WATER POLLUTION CONTROL RESEARCH SERIES • 15080HOL 01/72_
A Feasibility Demonstration of an Aerial
Surveillance Spill Prevention System
       U.S. ENVIRONMENTAL PROTECTION AGENCY

-------
          WATER POLLUTION CONTROL RESEARCH SERIES
The Water Pollution Control Research Series describes the
results and progress in the control and abatement of pollution
in our Nation's waters.  They provide a central source of
information on the research, development,  and demonstration
activities in the water research program of the Environmental
Protection Agency, through inhouse research and grants and
contracts with Federal, State, and local agencies,  research
institutions, and industrial organizations.

Inquiries pertaining to Water Pollution Control Research
Reports should be directed to the Chief, Publications Branch
(Water), Research Information Division, R&M,  Environmental
Protection Agency, Washington, DC  20460.

-------
A FEASIBILITY DEMONSTRATION OF AN AERIAL  SURVEILLANCE

               SPILL PREVENTION  SYSTEM
                         by
                    Robin  I. Welch
                Allan  D.  Marmelstein
                    Paul M. Maughan
                     Project Officer
                      John Riley

                      for the

          OFFICE OF  RESEARCH AND  MONITORING

            ENVIRONMENTAL  PROTECTION  AGENCY
                 Project  #15080  HOL
                Contract  #68-01-0145
                    January  1972

-------
              EPA Review Notice
This report has been reviewed by the Environmental
Protection Agency and approved for publication.
Approval does not signify that the contents neces-
sarily reflect the views and policies of the
Environmental Protection Agency, nor does mention
of trade names or commercial products constitute
endorsement or recommendation for use.

-------
                              ABSTRACT


Acquisition and interpretation of multispectral aerial photography and
thermal infrared imagery were performed to evaluate their utility in an
aerial surveillance spill prevention system.  The San Francisco Bay area
was used as a test site; major sub-areas were delineated which contained
facilities and activities that might lead to spills of oil and other
hazardous substances into waterways.

Results demonstrated that high quality, small scale (1/40,000 to 1/60,000),
color infrared photography can be used for regional surveillance, lead-
ing to classification of land use into areas where potential spill sources
exist.  High quality, large scale (1/5,000 to 1/10,000), color aerial
photography can be used for localized delineation of potential spill
sources.  Localized surveillance should be supported by low angle, oblique
telephotography and limited ground surveillance.

Recommendations are given for an operational spill surveillance system
using multiscale aerial photography obtained on a 9-inch film format.
Use of thermal infrared imagery is not indicated at this time, as addi-
tional information acquired is minimal compared to resources required
for its acquisition.

This  report was submitted in fulfillment of Project No. 15080 HOL,
Contract No. 68-01-0145, under the sponsorship of the Office of Research
and Monitoring, Environmental Protection Agency.

-------
                                CONTENTS
Section                                                          Page
  I      Conclusions                                               1
  II     Recommendations                                           5
  III    Introduction                                             15
  IV     Methods                                                  17
  V      Results                                                  45
  VI     Acknowledgements                                         97
  VII    References                                               99
  VIII   Glossary                                                101
  IX     Appendices                                              105

-------
                                FIGURES






                                                                Page








 1       SAN FRANCISCO TEST SITE                                  18



 2      ANTIOCH TEST AREA                                        19



 3      PITTSBURG TEST AREA                                      20




 4      NICHOLS TEST AREA                                        22



 5      AVON TEST AREA                                           23



 6      MARTINEZ TEST AREA                                       24



 7      RICHMOND TEST AREA                                       25



 8      RICHMOND TEST AREA                                       26



 9      OAKLAND ESTUARY TEST AREA                                27



10      HUNTERS POINT TEST AREA                                  29



11      HUNTERS POINT TEST AREA                                  30




12      PHOTOGRAPHY, AVON                                        46



13      PHOTOGRAPHY, AVON                                        48



14      PHOTOGRAPHY, AVON                                        50



15      PHOTOGRAPHY, AVON                                        51



16      PHOTOGRAPHY, AVON                                        52



17      PHOTOGRAPHY, AVON                                        53



18      PHOTOGRAPHY, AVON                                        55



19      PHOTOGRAPHY, AVON                                        56



20      PHOTOGRAPHY, AVON                                        57



21       PHOTOGRAPHY, AVON                                        58




22      PHOTOGRAPHY, AVON                                        60
                                  VI

-------
                              FIGURES (cont.)
                                                                 Page
23A     THERMOGRAM, AVON                                          61

