class 3 (Cluster 5)
Paris Twin
Spring Arbor
Olney New
Moses
West Frankfort New
Cedar
Gages
Catherine
Zurick
Storey
Serial Number
40
13
26
119
128
110
138
140
173
37
49
145
146
106
132
108
184
147
156
150
25
143
107
28
31
55
152
58
154
50
County
Jackson
Clark
Effingham
Williamson, Jackson
Wi 1 1 i amson
Saline
Grundy
Grundy
Putnam
Jackson
Johnson
Knox
Lake
Richland
Johnson, Williamson
Rock Island
Wabash
Lake
Lake
Lake
Edgar
Jackson
Richland
Franklin
Franklin
Lake
Lake
Lake
Lake
Knox
(continued)
157
-------
TABLE 24. (continued)
Name
Serial Number
County
Crab Orchard
East Loon
Summerset
Channel
Pierce
Sam Dale
Highland
Countryside
127
53
186
59
130
124
153
148
Williamson
Lake
Winnebago
Lake
Winnebago
Wayne
Lake
Lake
Class 4 (Cluster 6)
Carlyle
Mattoon
Canton
Commonwealth Edison-Dresden
We-Ma-Tuk
Cedar
Vandalia (City)
Rend
Shelbyville
Wolf
Stephen A. Forbes
Apple Canyon
McCullom
Centralia
Bangs
Round
Crystal
Diamond
Pana
Argyle
Paradise
Big
Upper Smith
Spring
Little Swan
Big
Raccoon
Quiver
New Pittsfield
Pistakee
14
24
36
137
33
39
27
29
114
17
78
144
159
79
56
66
68
57
113
157
15
176
180
67
185
134
80
90
100
65
Clinton, Bond, Fayette
Cumberland
Fulton
Grundy, Will
Fulton
Jackson
Fayette
Franklin, Jefferson
Shelby, Moultrie
Cook
Marion
Jo Daviess
Me Henry
Marion
Lake
Lake
Me Henry
Lake
Shelby, Christian
McDonough
Coles
Schuyler
Scott, Morgan
McDonough
Warren
Brown
Marion
Mason
Pike
Lake, McHenry
(continued)
158
-------
TABLE 24. (continued)
Name
Serial Number
County
Powerton Cooling
Pinckneyville
Lower Smith
182
99
181
Tazewell
Perry
Scott
Class 5 (Cluster 4)
Spring
Charleston
Clear
Chautauqua
Anderson
Fourth Lake
Meredosia
Calumet
Chain
Matanzas
Sugar Creek
Old Ben Mine
Harrisburg
Snyder's Hunting
Griswold
Fyre
Sanganois
Crane
Long
Moscow
Yorkey
Pekin
Kinneman
Lily
Lyerla-Autumnal
Tampier
Otter
Sunfish
Liverpool
Bath
Swan
Sahara Coal Company
Goose (Village Club)
8
16
86
87
34
151
10
18
83
89
178
32
111
142
158
170
9
85
177
164
179
116
168
136
120
23
167
126
88
166
93
175
139
Carrol 1
Coles
Mason
Mason
Fulton
Lake
Cass, Morgan
Cook
Mason
Mason
Schuyler
Franklin
Saline
Jackson
MeHenry
Mercer
Cass
Mason
Schuyler
Mason
Schuyler
Tazewel1
Massac
Cass
Union
Cook
Mason
Whiteside
Mason
Mason
Mercer
Saline
Grudy
(continued)
159
-------
TABLE 24. (continued)
Name
Serial Number
County
Class 6 (Cluster 1)
Horseshoe
Worl ey
Cattail
Bakers
Keithsburg
Open Pond (Marshy)
Lake of the Woods
West Frankfort Old
Marion
Sam Parr
Greenville New City
Nippersink
Decatur
Long
Fox
Grass
Petite
Hoi id ay
Third
DePue
Jack, Swan, and Grass
Crooked
Carbondale
Rice
Marie
Spri ng
Senachwine
Mound
Class 7 (Cluster 2)
Spri ng
Skokie Logoons
Turner
Class 8 (Cluster 3)
Saganashkee
Marshall
1
117
125
19
92
174
115
30
129
41
2
63
73
52
60
61
64
51
155
3
165
149
38
35
62
118
104
163
4
22
102
21
81
Alexander
Tazewel 1
Whites ide
Cook
Mercer
Saline
Tazewel 1
Franklin
Wi 1 1 i amson
Jasper
Bond
Lake
Macon
Lake
Lake
Lake
Lake
LaSalle
Lake
Bureau
Mason
Lake
Jackson
Fulton
Lake
Tazewel 1
Putnam
Mason
Bureau
Cook
Putnam
Cook
Marshal 1
(continued)
160
-------
TABLE 24. (continued)
Name Serial Number County
DuQuoin 98 Perry
Goose 100 Putman, Bureau
Swan 172 Putnam
Class 9 (Cluster 11)
Wonder 69 McHenry
Class 10 (Cluster 12)
Sawmill 103 Putnam
Class 11 (Cluster 9)
McGinnis 20 Cook
LaRue-Pine Hills 121 Union
Class 12 (Cluster 10)
Slocum 54 Lake
Trophic class number 2 (Cluster 7)--Terse description: (high SEC, low
TON, TPHOS, and CHLA). Most of the 19 Takes in this group are artificial
reservoirs in northern and southern Illinois of good to excellent water
quality; these lakes exhibit little use impairment from sediment-related
turbidity, algae, or aquatic macrophytes as assessed by IEPA (1978a). The
mean depth for the cluster averages 5.7 m and the maximum depth averages
13.8 m. The average detention time for this group is 3.45 years. These
lakes are generally clear with very low suspended solids levels. The summer
1977 sampling by IEPA produced mean values of 3.7 mg/liter, 2.7 mg/liter, and
2.06 m for TSS, VSS, and SEC, respectively, for this group.
Trophic class number 3 (Cluster 5)—Terse description: (average SEC,
TON, and CHLA, slightly low TPHOS). This class contains southern and
northern reservoirs and glacial lakes of good quality with little use
impairment from sediment-related turbidity. Algal bloom problems are slight
with aquatic vegetation problems ranging in severity from slight to moderate.
161
-------
TABLE 25. INTERPRETATION OF CLUSTERS DEVELOPED FROM SET THREE REGRESSION MODELS*
Observed
Problems
NES Data
(1973)
IEPA Data
(1977)
Hydrology
Set Three Models
Cluster •£
Number |
and Perceived r:
(Obs.) Quality «
I 0)
O 4->
I- >>
3 2
Morphology
3
1
2
3
4
5
6
7
8
9
10
11
12
8 ( 1)
7 (19)
5 (18)
6 (33)
4 (33)
1 (28)
2 ( 3)
3 ( 5)
11 ( 1)
12 ( 1)
At V A /
9 ( 2)
y \ t. y
10 ( 1)
1.0
1.6
2.1
2.4
3.4
3.1
4.0
3.0
3.0
4.0
3.0
4.0
1.0
1.3
1.7
2.3
3.1
2.8
3.5
3.0
3.0
4.0
1.5
4.0
1.0
1.7
2.4
2.1
1.9
2.8
1.7
2.3
3.0
l n
3 5
4.0
2.0
2.4
2.6
2.0
1.7
1.9
1.7
2.0
3.0
1 0
4.0
1.0
- 1.97 0.01 0.13
- 1.83 0.03 0.03
1.23 27 0.10 1.08 0.95 0.07 0.19
0.58 21 0.12 0.92 1.05 0.06 0.18
0.25 18 0.21 1.20 0.31 0.25 1.94
0.37 49 0.37 1.59 0.74 0.48 0.79
- 0.25 1.17 0.10
- 0.28 0.42 0.13
0.46 198 0.42 1.79 0.31 0.61 0.16
0.31 241 1.33 5.94 0.34 1.50 0.57
1
5
10
19
55
32
37
62
49
25
0
4
8
8
22
19
22
24
16
21
264
1,022
4,689
3,290
35,648
5,106
13,824
1,280
2,247
30,465
8.7
5.7
3.0
3.1
1.6
2.0
.8
1.2
2.0
0.5
1.2
24.4
13.8
9.8
8.9
3.8
5.2
1.8
4.5
10.4
1.5
2.7
56.9
81.4
28.1
34.3
29.2
41.8
.
9.2
12.6
9.0
4.8
1.72
3.45
1.14
1.35
1.28
1.86
.
.39
.62
.45
.24
7.42
1.84
1.03
0.68
0.35
0.43
0.28
0.32
0.31
0.17
0.27
0.34
1
10
2b
17
23
42
42
74
158
143
109
229
0.41
0.82
1.22
0.89
1.20
1.52
2.12
1.91
2.07
3.00
4.69
5.12
0.02
0.11
0.2U
0.15
0.36
0.38
0.7b
0.51
0.47
0.91
1.60
1.43
01
ro
Parameter means were constructed for 12 clusters.
Parameters, acronyms, and units are as follows:
Trophic Class - based on cluster means of estimated values for SEC,
TON, TPHOS, and CHLA, Scaled 1 - 12 with 1 = least eutrophic and
12 = most eutrophic.
Perceived quality - Summary of quality based upon analysis of
information from a variety of sources. Scaled 1-4 with 1 =
very good and 4 = very poor.
Observed Problems - Problems noted through field observations and
adversely affecting use. Scaled 1-4 with 1 = minimal problem and
4 = severe problem.
NES data (1973) - Data collected by U.S. EPA's National Eutrophication
Survey In October 1973. SEC - (meters). CHLA - (Mg/liter). TPHOS -
mg/liter). TON - (mg/Hter).
IEPA Data (1977) - Data collected by Illinois Environmental Protection
Agency during summer of 1977. TSS - total suspended solids (rig/liter).
VSS - volatile suspended solids (mg/liter). SEC - Secchi depth
(meters). TPHOS - total phosphorus (mg/liter). IN - inorganic nitrogen
(mg/liter). Algae - algal enumeration (cells x 1000/ml).
Morphology - Data from IEPA files. 7 - mean depth (meters). Zm -
maximum depth (meters).
Hydrology - Data from IEPA files. Detention time (years).
Set Three Models - cluster means based on parameter estimates generated
from the third set of regression models. SEC - Secchi depth (meters).
CHLA - chlorophyll & (ug/Iiter). TON - total organic nitroyen
(mg/liter). TPHOS - total phosphorus (mg/liter).
-------
Mean depths for the group average 3.0 m, while maximum depths average 9.8 m.
Water detention time averages 1.14 years for the group. According to IEPA
1977 summer sampling data, the average amount of total and volatile suspended
matter is three times that of trophic class 2, resulting in a 50 percent
decrease in Secchi transparency.
Trophic class number 4 (Cluster 6)--Terse description: (low SEC, TON,
TPHOS, and CHLA).The 33 water bodies in this category include reservoirs
from throughout the State, glacial lakes, and backwaters, ranging in water
quality from fair to good. Sediment-related turbidity is a slight to
moderate problem. Algal blooms are a slight problem. Aquatic macrophyte
problems are minimal. The mean depth of the cluster members is 3.1 m, while
the maximum depth averages 8.9 m. Average water detention time is 1.35
years. Average water transparency and volatile suspended solids
concentrations for this group in 1977 were approximately the same as trophic
class 3, while the inorganic suspensoid level is three times that of the
previous class.
Trophic class number 5 (Cluster 4)--Terse description: (low SEC,
slightly low TON and CHLA, average TPHOS). Most of the 20 water bodies
included in trophic class 5 are river backwaters or bottom land lakes that
are shallow and turbid. Mean depth for the group averages 1.6 m, while
maximum depth averages 3.8 m. Most are of poor to fair quality with severe
sediment-related turbidity problems. Algal problems are slight and aquatic
vegetation is minimal, largely a consequence of the light inhibition by
suspended sediment. Summer 1977 sampling data for total and volatile
suspended solids show a three-fold increase over the previous trophic class,
resulting in a three-fold reduction in Secchi transparency.
Trophic class number 6 (Cluster 1)—Terse description: (low SEC, average
TON, TPHOS, and CHLA). This class is represented by river backwaters,
reservoirs located throughout the State, and glacial lakes. Most of the
water bodies are of fair quality with sediment-related turbidity problems
ranging from slight to severe. Algal bloom problems range in severity from
slight to moderate. Use impairment from aquatic macrophytes is minimal.
Summer 1977 sampling results for total and volatile suspended solids and
Secchi depth for this class fall between the average values for trophic
classes 4 and 5.
Trophic class number 7 (Cluster 2)—Terse description: (low SEC,
moderately high TON and TPHOS, average CHLA). The three water bodies in this
class represent a glacial lake, a former sewage lagoon, and a river
backwater. These shallow water bodies are of poor quality. Severe
sediment-related turbidity and algal bloom problems exist. Vegetation
problems are minimal. This class, as well as the remaining five trophic
classes, had high suspended solids levels and low Secchi transparency in the
lEPA's 1977 summer sampling.
Trophic class number 8 (Cluster 3)--Terse description: (low SEC,
moderately high TON and TPHOS, very high CHLA). This group consists of three
river backwaters, a northern slough, and a southern reservoir. As presently
163
-------
constituted, quality of these shallow, short-detention water bodies ranges
from poor to fair. Sediment-related turbidity ranges from minimal to
moderate. Algal blooms and aquatic vegetation present slight use impairment
problems.
Trophic class number 9 (Cluster 11)—Terse Description: (low SEC,
moderately high TON and TPHOS, very high CHLA). The single water body in
this class is a shallow, short-detention, northern reservoir, Wonder Lake
(serial number 69). Sediment-related turbidity, algal blooms, and aquatic
macrophytes all present moderate use impairment problems for this lake.
Trophic class number 10 (Cluster 12)—Terse description: (low SEC, high
TON and TPHOS, very high CHLA).Sawmill Lake (serial number 103) is a river
backwater of poor quality as evidenced by severe use impairment problems from
sediment-related turbidity and algal blooms. Aquatic vegetation problems are
minimal.
Trophic class number 11 (Cluster 9)—Terse description: (low SEC, very
high TON and TPHOS, moderately high CHLA). This group consists of La
Rue-Pine Hills Swamp (serial number 121), in southern Illinois, and McGinnis
Slough (serial number 20), in northern Illinois. Quality of these shallow
water bodies is rated as fair with moderate to severe algal blooms, moderate
aquatic macrophyte problems, and minimal to slight sediment-related
turbidity.
Trophic class number 12 (Cluster 10)—Terse description: (low SEC, very
high TON, TPHOS, and CHLA). Slocum Lake (serial number 54) is a shallow,
short-detention, northern Illinois reservoir noted for its poor quality.
Severe sediment-related turbidity problems and algal blooms are known to
exist. Aquatic macrophytes present a minimal problem. Summer 1977 sampling
results indicate that practically all of the turbidity is caused by organic
particulate matter.
GENERAL DISCUSSION
Although LANDSAT had been used successfully several times in studies of
lake eutrophication, and statistically significant relationships between MSS
data and lake water quality parameters have been demonstrated, it was not at
all certain that such relationships or correlations would eventually be
obtained for Illinois lakes. While many of the lakes included in LANDSAT
investigations (e.g., Boland 1976) were characterized by low suspended solids
and color, Secchi depths greater than several meters, and low to high
nutrient concentrations, most Illinois lakes are characterized as turbid with
high sediment loads, low mean morphometric and Secchi depths, very high
nutrient levels, and short hydraulic retention times. Indeed, whereas many
of the lakes studied previously with LANDSAT have primary production that is
nutrient limited, most Illinois lakes are nutrient rich, but light poor, as a
result of turbidity from suspended inorganic and organic particulate matter.
A major consequence of sediment pollution in Illinois lakes, as it
relates to MSS data, is that the majority of observed chemical values are
represented within a small range. Further, some of the lakes have extreme
164
-------
levels for some contact-sensed parameters, which are not readily measured
with the multispectral scanner. Lake parameters operate over a very large
dynamic range in comparison with the range of the MSS. Consequently,
analysis of MSS data and lake properties is more sensitive to changes in
midrange values. In spite of this loss in statistical sensitivity, MSS data
were used to cluster lakes into physically significant groups and to estimate
trophic parameters from which clusters were developed.
It is not surprising, therefore, that Set One models, employing raw MSS
data to estimate trophic index parameters, are the least reliable, while Set
Three models, employing an expansion that makes the MSS data uniform over its
entire range, are best. That is, as the concentration increases at a given
wavelength, the optical density (reflectance) falls off nonlinearly, with the
slope rising fastest at low densities and falling off to near zero at high
densities; the optical density scale is logarithmic. The technique used in
developing Set Three models, of converting optical densities to ranks,
linearizes the slope over the entire range of lakes. Provided that the
optical density remains single-valued in the tails and that some
discrimination remains, conversion to ranks should be satisfactory, as has
been observed.
The application of clustering algorithms to the raw LANDSAT multispectral
data proved to be a viable approach, resulting in lake clusters with their
own spectral properties and physically significant characteristics. It
permited the selective comparison of four band readings, allowing an
objective differentiation of water types according to signatures based on a
spectral index. In general, increases in spectral responses in any MSS band
or combination of bands are associated with greater suspended or soluble
material levels and poorer water quality. Lakes having similar spectral band
DN readings are closely related in physical and chemical quality. Although
actual concentrations of non-sensible constituents are difficult to
accurately predict from the LANDSAT MSS data, the spectral responses of the
lakes demonstrate general trends with the contact-sensed data. The
application of a clustering algorithm to the spectral data in an unsupervised
mode has much intuitive appeal; it is simple and deals directly with the
spectral characteristics and physical water properties of the lakes.
The application of clustering algorithms to lake attributes whose values
have been estimated through regression models incorporating the LANDSAT MSS
bands as independent variables is attractive because a statistically
significant relationship has been demonstrated to exist between the contact-
sensed and remotely sensed data. However, at least in this study,
superimposed on the estimated values of the attributes is a substantial
amount of unexplained variation. Clustering on such data sets results in
groups or clusters whose memberships become increasingly suspect as the
unexplained variation associated with the regression modeling efforts
increases. Conversely, the more adequate the regression models, the more
confidence can be placed in the membership of the individual cluster.
Whether clustering on spectral attributes (e.g., LANDSAT bands) or
trophic attributes (e.g., trophic indicators estimated by regression on
165
-------
LANDSAT bands), the type of clustering method selected will, to some
undefined degree, mold the character of the final output. We know of no
systematic study that has attempted or is attempting to compare lake
classifications derived from different SAHN methods. Standardization and
transformation techniques also affect the resulting classifications; this is
another area in which little systematic work has been done using lakes.
While it is evident in light of the above discussion that each clustering
method has a bias, the clusters developed in this study can be interpreted in
"real world" terms. It is clear from our analyses that Illinois lakes
represent a much greater diversity of waters than is generally recognized.
Further, although a great many of these are "polluted" in the sense of high
nutrient enrichment, it is also true that some are "clean water" waters, with
aesthetically pleasing qualities. Further, Illinois lakes offer the
recreationist diversified opportunities ranging from canoeing, boating, and
swimming, to fishing, trapping, bird watching, and wildlife observation.