23B     THERMOGRAM, AVON                                          62

24      PHOTOGRAPHY, NICHOLS, AVON, MARTINEZ                      63

25      PHOTOGRAPHY, RICHMOND                                     65

26      THERMOGRAM, RICHMOND                                      66

27      THERMOGRAM, RICHMOND                                      67

28      THERMOGRAM, RICHMOND                                      69

29      THERMOGRAM, RICHMOND                                      70

30      THERMOGRAM, RICHMOND                                      71

31      THERMOGRAM, RICHMOND                                      72

32      PHOTOGRAPHY, RICHMOND                                     74

33      PHOTOGRAPHY, RICHMOND                                     75

34      PHOTOGRAPHY, RICHMOND                                     76

35      PHOTOGRAPHY, RICHMOND                                     78

36      PHOTOGRAPHY, OAKLAND ESTUARY                              79

37      PHOTOGRAPHY, MARTINEZ                                     80

38      PHOTOGRAPHY, NICHOLS                                      81

39      PHOTOGRAPHY, HUNTERS POINT                                83

40      PHOTOGRAPHY, PITTSBURG                                    84

41      PHOTOGRAPHY, PITTSBURG                                    88

42      PHOTOGRAPHY, AVON AND PITTSBURG                           89
                                  vn

-------
                                 TABLES
No.                                                             Page
1        Recommended Surveillance System                           2
2       Film/Filter Combinations                                 33
3       Remote Data Acquisition                                  36
4       Classification Scheme                                    42
5       Photogrammetric Measurements                             85
6       Results of Photointerpretation                           90

-------
                             SECTION I

                            CONCLUSIONS

The feasibility demonstration completed by Earth Satellite Corporation
for the Environmental Protection Agency has shown conclusively that
remote sensing techniques can be used to delineate areas where oil  and
other potentially hazardous materials are processed, stored, and trans-
ported.  Furthermore, these techniques can also be used for determining
where protective devices exist, where they should exist if they do not,
and where lack of sufficient protective measures has resulted in pre-
vious spill problems.

An aerial surveillance system developed for spill prevention purposes
should incorporate the following:

     0  Small scale  (1/40,000 to 1/60,000) color infrared/Wratten
        12 aerial photography for regional analysis in order to
        delineate oil refineries and other industrial  installa-
        tions which are potential sources of pollutants.

     0  Large scale  (1/5,000 to 1/10,000) color aerial photography
        of probable high hazard areas as determined by interpreta-
        tion of the small scale photography to provide detailed
        information on specific spill threats as well  as active
        spill sources.

Table 1 further specifies the use of small and large scale photography.
Large scale color photography was demonstrated to be extremely versatile
in the identification of various types of installations and activities
which have potential as spill sources.  Furthermore, such photography
is extremely useful for pinpointing existing protective features, as well
as installations which should be protected or in which protective faci-
lities have been compromised.

At the present time it has not been demonstrated that the use of thermal
infrared imagery will produce information on potential spill sources
which cannot be obtained by aerial photography with less cost and in a
more versatile manner.

It is important to note that a multiplicity of camera systems is routinely
available which can satisfy the specific demands of an aerial surveillance
system.  However, routinely available thermal scanners show a wide varia-
bility in both thermal and spatial resolution, whereas the data developed
as a part of this project indicate that resolution demands for thermal
data are high.

In order for an aerial surveillance system to be most useful, careful,
comprehensive photointerpretation must be performed.  Therefore, the
necessity exists to either train technicians in photointerpretation or
to employ qualified interpreters to perform comprehensive image analysis.

-------
                 Table  1.   Recommended Surveillance System.
          Low Altitude,  Large Scale
     High Industrial    Low  Industrial
       Concentration     Concentration
                           High Altitude,  Small  Scale
                          High. Industrial    Low Industrial
                           Concentration    Concentration
 Scale      1/5,000
             1/10,000
1/40,000
1/60,000
 Film             Aerial  Color
         (Kodak SO-397  or equivalent)
 Filter
Wratten 1A or Haze
                                     Color Infrared
                              (Kodak 2443 or equivalent)
         Wratten 12
 Camera*    9" x 9"  format,  roll  film magazine with  200  ft.  film capacity
 Lens*      8-1/4"  focal  length,  high
                    resolution
                                 6" focal  length, high
                                      resolution
*For oblique (hand-held)  photography:   70mm  format,  roll  film  magazine
 with 15 ft. capacity,  availability  of  80mm,  150mm,  and  250mm  focal  length
 lenses, as required.

-------
Furthermore, the use of phototnterpretation keys developed specifically
for use in interpreting photography for spill  prevention will provide
important guidelines for image interpreters assigned the task of detect-
ing and identifying potential spill sources.

-------
                              SECTION II

                            RECOMMENDATIONS

Recommendations stemming from this project for an operational  aerial
surveillance system are grouped into two major classes:

     0  Flight Planning and Data Acquisition
             Includes the selection of flight lines; coordination
             of aerial data acquisition with ground truth acquisi-
             tion; selection of sensors, films, and filters;
             specifications of flight parameters to include
             altitude, time of day and season; cost analysis;  and
             logistic support.

     0  Interpretation and Data Analysis
             Includes the methods and equipment for interpretation
             and mensuration of imagery; interpretation  and use of
             collateral data; use of interpretation aids; and  pre-
             sentation of results in a usable format.

Flight Planning and Data Acquisition

General Considerations

Following initial selection of areas for reconnaissance,  the  flight path
of the photographic aircraft must be determined in order that  desired
coverage be obtained.  Flight lines should be plotted on  topographic
maps of a scale commensurate with size of the area to be  covered and
complexity of the flight line.  U. S. Geological Survey  (USGS) 7-1/2
minute quadrangles (1/24,000 scale) or 15 minute quadrangles  (1/48,000
or 1/62,500 scale) are generally used for flight line plots.   These maps
usually indicate, by shading and symbols, the location of industrial
areas.  However, boundaries and placement of individual  structures
within these industrial areas must be considered only approximate and
cannot be relied on in mission planning.  At the time of mission planning
a current aeronautical chart for the area to be flown should  be consulted
to determine if restrictions exist pertaining to aircraft flights over
the target area.

Relevant ground truth can be a valuable aid to the image  interpreter.
Whenever possible, ground truth acquisition should follow large scale
overflights in close succession to optimize correlation  of information
from these two sources.  Obviously, the more dynamic the  phenomena, the
greater the importance of performing ground observations  soon  after
photointerpretation.  Ground observation should be designed to verify
photointerpretation findings on specific spill threats and to  determine
exactly the identity and condition of features and facilities  that appear
to be a potential threat or a source of past spill that  has not been
properly corrected.

-------
Color photographs taken on ground visits to target areas and low angle,
oblique color transparencies in 35mm or 70mm format are very useful
graphic aids to the image interpreter.  Ground and oblique photography
should be acquired under the direction of an interpreter who has been
working with the vertical imagery.  Although, interpretation of vertical
imagery alone by a skilled interpreter can yield an abundance of data,
the availability of ground truth data and oblique and ground photography
from previous studies will decrease both cost and time of the interpre-
tation.

In areas of denied access, low altitude oblique telephotography may  best
fulfill the need for ground truth.  A recommended supplemental  source
of intelligence on such an area would be visits to similar  facilities
employing the same materials and processes.

Aerial color reversal film appears to be of the greatest general appli-
cation in hazardous material surveys.  There are advantages to  interpret-
ing  original positive color transparencies, because of the  near-natural
color  rendition and high resolution of color films.  Both black-and-white
infrared and panchromatic films ordinarily used in aerial reconnaissance
are  usually reproduced as opaque positive prints or are copied  on film
as positive transparencies for interpretation purposes because  of the
special  skill necessary to interpret negatives.  On black-and-white
photography, pollutants in or out of water are harder to identify, in
that they are rendered in shades of gray rather than in near-natural
color.