More pragmatically, cursory comparisons of the 1973 and 1977 data bases
show few general changes, either for the better or worse, in these
impoundments. Since lakes are dynamic, and Illinois lakes with their
relatively short detention times particularly so, it is suggested that lake
hydraulic and morphologic factors and land use factors that determine lake
water quality, in part, have remained relatively unchanged during the period.
Stream water quality, by contrast, has improved overall since 1973 (Hudson et
al. 1978).
There are several advantages to using LANDSAT in place of conventional
lake monitoring. These include cost, resource requirements, timeliness of
the data, uniformity, objectivity, and flexibility of data formats.
The costs of purchasing the CCT's, extracting the lake pixels and
associated processing, and preparing statistical and photographic displays
for the group of approximately 150 lakes averaged under $200 per lake. A
substantial portion of the cost can be attributed to the batch operation mode
used to extract most of the lake MSS data. It was very time consuming and
laborious, requiring the expenditure of many manhours of effort. With the
current availability of the interactive programs described earlier in this
report, substantial reductions in extraction costs occur, accompanied by a
major reduction in turnaround time. A supplemental LANDSAT study of 60
Illinois lakes cost about $80 per lake for MSS processing. Total costs will
probably not exceed $250 per lake.
If relative lake information is required, standard, low-cost, readily
available, statistical computer programs can be used to analyze the spectral
data. If trophic indicator information is required, some contact-sensed
information would have to be acquired at the time of the satellite's passage
over the State. Using the methods described in this report, estimates of
average lake water quality could be obtained. We estimate the total per lake
cost of the present project at $500, and a total project cost of $75,000,
including preparation of this report.
166
-------
In contrast to this, the 1977 lake study involving 108 lakes cost $36,600
for sampling 3 sites per lake and lab analyses. To this must be added the
costs for the full-time limnologist, data processing and analyses, and report
preparation, which are at least equal to the sampling costs. Therefore, the
cost per lake for the conventional approach is at least one-third higher than
the LANDSAT approach. However, a word of caution is in order when examining
such comparisons. While both approaches (LANDSAT, conventional field
program) have the same general objective -- the classification of lakes
according to trophic state -- the quality and types of products generated,
the parameters measured and their associated accuracies and precisions, and
temporal efficiencies are generally not the same. In addition, this was a
prototype study for the Illinois Environmental Protection Agency. As such,
it had a "learning curve" associated with it, lacked the streamlining
characteristic of operational programs, and, therefore, incurred costs in
time and money that are above those for a program functioning at an optimal
level.
This study strongly supports the value and cost effectiveness of LANDSAT
for lake monitoring. When this is coupled with the sophistication available
in false-color imagery, many specific questions about the instantaneous or
dynamic conditions of selected lakes can be examined. This project has
served to increase our understanding of a substantial number of Illinois
water bodies. For the first time a quantitative, geographically
comprehensive understanding of these bodies is available. As our
understanding improves, rational decisions concerning the uses and fates of
these lakes can be made, costs determined, programs implemented, and benefits
accrued and assessed.
Problems
By virtue of its repetitive coverage, synoptic overview, and ability to
generate permanent records amenable to automated image-processing techniques,
the LANDSAT MSS is attractive for purposes of environmental assessment and
monitoring. In this study LANDSAT provided a view of the past; it does
provide a monitoring strategy that is objective, uniform, frequent, resource
tolerant, and cost effective for the future. However, LANDSAT is not a
panacea; it has limitations and problems associated with. it. A number of
problems were encountered during the study. Several were solved; others, by
virtue of their nature, defied solution. Not all are mutually exclusive.
The specific problems are discussed in the following subsections.
Availability of Imagery--
Ten LANDSAT MSS scenes are required for complete areal coverage of
Illinois. This necessitates that the satellite make three passes over the
State, once on each of three paths (Figure 19). In 1973, only LANDSAT-1
coverage was available for the Illinois lakes. Excluding side overlap, the
satellite provided coverage on a repetitive 18-day basis. It was our
intention to use MSS data acquired concurrently with one of the three NES
sampling rounds, preferably the summer round. The occurence of substantial
areas of cloud cover in numerous LANDSAT scenes acquired between April and
November of 1973 restricted us to the scenes of October 14, 15, and 16, 19/J.
167
-------
Cloud cover will continue to plague water and terrestrial studies dependent
upon LANDSAT data. Unlike aircraft, LANDSAT is locked into a fixed flyover
schedule. Although the problem of obtaining good, cloud-free imagery has
been reduced by the presence of two currently operational satellites
(LANDSAT-2 and LANDSAT-3), it is still a problem of substantial magnitude.
The CCT's for the 10 Illinois scenes, not available at the EROS data
Center, had to be generated at NASA-Goddard. This process took several
months, with the CCT's for only eight scenes being generated. The long delay
in delivery was a consequence of several factors:
1. NASA-Goddard software changes. The generation of our "historic"
CCT's from the high-density tapes required that NASA replace its
currently operational software with that developed at the start
of the LANDSAT program. Difficulties were encountered as a result
of hardware and programmatic changes.
2. Deterioration of high density tapes. Although only about three
years old, some of the high-density tapes had degraded in quality
to the point where reading them became difficult or impossible.
3. Unavailability of spacecraft data. The scene data for scene
1449-16084 were available, but the spacecraft data necessary
to calibrate the scene could not be found.
One of the scenes (1449-16084), arrived without the proper internal
calibration data. Much time and effort was expended in attempts to calibrate
and use the data from this scene. The black-and-white imagery available from
EROS for the 10 scenes (Figures 20, 22-30) was originally generated at
NASA-Goddard using an electron beam recorder and data on the density tapes.
If the difficulties experienced in having "historic" CCT's generated from
high density tapes is typical of that experienced by other investigators, the
question to ponder becomes: "What will the value of the archived tapes be a
decade from now?"
Availability of Interactive Image-Processing System--
At the start of the project, we were limited to an image-processing
system that operated in a batch mode (see Section 6). By virtue of its
nature, the process of extracting the data for about 150 lakes (many
appearing in two or more scenes) using the batch mode is inefficient both
timewise and costwise. The development and refinement of the interactive
image-processing system (LAKELOC) has greatly speeded up the extraction
process. What initially took hours now requires but a few minutes.
Placement at the system's controls of an individual knowledgeable about the
lakes of Illinois could improve its efficiency even more.
Environmental Factors--
It is well recognized that the atmosphere can have substantial effects on
the signal returning from lakes. While the MSS data were adjusted to a
168
-------
common date, October 15, no concentrated effort was mad? to remove
atmospheric effects through the use of theoretical modeling or empirical
modeling involving terrestrial or aquatic calibration points. An unknown
amount of variation, attributable to the variation in intensity of
atmospheric effects across the LANDSAT scenes, exists in the MSS lake data.
During the period of October 11-14, 1973, much of Illinois experienced
moderate rains (Figure 64). The amount of precipitation appears to be normal
for this season. It is very likely that many of the lakes experienced large
influxes of suspended sediments. In some cases, a substantial proportion of
each lake's volume was displaced by the incoming waters. The combination of
inorganic sediment influx and volume displacement did not disrupt the average
characteristics of the water bodies if comparisons of the NES data for the
spring and summer of 1973 with 1577 data are representative of the true
quality.
At the time of satellite flyover (October 14-16), the solar angle was
small, measuring between 35 and 38 degrees. This effectively reduces the
amount of light entering the water body, and consequently the intensity of
the return signal is less than at higher solar angles. This makes the
spectral distinction between lake types a more difficult task because there
are fewer DN levels.
The season during which the spectral data are collected has some
relationship to the degree of success experienced by LANDSAT-oriented lake
classification projects. As suggested by Rogers (1977) and others, success
is most likely when the secondary manifestations of eutrophication (e.g.,
turbidity, algal blooms, macrophyte beds) are most evident. In Illinois
water bodies, July and August are probably the best period for attempting
satellite-based classifications. Unfortunately, this project was limited by
cloud cover and NES sampling dates to the middle of October.
Contact-Sensed Data--
The NES data were collected without any thought being given to their use
in a satellite-related lake classification project. Hence, sample site
selections were not always optimal in location or number for use with LANDSAT
data. In addition, field notes regarding phenomena of special interest
(e.g., tubidity plumes, macrophyte beds) were not adequately documented for
purposes of this project. Some parameters of particular interest to us
(e.g., total suspended solids, volatile suspended solids) were not measured
by the NES.
LANDSAT MSS—
The LANDSAT multispectral scanner was not designed specifically with
lakes in mind. It is, both spectrally and spatially speaking, a low-
resolution sensor. While its gain can be increased in the GRN and RED bands,
this project was restricted to the normal gain settings. The "blocky"
appearance often evident along the land-water interface can be disconserting
to the neophyte user of LANDSAT imagery. Even more distracting, as well as
detrimental, is the "sixth line" banding or striping so apparent in MSS
169
-------
imagery of water bodies. Although the trained eye can overlook the striping,
it is distracting. More importantly, it represents data values that increase
the variance, thereby adding to the uncertainty of satellite-derived trophic
indicator estimates and lake classification products (e.g., thematic
photomaps). Attempts to destrip MSS data for water bodies have only been
moderately successful.
Multivariate Trophic Indices—
In Section 6 the use of principal components analysis was demonstrated as
a means of developing two multivariate trophic state indices (i.e., PC1F5,
PC1Y5). The concept has been used previously to rank lakes (e.g., Boland
1976, Brezonik and Shannon 1971, Sheldon 1972). We suggest that a word of
caution is in order when examining the resultant ordinations of Illinois
lakes. Many Illinois water bodies are turbid primarily because of inorganic
suspended solids that drastically reduce light levels and affect algal and
macrophyte productivity. Still other Illinois lakes have turbidity problems
caused by algae. It may be more proper to segregate the water bodies into
two or more basic groups and then apply the principal components analysis or
some other ordination technique. This is one area that should be explored
further.
Illinois Lakes Data Base--
Prior to the commencement of this project, the IEPA had assembled very
little information on the chemical and biological quality of the State's
lakes. With the exception of some scattered reports that were available from
other State or Federal agencies on specific lakes, no comprehensive inventory
of lake data existed. This information was required to process the MSS tapes
and to interpret and evaluate the classification results. The task of
assembling and verifying the necessary data was laborious and took several
months. Furthermore, quality control and comparability problems arise when
using data from several different sources.
At the onset of this project, the IEPA had no formal program for the
monitoring and classification of lakes. While much information was initially
available from numerous individuals within the Agency, it took a substantial
amount of time to consolidate Agency expertise on the condition and nature of
Illinois water bodies.
Model Development—
The development of the regression models was no trivial task. The
difficulty was caused, in part, by the nonnormality of the ground truth and
MSS data, the very limited range of DN levels in the four LANDSAT-1 MSS
bands, and the receipt of faulty MSS data for scene 1449-16084. The models
are specific to the set of scenes from which they were developed. It is
unlikely that they will yield satisfactory results if applied to other
LANDSAT scenes. However, the overall approach is applicable to lakes in
other geographic areas and to satellite MSS data collected at other times.
170
-------
Whether employing regression analysis or clustering techniques, the use
of average DN values for each spectral band for a lake does not account for
the spectral variability for a specific portion of a lake and its associated
trophic condition. Thus, lake characterization using average spectral
responses affords, at best, the average lake response but provides no
information about its variability. This information is revealed through the
analysis of band DN histograms, or by optical techniques that convert the
differing DN values into a false-color image. Information regarding
variability within lakes was not obtained in this study.
171
-------
REFERENCES
Anderberg, M. R. 1973. Cluster Analysis for Applications. New York:
Academic Press. 373 pp.
Anderson, T. W. 1958. An Introduction to Multivariate Statistical Analysis.
New York: John Wiley and Sons. 374 pp.
Army Corps of Engineers. 1969. Fluvial Sediment in the Upper Mississippi
River Basin: Upper Mississippi River Comprehensive Basin Study.
Appendix 6. (Draft Number 2).
Atkins, W. R. G., and H. H. Poole. 1952. An experimental study of the
scattering of light by natural waters. Proceedings of the Royal Society
(London). Series B. 140:321-338.
Barr, A. J., J. H. Goodnight, J. P. Sail, and J. T. Helwig. 1976. A User's
Guide to SAS 76. Raleigh, North Carolina: Sparks Press. 329 pp.
Beeton, A. M. 1965. Eutrophication of the St. Lawrence Great Lakes.
Limnology and Oceanography. 10(2):240-254.
Bethel, J. P. (ed). 1969. Webster's Geographical Dictionary. Springfield,
Massachusetts: G. and C. Merriam Company. 1293 pp.
Bennett, G. W. 1960. Sport fishing on lakes. Atlas of Illinois Resources.
III. Forest, Wildlife, and Recreation. Springfield, Illinois:
Department of Registration and Education. 47 pp.
Birge, E. A., and C. Juday. 1929. Transmission of solar radiation by the
water of inland lakes. Transactions Wisconsin Academy Science Arts
Letters. 24:509-580.
Birge, E. A., and C. Juday. 1930. A second report on solar radiation and
inland lakes. Transactions Wisconsin Academy Science Arts Letters.
25:285-335.
Birge, E. A., and C. Juday. 1931. A third report on solar radiation and
inland lakes. Transactions Wisconsin Academy Science Arts Letters.
26:383-425.
Birge, E. A., and C. Juday. 1932. Solar radiation and inland lakes, fourth
report. Transactions Wisconsin Academy Science Arts Letters.
27:523-562.
172
-------
Blackwell, R. J., and D. H. P. Boland. 1975. The trophic classification of
lakes using ERTS multispectral scanner data. Proceedings American
Society of Photogrammetry 41st Annual Meeting. March 9-14, 1975.
Washington, D. C.
Blackwell, R. J., and D. H. P. Boland. 1979. Trophic Classification of
Selected Colorado Lakes: Lake Classification through the Amalgamation of
Contact-Sensed Data and Digitally Processed Multispectral Scanner Data
Acquired by Satellite and Aircraft. JPL Publication Number 78-100. EPA
Publication Number EPA-600/4-79-005. Joint Publication of Jet Propulsion
Laboratory, Pasadena, California and U.S. Environmental Protection
Agency, Las Vegas, Nevada. 223 pp.
Boesch, D. F. 1977. Application of Numerical Classification in Ecological
Investigations of Water Pollution. U.S. EPA Ecological Research Series
Report EPA-600/3-77-033. 125 pp.
Boland, D. H. P. 1976. Trophic Classification of Lakes Using LANDSAT-1
(ERTS-1) Multispectral Scanner Data. U.S. EPA Ecological Research Series
Report. EPA-600/3-76-037. 263 pp.
Boland, D. H. P., and R. J. Blackwell. 1975. The LANDSAT-1 multispectral
scanner as a tool in the classification of inland lakes. Proceedings
NASA Earth Resources Survey Symposium. First Comprehensive Symposium
on the Practical Application of Earth Resources Survey Data. June 9-12,
1975. Shamrock Hilton. Houston, Texas. NASA TM X-58168. JSC-09930.
I-A:419-442.
Boland, D. H. P., and R. J. Blackwell. 1978. Trophic classification of
Colorado lakes utilizing contact data, LANDSAT and aircraft-acquired
multispectral scanner data. Proceedings 4th Joint Conference on
Sensing of Environmental Pollutants. November 6-11, 1977. New
Orleans Hilton. Washington, D.C.: American Chemical Society.
pp:888-894.
Bradbury, J. C., and E. Atherton. 1965. The Precambrian basement of
Illinois. Illinois State Geological Survey Circular Number 382.
Champaign-Urbana, Illinois. 13 pp.
Brezonik, P. L., and E. E. Shannon. 1971. Trophic State of Lakes in North
Central Florida. Publication 13. Florida Water Resources Research
Center. Gainesville: University of Florida. 102 pp.
Clifford, H. T., and W. Stephenson. 1975. An Introduction to Numerical
Classification. New York: Academic Press. 229 pp.
Collins, J. R. 1925. Change in the infrared absorption spectrum of water
with temperature. Physics Review. 26:771.
173
-------
Curran, R. J. 1972. Ocean color determination through a scattering
atmosphere. Applied Optics. 11(8):1857-1866.
Dagnelie, P. 1971. Some ideas on the use of multivariate statistical
methods in ecology, pp. 167-180. In: G. P. Patil, E. C. Pielou, and
W. E. Waters (eds). Statistical Ecology. Vol. 3. University Park,
Pennsylvania: Pennsylvania State University Press.
Davis, F. J. 1941. Surface loss of solar and sky radiation by inland lakes.
Transactions Wisconsin Academy Science Arts Letters. 33:83-93.
Dawson, L. H., and E. 0. Hulburt. 1937. The scattering of light by water.
Journal Optical Society America. 27:199-201.
Donaldson, J. R. 1969. The classification of lakes. In: Proceedings of
the Eutrophication-Biostimulation Assessment Workshop. June 19-21, 1969.
Sponsored by University of California (Berkeley) and U.S. Environmental
Protection Agency, pp. 171-185.
Dunn, 0. J., and V. A. Clark. 1972. Applied Statistics: Analysis of
Variance and Regression. New York: John Wiley and Sons. 387 pp.
Duntley, S. Q. 1963. Light in the sea. Journal Optical Society America.
53:214-233.
Edmondson, W. T. 1974. Review of Environmental phosphorus handbook. E. J.
Griffith, A. Beeton, J. M. Spencer, and D. T. Mitchell (eds). New York:
Wiley-Interscience. 1973. 718 pp. In: Limnology and Oceanography.
19(2):369-375.
Environmental Data Service. 1973. Climatological Data: Illinois. October
1973. Asheville, North Carolina: U.S. Department of Commerce.
Ewan, T. 1894. On the absorption spectra of dilute solutions. Proceedings
Royal Society (London). 57:117-161.
Fenneman, N. M. 1928. Physiographic divisions of the United States.
Annals Association of American Geographers. 18:264-353.
Fenneman, N. M. 1938. Physiography of Eastern United States. New York:
McGraw-Hill. 691 pp.
Fehrenbacher, J. B., G. 0. Walker, and H. L. Wascher. 1967. Soils of
Illinois. University of Illinois College of Agriculture. Bulletin
Number 725. Champaign-Urbana, Illinois. 47pp.
174
-------
Fisher, L. T., and F. L. Scarpace. 1975. Trophic status of inland lakes
from LANDSAT. Proceedings NASA Resources Survey Symposium. First
Comprehensive Symposium on the Practical Application of Earth Resources
Survey Data. June 9-12, 1975. Shamrock Hilton. Houston, Texas. NASA
TM X-58168. JSC-09930. I-A:443-451.
Flemal, R. C. 1972. Pleistocene geology between Ue Kalb and Danville,
Illinois. Geology Guidebook Series Number 3. Department of Geology.
De Kalb, Illinois: Northern Illinois University. 48 pp.
Fruh, E. G., K. M. Stewart, G. F. Lee, and G. A. Rohlich. 1966.
Measurements of eutrophication and trends. Water Pollution Control
Federation Journal. 38(8):1237-1258.
Gerd, W. V. 1957. An attempt at the biolimnological regionalization of the
USSR. Transactions of the Sixth Conference on the Biology of Inland
Waters. Edited by E.N. Pavlovskii. Original in Russian. Israel Program
for Scientific Translations. Jerusalem: IPST Press. 1969. pp. 138-
145.