The  use  of  color photography for aerial surveillance and reconnaissance
projects is becoming more common for the following reasons:

      0   The information content of color film is usually higher
         than that  of black-and-white for many applications, due
         in  part to high color fidelity and the wide variety of
         interpretative methods which can be used on positive
         transparencies.

      0   The quality of color films has improved greatly in
         recent years, and there is presently a greater latitude
         of  acceptable exposure, thus making them easier to  use.

      0   Color aerial films are being used by more commercial
         aerial survey firms, thereby insuring a more certain
         and steady supply.

      0  Although higher cost of color imagery has long been
        presented  as a drawback, the additional cost is usually
        only a small part of the overall project cost, and  the
        purchase of color photography is justifiable by increased
        data accuracy and output.  Color photography can often
        be flown at smaller scales than panchromatic photography
        and still  provide the same information.

-------
For the above stated reasons, it is recommended that conventional aerial
color film be used for acquiring aerial photographs at medium and low
altitudes (below 20,000 feet) for an aerial surveillance spill preven-
tion system.  Over industrial areas, the haze-penetrating characteristics
of color infrared film (when exposed through a Wratten 12 filter) and
its enhancement of the land/water interface make it the best choice for
small scale photography when such photography is to be acquired at
altitudes above 20,000 feet.

A number of high quality cameras, designed or adaptable to aerial use,
can be used with the specified film/filter combinations.  The camera
recommended for a spill prevention system is a 9-inch format mapping
camera with lens focal length appropriate for the scale desired and the
operating ceiling of the aircraft.

As will become apparent, the critical relevant parameters evaluated were
film/filter and scale, rather than format.  Format merely affects area
coverage per frame at any given scale; hence, recommendations relative
to format were inferred from the nature of the problem and from an
evaluation of the effects of area coverage on the working format.  In
some instances, for example where a need for vertical photography over
a small area exists, some other format may be the most appropriate
choice; nonetheless, in the general case, it is felt that a 9-inch for-
mat system will be of greatest utility.  The rationale for recommending
the 9-inch format includes the following considerations:

     0  Image acquisition cost is decreased with large format
        imagery, due to reduction in the number of photographs
        necessary to cover any given area, since, at constant
        scale, area coverage per frame increases with format
        size.  Thus, at any given scale, 70mm format covers
        only 6% of the area covered by 9-inch format.

     0  Both interpretation and mensuration of the aerial
        imagery is enhanced by displacement of the image away
        from the vertical aspect.  With increasing format size
        this radial displacement also increases, presenting,
        in effect, a slightly oblique view of objects imaged
        on the outer one-third of each frame.  At least one
        method of height determination is simplified by con-
        siderable radial displacement.  With successive frames
        and overlapping flight lines, one achieves both the
        vertical and oblique view of the object.

     0  Large format size enables the interpreter to more
        accurately analyze the effect of topography on target
        areas.  Within a reasonable range for topographic
        analysis, area coverage is of more importance to the
        interpreter than is scale.

-------
     0  Nearly all  commercial  aerial  photographic companies
        have 9-inch format camera systems, whereas few have
        70mm systems.

     0  Generally,  9-inch format camera systems provide annota-
        tion between frames of date,  altitude, frame number,
        and, in some cases, other ancillary data.  Such between-
        frame annotation is not provided in 70mm systems.

For oblique photography from a light  aircraft, the camera should be of
high quality, light weight, and compact design, have interchangeable
lenses, and be of about 70mm format.   A number of cameras which meet
these criteria, such as Hasselblad or Bronica, are readily available.
A camera of this type, rather than a  9-inch format mapping camera, may
be used in a vertical  mode, although  area coverage will be decreased,
necessitating either reduced coverage or more flight lines.

The tendency of long focal length lenses to decrease vertical exaggera-
tion or "flatten" images leads to a recommendation of the use of short
focal length lenses and low flight altitude rather than long focal length
lenses and high flight altitude.  Lens focal length and camera configura-
tion recommendations are presented in the following paragraphs.
 Regional Vertical Coverage

 Regional surveillance is that which will be used to classify land use
 activities and permit an interpreter to develop information which can
 be  used to assign priorities to follow-up, detailed coverage.  The infor-
 mation to be derived from regional analysis includes:

             0  Land use category to a 10-acre minimum size
             0  Number of storage tanks
             0  Major leaks on land and in water, or spill
                stains (300 feet in diameter or greater)
             0  Sources of waste products
             0  Major outfall discharge points
             0  Surface drainage patterns
             0  Water boundaries
             0  Dispersion of drainage in waterways
             0  Ship-loading facilities
             0  Major surface pipeline routes

These items are considered to be strategic data in that they usually
need to be obtained only at infrequent intervals (annually or longer,
depending upon needs and enforcement capabilities); hence, the stipula-
tion that such coverage is best obtained when atmospheric conditions
are, at the least, near optimum.  Optimum conditions here refer to
periods of no cloud coyer and very light to no haze.

-------
The system specifications for regional surveillance are as follows:

        0  Vehicle
             -- Fixed-wing aircraft
             -- Single or multi-engine with turbo chargers or jet
                engines
             -- Camera hatch in fuselage
             -- Precision navigation instruments (as specified
                for instrument flight rules by the Federal Aviation
                Agency)
             -- Electrical system and mounting arrangement to
                accommodate an aerial camera
             — Service ceiling, 25,000 feet

        0  Camera
             -- High resolution aerial camera with removable film
                magazine such as Wild RC-8 or equivalent
             -- 9-inch format
             -- 6-inch focal length

        0  Film/Filter Combination
             -- Kodak Aerochrome Infrared Type 2443 or equivalent
                (color infrared with Wratten 12 filter)

        0  Film Processing
             -- Kodak RT Color Processor, Model 1411 or as specified
                by the film manufacturer

        0  Scale
             -- 1/40,000 over areas with dense industrial development
                and characterized by persistent atmospheric haze
             -- 1/60,000 over areas with scattered industrial develop-
                ment and generally minimal atmospheric haze

        0  Atmospheric and Sun Angle Restrictions
             -- No cloud cover over target area
             -- Light to medium haze or less
             -- Sun angle above 30°

        0  Stereoscopic Overlap and Sidelap
             -- 60% forward lap (55% to 65% acceptable)
             — 30% sidelap (10% to 50% acceptable)

        0  Angle of Coverage
             -- Vertical coverage required with no more than 3° out
                of the vertical acceptable
Local Vertical Coverage

Local vertical coverage can be used to identify specific spill threats
and assess the need for further follow-up by ground observation.  The

-------
information to be derived from local analysis includes:
     o
        Specific activities currently being carried on at each site
     0  Minor leaks on land and in water, or minor spill  stains
        (10 feet diameter or greater)
     0  General  identification of waste products
     °  Minor outfalls
     0  Status of drainage features (active or inactive)
     0  Pumping and regulating stations
     0  Surface pipeline routes
     0  Condition and integrity of dikes
     °  Drainage facilities associated with revetments around
        tanks and other facilities
     0  Condition and integrity of revetments
     0  Volume of lagoons or storage piles
     0  General conditions of pipelines and supporting structures
     0  Transshipment points
     0  General condition of storage tanks
     0  Presence or absence of protective features and the state
        of repair of each (e.g., rails, fences, railcar barricades,
        dams, floating booms, pilings, channel markers)
     0  Numbers and types of vehicles, vessels and rail cars

 Because of their transient or changing nature, many of these  items are
 considered to be tactical data.  Therefore, remote sensing designed to
 obtain these data usually must be accomplished at more frequent intervals,
 perhaps under less optimum weather conditions than specified  for acquisi-
 tion of strategic data.