Gunning, G. E. 1966. Illinois. Limnology in North America. D.G. Frey
(ed). Madison: University of Wisconsin Press. 734 pp.
Guthrie, D., C. Avery, and K. Avery. 1973. Statistical Interactive
Programming System (*SIPS) User's Reference Manual. Technical Report
Number 36. Department of Statistics. Corvallis: Oregon State
University. Ill pp.
Hansen, K. 1962. The dystrophic lake type. Hydrobiologia. 19(2):182-191.
Harding, S. T. 1942. Lakes. Hydrology. 0. E. Meinzer (ed). New York:
McGraw-Hill. 712 pp.
Hasler, A. D. 1947. Eutrophication of lakes by domestic drainage. Ecology.
28:383-395.
Hasler, A. D., and B. Ingersoll: 1968. Dwindling lakes. Natural History.
77(9):8-19
Hocking, R. R. 1976. The analysis and selection of variables in linear
regression. Biometrics. 32:1-49.
Horowitz, H. M., R. F. Nalepka, P. D. Hyde, and J. P. Morgenstern. 1971.
Estimating the proportions of objects within a single resoluton element
of a multispectral scanner. Proceedings 7th International Symposium on
Remote Sensing. May 17-21, 1971. Ann Arbor, Michigan. 2:1307-1320.
Hotelling, H. 1933a. Analysis of a complex of statistical variables into
principal components (I. Introduction). The Journal of Educational
Psychology. 24:417-441.
175
-------
Hotelling, H. 19335. Analysis of a complex of statistical variables into
principal components (II). The Journal of Educational Psychology.
24:498-520.
Hotelling, H. 1936. Simplified calculation of principal components.
Psychometrika. l(l):27-35.
Hudson, L. D., D. J. Schaeffer, W. J. Tucker, and W. H. Ettinger. 1978. The
1976 Illinois drought: Evidence for improved water quality. Journal of
Environmental Management. (In Press).
Hulburt, E. 0. 1945. Optics of distilled and natural water. Journal
Optical Society America. 35(11):698-705.
Hulstrom, R. L. 1973. The cloud bright spot. Photogrammetric Engineering.
39(4):370-376.
Hutchinson, G. E. 1957. A Treatise on Limnology. Volume I. Geography,
Physics, and Chemistry. New York: John Wiley. 1015 pp.
Hutchinson, G. E. 1967. A Treatise on Limnology. Volume II. Introduction
to Lake Biology and th'e Limnoplankton. New York: John Wiley. 1115 pp.
Hutchinson, G. E. 1973. Eutrophication: The scientific background of a
contemporary practical problem. American Scientist. 61:269-279.
IDOC. 1972. Illinois Surface Water Inventory. Special Fisheries Report No.
40. State of Illinois Department of Conservation. Division of
Fisheries. Springfield, Illinois.
IEPA. 1975. State of Illinois Environmental Protection Agency Water
Pollution Control Program Plan for FY-1976. State of Illinois
Environmental Protection Agency. Springfield, Illinois.
IEPA. 1976. Phase 1 of the Water Quality Management Basin Plan for the Big
Muddy River Basin. Volume 1. Existing Water Quality. State of Illinois
Environmental Protection Agency. Springfield, Illinois.
IEPA. 1978a. Assessment and Classification of Illinois Lakes. 208 Water
Quality Management Planning Program Staff Report. State of Illinois
Environmental Protection Agency. Springfield, Illinois.
IEPA. 1978b. Clean Lakes Strategy for Illinois. 208 Water Quality
Management Planning Progress Staff Report. State of Illinois
Environmental Protection Agency. Springfield, Illinois.
Illinois Cooperative Crop Reporting Service. 1974. Illinois Agricultural
Statistics Annual Summary 1974. Bulletin 74-1. In cooperation with the
Illinois Department of Agriculture and the U.S. Department of
Agriculture. 105 pp.
176
-------
Ivimey-Cook, R. B., M. C. F. Proctor, and D. L. Wigston. 1969. On the
problem of the R/Q terminology in multivariate analyses of biological
data. Journal of Ecology. 57:673-675.
James, H. R., and E. A. Birge. 1938. A laboratory study of the absorption
of light by lake waters. Transactions Wisconsin Academy Science Arts
Letters. 31:1-154.
Jarnefelt, H. 1958. On the typology of northern lakes. In: Proceedings
International Association Theoretical Applied Limnology. Volume XIII.
Congress in Finland 1956. T. T. Macan (ed). pp. 228-235.
Jerlov, N. G. 1968. Optical Oceanography. Elsevier Oceanography Series.
Volume 5. Amsterdam: Elsevier. 194 pp.
Johnson, S. C. 1967. Hierarchical clustering schemes. Psychometrika.
32(3):241-254.
Juday, C., and E. A. Birge. 1933. The transparency, the color, and the
specific conductance of the lake water of northeastern Wisconsin.
Transactions Wisconsin Academy Science Arts Letters. 28:205-259.
Keniston, J. A. 1978. Program Cluster. An Aid to Numerical Classification.
Newport, Oregon: Oregon State University Marine Science Center
(unpublished document). 32 pp.
Kofoid, C. A. 1903. Plankton studies. IV. The plankton of the Illinois
River 1894-99, with introductory notes upon the hydrography of the
Illinois River and its basin. I. Quantitative investigations and
general results. Illinois Laboratory Natural History Bulletin.
6(2):95-635.
Larkin, P. A., and T. G. Northcote. 1958. Factors in lake typology in
British Columbia, Canada. Proceedings International Association
Theoretical Applied Limnology (Verh. Internat. Ver. Limnol.). 13:252-263.
Leighton, M. M., G. E. Ekblaw, and L. Horberg. 1948. Physiographic
divisions of Illinois. Illinois State Geological Survey, Report of
Investigation. 129:16-33.
Lindeman, R. L. 1942. The trophic-dynamic aspect of ecology. Ecology.
23(4):399-418.
Lueschow, L. A., J. M. Helm, D. R. Winter, and G. W. Karl. 1970. Trophic
nature of selected Wisconsin lakes. Wisconsin Academy Sciences Arts
Letters. 58:237-264.
McCarthy, J. T. 1972. Alkalinity, Sulfate, and Chloride Production in the
Streams of Illinois. Unpublished M.S. Thesis. Department of Geology.
De Kalb, Illinois: Northern Illinois University. 89 pp.
177
-------
McCloy, K. R. 1977. The vector classifier. Proceedings llth International
Symposium of Remote Sensing of Environment. April 25-29, 1977. Ann
Arbor, Michigan: Environmental Research Institute of Michigan.
1:535-543.
McCluney, W. R. 1976. Remote measurement of water color. Remote Sensing of
Environment. 5:3-33.
McDannald, A. H. (ed). 1952. The New Modern Encyclopedia: A Library of
World Knowledge. New York: W. H. Wise and Company. 1494 pp.
Margalef, R. 1958. "Trophic" typology versus biotic typology as exemplified
in the regional limnology of northern Spain. Proceedings International
Association Theoretical Applied Limnology. 13:339-349.
Moore, D., and L. Haertel. 1975. Remote Sensing of Water Quality in Prairie
Lakes. Completion Report. Project Number B-039-SDAK. Brookings, South
Dakota: South Dakota State University Remote Sensing Institute. 61 pp.
Morrison, D. F. 1967. Multivariate Statistical Methods. New York: McGraw-
Hill. 338 pp.
Moyle, J. B. 1945. Classification of lake waters upon the basis of
hardness. Proceedings Minnesota Academy Science. 13:8-12.
Moyle, J. B. 1946. Some indices of lake productivity. Transactions of
the American Fisheries Society. 76:322-334.
NASA. 1972. NASA Earth Resources Technology Satellite Data Users Handbook.
National Aeronautics and Space Administration. Goddard Space Flight
Center. Document Number 71SD4249. Subject to revision.
Naumann, E. 1919. Na'gra Synpunkte angaende planktons okologi. Med.
sarskild hansyn till fytoplankton. Svensk Botanisk Tidskrift.
13:129-158.
Naumann, E. 1931. Limnologische terminologie. E. Adberhalden: Handbuch
der biologischen Arbeitsmethoden. Berlin and Wein, Urban and
Schwarzenberg, Abt IX, Teil 8:776 pp.
O'Donnell, D. J. 1935. Annotated list of the fishes of Illinois. Illinois
Natural History Survey Bulletin. 20(5):473-500.
Odum, E. P. 1959. Fundamentals of Ecology. Philadelphia: W. B. Saunders
Company. 546 pp.
Orloci, L. 1975. Multivariate Analysis in Vegetation Research. The Hague:
Junk. 276 pp.
Page, J. L. 1949. Climate of Illinois. Urbana: University of Illinois
Press. 276 pp.
178
-------
Piech, K. R., and J. E. Walker. 1971. Aerial color analyses of water
quality. Journal Survey Mapping Division. Proceedings American Society
Civil Engineers. SU2(8507):185-197.
Pielou, E. C. 1969. An Introduction to Mathematical Ecology. New York:
Wiley-Interscience. 286 pp.
Pennak, R. W. 1958. Regional lake typology in northern Colorado, U.S.A.
Proceedings International Association Theroetical Applied Limnology.
13:262-283.
Raman, C. V. 1922. On the molecular scattering of light in water and the
colour of the sea. Proceedings Royal Society (London). Ser. A. 101:
64-79.
Rand McNally. 1971. Rand McNally Road Atlas. United States, Canada,
Mexico. Chicago: Rand McNally. 118 pp.
Rawson, D. S. 1956. Algal indicators of trophic lake types. Limnology and
Oceanography. 1(1):18-25.
Rawson, D. S. 1960. A limnological comparison of twelve large lakes in
northern Saskatchewan. Limnology and Oceanography. 5(2):195-211.
Rogers, R. H., and K. Peacock. 1973. A technique for correcting ERTS data
for solar and atmospheric effects. Symposium on Significant Results
Obtained from the Earth Resources Technology Satellite-1. March 5-9,
1973. NASA (Goddard Space Flight Center). 1-6:1115-1122.
Rogers, R. H., N. J. Shah, J. B. McKeon, and V. E. Smith. 1976. Computer
mapping of water quality in Saginaw Bay with LANDSAT digital data.
Proceedings of ACSM-ASP convention. February 22-28, 1976. Washington
Hilton. NASA-CR-146057. NTIS E76-10124.
Rogers, R. H. 1977. Application of LANDSAT to the Surveillance and Control
of Lake Eutrophication in the Great Lakes Basin. Final Report.
Prepared for National Aeronautics and Space Administration. Goddard
Space Flight Center. 169 pp.
Round, F. E. 1958. Algal aspects of lake typology. Proceedings
International Association Theoretical Applied Limnology. 13:306-310.
Rowe, K., and J. Barnes. 1976. Statistical Interactive Programming System
(SIPS) Command Reference Manual. Statistical Computing Report Number
3. Corvallis, Oregon: Oregon State University Department of Statistics.
66 pp.
Ruttner, F. 1963. Fundamentals of Limnology. Third edition. Translated by
D. G. Frey and F. E. J. Fry. Toronto: University of Toronto Press.
295 pp.
179
-------
Sawyer, W. R. 1931. The spectral absorption of light by pure water and Bay
of Fundy water. Contribution to Canadian Biology and Fisheries being
studied from the Biological Stations of Canada. Series D.
Hydrographic. 7(8):73-89.
Schaeffer, D. J., M. C. Long, and K. G. Janardan. 1974. Statistical
analysis of the recovery of coliform organisms on Gelman and Millipore
membrane filters. Applied Microbiology. 28(4):605-607.
Scherz, J. P., D. R. Crane, and R. H. Rogers. 1975. Classifying and
monitoring water quality by use of satellite imagery. Proceedings of
ASP-ACSM Fall Convention. October 26-31, 1975. Phoenix, Arizona.
Scherz, J. P., D. R. Graff, and W. C. Boyle. 1969. Photographic
characteristics of water pollution. Photogrammetric Engineering.
35:38-43.
Seal, H. L. 1964. Multivariate Statistical Analysis for Biologists.
London: Methuen. 207 pp.
Sheldon, A. 1972. A quantitative approach to the classification of inland
waters. In: Natural Environments. Studies in Theoretical and Applied
Analysis. J. V. Krutilla (ed). Baltimore: Johns Hopkins University
Press, pp. 205-261.
Simonson, R. W. 1962. Soil classification in the United States. Science.
137:1027-1034.
Sneath, P. H. A., and R. R. Sokal. 1973. Numerical Taxonomy: The
Principles and Practice of Numerical Classification. San Francisco: F.
H. Freeman. 573 pp.
Snedecor, G. W., and W. G. Cochran. 1967. Statistical Methods. Sixth
edition. Ames: Iowa State University Press. 593 pp.
Soil Survey Staff. 1960. Soil Classification: A Comprehensive System -
Seventh Approximation. Washington, D.C.: U.S.. Department of
Agriculture Soil Conservation Service.
Sokal, R. R. 1974. Classification: Purposes, principles,
progress, prospects. Science. 185:1115-1123.
Strahler, A. N. 1969. Physical Geography. New York: John Wiley and Sons.
733 pp.
Strong, A. E. 1973. ERTS-1 observes algal blooms in Lake Erie and Utah
Lake. Symposium on Significant Results Obtained from The Earth Resources
Technology Sateltite-1. March 5-9, 1973. NASA (Goddard Space Flight
Center). 1-6:1605-1612.
180
-------
Taylor, W. D., L. R. Williams, S. C. Hern, and V. W. Lambou. Phytoplankton
Water Quality Relationships in U.S. Lakes, Part VII: Comparison of Some
New and Old Indices and Measurements of Trophic State. Working Paper
Number 711. National Eutrophication Survey. U. S. Environmental
Protection Agency Office of Research and Development. (In press).
Thornbury, W. D. 1965. Regional Geomorphology of the United States. New
York: John Wiley and Sons. 609 pp.
Upper Mississippi River Basin Study Commission. 1970. Upper Mississippi
River Basin Report. Volume III. p. D-21.
U.S. EPA. 1975. National Eutrophication Survey Methods 1973-1975. National
Eutrophication Survey Working Paper Number 175. Las Vegas, Nevada:
National Environmental Research Center, Las Vegas, Nevada, National
Environmental Research Center, Corvallis, Oregon. U.S. Environmental
Protection Agency. 91 pp.
Uttormark, P. D., and J. P. Wall. 1975. Lake Classification—A Trophic
Classification of Wisconsin Lakes. U.S. EPA Ecological Research Series
Report EPA-660/3-75-033. 165 pp.
Veatch, J. 0., and C. R. Humphrys. 1966. Water and Water Use Terminology.
Kaukauna, Wisconsin: Thomas Printing and Publishing. 375 pp.
Vollenweider, R. A. 1968. Scientific Fundamentals of the Eutrophication of
Lakes and Flowing Waters with Particular Reference to Nitrogen and
Phosphorus as Factors in Eutrophicaton. Technical report prepared for
the Organization for Economic Cooperation and Development. Paris,
France. 159 pp.
Weber, C. A. 1907. Aufbau and Vegetation der Moore Norddeutschlands.
Beiblatter Botanische Jahrbucker fur Systematik. Pflanzengeschichte und
Pflanzengeographie. 90:19-34. Supplement to Bot. Jahrb. 40.
Welch, P. S. 1952. Limnology. Second edition. New York: McGraw-Hill
538 pp.
Whittaker, R. H. 1967. Gradient analysis of vegetation. Biological Review.
42:207-264.
Whittaker, R. H., and H. G. Gaich, Jr. 1973. Evaluation of ordination of
communities. Handbook of Vegetation Science. R. H. Whittaker (ed).
The Hague: Junk. 5:289-321.
Whipple, G. C. 1898. Classifications of lakes according to temperature.
America Naturalist. 32(373):25-33.
Williams, W. T. 1971. Principles of clustering. Ann. Rev. Ecol. Syst.
2:303-326.
181
-------
Williams, W. T., and J. M. Lambert. 1961. Multivariate methods in plant
ecology. I. Association analysis in plant communities. Journal
of Ecology. 47:83-101.
Willman, H. B., and J. C. Frye. 1970. Pleistocene Stratigraphy of Illinois.
Illinois State Geological Survey Bulletin Number 94. Champaign-Urbana,
Illinois. 204 pp.
Winner, R. W. 1972. An evaluation of certain indices of eutrophy and
maturity in lakes. Hydrobiologia. 40(2):223-245.
Work, E. A., and D. S. Gilmer. 1976. Utilization of satellite data for
inventorying prairie ponds and lakes. Photogrammetric Engineering and
Remote Sensing. 42(5):685-694.
Yarger, H. L., and J. R. McCauley. 1975. Quantitative water quality with
LANDSAT and Skylab. Proceedings NASA Earth Resources Survey Symposium.
First Comprehensive Symposium on the Practical Application of Earth
Resources Survey Data. June 9-12, 1975. Shamrock Hilton. Houston,
Texas. NASA TM X-58168. JSC-09930. I-A:347-370.
Zafar, A. R. 1959. Taxonomy of lakes. Hydrobiologia. 13(3):287-299
Zumberge, J. H. 1952. The Lakes of Minnesota: Their Origin and
Classification. Minnesota Geological Survey Bulletin Number 35.
99 pp.
182
-------
APPENDIX
MEASURED AND ESTIMATED WATER QUALITY DATA FOR SELECTED
ILLINOIS WATER BODIES
Table Page
A-l Unadjusted and adjusted multispectral scanner data for
Illinois water bodies (October 14-16, 1973) 184
A-2 Trophic indicator mean values for 31 water bodies sampled
October 16-19, 1973 210
A-3 Trophic indicator annual mean values for 31 water bodies
sampled three times during 1973 212
A-4 Trophic indicator and multivariate trophic index estimates
for 145 Illinois lakes based on set three regression
models 214
A-5 Group and individual lake water quality and ancillary data
values for selected Illinois water bodies classified by
complete linkage clustering on four spectral attributes
(GRN, RED, IR1, IR2) 220
183
-------
TABLE A-l. UNADJUSTED AND ADJUSTED MULTISPECTRAL SCANNER DATA FOR ILLINOIS WATER BODIES
(OCTOBER 14, 15, and 16, 1973)*
Name of Water Body
Anderson
NES LANDSAT LANDSAT
Serial STORET Scene Scene Area
Number Number Date Number** Pixels (ha)
34 10-15-73 1449-16084 969 620
10-16-73 1450-16142 944 603
LANDSAT MSS Mean DN
Value and (Deviation)
URN RED
41.89 30.16
(3.56) (3.19)
42.19 30.63
(2.19) (1.81)
+43.26 +30.22
IR1
22.63
(5.66)
23.49
(3.52)
+22.08
IR2
9.02
(5.02)
9.09
(4.96)
+9.30
Apple Canyon
Argyle
Bakers
Bangs
Bath
144
157
19
56
166
10-16-73 1450-16140 240 153
10-16-73 1450-16142 49 31
10-14-73 1448-16023 78
10-14-73 1448-16023 191 122
10-15-73 1449-16084 97 62
10-16-73 1450-16142 73 47
33.22 18.82
(3.15) (4.11)
+36.72 +21.03
39.16 26.36
(1.82) (1.95)
+41.05 +26.90
15.95 10.81
(6.38) (8.13)
+17.44 +10.39
20.32 14.20
(5.51) (7.99)
+20.13 +12.54
50 46.42 31.56 29.03 11.84
(2.39) (2.12) (3.04) (5.09)
+45.51 +30.92 +28.48 +12.13
37.36 21.40
(2.17) (2.78)
+36.74 +20.60
40.33 29.74
(3.86) (4.06)
42.10 31.68
(2.41) (2.73)
+43.20 +31.04
14.83 6.20
(5.32) (6.29)
+15.03 +7.15
28.19
(6.92)
27.71
(3.45)
+24.68
16.02
(7.35)
14.78
(5.99)
+12.91
(continued)
-------
TABLE A-l. (continued)
00
en
Name of Water
Biq
U 1 VJ
Rid
L* 1 M
NES LANDSAT
Serial STORE! Scene
Body Number Number Date
134 10-15-73
10-16-73
176 10-15-73
10-16-73
LANDSAT
Scene
Number** Pixels
1449-16084 149
1450-16142 127
1449-16084 117
1450-16142 106
LANDSAT MSS Mean DN
Area Value and (Deviation)
(ha) GRN
95 39.
(3.
81 39.
(2.
+41.
75 40.
(3.
68 42.
(2.
+43.
66
56)
79
13)
51
66
95)
30
51)
34
RED
26
(3
28
(2
+28
29
(3
31
(2
+31
.65
.60)
.11
.23)
.26
.80
.44)
.77
.84)
.11
IR1
25
(9
22
(6
+21
26
(8
25
(4
+23
.18
.85)
.48
.08)
.46
.55
.80)
.03
.63)
.03
IR2
13.
(8.
11.
(6.
+11.
11.
(6.
11.
(6.
+10.
56
11)
77
81)
00
70
53)
21
27)
65
Bloomington
Bracken
Calumet
Canton
71 1703 10-15-73 1449-16084 374 238
145 10-16-73 1450-16142 83 53
18 10-14-73 1448-16023 631 403
36 10-15-73 1449-16084 151 97
10-16-73 1450-16141 128 82
38.87 27.00 24.16 11.78
(3.65) (3.94) (11.50) (8.42)
35.53 19.84
(2.05) (2.74)
+38.40 +21.82
52.22 41.40
(3.59) (3.95)
+51.13 +40.92
41.28 28.19
(4.02) (3.83)
42.40 30.48
(2.74) (3.57)
+43.41 +30.10
18.28 12.40
(5.64) (7.75)
+18.87 +11.40
25.44 10.41
(4.78) (6.48)
+25.03 +10.87
26.95
(12.09)
21.39
(4.95)
+20.79
12.34
(7.23)
10.75
(7.25)
+10.35
(continued)
-------
TABLE A-l. (continued)
CO
CTi
NES LANDSAT LANDSAT LANDSAT MSS Mean DN
Serial STORE! Scene Scene Area Value and (Deviation)
Name of Water Body Number Number Date Number** Pixels (ha) GRN RED IR1 IR2
Carbondale
Carlyle
Catherine
Cattail
Cedar
38 10-14-73 1448-16035 45 29 47.
(2.
+46.
10-15-73 1449-16093 42 27 45.
(2.
14 1706 10-14-73 1448-16032 15,435 9,864 46.
(2.
+45.
58 10-14-73 1448-16023 95 61 35.
(1.
+35.
125 10-16-73 1450-16140 30 19 36.
(2.
+39.
39 10-14-73 1448-16035 113 72 43.
(4.
+42.
10-15-73 1449-16093 121 77 40.
(4.
22
38)
28
09
42)
30
36)
39
77
90)
20
83
22)
35
59
54)
77
28
18)
34.
(4.
+34.
34.
(3.
31.
(2.
+31.
19.
(1.
+18.
26.
(2.
+26.
32.
(7.
+31.
28.
(6.
62
89)
03
19
77)
66
98)
02
65
84)
82
36
44)
90
15
35)
52
87
36)
28.
(6.
+28.
27.
(5.
18.
(5.
+18.
14.
(5.
+15.
26.
(2.
+23.
23.
(7.
+23.
21.
(7.
75
09)
22
90
58)
18
08)
20
96
18)
15
29
79)
80
66
33)
40
76
03)
13.
(7.
+13.
13.
(5.
6.
(5.
+6.
6.
(5.
+7.
19.
(5.
+16.
13.
(7.
+13.
12.
(7.
79
71)
85
33
58)
01
22)
99
28
93)
22
69
87)
02
11
35)
25
85
45)
Cedar
55
1759 10-14-73 1448-16023 200
128 34.68 19.65
(2.71) (4.17)
+34.14 +18.82
13.58 6.86
(6.99) (7.42)
+13.85 +7.74
(continued)
-------
TABLE A-l. (continued)
Name of Water Body
NES
Serial STORE!
Number Number
LANDSAT LANDSAT
Scene Scene Area
Date Number** Pixels (ha)
LANDSAT MSS Mean DN
Value and (Deviation)
GRN RED IR1 IR2
00
Centralia
Chain, Ingram,
Sangamon, Stafford,
Stewart, Snicarte
Channel
Charleston
Chautauqua
Clear
79
83
59
16
87
10-14-73 1448-16032 144 92
10-15-73 1449-16084 427 273
10-16-73 1450-16142 467
10-14-73 1448-16023 221 141
36.47 21.86 17.86 11.38
(2.37) (2.98) (6.12) (7.84)
+35.87 +21.06 +17.90 +11.72
40.27 29.10
(3.28) (2.84)
298 41.21 30.74
+42.55 +30.30
25.50 11.59
(6.72) (6.01)
25.88 12.03
+23.55 +11.16
1708 10-14-73 1448-16032 189 121
37.51 21.85 16.72 6.72
(1.74) (2.71) (5.13) (6.49)
+36.88 +21.05 +10.82 +7.61
42.39 30.65 22.51 9.91
(2.53) (2.34) (3.51) (6.02)
+41.61 +30.00 +22.31 +10.43
10-15-73 1449-16084 1,610 1,030 42.66 32.41 26.91 11.78
(3.99) (4.50) (4.92) (3.73)
10-16-73 1450-16142 1,604 1,025 42.30 32.09 26.17 11.18
(2.75) (3.54) (2.29) (3.27)
+43.34 +31.36 +23.73 +10.63
86
10-15-73 1449-16084 977 625
10-16-73 1450-16142 968 619
41.10 30.26
(3.35) (2.55)
41.08 30.46
(2.12) (1.91)
+42.45 +30.09
25.67
(5.50)
25.40
(2.94)
+23.26
11.71
(4.89)
11.30
(4.46)
+10.70
(continued)
-------
TABLE A-l. (continued)
00
00
Name of Water Body
Coal City Recreation
Club
Commonwealth Edison-
Dresden Nuclear
Countryside
Crab Orchard
Crane
NES LANDSAT
Serial STORET Scene
Number Number Date
138 10-14-73
10-15-73
137 10-14-73
10-15-73
148 10-14-73
127 1712 10-14-73
85 10-15-73
10-16-73
LANDSAT
Scene
Number** Pixels
1448-16023 53
1449-16082 51
1448-16023 720
1449-16082 720
1448-16023 80
1448-16035 4,198
1449-16084 1,212
1450-16142 1,589
LANDSAT MSS Mean DN
Area Value and (Deviation)
(ha) GRN RED IR1 IR2
34 48
(5
+47
33 48
(5
460 41
(3
+40
460 41
(3
51 38
(1
+37
2,683 39
(2
+38
776 43
(4
1,015 44
(3
+45
.03
.31)
.07
.43
.35)
.26
.40)
.51
.31
.49)
.16
.34)
.51
.15
.29)
.47
.71
.90)
.83
.84)
.19
28
(4
+28
28
(4
29
(3
+28
29
(3
22
(1
+21
24
(2
+23
34
(6
35
(5
+34
.90
.38)
.22
.41
.21)
.34
.42)
.66
.56
.56)
.34
.40)
.55
.32
.67)
.56
.11
.41)
.54
.43)
.04
22
(6
+22
22
(6
17
(4
+17
17
(4
18
(4
+18
17
(5
+17
27
(6
27
(3
+24
.90
.48)
.68
.56
.13)
.32
.94)
.39
.53
.93)
.15
.53)
.18
.36
.33)
.43
.69
.39)
.79
.41)
.73
16
(7
+16
16
(7
6
(5
+7
7
(5
8
(6
+9
6
(6
+7
11
(4
12
(4
+11
.73
.83)
.44
.54
.29)
.71
.31)
.60
.18
.19)
.84
.99)
.48
.76
.33)
.65
.63
.47)
.34
.16)
.36
Crooked
149
10-14-73 1448-16023
84 54 45.82 26.64 24.54 9.74
(1.80) (1.62) (4.15) (5.75)
+44.93 +25.92 +24.23 +10.28
(continued)
-------
TABLE A-l. (continued)
00
IO
NES LANDSAT LANDSAT
Serial STORET Scene Scene
Name of Water Body Number Number Date Number** Pixels
Crystal
Dawson
Decatur
Deep
DePue
68 10-15-73 1449-16082 146
70 10-15-73 1449-16084 105
73 1714 10-14-73 1448-16032 1,778
10-15-73 1449-16084 1,862
150 10-14-73 1448-16023 136
3 1752 10-15-73 1449-16082 360
10-16-73 1450-16140 373
LANDSAT MSS Mean DN
Area Value and (Deviation)
(ha) GRN RED IR1 IR2
93 36
(1
67 37
(3
1,136 39
(2
+39
1,192 38
(3
87 36
(1
+36
230 41
(2
238 39
(2
+41
.69
.98)
.47
.23)
.7-8
.33)
.08
.38
.62)
.80
.49)
.19
.06
.51)
.38
.53)
.21
20
(2
25
(4
27
(2
+27
27
(3
19
(2
+19
29
(2
29
(2
+29
.78
.90)
.26
.09)
.83
.35)
.13
.30
.28)
.86
.11)
.03
.52
.18)
.80
.37)
.57
15
(5
27
(13
18
(4
+18
21
(8
12
(4
+13
22
(4
23
(3
+22
.21
.93)
.70
.88)
.60
.59)
.60
.74
.45)
.85
.98)
.16
.53
.30)
.90
.38)
.33
8
(6
15
(13
7
(6
+8
9
(6
6
(6
+7
10
(6
11
(5
+11
.02
.39)
.18
.73)
.49
.03)
.29
.84
.68)
.03
.00)
.00
.59
.23)
.84
.61)
.04
Devil's Kitchen
128
10-14-73 1448-16035 370
236 34.07 17.48
(1.81) (2.07)
+33.55 +16.61
15.30 9.89
(6.46) (7.85)
+15.48 +10.41
(continued)
-------
TABLE A-l. (continued)
NES LANDSAT LANDSAT LANDSAT MSS Mean DN
Serial STORET Scene Scene Area Value and (Deviation)
Name of Water Body Number Number Date Number** Pixels (ha) GRN RED IR1 IR2
Diamond
DuQuoin
Dutchman
East Loon
Evergreen
Fourth
Fox
57 10-14-73 1448-16023 95 61 39.
(2.
+38.
98 10-14-73 1448-16035 86 55 34.
(3.
+34.
10-15-73 1449-16093 81 52 32.
(2.
49 10-14-73 1448-16035 26 17 37.
(2.
+37.
53 1757 10-14-73 1448-16023 105 67 35.
(1.
+34.
72 10-15-73 1449-16084 466 298 36.
(3.
151 10-14-73 1448-16023 200 128 42.
(2.
+41.
60 1755 10-14-73 1448-16023 1,159 741 39.
(2.
+39.
03
00)
35
59
28)
05
97
61)
80
45)
16
27
74)
71
88
70)
37
14)
59
92
27)
22
22.
(2.
+21.
20.
(4.
+20.
19.
(3.
21.
(2.
+20.
20.
(2.
+20.
23.
(5.
30.
(1.
+29.
26.
(1.
+25.
75
39)
97
81
05)
00
76
87)
42
51)
62
85
32)
04
40
07)
04
57)
38
67
98)
95
15.
(5.
+15.
18.
(7.
+18.
17.
(6.
18.
(6.
+18.
15.
(5.
+16.
21.
(11.
21.
(3.
+21.
17.
(3.
+18.
63
45)
79
53
41)
54
55
59)
42
95)
43
88
91)
03
80
99)
63
89)
47
99
85)
02
6.
(6.
+7.
10.
(8.
+11.
10.
(7.
11.
(8.
+11.
8.
(6.
+9.
10.
(7.
8.
(6.
+9.
6.
(5.
+7.
74
86)
63
97
42)
36
16
52)
65
30)
96
53
84)
21
77
45)
53
27)
21
09
06)
06
(continued)
-------
TABLE A-l. (continued)
Name of Water Body
NES
Serial STORE!
Number Number
LANDSAT LANDSAT
Scene Scene Area
Date Number** Pixels (ha)
LANDSAT MSS Mean DN
Value and (Deviation)
GRN RED IR1 IR2
Fyre 170
Gages 152
George 108
Glen 0. Jones 110
Goose (Sparland 82
Conservation Area)
Goose 101
10-16-73 1450-16140 60 38
10-14-73 1448-16023 83 53
10-16-73 1450-16140 89 57
10-14-73 1448-16035 61 39
10-15-73 1449-16084 604 387
10-15-73 1449-16082 1,088 696
10-16-73 1450-16140 1,160 741
32.94 19.18
(2.40) (3.25)
+36.52 +21.31
36.16 20.38
(1.76) (2.12)
+35.57 +19.56
17.84 11.76
(6.80) (8.19)
+18.60 +10.99
14.01 6.28
(5.09) (5.43)
+14.25 +7.22
35.43 19.55 15.87 10.62
(1.68) (2.18) (6.00) (8.21)
+38.33 +21.60 +17.39 +10.27
36.91 19.86
(2.53) (3.16)
+36.30 +19.03
15.32 9.02
(6.15) (7.17)
+15.50 +9.64
42.58 34.18 27.51
(3.55) (4.15) (5.09)
42.67 31.37
(1.96) (1.73)
40.60 31.93
(2.00) (1.77)
+42.10 +31.22
24.88
(2.82)
28.04
(2.03)
+24.88
12.80
(4.53)
11.06
(4.03)
16.56
(3.63)
+14.04
(continued)
-------
TABLE A-l. (continued)
vo
ro
Name of Water Body
Goose (Village Club)
Grass
Greenville New City
(Governor Bond)
Griswold
Harrisburg
Highland (Old Taylor
NES LANDSAT
Serial STORET Scene
Number Number Date
139 10-14-73
10-15-73
61 1756 10-14-73
2 10-14-73
158 10-15-73
111 10-14-73
153 10-14-73
LANDSAT
Scene
Number**
1448-16023
1449-16082
1448-16023
1448-16032
1449-16082
1448-16035
1448-16023
Pixels
116
117
1,025
482
30
94
65
Area
(ha)
74
75
655
308
19
60
42
LANDSAT MSS Mean DN
Value and (Deviation)
GRN RED IR1
IR2
45.05 31.40 24.10 17.05
(6.79) (7.63) (7.98) (8.10)
+44.18 +30.76 +23.81 +16.73
45.01 31.92 25.13 17.97
(6.72) (8.03) (8.14) (7.39)
37.61 25.95 17.38
(1.69) (2.20) (4.11)
+36.98 +25.22 +17.45
38.52 24.90 18.43
(2.41) (2.76) (5.73)
+37.86 +24.15 +18.44
36.59 22.49 19.66
(2.39) (2.45) (5.27)
36.08 22.34 18.10
(3.63) (4.09) (7.12)
+35.50 +21.55 +18.13
33.52 17.87 12.61
(1.58) (1.60) (4.65)
+33.02 +17.01 +12.93
7.00
(5.79)
+7.86
8.90
(7.22)
+9.54
11.69
(6.09)
10.64
(7.64)
+11.07
5.80
(5.89)
+6.80
Holiday
51
1754 10-15-73 1449-16082 186
119 35.36 22.09 17.15 8.65
(1.98) (2.14) (4.97) (6.19)
(continued)
-------
TABLE A-l. (continued)
NES LANDSAT
Serial STORET Scene
Name of Water Body Number Number Date
Horseshoe 1
Jack, Swan, Grass 165
Keithsburg National 92
Wildlife Refuge
Kinkaid 40
Kinneman 168
10-15-73
10-15-73
10-16-73
10-16-73
10-14-73
10-15-73
10-14-73
LANDSAT
Scene
Number** Pixels
1449-16093 262
1449-16084 927
1450-16142 859
1450-16142 126
1448-16035 1,338
1449-16093 1,371
1448-16035 36
LANDSAT MSS Mean DN
Area Value and (Deviation)
(ha) GRN RED IR1 IR2
168 38.
(5.
593 39.
(3.
549 39.
(1.
+41.
81 36.
(1.
+39.
855 37.
(4.
+36.
876 42.
(4.
23 50.
72
13)
32
22)
49
97)
29
85
77)
37
43
71)
80
03
20)
08
25
(6
27
(2
28
(1
+28
25
(1
+26
22
(6
+21
21
(5
33
.28
.47)
.39
.41)
.57
.86)
.62
.34
.62)
.10
.52
.75)
.73
.13
.29)
.44
23.
(8.
23.
(6.
23.
(3.
+21.
25.
(4.
+23.
17.
(7.
+17.
16.
(7.
30.
12
29)
62
90)
27
63)
94
76
61)
48
22
82)
30
07
27)
44
12.
(8.
10.
(5.
9.
(4.
+9.
19.
(6.
+16.
9.
(7.
+9.
8.
(7.
20.
57
39)
56
89)
99
93)
87
80
16)
09
26
71)
85
76
77)
33
Lake of Egypt
132
10-14-73 1448-16035 1,186 758
(4.69) (5.00)
+49.05 +32.83
39.75 23.72
(4.67) (4.66)
+39.05 +22.95
(5.87) (4.93)
+29.55 +19.62
17.03 9.41
(6.48) (7.54)
+17.12 +9.99
(continued)
-------
TABLE A-l. (continued)
ID
NES LANDSAT LANDSAT
LANDSAT MSS Mean DN
Serial STORE! Scene Scene Area Value and (Deviation)
Name of Water Body Number Number Date Number** Pixels (ha) GRN RED IR1 IR2
Lake of the Woods 115
LaRue-Pine Hills 121
Ecological area
Lily 136
10-15-73 1449-16084 65 42 40.
(3.
10-16-73 1450-16142 58 37 40.
(2.
+41.
10-14-73 1448-16035 18 12 37.
(2.
+37.
10-15-73 1449-16084 102 65 43.
(4.
10-16-73 1450-16142 85 54 44.
(2.
+44.
65
99)
44
90)
99
94
37)
30
84
12)
32
08)
81
29
(3
29
(2
+29
24
(3
+23
34
(4
36
(2
+34
.42
.97)
.96
.70)
.70
.05
.66)
.29
.94
.89)
.69
.99)
.93
29.
(6.
27.
(3.
+24.
29.
(4.
+28.
36.
(10.
30.
(3.
+26.
51
84)
75
63)
70
22
05)
66
76
50)
17
95)
19
15.
(6.
13.
(5.
+12.
23.
(5.