 The system specifications for obtaining detailed localized information
 are as follows:

        0  Vehicle
             -- Fixed-wing aircraft
             — Single or multi-engine, reciprocating or turbo-prop
                type
             -- Camera hatch in fuselage
             -- Instrumented for visual flight rules  (as specified
                by the Federal Aviation Agency)
             — Electrical system and mounting arrangement to
                accommodate an aerial camera
             -- Service ceiling, 12,000 feet

        0  Camera
             -- High resolution aerial camera with removable  film
                magazine such as Zeiss RMK 21/23 or equivalent
             -- 9-inch format
             -- 8-1/4-inch format length

        0  Film/Filter Combination
             -- Kodak Aerial Ektachrome Type SO-397 or equivalent
             -- Suitable haze filter, such as Wratten 1A or HF-2, -3,  or
                -4
                                  10

-------
        0  Film Processing
             -- Kodak RT Color Processor, Model 1411  or as specified
                by film manufacturer

        0  Scale
             -- 1/5,000 over areas with dense industrial  development
                and numerous potential spill sources
             -- 1/10,000 over areas with scattered industrial  develop-
                ment and few potential spill sources  (if in doubt,
                the larger scale should be specified)

        0  Atmospheric and Sun Angle Restrictions
             -- No cloud cover over target area
             -- Light to medium haze or better (this  requirement may
                be modified depending upon urgency for large scale
                coverage and frequency of acceptable  atmospheric
                conditions)
             ~ Sun angle above 30°

        0  Stereoscopic Overlap and Sidelap
             -- 60% forward lap (55% to 65% acceptable)
             — 30% sidelap 00% to 50% acceptable)

        0  Angle of Coverage
             — Vertical coverage required with no more than 3° out
                of the vertical acceptable


Local Oblique Coverage

Oblique aerial photographs have been shown to be very useful for document-
ing specific problem features and to provide images of facilities which
are difficult to evaluate in vertical views because of overhanging  struc-
tures or vegetation.  Low altitude stereoscopic coverage is used to iden-
tify features which are not easily recognizable or whose condition  cannot
be evaluated from the vertical view.  However, oblique coverage is
difficult to specify without prior knowledge of the target area.

Oblique photography should not be considered as an end in itself, but
only as support for vertical coverage.  Its most advantageous  use is as
a substitute for ground truth where low angle oblique telephotography
is more practical than a ground visit.  For example,  if vertical surveil-
lance suggests the presence of a small oil leak, low angle oblique
photography using ultraviolet or plus-blue panchromatic photography,
which has been shown to be useful  in detection of thin oil films on
water, may be indicated.  In any case of oblique photography,  complete
documentation must be obtained, including date, altitude, location,
angle of view, film/filter, etc.

Oblique photography can be obtained from the photographic aircraft
following vertical coverage by use of a hand-held 70mm camera  with  eye-
                                  11

-------
level viewfinder and appropriate lens (80mm, 150mm,  or 250mm focal  length,
as required),  Care must be exercized in obtaining oblique photography
so as to take advantage of the correct angle of illumination, to position
the aircraft in such a manner that the ground scene is not obstructed,
and to insure that aircraft motion is compensated in making the exposure.


Image Acquisition Procedures

The flight crew should be carefully instructed to use the flight maps,
flight parameters, and system specifications supplied by  the photointer-
preter in order to obtain coverage as required.   It should  also be the
responsibility of the flight crew to determine whether the  prevailing
weather conditions are suitable for coverage over the areas  scheduled.

Exposure settings for lens and shutter should be determined by  reference
to manufacturer's specifications, but modified as necessary for the pre-
vailing atmospheric and illumination conditions.

Color film processing should be performed as soon as  possible after expo-
sure to reduce  the effects of color balance shift which can result from
prolonged delay after exposure, particularly where film is  subjected  to
high temperature  environments.
 Interpretation and Data  Analysis

 An important first consideration  in commencing an  image  interpretation
 effort is organization.   The  particular  system used  is not critical,  as
 long as it is applicable to the user organization.   Film should  be  labeled
 with acquisition data,  logged in,  and  stored  for easy retrieval.  System-
 atic interpretation is  necessary  to ensure  that all  image areas  are
 scanned and that specific areas are interpreted according to their  indivi-
 dual priorities.  All  interpreters should use the  same symbols,  codes,
 and record-keeping procedures to  facilitate data storage and retrieval.