+22.
17.
(6.
16.
(5.
+14.
85
52)
77
29)
27
27
02)
21
62
31)
69
71)
12
Lincoln Trail State 13
Park
Little Grassy 119
Little Swan 185
10-14-73 1448-16032 73 47
10-14-73 1448-16035 522 334
10-16-73 1450-16142 151 97
34.34 17.72
(2.54) (2.56)
+33.81 +16.86
35.30 18.56
(2.22) (2.45)
+34.74 +17.71
40.97 31.39
(1.34) (1.39)
+42.37 +30.31
15.60 10.30
(6.01) (6.35)
+15.76 +10.77
14.27 8.39
(6.53) (7.72)
+14.50 +9.09
23.14 12.13
(3.47) (4.59)
+21.86 +11.23
(continued)
-------
TABLE A-l. (continued)
C7I
NES LANDSAT
Serial STORE! Scene
Name of Water Body Number Number Date
Liverpool
Long
Long
Lower Smith
88 10-15-73
10-16-73
52 1725 10-14-73
177 10-15-73
10-16-73
181 10-15-73
10-16-73
LANDSAT
Scene
Number** Pixels
1449-16084 36
1450-16142 30
1448-16023 242
1449-16084 77
1450-16142 68
1449-16084 79
1450-16142 60
LANDSAT MSS Mean DN
Area Value and (Deviation)
(ha) GRN RED IR1 IR2
23 42
(3
20 43
(2
+43
155 36
(1
+35
49 43
(5
44 44
(3
+44
51 40
(3
38 42
(2
+43
.17
.93)
.03
.32)
.87
.58
.76)
.98
.12
.08)
.23
.77)
.75
.91
.58)
.39
.65)
.41
30
(4
33
(2
+32
23
(1
+22
32
(5
34
(5
+32
29
(3
32
(2
+31
.77
.13)
.83
.08)
.71
.53
.71)
.76
.66
.83)
.11
.16)
.93
.77
.36)
.18
.67)
.43
34
(7
29
(3
+25
16
(4
+16
30
(9
27
(5
+24
30
(10
25
(3
+23
.57
.57)
.23
.86)
.62
.40
.44)
.52
.70
.20)
.80
.42)
.73
.04
.33)
.26
.52)
.17
19.
(7.
16.
(5.
+14.
6.
(5.
+7.
14.
(7.
13.
(6.
+12.
15.
(7.
14.
(6.
+12.
86
41)
96
76)
29
42
82)
35
51
28)
61
52)
17
24
59)
69
85)
85
Lyerla-Autumnal
Flooding Ponds
120
10-14-73 1448-16035 345
221 42.31 27.46
(3.68) (3.30)
+41.53 +26.75
24.34 13.71
(7.04) (8.03)
+24.04 +13.78
(continued)
-------
TABLE A-l. (continued)
CTv
Name of Water Body
Marie
Marion
Marshall County Pu-
lic Hunting and
Fishing Area
Matanzas
Mattoon
McCullom
NES LANDSAT
Serial STORE! Scene
Number Number Date
62 1727 10-14-73
129 10-14-73
81 10-15-73
10-16-73
89 10-15-73
10-16-73
24 10-14-73
159 10-15-73
LANDSAT
Scene
Number** Pixels
1448-16023 347
1448-16035 81
1449-16084 1,131
1450-16142 1,138
1449-16084 242
1450-16142 236
1448-16032 611
1449-16082 168
LANDSAT MSS Mean DN
Area Value and (Deviation)
(ha) GRN RED IR1 IR2
222 39.
(2.
+38.
52 34.
(2.
+34.
724 39.
(3.
729 38.
(1.
+40.
155 41.
(3.
151 42.
(2.
+43.
391 42.
(2.
+41.
107 37.
(2.
18
19)
50
86
05)
32
09
22)
95
86)
90
79
45)
08
07)
18
07
83)
30
08
02)
25
(2
+24
20
(2
+19
27
(2
28
(2
+28
31
(2
31
(1
+31
28
(2
+27
21
(2
.04
.11)
.30
.24
.27)
.42
.38
.60)
.87
.07)
.85
.60
.60)
.85
.80)
.17
.13
.69)
.43
.39
.35)
19
(4
+19
18
(6
+18
23
(5
24
(2
+22
24
(5
25
(3
+23
18
(6
+18
17
(5
.07
.79)
.05
.59
.52)
.59
.48
.70)
.95
.81)
.98
.57
.54)
.08
.14)
.06
.31
.16)
.33
.40
.69)
7.
(6.
+8.
10.
(7.
+11.
10.
(5.
11.
(4.
+11.
11.
(6.
11.
(5.
+10.
8.
(7.
+9.
10.
(7.
76
23)
53
92
37)
32
51
09)
94
54)
11
43
13)
12
33)
59
73
57)
39
38
24)
(continued)
-------
TABLE A-l. (continued)
NES LANDSAT
Serial STORE! Scene
Name of Water Body Number Number Date
McGinnis (Orland) 20
Meredosia 10
Mesa 184
Moscow 164
10-14-73
10-15-73
10-15-73
10-16-73
10-14-73
10-15-73
10-16-73
LANDSAT
Scene
Number** Pixels
1448-16023 179
1449-16082 160
1449-16084 1,100
1450-16142 965
1448-16032 56
1449-16084 164
1450-16142 148
LANDSAT MSS Mean DN
Area Value and (Deviation)
(ha) CRN RED IR1 IR2
114 37.
(2.
+37.
102 38.
(1.
704 46.
(5.
617 46.
(4.
+46.
36 37.
(2.
+36.
106 40.
(3.
95 40.
(2.
+42.
99
42)
35
35
50)
36
37)
85
23)
66
26
04)
64
09
53)
94
18)
35
23.71
(2.
+22.
24.
(1.
38.
(7.
39.
(6.
+37.
20.
(2.
+19.
28.
(3.
29.
(1.
+29.
42)
94
04
46)
49
29)
52
45)
14
80
22)
99
15
32)
72
82)
51
27.59
(4.04)
+27.12
28.
(2.
28.
(6.
28.
(4.
+25.
15.
(5.
+15.
25.
(7.
25.
(3.
+23.
03
52)
97
26)
33
60)
06
46
82)
63
85
21)
06
64)
05
13.36
(4.88)
+13.47
12.
(4.
12.
(6.
12.
(5.
+11.
8.
(6.
+9.
12.
(7.
12.
(5.
+11.
89
05)
81
15)
01
25)
15
39
83)
09
95
11)
13
45)
23
Moses
28
10-14-73 1448-16035 75
48 33.86 19.55 18.06 11.90
(2.55) (3.11) (6.95) (8.38)
+33.35 +18.72 +18.09 +12.18
(continued)
-------
TABLE A-l. (continued)
00
Name of Water Body
Mound
Murphysboro
New Pittsfield
NES LANDSAT
Serial STORET Scene
Number Number Date
163 10-15-73
10-16-73
37 10-14-73
10-15-73
100 10-16-73
LANDSAT
Scene
Number
1449-16084
1450-16142
1448-16035
1449-16093
1450-16142
Pixels
353
329
74
85
113
Area
(ha)
226
210
47
54
72
LANDSAT MSS Mean DN
Value and (Deviation)
CRN
36.
(3.
36.
(1.
+39.
37.
(1.
+36.
35.
(1.
45.
85
04)
46
52)
08
29
60)
67
04
81)
65
RED
24.
(2.
24.
(1.
+25.
21.
(2.
+20.
19.
(1.
36.
57
91)
90
80)
76
56
02)
76
76
93)
75
IR1
24.
(9.
21.
(3.
+20.
16.
(4.
+16.
17.
(6.
25.
13
27)
47
74)
84
21
80)
34
32
71)
29
IR2
11.76
(6.70)
10.75
(5.82)
+10.35
9.04
(6.24)
+9.66
11.18
(8.53)
11.67
Nipper sink
Old Ben Mine
Olney East Fork
63
10-14-73 1448-16023 310
32 1765 10-14-73 1448-16035
44
106
10-14-73 1448-16032 448
(3.39) (5.01)
+45.78 +34.98
198 38.24 25.94
(2.17) (2.57)
+37.59 +25.21
28 37.97 24.45
(3,06) (2.82)
+37.33 +23.70
286 39.25 22.69
(3.37) (3.65)
+38.57 +21.91
(4.54) (6.00)
+23.19 +10.94
17.51 6.61
(3.95) (5.13)
+17.57 +7.52
20.59 11.63
(6.99) (6.89)
+20.49 +11.94
17.28 9.47
(6.97) (7.69)
+17.35 +10.04
(continued)
-------
TABLE A-l. (continued)
VO
NES LANDSAT LANDSAT LANDSAT MSS Mean DN
Serial STORE! Scene Scene Area Value and (Deviation)
Name of Water Body Number Number Date Number** Pixels (ha) 6RN
Olney New 107
Open Pond (Marshy) 174
Otter 167
10-14-73 1448-16032 68 44 37.
(2.
+36.
10-14-73 1448-16035 152 97 40.
(4.
+40.
10-15-73 1449-16084 117 75 36.
(3.
10-16-73 1450-16142 93 59 36.
(2.
+39.
54
80)
91
94
79)
20
69
43)
92
35)
42
RED
22
(3
+21
27
(7
+26
24
(3
25
(2
+26
.38
.01)
.59
.26
.72)
.55
.22
.35)
.43
.89)
.17
IR1
20.
(7.
+19.
24.
(8.
+24.
26.
(8.
23.
(4.
+21.
01
27)
94
74
07)
42
36
43)
08
73)
83
IR2
12.
(7.
+13.
17.
(7.
+17.
16.
(7.
15.
(7.
+13.
99
36)
14
62
67)
23
45
39)
23
02)
19
Pana
Paradise
Paris Twin
113
15
25
10-14-73 1448-16032 118 75
10-14-73 1448-16032 88 56
10-14-73 1448-16032 111 71
41.68 28.06
(5.47) (7.81)
+40.92 +27.36
40.70 27.95
(1.85) (1.44)
+39.97 +27.25
34.98 20.78
(2.71) (3.01)
+34.43 +19.97
20.73 12.42
(7.68) (7.72)
+20.62 +12.64
19.59 9.10
(5.33) (7.43)
+19.54 +9.71
18.35 12.22
(6.61) (7.27)
+18.37 +12.47
(continued)
-------
TABLE A-l. (continued)
o
o
Name of Water Body
Pekin
k
Petersburg
Petite
Pierce
Pinckneyville
Pistakee
NES LANDSAT
Serial STORET Scene
Number Number Date
116 10-15-73
10-16-78
169 10-15-73
64 10-14-73
130 10-15-73
10-16-73
99 10-14-73
10-15-73
65 1733 10-14-73
LANDSAT
Scene
Number** Pixels
1449-16084 97
1450-16142 58
1449-16084 89
1448-16023 103
1449-16082 91
1450-16140 94
1448-16035 77
1440-16093 77
1448-16023 1,067
LANDSAT MSS Mean DN
Area Value and (Deviation)
(ha) GRN RED IR1 IR2
62 41.27 30.62 29.44
(4.14) (4.22) (5.76)
37 42.12 32.32 27.18
(2.82) (2.79) (3.97)
+43.21 +31.53 +24.35
57 37.06 22.38 25.47
(5.63) (8.23) (15.49)
66 39.10 25.50 18.30
(1.76) (1.80) (4.36)
+38.42 +24.76 +18.32
58 37.74 21.96 17.86
(2.04) (2.27) (5.59)
60 33.07 19.39 15.62
(2.06) (2.18) (5.21)
+36.61 +21.47 +17.23
49 41.50 26.28 18.48
(3.06) (2.93) (5.06)
+40.75 +25.56 +18.49
49 39.16 24.51 18.40
(3.00) (2.93) (5.56)
682 37.73 23.79 16.88
(1.72) (2.08) (4.50)
+37.08 +23.02 +16.97
17.00
(6.64)
14.43
(5.79)
+12.69
12.40
(7.55)
7.81
(6.62)
+8.57
9.84
(6.83)
8.14
(7.20)
+8.70
8.47
(7.09)
+9.16
8.87
(7.02)
6.55
(5.89)
+7.46
(continued)
-------
TABLE A-l. (continued)
ro
o
Name of Water Body
Powerton Cooling
Quiver
Raccoon
Rend
Rice
NES LANDSAT
Serial STORET Scene
Number Number Date
182 10-15-73
10-16-73
90 10-15-73
10-16-73
80 1762 10-14-73
29 1735 10-14-73
35 10-15-73
10-16-73
LANDSAT
Scene
Number**
1449-16084
1450-16142
1449-16084
1450-16142
1448-16032
1448-16035
1449-16084
1450-16142
LANDSAT MSS Mean DN
Area Value and (Deviation)
Pixels (ha) GRN RED IR1 IR2
893 572 36
(3
928 593 36
(2
+38
216 138 40
(3
210 134 39
(2
+41
366 234 45
(3
+44
9,765 6,242 37
(3
+37
976 625 37
(3
947 605 37
(1
+39
.89
.81)
.33
.34)
.99
.13
.56)
.90
.61)
.59
.23
.00)
.36
.71
.20)
.08
.91
.18)
.00
.65)
.48
22
(4
22
(2
+23
30
(4
30
(3
+30
31
(3
+31
24
(3
+23
24
(2
25
(1
+26
.33
.36)
.54
.90)
.92
.59
.23)
.43
.15)
.06
.66
.81)
.02
.52
.86)
.77
.90
.62)
.28
.85)
.06
16
(7
16
(3
+17
26
(6
22
(5
+21
21
(4
+20
15
(5
+15
21
(6
20
(3
+20
.63
.54)
.04
.78)
.49
.47
.11)
.59
.24)
.53
.04
.47)
.91
.18
.68)
.36
.96
.77)
.70
.74)
.36
6
(4
5
(4
+7
13
(7
11
(6
+10
8
(5
+9
6
(6
+7
9
(6
8
(5
+8
.94
.75)
.90
.54)
.28
.84
.33)
.29
.96)
.70
.29
.89)
.00
.39
.27)
.32
.98
.03)
.68
.62)
.99
(continued)
-------
TABLE A-l. (continued)
ro
o
ro
Name of Water Body
Round
Saganashkee
Sahara Coal Company
NES LANDSAT
Serial STORET Scene
Number Number Date
66 10-14-73
21 10-14-73
10-15-73
175 10-14-73
LANDSAT
Scene
Number**
1448-16023
1448-16023
1449-16082
1448-16035
LANDSAT MSS Mean DN
Pixels
147
229
262
153
Area
(ha)
94
146
167
98
Value and (Deviation)
GRN
37
(1
+36
36
(1
+35
37
(1
49
(3
+48
.21
.57)
.59
.34
.48)
.75
.27
.52)
.09
.16)
.10
RED
21
(2
+20
22
(1
+21
23
(2
40
(4
+40
.39
.23)
.59
.27
.94)
.48
.69
.02)
.58
.11)
.09
IR1
13.11
(5.25)
+13.40
17.49
(4.12)
+17.55
19.30
(5.12)
34.35
(4.25)
+33.52
IR2
5
(6
+6
6
(5
+7
9
(6
24
(3
+23
.75
.39)
.76
.97
.73)
.83
.24
.62)
.19
.61)
.03
St. Mary's
Sam Dale State
Sam Parr State
146
124
41
10-14-73 1448-16023 65 42
10-14-73 1448-16032 111 71
10-14-73 1448-16032 82 52
41.72 23.98
(2.84) (2.47)
+40.96 +23.22
35.02 20.13
(2.56) (3.29)
+34.47 +19.31
36.23 21.18
(2.43) (2.82)
+35.64 +20.37
21.86 11.81
(4.85) (7.61)
+21.69 +12.10
16.93 8.91
(6.58) (7.28)
+17.02 +9.54
21.18 14.31
(6.07) (7.46)
+21.05 +14.31
(continued)
-------
TABLE A-l. (continued)
NES LANDSAT LANDSAT LANDSAT MSS Mean DN
Serial STORE! Scene Scene Area Value and (Deviation)
Name of Water Body Number Number Date Number** Pixels (ha) GRN RED IR1 IR2
Sand 147 10-14-73 1448-16023 61 39 35.
(1.
+35.
Sanganois Conserva- 9 10-15-73 1449-16084 928 594 43.
tion Area (Muscooten, (5.
Sangamon, Treadway) 10-16-73 1450-16142 666 426 45.
(3.
+45.
o Sangchris 11 10-15-73 1449-16084 1,701 1,089 38.
(3.
Sara 26 10-14-73 1448-16032 308 197 35.
(2.
+34.
Sawmill 103 10-15-73 1449-16082 359 230 39.
(1.
Senachwine 104 10-15-73 1449-16082 2,508 1,603 41.
(2.
10-16-73 1450-16140 1,892 1,209 40.
(2.
+41.
85
74)
27
37
68)
11
99)
39
07
83)
18
76)
63
64
86)
87
15)
31
14)
89
19.
(1.
+18.
34.
(8.
36.
(5.
+34.
26.
(6.
19.
(3.
+19.
28.
(1.
30.
(1.
31.
(1.
+30.
59
66)
76
14
25)
60
42)
86
60
02)
89
36)
06
52
65)
45
78)
05
98)
55
13.
(5.
+14.
33.
(7.
29.
(4.
+26.
24.
(12.
16.
(7.
+16.
24.
(3.
21.
(2.
25.
(2.
+23.
85
43)
10
71
47)
89
18)
02
72
26)
48
05)
59
13
32)
78
86)
21
23)
14
6.
(6.
+7.
16.
(6.
15.
(5.
+13.
14.
(9.
10.
(8.
+11.
10.
(4.
7.
(3.
12.
(3.
+11.
39
38)
32
49
44)
22
86)
19
70
51)
92
39)
32
60
24)
99
73)
19
60)
27
(continued)
-------
TABLE A-l. (continued)
PO
o
Serial
Name of Water Body Number
Shelbyville 114
Skokie Lagoons 22
Slocum 54
Snyder's Hunting Club 142
South Wilmington 140
Fireman's Beach and
Park
NES LANDSAT
STORE! Scene
Number Date
1739 10-14-73
10-15-73
10-14-73
10-15-73
1758 10-14-73
10-14-73
10-15-73
10-14-73
10-15-73
LANDSAT
Scene
Number**
1448-16032
1449-16084
1448-16023
1449-16082
1448-16023
1448-16035
1449-16093
1448-16023
1449-16082
Pixels
6,796
7,715
122
131
138
27
27
155
158
Area
(ha)
4,343
4,938
80
84
88
17
17
99
101
LANDSAT MSS Mean DN
Value and (Deviation)
6RN RED IR1 IR2
37
(3
+36
36
(4
41
(2
+40
46
(3
40
(1
+39
36
(2
+36
34
(1
47
(3
+46
46
(3
.59
.03)
.96
.76
.07)
.31
.91)
.56
.35
.87)
.71
.85)
.98
.81
.31)
.20
.48
.79)
.02
.42)
.09
.65
.74)
24.
(4.
+23.
24.
(5.
27.
(3.
+27.
31.
(4.
22.
(1.
+21.
23.
(3.
+22.
22.
(2.
28.
(4.
+27.
28.
(4.
24
12)
48
50
19)
86
79)
16
35
02)
68
50)
90
25
3D
48
29
10)
26
90)
57
49
65)
15
(6
+15
19
(11
27
(5
+26
30
(4
26
(2
+26
20
(6
+20
19
(6
20
(7
+20
21
(6
.26
.23)
.44
.37
.03)
.42
.11)
.96
.87
.81)
.47
.84)
.06
.14
.55)
.06
.37
.28)
.87
.11)
.75
.72
.93)
7.10
(7.07)
+7.95
9.60
(7.75)
17.44
(7.41)
+17.07
19.80
(6.37)
9.46
(4.72)
+10.03
14.33
(7.58)
+14.33
13.74
(8.27)
14.56
(7.48)
+14.53
15.11
(7.74)
(continu ed)
-------
TABLE A-l. (continued)
Name of Water Body
Spring
NES LANDSAT
Serial STORE! Scene
Number Number Date
4 10-15-73
10-16-73
LANDSAT
Scene
Number**
1449-16082
1450-16140
LANDSAT MSS Mean DN
Pixels
211
214
Area
(ha)
135
137
Value and (Deviation)
GRN
39.
(2.
37.
(1.
+40.
54
18)
93
82)
15
RED
27
(2
27
(2
+28
.96
.31)
.86
.20)
.06
IR1
24
(4
24
(3
+22
.04
.31)
.72
.68)
.84
IR2
12.14
(6.64)
13.64
(5.64)
+12.19
Spri ng
o Spring
tn
Spring
Spring Arbor
8
67
118
143
10-16-73 1450-16140 1,689 1,079
10-16-73 1450-16142 151 97
10-15-73 1449-16084 845 541
10-16-73 1450-16142 790 505
10-14-73 1448-16035 24 15
10-15-73 1449-16093 27 17
40.43 31.09
(2.80) (3.31)
+41.98 +30.58
40.97 31.39
(2.50) (2.22)
+42.37 +30.81
37.32 23.74
(3.30) (3.35)
37.02 23.95
(1.99) (2.41)
+39.49 +25.02
37.45 21.99
(2.08) (2.27)
+36.82 +21.20
34.62 19.66
(1.93) (2.39)
23.79 10.51
(3.62) (5.23)
+22.26 +10.20
23.14 12.13
(4.21) (6.81)
+21.86 +11.23
21.54
(7.04)
20.21
(3.83)
+20.06
20.58
(5.47)
+20.48
19.14
(6.08)
10.90
(7.64)
8.61
(5.96)
+9.00
14.37
(7.17)
+14.36
14.22
(7.31)
Springfield
112 1742 10-15-73 1449-16084 2,435 1,558 37.48 23.96 18.02 8.18
(continued)
-------
TABLE A-l. (continued)
ro
o
01
Name of Water Body
Stephen A. Forbes
Storey
Sugar Creek (Curry)
Summerset
NES LANDSAT LANDSAT LANDSAT MSS Mean DN
Serial STORE! Scene Scene Area Value and (Deviation)
Number Number Date Number** Pixels (ha) GRN RED IR1 IR2
78 10-14-73 1448-16032 272 174 36
(3
+35
50 1751 10-16-73 1450-16142 70 45 37
(2
+39
178 10-15-73 1449-16084 59 38 45
(4
10-16-73 1450-16142 54 35 45
(2
+46
186 10-15-73 1449-16082 144 92 35
(2
10-16-73 1450-16142 147 94 31
(1
+35
.18
.95)
.59
.38
.81)
.75
.15
.55)
.96
.29)
.01
.17
.31)
.75
.68)
.65
20
(3
+19
22
(2
+23
36
(5
38
(3
+36
19
(2
18
(2
+20
.79
.99)
.98
.48
.92)
.88
.31
.57)
.11
.47)
.04
.85
.51)
.34
.20)
.66
16
(6
+16
19
(6
+19
31
(4
32
(2
+27
15
(4
14
(4
+16
.43
.63)
.55
.62
.69)
.07
.63
.85)
.42
.64)
.58
.95
.76)
.30
.40)
.42
9.
(7.
+9.
11.
(8.
+10.
15.
(4.
16.
(4.
+14.
7.
(5.
8.
(5.
+8.
24
95)
84
42
09)
78
47
83)
98
62)
30
86
43)
00
74)
61
Sunfish
Swan
126
93
10-16-73 1450-16140 89 57
10-16-73 1450-16140 35 22
38.12 28.20
(1.63) (1.66)
+40.29 +28.33
23.31 13.34
(4.96) (7.49)
+21.97 +12.00
42.51 32.57 29.99 21.59
(2.53) (2.57) (3.41) (5.25)
+43.49 +31.73 +26.08 +17.23
(continu ed)
-------
TABLE A-l. (continued)
ro
o
NES LANDSAT
Serial STORE! Scene
Name of Water Body Number Number Date
Swan
Tampier
Third
Thunderbird
Turner
Upper Smith
172 10-15-73
23 10-14-73
10-15-73
155 10-14-73
173 10-15-73
10-16-73
102 10-15-73
180 10-15-73
10-16-73
LANDSAT
Scene
Number** Pixels
1449-16082 198
1448-16023 140
1449-16082 147
1448-16023 99
1449-16082 57
1450-16140 60
1449-16062 147
1449-16084 184
1450-16142 160
LANDSAT MSS Mean DN
Area Value and (Deviation)
(ha) GRN RED IR1 IR2
127 39.
(2.
90 39.
(3.
+38.
94 39.
(3.
63 34.
(1.
+34.
37 37.
(3.
38 34.
(2.
+38.
94 39.
(1.
118 40.
(3.
102 41.
(2.
+42.
70
00)
22
46)
54
15
14)
60
58)
06
57
14)
98
14)
00
54
89)
83
21)
41
14)
69
28.
(1.
25.
(4.
+24.
25.
(3.
20.
(2.
+19.
21.
(3.
19.
(2.
+21.
27.
(1.
28.
(2.
29.
(1.
+29.
17
95)
38
3D
64
27
70)
16
36)
34
82
35)
28
65)
39
92
76)
73
93)
96
72)
70
23.
(4.
20.
(6.
+20.
21.
(5.
16.
(5.
+16.
17.
(7.
16.
(5.
+17.
24.
(3.
24.
(7.
23.
(4.
+21.
17
08)
94
62)
64
15
58)
08
48)
22
45
43)
29
52)
65
12
42)
59
80)
22
03)
91
10.
(6.
11.
(7.
+11.
12.
(7.
7.
(7.
+8.
11.
(6.
12.
(7.
+11.
11.
(5.
11.
(6.
10.
(6.
+10.
89
60)
31
57)
66
05
14)
21
08)
05
64
88)
04
19)
17
24
10)
15
52)
65
17)
29
(continued)
-------
TABLE A-l. (continued)
00
Name of Water Body
Vandal ia City
We-Ma-Tuk
West Frankfort New
West Frankfort Old
West Loon
Wi 1 dwood
NES LANDSAT LANDSAT LANDSAT MSS Mean DN
Serial STORET Scene Scene Area Value and (Deviation)
Number Number Date Number** Pixels (ha) 6RN RED IR1 IR2
27 1764 10-14-73 1448-16032 394 252 39.
(4.
+38.
33 1761 10-15-73 1449-16084 117 75 40.
(4.
10-16-73 1450-16142 86 55 38.
(3.
+40.
31 10-14-73 1448-16035 100 64 34.
(1.
+34.
30 10-14-73 1448-16035 87 56 33.
(1.
+33.
156 10-14-73 1448-16023 109 70 35.
(3.
+35.
162 10-15-73 1449-16084 124 79 39.
(3.
15
07)
47
75
41)
04
72)
24
72
85)
18
59
53)
09
66
07)
09
52
25)
24,
(4.
+23.
29.
(4.
26.
(4.
+27.
19.
(2.
+19.
18.
(2.
+18.
19.
(4.
+18.
22,
(3,
,16
,30)
,40
,45
,96)
,81
,45)
,25
,92
.21)
,09
,89
.03)
,05
,56
,22)
,73
,62
.83)
17.
(6.
+17.
30.
(12.
21.
(6.
+20.
14.
(5.
+14.
17.
(7.
+17.
13.
(7.
+13.
16.
(6.
47
78)
53
70
28)
26
31)
71
70
60)
91
42
02)
48
00
51)
30
73
79)
9.
(8.
+10.
16.
(8.
13.
(7.
+11.
7.
(6.
+8.
10.
(7.
+10.
6.
(7.
+7.
10.
(7.
46
01)
03
56
23)
12
89)
86
39
68)
20
03
78)
53
41
17)
34
24
09)
Wolf
17
10-14-73 1448-16023 702
449 36.17 21.73
(4.75) (6.87)
+35.58 +20.93
15.85 9.08
(8.82) (8.44)
+16.00 '9.69
(continued)
-------
TABLE A-l. (continued)
o
VD
NES LANDSAT
Serial STORET Scene
Name of Water Body Number Number Date
Wonder
Wo r ley
Yorkv
i ** • " j
69 1750 10-15-73
117 10-15-73
10-16-73
179 10-15-73
10-16-73
LANDSAT
Scene
Number**
1449-16082
1449-16084
1450-16142
1449-16084
1450-16142
Pixels
442
137
158
257
236
Area
(ha)
283
88
101
165
151
LANDSAT MSS Mean DN
Value and (Deviation)
URN
37
(1
39
(2
39
(1
+41
40
(3
42
(2
+43
.29
.95)
.45
.94)
.17
.9.5)
.06
.37
.51)
.19
.17)
.26
RED
23
(2
27
(2
28
(2
+28
29
(3
31
(1
+31
.90
.03)
.77
.34)
.78
.01)
.78
.39
.61)
.79
.93)
.12
IR1
19
(4
27
(6
27
(3
+24
27
(8
25
(3
+23
.75
.35)
.18
.81)
.57
.25)
.59
.23
.39)
.40
.25)
.26
IR2
9.50
(6.22)
13.00
(5.88)
13.77
(4.97)
+12.27
11.81
(5.95)
11.92
(5.53)
+11.10
Zurich
154
10-14-73 1448-16023 146
93 35.01 19.38
(2.08) (2.48)
34.46 18.54
12.08 5.24
(4.92) (5.36)
12.43 6.31
*A water body's mean DN value for a particular band was calculated by summing the band's UN value for
each pixel and then dividing the sum by the total number of pixels.
**The multispectral scanner data from LANDSAT Scene 1449-16084 are known to be in error, a consequence
of missing calibration data, and were not used in model development or for predictive purposes.
+Multispectral scanner data adjusted to a common date (October 15th) through the use of regression
models.
-------
TABLE A-2. TROPHIC INDICATOR MEAN VALUES FOR 31 WATER BODIES SAMPLED OCTOBER 16-19, 1973
Name of Water
Body
Baldwin
Bloomington
Carlyle
Cedar
Charleston
Coffeen
Crab Orchard
Decatur
DePue
East Loon
Fox
Grass
Highland Silver
Hoi iday
Horseshoe
Long
Lou Yaeger
Marie
Old Ben Mine
Pistakee
Raccoon
Rend
Sangchris
Shelbyville
Slocum
Springfield
Storey
Vandal i a
Serial
Number
105
71
14
55
16
94
127
73
3
53
60
61
77
51
131
52
95
62
32
65
80
29
11
114
54
112
50
27
NES
STORET
Number
1763
1703
1706
1759
1708
1711
1712
1714
1752
1757
1755
1756
1740
1754
1766
1725
1726
1727
1765
1733
1762
1735
1753
1739
1758
1742
1751
1764
CHLA
(yg/liter)
11.9
56.8
19.9
5.6
18.0
5.8
46.7
21.4
42.4
26.8
37.4
46.1
5.8
67.0
99.5
61.2
8.6
70.7
24.6
66.5
10.6
15.6
15.6
12.8
241.4
17.4
30.0
13.5
COND
(umhos/cm)
457
389
359
302
519
521
235
435
590
354
472
501
191
473
547
581
200
467
1376
467
188
281
399
384
622
318
408
170
SEC
(m)
0.94
0.79
0.48
2.77
0.25
1.37
0.36
0.46
0.15
0.91
0.36
0.31
0.41
0.46
0.56
0.31
0.43
0.56
0.48
0.31
0.41
0.61
0.89
0.48
0.31
0.56
0.89
0.71
TPHOS
(mg/liter)
0.045
0.044
0.091
0.053
0.207
0.055
0.114
0.143
0.499
0.087
0.212
0.347
0.107
0.173
0.098
0.785
0.150
0.286
0.930
0.216
0.145
0.065
0.047
0.063
1.330
0.095
0.125
0.175
TON
(mg/liter)
0.714.
0.707
0.742
1.105
1.200
0.469
0.843
0.590
2.020
1.380
0.970
1.773
0.651
1.200
2.375
2.074
0.651
1.896
1.690
1.635
0.900
0.971
0.547
0.595
5.940
0.595
0.980
1.019
(continued)
-------
TABLE A-2. (continued)
Name of Water
Body
Vermilion
We-Ma-Tuk
Wonder
Serial
Number
122
33
69
NES
STORE!
Number
1748
1761
1750
CHLA
(ug/liter)
27.0
8.3
198.0
COND
(pmhos/cm)
464
758
523
SEC
(m)
0.36
1.07
0.46
TPHOS
(mg/liter)
0.117
0.103
0.423
TON
(my/liter)
0.740
0.588
1.788
-------
TABLE A-3. TROPHIC INDICATOR ANNUAL MEAN VALUES FOR 31 WATER BODIES SAMPLED THREE TIMES DURING 1973
Name of Water
Body
Baldwin
Bloomington
Carlyle
Cedar
Charleston
Coffeen
Crab Orchard
Decatur
DePue
East Loon
Fox
Grass
Highland Silver
Hoi iday
Horseshoe
Long
Lou Yaeger
Marie
Old Ben Mine
Pistakee
Raccoon
Rend
Sangchris
Shelbyville
Slbcum
Springfield
Storey
Vandalia
Serial
Number
105
71
14
55
16
94
127
73
3
53
60
61
77
51
131
52
95
62
32
65
80
29
11
114
54
112
50
27
NES
STORET
Number
1763
1703
1706
1759
1708
1711
1712
1714
1752
1757
1755
1756
1740
1754
1766
1725
1726
1727
1765
1733
1762
1735
1753
1739
1758
1742
1751
1764
CHLA
(yg/liter)
11.3
26.2
17.4
5.8
12.0
7.7
59.9
43.0
58.8
22.3
63.8
83.5
5.8
51.2
182.2
49.3
10.7
39.5
31.4
75.9
19.2
23.5
19.3
17.2
221.1
13.0
17.2
11.3
COND
(ymhos/cm)
499
448
379
380
564
514
256
518
703
446
525
564
205
594
567
596
207
488
645
524
260
306
398
430
676
373
455
179
SEC
(m)
0.99
0.89
0.56
2.54
0.23
1.12
0.46
0.51
0.25
1.27
0.36
0.48
0.28
0.38
0.43
0.43
0.25
0.81
0.56
0.36
0.41
0.74
0.64
0.99
0.33
0.43
1.04
0.56
TPHOS
(rug/liter)
0.043
0.064
0.088
0.035
0.155
0.045
0.125
0.126
0.542
0.102
0.225
0.280
0.225
0.180
0.164
0.580
0.214
0.148
0.760
0.203
0.119
0.070
0.069
0.075
0.819
0.123
0.093
0.150
TON
(mg/liter)
0.720
0.717
0.885
1.086
0.785
0.533
1.162
0.682
1.478
1.239
1.823
1.874
0.960
1.441
2.832
1.638
0.674
1.476
1.308
1.680
1.091
0.960
0.584
0.654
4.398
0.601
0.976
0.882
(continued)
-------
ro
TABLE A-3. (continued)
""• " — — ' — ' • — ~ ~" " " " ~" ------ — - -
Name of Water
Body
Vermilion
We-Ma-Tuk
Wonder
Serial
Number
122
33
69
NES
STORE!