 The equipment and physical facilities  needed  for an  image interpretation
 shop depend largely upon image quality and  final product desired.   While
 high-quality transparencies are best exploited with  zoom-sterescopic
 systems,  no amount of magnification or optical resolution can  compensate
 for poor quality transparencies.   Likewise, for most uses, light tables
 need only be suitable  in illumination  qualities and  comfortable  for the
 interpreter.  If paper  prints are  to be  used, stereoscopes should be
 selected  on the  basis of format and desired magnification range.

 All  available collateral  data sources  should  be made available to inter-
 preters during the read-out process.   Common  sources of  collateral  data
 are publications,  photointerpretation  keys, comparative  photographic
 coverage,  and ground truth data.   If ground or aerial photography in
 35mm slide format  has been acquired, both slides and viewing equipment
 should  be readily  available.
                                   12

-------
The amount and detail of photogrammetry depends on the existence of a
need for ground-distance measurements.  Measurement of heights of ver-
tical or near vertical objects and length and width of horizontal objects
is relatively stmple If object boundaries are discrete and the inter-
preter has sufficient knowledge of the conditions under which the imagery
was acquired.  Such measurements should be easily obtainable using the
recommended system; thus, capacities of storage tanks and vertically-
walled revetments can easily be determined.  However, measurement of the
holding capacity of banked earth revetments, catchment basins, and other
irregular features, as well as establishment of a datum from which to
measure, requires a cartographic camera, ground survey support, special
image preparation equipment, complex plotting instruments, and technical
expertise.  Although the recommended system can be used to determine
where such measurements are required, additional aerial photography
and ground support, as detailed previously, would be required.  Costs
associated with these additional measurements run $60 to $100 per frame;
thus, care should be exercised in their use.

Simple mensuration as described can be performed on vertical imagery
with a minimal inventory of equipment, including a photointerpreter's
magnifier, seven power, with reticle graduated in thousandths of feet
and five one-hundredths of millimeters; a photointerpreter's rule with
a range of at least 6 inches or 15 centimeters, graduated in the above
state increments; a parallax wedge for height determination; and a simple
calculating device, preferably a photointerpreter's slide rule or circular
rule.  Accuracy of photogrammetric measurements will depend on the reso-
lution of the imagery, the accuracy of collateral image data, and the
skill and precision of the image interpreter.
                                  13

-------
                             SECTION III

                             INTRODUCTION

This report represents the results of an Environmental  Protection Agency
project to demonstrate the feasibility of utilizing state-of-the-art
technology in remote sensing aerfal reconnaissance for spill prevention
surveillance.  Specifically, the objective was to determine the extent
to which aerial remote sensing systems might be used to detect potential
sources of water pollution by oil and other hazardous materials.
Hopefully, such prior identification would permit appropriate action to
be taken to prevent these potential sources from becoming actual.  Remote
sensing aerial reconnaissance seemed, and has proven to be, an excellent
choice for such surveillance, because many potential and actual pollu-
tion sources occur in highly industrialized areas where, for various
reasons, ground access, hence source identification, is extremely diffi-
cult.

This project utilized multispectral photography and thermal imagery.
Many other remote sensing devices are available, but their highly spe-
cialized capabilities are not presently applicable to a project of this
nature; such devices include magnetometers, radiometers, scatterometers,
spectrometers, and others, some of which may ultimately be of use in
spill surveillance.

For this initial demonstration project, the highly industrialized but
geographically and culturally-varied San Francisco Bay area was chosen
as the test site.  Within this site, several target areas (later to be
described) were selected which contained representative installations,
in varied settings, of the type toward which a preventive surveillance
effort might logically be directed.

The study was conducted in three phases.  These phases, and a brief des-
cription of the tasks accomplished in each, were:

     PHASE I:  Preliminary Studies - A mid-altitude reconnaissance
     flight was made over the entire San Francisco Bay test site,
     thereby obtaining results which led to a flight plan for
     missions to obtain high altitude, small scale photography.
     This photography was obtained and interpreted to select the
     specific test areas for detailed study.  A ground truth plan
     was also developed.