Number
1748
1761
1750
CHLA
(ng/Hter)
31.1
8.0
98.5
COND
(umhos/cm)
546
907
610
SEC
(m)
0.48
0.86
0.36
TPHOS
(mg/liter)
0.112
0.071
0.430
TON
(my/ liter)
0.821
0.730
1.599
-------
TABLE A-4. TROPHIC INDICATOR AND MULTIVARIATE TROPHIC INDEX ESTIMATES FOR 145 ILLINOIS LAKES
BASED ON SET THREE REGRESSION MODELS
ro
Trophic Indicators
Lake Name
Anderson
Apple Canyon
Argyl e
Bakers
Bangs
Bath
Big
Big
Bracken
Calumet
Canton
Carbondale
Carlyle
Catherine
Cattail
Cedar
Cedar
Centrali a
Chain
Channel
Charleston
Chautauqua
Clear
Coal City Recreation Club
Commonwealth
Edison-Dresden Nuclear
Serial
Number
34
144
157
19
56
166
134
176
145
18
36
38
14
58
125
39
55
79
83
59
16
87
86
138
137
SEC
(m)
0.3
1.4
1.1
0.2
1.0
0.3
0.4
0.3
1.7
0.3
0.5
0.2
0.5
1.1
0.8
0.5
1.4
1.5
0.3
0.7
0.3
0.2
0.3
24.9
0.3
Multivariate
Trophic Indices
CHLA TON TPHOS PC1Y5 PC1F5
(yg/liter) (mg/liter) (nig/liter) (dimensionless)
32
35
3
18
38
3
18
12
42
29
6
1
14
67
3
2
8
24
16
91
19
15
22
2
11
2.29
0.93
0.60
4.62
0.87
1.37
0.98
1.30
0.78
2.51
0.78
2.92
0.66
1.35
1.03
0.91
1.23
1.04
1.47
1.22
1.33
1.59
1.49
0.44
0.72
0.22
0.12
0.17
1.02
0.11
0.64
0.22
0.26
0.16
0.08
0.09
1.18
0.06
0.19
2.49
0.45
0.07
0.22
0.40
0.22
0.31
0.35
0.36
0.36
0.12
0.70
-1.11
-2.45
0.14
-1.20
-0.66
-0.32
0.06
-1.27
-3.55
-1.45
-0.68
-1.61
-0.17
-0.90
-0.80
-2.55
-1.15
0.43
0.89
0.78
0.53
0.70
-7.43
-0.26
2.54
0.62
-0.68
4.24
0.27
1.90
1.51
2.26
0.38
0.64
0.50
3.09
-0.33
1.32
0.73
0.97
-0.12
0.61
2.67
1.99
2.65
2.85
2.83
-4.83
0.69
(continued)
-------
TABLE A-4. (continued)
ro
i—>
en
Trophic Indicators
Lake Name
Countryside
Crab Orchard
Crane
Crooked
Crystal
Decatur
Deep
DePue
Devil 's Kitchen
Diamond
DuQuoi n
Dutchman
East Loon
Fourth Lake
Fox
Fyre
• ,/•*•
Gages
George
Glen 0. Jones
Goose
Goose (Village)
Grass
Greenville New City
Griswold
Harrisburg
Highland
Hoi iday
Serial
Number
148
127
85
149
68
73
150
3
128
57
98
49
53
151
60
170
152
108
110
101
139
61
2
158
111
153
51
SEC
(m)
0.9
0.5
0.2
0.6
1.2
0.3
2.4
0.2
4.4
1.1
0.6
2.3
1.0
0.2
0.3
1.1
1.3
1.9
3.2
0.2
1.9
0.3
0.5
0.8
0.9
1.8
0.5
Multivariate
Trophic Indices
CHLA TON TPHOS PC1Y5 PC1F5
(ng/liter) (mg/liter) (mg/liter) (dimensionless)
71
69
1
108
22
32
2
23
30
35
62
37
30
36
70
50
20
30
19
16
18
38
38
45
34
10
51
1.08
1.00
1.65
1.77
0.80
1.07
0.66
1.47
1.47
0.69
3.80
0.83
1.40
1.32
1.11
1.19
0.96
0.67
0.73
2.18
0.61
1.35
1.06
1.36
1.38
1.74
1.72
0.16
0.18
0.33
0.29
0.06
0.20
0.03
0.39
0.11
0.07
0.44
0.20
0.12
0.26
0.33
0.24
0.09
0.07
0.04
0.60
2.54
0.25
0.17
0.42
0.27
0.12
0.24
0.20
0.71
-0.61
0.20
-1.95
1.01
-4.24
1.13
-2.08
-1.53
1.48
-1.48
-0.76
1.31
1.76
-0.11
-1.99
-2.02
-2.96
1.22
-4.60
1.37
0.24
0.41
-0.12
-2.58
1.04
1.66
1.72
2.26
2.81
-0.15
2.14
-1.83
2.99
0.35
-0.09
3.35
0.20
1.18
2.87
2.41
1.52
-0.09
-0.28
-0.81
3.74
-2.49
2.44
1.68
2.00
1.59
0.05
2.49
(continued)
-------
TABLE A-4. (continued)
ro
i—'
Ol
Lake Name
Horseshoe
Jack, Swan, Grass
Keithsburg
Klnkaid
Kinneman
Lake of Egypt
Lake of the Woods
Lame-Pine
Lily
Lincoln Trail
Little Grassy
Little Swan
Liverpool
Long
Long
Lower Smith
Lyerla-Autumnal
Marie
Marion
Marshall
Matanzas
Mattoon
McCollum
McGinnis
Meredosia
Mesa
Moscow
Moses
Serial
Number
1
165
92
40
168
132
115
121
136
13
119
185
88
52
177
181
120
62
129
81
89
24
159
20
10
184
164
28
SEC
(m)
0.4
0.3
1.0
6.7
2.7
1.6
0.3
2.0
0.3
4.5
3.4
0.4
0.4
0.4
0.3
0.5
0.7
0.4
1.1
0.3
0.3
0.6
1.4
0.2
0.2
2.0
0.3
1.6
Trophic
Indicators
Multivariate
Trophic Indices
CHLA TON TPHOS PC1Y5 PC1F5
(yg/liter) (mg/liter) (mg/liter) (dimensionless)
48
39
6
28
30
15
15
5
9
30
12
5
1
58
1
0
21
80
73
26
13
8
30
104
4
22
19
50
1.81
1.38
0.94
0.40
1.42
0.59
1.75
1.98
1.50
1.35
0.95
0.95
1.31
1.51
1.44
0.92
1.17
1.29
2.27
1.55
1.34
0.62
0.85
10.59
1.84
0.76
1.29
2.44
0.98
0.30
2.34
0.01
6.09
0.04
0.74
41.56
0.98
0.12
0.04
0.21
1.05
0.30
0.40
0.36
0.92
0.26
0.48
0.48
0.28
0.05
0.10
3.63
0.26
0.04
0.35
0.56
1.32
1.21
-1.15
-4.59
-6.18
-2.55
0.55
-0.59
-1.71
-2.11
-3.20
-0.71
-1.53
1.33
-1.00
-1.65
-0.64
1.59
0.91
1.08
0.23
-1.79
0.37
3.69
-1.14
-2.41
0.17
0.22
3.19
2.92
0.40
-2.56
-2.26
-0.76
2.99
0.65
1.38
0.23
-1.68
1.27
1.24
2.39
1.82
0.64
1.53
2.73
2.64
3.05
2.41
-0.16
0.37
6.85
2.06
-0.42
2.30
2.10
(continued)
-------
TABLE A-4. (continued)
INJ
I—'
—I
Trophic Indicators
Lake Name
Mound
Murphysboro
New Pittsfield
Nipper sink
Old Ben Mine
Olney East Fork
Olney New
Open Pond
Otter
Pana
Paradise
Paris Twin
Pekin
Petite
Pierce
Pinckneyville
Pistakee
Powerton Cooling
Quiver
Raccoon
Rend
Rice
Round
Saganashkee
Sahara Coal Company
Sam Dale
Sam Parr
Sand
Sanganois
Serial
Number
163
37
100
63
32
106
107
174
167
113
15
25
116
64
130
99
65
182
90
80
29
35
66
21
175
124
41
147
9
SEC
(m)
0.4
1.7
0.3
0.3
0.6
1.7
1.6
1.3
0.7
0.9
0.4
1.5
0.3
0.4
1.0
0.6
0.4
0.5
0.3
0.4
0.5
0.3
1.4
0.4
0.7
1.0
1.6
1.6
0.3
CHLA
dig/liter)
50
34
1
51
41
29
38
2
15
4
22
36
2
50
54
48
67
78
9
16
12
70
2
76
267
66
58
21
3
TON
(mg/liter)
1.34
1.25
1.15
1.27
1.35
0.66
1.05
0.87
1.00
0.67
0.92
1.62
1.30
1.08
0.95
0.91
1.24
0.97
1.03
0.94
0.85
1.40
0.66
1.62
2.23
1.90
1.63
0.99
1.67
Multivariate
Trophic Indices
TPHOS PC1Y5 PC1F5
(mg/liter) (dimensionless)
0.33
0.20
0.15
0.28
0.49
0.06
0.56
3.61
0.62
0.23
0.15
0.49
0.55
0.18
0.13
0.13
0.25
0.22
0.22
0.10
0.10
0.30
0.05
0.33
30.82
0.23
1.49
0.08
0.65
1.18
-0.95
-1.54
1.58
0.56
-2.00
-0.60
-1.72
-0.48
-1.99
0.01
-0.31
-0.69
0.77
-0.32
0.10
1.16
0.94
-0.05
-0.60
1.07
1.83
-3.65
1.84
-6.51
0.48
0.37
-2.22
-1.50
2.72
0.92
1.10
2.48
2.16
-0.28
0.95
-0.17
1.23
-0.23
1.50
1.45
1.81
1.99
1.19
1.43
2.16
1.79
1.77
1.23
0.35
3.11
-1.54
3.14
-2.13
2.23
1.86
-0.18
1.73
(continued)
-------
TABLE A-4. (continued)
ro
i—i
oo
Trophic Indicators
Lake Name
Sara
Sawmill
Senachwine
Shelbyville
Skokie Lagoons
Sloe urn
Snyder's Hunting
South Wilmington
Spri ng
Spri ng
Spri ng
Spri ng
Spring Arbor
St Mary's
Stephen A. Forbes
Storey
Sugar Creek
Summerset
Sunfish
Swan
Swan
Tampier
Third
Thunderbird
Turner
Upper Smith
Vandal i a
We-Ma-Tuk
West Frankfort New
Serial
Number
26
103
104
114
22
54
142
140
4
8
67
118
143
146
78
50
178
186
126
93
172
23
155
173
102
180
27
33
31
SEC
(m)
2.4
0.2
0.2
0.6
0.6
0.3
0.9
7.9
0.2
0.2
0.4
0.4
2.6
1.4
1.4
1.0
0.3
0.8
0.4
0.8
0.2
0.6
0.7
2.2
0.2
0.3
1.1
0.6
1.1
CHLA TON
(yg/liter) (mg/liter)
21
51
58
7
5
253
15
1
30
16
3
81
31
67
34
45
15
64
11
9
40
30
79
11
46
19
18
10
26
1.15
3.06
1.74
0.80
1.64
6.47
1.67
0.44
2.18
1.30
0.95
1.29
1.36
0.92
1.10
0.84
2.10
1.51
1.10
0.92
2.09
1.07
2.37
0.64
2.71
1.09
0.70
0.83
1.49
Multivariate
Trophic Indices
TPHOS PC1Y5 PC1F5
(mg/liter) (dimensionless)
0.15
0.91
0.45
0.07
5.46
1.48
0.96
0.20
1.01
0.28
0.21
0.26
0.97
0.32
0.11
0.16
1.18
0.19
0.42
3.09
0.67
0.41
0.29
0.10
0.95
0.20
0.07
0.25
0.11
-1.82
3.00
2.72
-1.65
-0.64
4.31
-0.15
-6.15
1.73
0.73
-0.71
1.53
-0.80
-0.70
-1.14
-0.68
-2.00
0.33
-0.00
-2.84
2.11
-0.03
1.18
-2.67
2.51
-0.04
-1.70
-0.91
-1.17
0.25
5.04
3.96
0.03
1.67
6.78
1.45
-3.75
3.82
2.61
1.27
2.81
0.75
1.14
0.74
0.95
1.72
1.83
1.81
-0.52
4.00
1.65
2.73
-0.84
4.59
1.89
-0.01
0.81
0.92
(continued)
-------
TABLE A-4. (continued)
ro
Trophic Indicators
Lake Name
West Frankfort Old
West Loon
Wolf
Wonder
Worl ey
Yorkey
Zurich
Serial
Number
30
156
17
69
117
179
154
SEC
(m)
1.2
2.0
1.2
0.4
0.3
0.3
1.6
CHLA TON
(ug/liter) (nig/liter)
79
3
13
77
23
9
1
3.22
0.85
0.98
1.74
1.92
1.29
0.99
Multi
Trophic
variate
Indices
TPHOS PC1Y5 PC1F5
(my/liter) (dimensionless)
0.40
0.04
0.08
0.37
0.88
0.30
0.09
0.98
-3.74
-1.80
2.03
1.05
-0.13
-3.86
2.90
-1.21
0.21
3.37
3.37
2.14
-1.30
-------
TABLE A-5. GROUP AND INDIVIDUAL LAKE WATER QUALITY AND ANCILLARY DATA VALUES FOR SELECTED
ILLINOIS WATER BODIES CLASSIFIED BY COMPLETE LINKAGE CLUSTERING ON FOUR SPECTRAL
ATTRIBUTES (CRN, RED, IR1, IR2)
ro
ro
o
LAKE NAME -No.
Centra Ha -079
Harrlsburg -111
Fyre -170
Apple Canyon -144
McCullom -159
Dutchman -049
Thunder-bird -173
Bracken -145
la mean
Countryside -148
Pierce -130
George -108
Olney East Fork -106
Lake of Egypt -132
Vandalla -027
Ib mean
Bangs -056
Crystal -068
Channel -059
Diamond -057
Long -052
P1 stakes -065
Rend -029
Shelbyvllle -114
Holiday -051
Ic mean
K1nca1d -040
Id mean
Cluster 1 mean
SPECTRAL INDEX
SIGNATURES
Raftoctanca Scaled %
6RN RED IR1 IR2
16 18 26 32
14 20 27 28
20 19 29 28
21 18 24 24
23 20 24 24
23 16 28 34
25 21 24 32
30 21 31 30
22 19 27 29
25 20 27 19
26 22 26 19
30 21 24 24
31 22 23 22
33 26 22 22
30 28 24 22
29 23 24 21
21 16 12 5
20 17 13 10
22 18 21 8
30 22 16 8
17 25 19 6
23 26 22 7
23 29 14 6
22 28 14 10
13 23 22 14
21 23 17 10
50 19 17 15
50 19 17 15
25 22 22 19
GROUND TRUTH
(1977)
Suipendtd
wlidj me/I
Tot Vol.
15 3
9 6
4 2
9 4
11 10
7 7
3 0
4 2
8 3
7 4
16 14
3 3
14 11
8 8
10 9
1 0
1 0
10 6
S«c.
Depth
m.
.71
.74
1.65
1.04
.68
1.40
2.71
1.59
.90
1.40
.60
2.26
.80
.87
1.13
1.97
1.97
1.27
OBSERVED PROBLEMS
1- Minimum,4-Savare
ME < I "S.
332
2 1 3
2 2 3
223
2 2 3
1 1 3
1 2 2
2 2 3
1.9 1.9 2.7
2 3 3
1 2 2
1 2 3
2 1 3
4 2 1
2.0 2.0 2.4
223
1 2 2
1 3 2
2 3 2
232
3 3 1
2 1 2
2 2 1
3 3 1
2.0 2.4 1.8
1 1 2
1 1 2
1.9 2.1 2.3
NES DATA
1973
MC. chla
m. ug/l
-
.71 13
.71 13
.31 61
.31 66
.61 15
.48 12
.43 39
-
.49 33
MORPHOLOGY
Depth) matert)
Mam Max.
2.1 7.5
3.1 9.2
4.3 11.0
9.2 21.4
1.2 2.9
2.3 8.8
10.7 18.3
4.2 10.1
4.6 11.2
2.3 3.7
3.8 11.0
7.0 17.7
4.6 12.2
5.6 15.9
4.2 11.3
HYDRAULIC FACTORS
U I* il I!
11 II iJ II
4.7 .8 .18 24
26.5 .7 .17 10
65.1 3.2 .07 4
33.8 1.7 .10 4
14.9 .7 .27 17
25.8 .7 .19 10
7.5 .4 .35 68
12.6 .6 .26 20
26.0 1.0 .20 12
30.3 1.5 .18 8
7.6 .4 .43 50
24.5 1.2 .16 5
42.8 1.4 .07 12
76.3 2.2 .03 5
16.2 .6 .18 8
4.6 12.0 32.9 1.3 .18 9
4.2 7.6 130.1 6.4 .04 2
4.1 12.5 30.3 1.5 .01 8
4.2 12.2 - ...
2.9 7.3 109.0 5.4 .05 2
4.0 9.2 5.5 .3 .45 46
1.8 9.2
3.0 9.5 18.1 .6 .05 35
5.8 19.8 9.7 .4 .06 13
2.7 5.8 2.1 .1 .10 296
3.6 10.3
8.6 24.4
8.6 24.4
4.4 11.6
43.5 2.0 .11 58
56.9 1.7 .04 7
56.9 1.7 .04 7
34.9 1.5 .16 30
BIOTA
§1
*i
<°
19.7
.9
.1
6.9
3.4
1.4
.9
.2
.4
1.2
.6
.1
2.6
2.5
1.4
.3
.3
2.5
(continued)
-------
TABLE A-5. (continued)
LAKE NAME -No.
SPECTRAL INDEX
SIGNATURES
Reflectance Scaled %
GRN RED IR1 IR2
GROUND TRUTH
(1977)
Suspended
solids mg/l
Tot. Vol.
Sec.
Depth
m.
OBSERVED PROBLEMS
1- Minimum, 4-Severe
NES DATA
1973
tec. chla
m. "9/1
MORPHOLOGY
Depth! meters)
Mean Max.
HYDRAULIC FACTORS
1! 1! II ll
BIOTA
Algal count
1000/ml.
ro
ro
Petite
Pinckneyvnie
Marie
Crab Orchard
Powerton Cooling
Grass
Nlppersink
2a mean
Saganashkee
Wonder
Greenville New C
R1ce
Spring
Mound
2b mean
Grlswold
Old Ben Mine
Olney New
St. Marys
Storey
2c mean
Sam Parr
Spring Arbor
Snyders Hunting
2d mean
Cluster 2 mean
-064
-099
-062
-127
-182
-061
-063
-021
-069
-002
-035
-118
-163
-158
-032
-107
-146
-050
-041
-143
-142
30
34
30
30
33
22
25
29
24
24
27
36
36
34
30
20
24
22
44
37
29
15
9
8
11
27
34
32
32
29
30
35
35
32
29
30
31
39
35
38
34
24
29
20
27
30
27
15
13
23
17
29
28
28
31
24
24
24
24
26
33
35
28
38
36
40
35
34
38
36
44
34
37
41
32
33
35
33
14
15
13
8
6
9
7
10
18
19
19
16
16
24
19
32
34
41
35
27
34
48
47
44
46
23
.
3
-
26
-
-
-
14
54
49
15
-
-
-
39
27
-
4
-
15
15
3
-
-
3
22
.
1
-
17
-
.
-
9
33
16
6
-
-
-
18
21
-
2
-
9
11
2
-
-
2
12
.
2.49
-
.52
.
-
-
1.51
.14
.31
.73
-
-
-
.39
.47
-
1.03
-
.59
.70
1.65
-
-
1.65
.88
3
1
2
3
1
4
4
2.6
1
3
2
3
4
-
2.6
2
3
2
-
3
2.5
2
2
2
2.0
2.5
3
2
3
3
2
3
3
2.7
2
3
2
3
2
-
2.4
2
3
2
-
3
2.5
3
2
-
2.5
2.6
1
3
2
2
1
3
1
1.9
2
3
1
1
2
-
1.8
3
3
3
-
3
3.0
2
3
4
3.0
2.3
_
.
.56
.36
_
.31
.40
_
.46
_
.
.
-
.46
„
.49
-
-
.49
„
_
-
-
.43
_
_
70
47
_
46
54
_
198
_
-
.
-
198
_
24
.
-
24
_
.
-
-
77
2.3
3.7
2.8
2.8
0.8
.9
2.2
1.2
3.8
4.0
.8
1.4
.6
2.2
2.2
1.2
3.4
3.1
4.2
2.8
3.1
3.7
1.2
2.7
2.4
6.7
8.5
10.7
7.5
1.8
2.1
6.2
2.7
10.4
7.5
1.4
3.4
-
5.1
3.7
1.5
9.8
3.7
11.0
5.7
7.0
15.3
2.4
8.2
6.1
_
19.8
_
17.7
_
.
18.7
11.0
12.7
13.9
17.4
44.1
-
19.8
29.1
_
21.5
30.6
12.4
23.4
13.9
73.2
6.8
31.2
23.1
_
.6
_
.5
_
_
.6
.5
.6
.6
.9
2.1
-
1.0
1.4
.
.7
1.5
.6
.9
.5
2.1
.2
1.0
.9
_
.20
.08
.
_
.14
.34
.08
.05
.06
.
-
.13
.06
.
.06
.05
.09
.07
.09
.07
.55
.25
.14
-
26
22
„
_
24
58
20
9
37
14
-
23
9
.
24
8
21
14
29
5
75
14
25
_
.4
25.6
_
_
13.0
1.8
2.2
11.8
.
-
5.3
.4
_
3.8
5.3
3.2
9.5
_
-
9.5
7.7
(continued)
-------
TABLE A-5. (continued)
LAKE NAME -No.
SPECTRAL INDEX
SIGNATURES
R«f toctanc* Scaled %
CRN RED IR1 IR2
GROUND TRUTH
(1977)
Suipendtd
K>lid» mg/l
Tot Vol.
SM.
Depth
m.
OBSERVED PROBLEMS
1- Mintmum,4-S«v«fe
»? < Z a
NES DATA
1973
we chta
m. ufl/l
MORPHOLOGY
D«pth(rmtwt)
Mom Max.
HYDRAULIC FACTORS
•
S« I, li |l
II 11 it II
BIOTA
IE
if
ro
ro
ro
East Loon
Sam Dale
Summerset
Third
Glen 0. Jones
Mesa
Stephen A Forbes
Wolf
3a mean
Devils Kitchen
Lincoln Trail
Little Grassy
3b mean
Marlon
Paris East
Moses
Murphysboro
Sara
DuQuoln
West Frankfort 0
3c mean
Sand
West Loon
Deep
Zurlck
Round
Catherine
Gages
Cedar
West Frankfort N
Highland
3d mean
Cluster 3 mean
-053
-124
-186
-155
-110
-184
-078
-017
-128
-013
-119
-129
-025
-028
-037
-026
-098
-030
-147
-156
-150
-154
-066
-058
-152
-055
-031
-153
10
8
12
6
18
20
14
14
13
3
5
10
6
7
8
2
11
9
0
1
5
13
12
18
8
20
12
14
6
7
0
11
10
14
11
13
11
10
14
14
18
13
0
1
5
2
12
14
9
13
10
13
6
11
9
9
10
8
16
9
12
9
10
2
9
10
17
22
17
16
15
15
20
17
18
14
16
10
13
29
28
27
23
20
24
24
25
8
4
3
0
5
13
9
7
12
2
6
15
17
19
9
10
20
17
21
20
17
25
27
17
23
30
37
35
29
30
23
25
30
6
6
4
0
3
5
5
9
11
3
5
16
2
16
.
.
6
.
12
8
9
0
2
2
1
6
7
12
18
2
13
3
9
10
2
2
5
9
.
7
3
3
-
6
7
2
7
.
..
6
.
9
8
6
0
2
1
1
4
4
9
15
1
12
3
7
8
2
2
5
9
.
12
3
2
-
5
6
1.53
.81
.
1.00
.
.89
1.47
1.14
4.52
2.20
3.20
3.31
.84
1.18
.47
.51
1.83
.59
1.31
.95
.84
2.08
2.15
1.24
.92
.
.64
1.30
1.42
-
1.32
1.43
1
3
1
2
1
1
2
2
1.6
1
1
1
1.0
3
2
2
2
1
3
1
2.0
2
1
1
1
2
1
2
1
1
1
1.3
1.5
3
3
2
4
3
1
2
1
2.4
1
2
1
1.3
4
2
2
3
2
2
3
2.6
2
2
1
1
2
3
3
2
2
2
2.0
2.2
3 .91
3
2
2
2
4
3
3
2.7 .91
1
3
1
1.7
3
2
3
1
3
3
4
2.7
3
3
3
2
2
2
3
3 2.77
3
3
2.7 2.77
2.6 1.84
27
,
.
.
.
-
.
-
27
.
.
-
-
m
.
.
-
-
.
-
-
_
.
-
.
-
_
-
6
-
-
6
17
1.8
2.4
6.1
5.7
4.4
4.0
4.3
2.1
3.8
11.0
3.8
7.8
7.5
4.3
3.1
2.7
4.3
6.1
2.1
2.4
3.6
3.0
6.4
5.9
3.3
3.5
5.1
3.2
1.2
2.4
3.2
3.7
4.1
7.9
5.5
14.3
18.9
9.1
7.9
8.5
6.4
9.8
27.4
12.5
23.5
21.1
7.0
8.1
6.4
9.8
15.8
9.1
6.1
8.9
10.1
11.3
15.8
9.8
10.7
12.2
14.6
12.2
4.6
10.7
11.2
11.3
4.7
10.2
38.6
90.0
48.1
67.2
16.2
-
39.3
76.1
26.2
81.2
61.2
22.6
4.0
23.5
35.4
47.2
7.4
13.9
22.0
145.0
211.3
196.2
60.5
150.7
.
36.1
48.7
10.7
50.3
90.9
53.8
.2
.3
1.9
4.4
1.3
2.2
.6
-
1.5
2.1
.9
2.3
1.8
.6
.2
.7
1.1
1.9
.2
.4
.7
7.1
10.4
9.7
3.0
7.4
.
1.7
1.7
.3
2.5
4.2
2.3
.70
.35
.95
.43
.10
.07
.18
-
.40
.04
.18
.04
.09
.18
.74
.19
.13
.07
.43
.38
.30
.03
.03
.03
.07
.03
-
.15
.10
.32
.43
.13
.32
54
50
10
3
5
6
31
-
9
5
15
5
6
17
32
27
11
5
86
46
32
2
1
1
4
2
.
7
5
59
5
7
19
2.3
7.2
-
-
.8
-
2.9
.8
2.8
.1
.2
1.0
.4
.4
1.5
23.7
6.9
.2
1.0
2.8
5.2
.1
1.3
.2
.5
.5
-
1.5
.2
.2
-
.6
2.3
(continued)
-------
TABLE A-5. (continued)
LAKE NAME -No.
SPECTRAL INDEX
SIGNATURES
Reflectance Scaled %
CRN RED IR1 IR2
GROUND TRUTH
(1977)
Suspended
solidi mg/l
Tot Vol.
Sec.
Depth
m.
OBSERVED PROBLEMS
1- Minimum,4-Severe
NES DATA
1973
sec. chla
m. ug/l
MORPHOLOGY
Depth(meteri) '
Mean Max.
HYDRAULIC FACTORS
U It M M
£ §• » E ^-£ =-i
S 8 O '.P > a* u. *
BIOTA
Algal count
1000/ml.
ro
r\>
CO
Charleston
Spring
Quiver
Little Swan
Spring
Chain
Clear
Moscow
4a mean
Oepue
Marshall
B1g
Fourth
Jack, Swan, Gras
4b mean
Anderson
Upper Smith
Canton
Raccoon
Senachwlne
4c mean
Decatur
Fox
Mattoon
Paradise
Commonwealth Edi
Carlyle
4d mean
Cluster 4 mean
-016
-008
-090
-185
-067
-083
-086
-164
-003
-081
-134
-151
-165
-034
-180
-036
-080
-104
-073
-060
-024
-015
-137
-014
48
50
47
52
52
53
52
52
51
45
44
47
47
46
46
57
54
57
63
49
56
34
34
46
34
46
68
44
49
55
57
55
56
58
56
55
53
56
53
50
48
53
49
51
56
54
55
59
57
56
43
38
45
44
53
59
47
53
47
47
43
45
45
53
51
50
48
48
50
43
43
45
46
46
45
40
40
44
43
29
27
28
34
24
27
34
41
25
23
26
29
29
29
26
29
27
26
29
28
17
21
24
18
24
24
16
10
18
12
4
18
20
5
4
11
21
63
104
127
-
17
-
.
-
78
_
120
-
-
-
120
95
-
15
22
97
57
36
.
12
33
.
-
27
62
11
83
29
.
4
.
_
-
32
_
27
-
-
-
27
37
-
7
6
40
23
12
.
4
11
_
-
9
23
.23
.19
.07
.
.71
-
.
-
.30
_
.11
-
.
-
.11
.12
-
.75
.59
.10
.39
.28
_
.55
.23
.
-
.34
.28
4
3
4
2
3
4
4
3.4
3
4
-
2
4
3.3
3
2
3
3
4
3.0
4
3
2
4
_
2
3.0
3.2
1
2
2
3
2
2
2
2.0
3
1
-
1
-
1.7
3
2
3
2
2
2.4
2
3
2
1
_
2
2.0
2.1
1
2
1
3
1
1
1
1.4
1
2
.
3
1
1.7
1
2
2
3
1
1.8
1
1
2
2
_
1
1.4
1.6
.25
.
.
.
-
.
.
-
.25
.15
.
.
-
-
.15
_
-
-
.41
-
.41
.46
.36
-
_
_
.48
.43
.37
18
_
.
-
.
„
-
18
58
_
.
1.2
-
58
„
-
-
10
-
10
43
38
.
_
.
20
33
31
.9
.9
.8
2.4
3.2
-
.5
1.4
.9
.3
.
1.8
1.2
.8
1.1
1.2
4.3
1.2
.3
1.6
2.2
1.7
3.2
2.3
1.8
3.3
2.4
1.7
4.3
4.3
1.2
9.5
10.7
.
.9
5.1
1.8
1.5
_
45.0
-
1.6
1.6
1.8
10.7
3.7
1.5
3.9
7.0
6.7
10.7
7.0
4.9
10,. 7
7.8
4.9
.1
_
10.9
6.8
.
_
.
5.9
.
_
.
2.2
-
45.0
43.4
-
11.1
3.9
-
19.5
1.1
.4
9.6
3.5
.
5.0
3.9
11.7
.0
_
.5
.3
_
_
_
.3
.
_
_
.12
-
2.2
2.1
-
.6
.1
-
.9
.1
_
.4
.1
_
.2
.2
.6
8.50
_
.30
.42
-
_
_
3.07
.
_
.
6
-
.12
.10
-
.26
.60
-
.32
5.00
_
.23
.80
_
.06
1.52
1.49
2,083
.
35
56
.
»
_
725
.
»
.
.
-
6
18
-
57
129
-
98
112
»
13
37
.
26
54
234
.2
5.4
.2
-
2.4
-
.
.
2.1
.
1.0
_
-
1.0
71.3
-
3.7
1.5
.7
19.3
5.8
_
6.2
.3
.
-
4.1
6.6
(continued)
-------
TABLE A-5. (continued)
ro
no
LAKE NAME -No.
Argyle -157
Pana -113
He-Ma-Tuk -033
Cedar -039
Sunflsh -126
5a mean
Dtter -167
f ampler -023
Horseshoe -001
5b mean
Sawmill -103
'urner -102
Swan -172
Spring -004
.ake of the Wood -115
Jorley -117
-yerala-Autumnal -120
>ooked -149
>c mean
:atta11 -125
Celthsburg -092
Ipen Pond -174
ikokle Lagoons -022
id mean
1cG1nn1s -020
i locum -054
.aRue-Plne -121
ie mean
Muster 5 mean
SPECTRAL INDEX
SIGNATURES
Reflectance Scaled %
CRN RED IR1 IR2
44 42 37 37
44 44 39 36
40 44 39 33
40 50 44 39
40 48 45 34
42 46 41 36
36 39 45 41
34 36 41 34
32 36 51 37
34 38 46 37
37 49 55 26
36 47 55 29
37 48 51 27
36 47 55 35
50 54 58 36
45 50 58 36
47 42 55 45
66 38 56 24
44 47 55 32
35 42 54 58
35 39 52 58
40 41 57 65
42 43 69 64
38 41 58 61
30 31 74 39
39 22 65 22
24 27 77 95
31 27 72 52
40 42 54 41
GROUND Tl
11977)
Suspended
wild. my/I
Tot Vol.
4 2
4 3
33 21
14 9
27 14
27 18
27 16
31 16
31 16
45 26
114 72
37 22
65 40
25 21
25 21
32 22
*UTH
Sec.
Depth
m.
1.90
1.94
.33
1.37
.33
.33
.33
.30
.30
.08
.09
.25
.14
.34
.34
.50
OBSERVED PROBLEMS
1- Minimum,4-Severe
i! * 11
*6 < f-fi.
322
332
223
1 2 1
2 2 1
2.2 2.2 1.8
3 - 3
1 2 1
333
2.3 2.5 2.3
4 1 1
1 3
3 1 1
4 2 1
4 3 1
3 3 3
232
3.3 2.0 1.7
323
332
4 4 1
3.3 3.0 2.0
1 3 4
4 4 1
244
2.3 3.7 3.0
2.7 2.5 2.0
NES DATA
1873
we. chla
m. u«/l
1.07 8
_
-
-
.31 241
.31 241
.69 125
MORPHOLOGY
Depth(meteri)
Mem Max.
5.3 11.6
4.2 11.6
1.8 7.6
7.0 12.2
1.8
4.6 9.0
.6 1.2
1.5 4.9
1.1 1.7
1.1 2.6
1.4 3.4
.5 .9
.5 .9
.9 3.1
4.3 9.8
1.5 3.6
.3 1.5
.8 1.8
.8 1.8
.6 1.7
.5 1.2
1.2 2.7
.6 1.8
.8 2.0
1.8 4.3
HYDRAULIC FACTORS
1* !' '1 Si
h II i£ II
12.1 .6 .45 32
.2 .0 4.00 758
2.2 .1 .14 289
39.4 1.1 .07 10
13.5 .5 1.17 272
15.9 .8 .28 40
15.9 .8 .28 40
44.1 2.1 .08 14
15.9 .5 .27 24
30.0 1.1 .12 19
9.0 .5 .48 70
4.8 .2 .70 53
6.9 .3 .59 62
13.1 .7 .72 143
BIOTA
Algal count
1000/ml.
.2
13.3
1.3
4.9
8.3
14.2
11.3
2.9
2.9
.7
7.7
13.8
7.4
30.5
30.5
11.4
(continued)
-------
TABLE A-5. (continued)
ro
LAKE NAME -No.
B1g -176
fatanza -089
Vorkey -179
"hautauqua -087
Goose -101
Bath -166
Pekln -116
Lower Smith -181
Crane -085
Long -177
-Iverpool -088
Bakers -019
6a mean
L1ly -136
Sanganols -009
Carbondale -038
Sugar Creek -178
Meredosla -010
New Plttsfleld -100
6b mean
Coal City Recrea -138
South Wilmington -140
Soose Village -139
Swan -093
6c mean
Calumet -018
dnneman -168
Sahara Coal Comp -175
6d mean
Cluster 6 mean
RC:mr/4848/sp
SPECTRAL INDEX GROUND TRUTH OBSERVED PRO
SIGNAIUHtS |1»//| 1.Mlnl««,m.«
Raf lactanca Scaled * Suspended Sac. •
iolidsmg/1 Depth ^f &
CRN RED IR1 IR2 Tot Vol. m. «l <
57 60 50 26
56 60 50 26 68 20 .11 4 2
57 60 51 29 - - - 2
57 61 54 26 - - - 4 2
53 61 59 28 4 1
56 59 58 39
56 61 57 38 - - - 4 1
57 61 51 39 - - - 2 1
67 72 58 30 - - - 4
65 67 58 35
60 66 63 48 4
69 59 76 35 - - - 1 2
59 62 57 33 68 20 .11 3.2 1.5
65 75 65 47 3 2
68 75 64 41 - - - 4 2
67 72 73 42 2 2
72 80 72 48 - - -
75 84 60 29 150 39 .11 4 3
71 76 51 28 - - - 3 2
70 77 68 39 150 39 .11 3.2 2.2
85 49 48 61 - - - 2 1
75 49 44 53 2 1
66 63 60 70 - - - 2 1
58 62 65 65 4 2
71 56 54 62 - - - 2.5 1.3
100 100 60 27 - - - 4 1
89 67 81 80
83 97 100 100
91 88 80 69 - - - 4 1
67 68 62 44 58 17 .11 3.1 1.6
BLEMS NESDATA MORPHOLOGY HYDRAULIC FACTORS
— ueptn(niater») || g 5 a is o>5
|| tac. chla filfc * £• 1 2
I •& m. ug/l Maan Max. sffi'S > 5? E £
2 - 1.5 2.4
1 - - 1.2 4.6 - - - -
1 - .6 1.2
1 . . ....
1 - - .5 .9 - - - -
2 .6 1.2
1 - .6 1.2
4 - 1.1 4.9 23.1 1.1 .15 11
BIOTA
s
§1
ll
-
1.5 - - .9 2.3 23.1 1.1 .15 11
1 - - 1.2 2.1
1 - - -
2 - 2.4 6.4 17.2 .5 .25 22 .2
1 .8 1.5 164.4
1 - 3.8 10.3 12.7 - .25 30 .8
1.2 - - 2.0 5.0 14.9 .5 .25 26
2 - - 8.7 17.3 - ...
1 . . . . . ...
1 - 7.6 19.8 - ...
1 - - - - .3 .01
1.3 - - 8.0 18.6 .3 .01
1 - 2.4 3.1
1 - - 2.4 3.1 - - - -
1.4 - - 2.4 5.5 13.2 .56 .22 21
55.1
-
-
55.1
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing}
1. REPORT NO.
EPA-600/3-79-123
3. RECIPIENT'S ACCESSION-NO.
TITLE AND SUBTITLE
TROPHIC CLASSIFICATION OF SELECTED ILLINOIS WATER
BODIES: Lake Classification Through Amalgamation of
LANDSAT Multispectral Scanner and Contact-Sensed Data
5. REPORT DATE
December 1979
6. PERFORMING ORGANIZATION CODE
. AUTHOR(S)
D. H. P. Boland, D. J. Schaeffer*, D. F. Sefton*,
R. P. Clarke*, and R. J. Blackwell**
8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Environmental Monitoring and Support Laboratory
Office of Research and Development
U.S. Environmental Protection Agency
Las Vegas, Nevada 89114
10. PROGRAM ELEMENT NO.
1BD613
11. CONTRACT/GRANT NO.
12. SPONSORING AGENCY NAME AND ADDRESS
U.S. Environmental Protection Agency—Las Vegas
Office of Research and Development
Environmental Monitoring and Support Laboratory
Las Vegas, Nevada 89114
13. TYPE OF REPORT AND PERIOD COVERED
Final 1976-1978
14. SPONSORING AGENCY CODE
EPA/600/07
15. SUPPLEMENTARY NOTES
Jointly sponsored with the Division of Water Pollution Control, Illinois Environmental
Protection Agency, Springfield, IL 62706 *Illinois EPA **Jet Propulsion Laboratory
16. ABSTRACT
A project was initiated to determine the feasibilwty of assessing and classifying
a group of Illinois lakes through the utilization of a combination of contact- and
satellite-acquired data. LANDSAT multispectral scanner (MSS) digital multidate data
for 145 Illinois lakes were extracted from computer-compatible tapes and adjusted
through regression analysis to a common acquisition date. Next, MSS lake pixel counts
were converted to lake surface area estimates. Regression models employing transformed
Mss bands as independent variables were developed for the estimation of several water
quality parameters and two multivariate trophic state indices. The water quality
parameter estimates were then used to develop lake rankings that, when evaluated, were
found to be in general agreement with ancillary data. Complete linkage-based cluster
analyses of the raw MSS data and the LANDSAT-derived water quality parameter estimates
for the 145 lakes resulted in the identification of physically significant lake groups,
each of which may be characterized by its overall quality, physical and chemical
properties, biology, watershed factors, morphology, and use impairment. LANDSAT, when
used with the appropriate calibration data, is capable of providing a comprehensive,
objective, rapidly acquired, synoptic view of lacustrine quality and use impairment.
It is a flexible, cost-effective means for monitoring lakes on a Statewide basis.
7.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.IDENTIFIERS/OPEN ENDED TERMS
c. cos AT I Field/Group
water quality, limnology, lakes, artificial
satellites, remote sensing, multivariate
analysis, mapping, classifiers
eutrophication, LANDSAT
image processing,
trophic state index,
multispectral scanner,
Illinois lakes,
lake classification
08H
20F
8. DISTRIBUTION STATEMENT
RELEASE TO PUBLIC
19. SECURITY CLASS (ThisReport)'
UNCLASSIFIED
21. NO. OF PAGES
244
2O. SECURITY CLASS (This page)
UNCLASSIFIED
22. PRICE
EPA Form 222O-1 (9-73)
*U.S. GOVERNMENT PRINTING OFFICE: 1980-685-124
------- |