     PHASE II:  Program Implementation - An evaluation was made of
     various scales and sun angles, on the basis of which optimum
     values for each were selected.  Low altitude, large scale
     photographic flights and day and night thermal infrared flights
     were planned and executed, as was ground truth coverage of all
     test areas.
                                  15

-------
PHASE III:  Data Analysts - The large scale photography  and
thermal  imagery was interpreted and correlated with ground
truth.  Photointerpretation keys were developed and a photo-
interpretation test was administered to skilled interpreters
The results obtained, and the conclusions drawn from them
have led to recommendations for an operational aerial sur-
veillance system.
                             16

-------
                             SECTION IV

                              METHODS

Site Selection

The San Francfsco Bay regton was chosen as the test site because it
represented geographically most of the situations wherein one might
expect to find potential sources of oil and hazardous materials in
juxtaposition to coastal and inland waterways.

Two methods were readily available for surveying the San Francisco Bay
test site in order to identify specific candidate test areas:  interpre-
tation of existing photographs and aerial  reconnaissance.  Initially,
a set of contact black-and-white scale (1/80,000) prints were purchased
from open files of the U. S. Geological Survey.   This stereoscopic
coverage was interpreted in order to delineate areas of interest for
further survey.

Within the San Francisco Bay test site, several  sub-areas along the
San Francisco Bay and Sacramento - San Joaquin River waterfront were
overflown and delineated as candidate test areas for continued surveil-
lance.  From this reconnaissance flight, major sub-areas of the region
were selected for high altitude overflight in order to obtain small
scale photography for use in final selection of test areas.  Selection
of the test areas was coordinated with interpretation of recently
acquired National Aeronautics and Space Administration (NASA) photography
of the same region on the same type of film, but at a scale of 1/60,000
on a 9-inch format.  This evaluation led to selection of test areas
adjacent to waterways at Antioch, Pittsburg, Nichols, Avon, and Martinez,
all on the south shore of the Sacramento River.   On the San Francisco
Bay shore, several additional large areas were chosen, including the
waterfronts at Richmond, Oakland Estuary, and Hunters Point in the City
of San Francisco (Figure 1).


Antioch (Figure 2)

The river waterfront at Antioch embodies several installations which are
potential spill sources.  Immediately adjacent to the waterfront is a
paper and cardboard plant with a waste water outfall into a small slough.
There is also a food canning plant on the waterfront, with direct outfalls
into the river.  Both installations are close enough to the waterfront
that spills could occur during routine operations.  Neither of the above-
mentioned plants has any apparent protection against flooding or storm
drainage.


Pittsburg (Figure 3)

The waterfront at Pittsburg contains numerous small industrial installa-
tions, as well as a very large steel mill.  Situated on this waterfront
                                  17

-------
                                                 l\
                                               Pittsburg
FIGURE 1.   SAN  FRANCISCO  TEST  SITE
                     18

-------



re

C
o
Q
	-I
           '
     FIGURE 2.  ANTIOCH TEST AREA

-------
                             o
\
               in
               z
               §
               o
               cr
               CO
              pu K i S !
              s u AA o • 5
     r c
                           02   / y*Q
                           *^^     -^  •--
                          • r_f*.    . * *
                      * '   If*   */   '»
                      ?' •  e --/:,'
                                   o
                      "  -   '  / J>  '
           .,  -
:  m^  jiB	
   15"    , "--:v:-.:^
          ,     ^
 ~r .««£   [  7  ' • ..
         FIGURE 3.  PITTSBURG TEST AREA
                         20

-------
is a major power generating facility (which depends upon natural gas as
its primary fuel source, but also contains several "
holding fuel oil for
and a small boat harbor, all of which may contribute spill  prot
Nichols (Figure 4]
2rating facility (which depends  upon  natural  gas  a:
rce,  but also contains  several  large  storage  tanks
 stand-by operations],  a sewage  treatment plant,
aor,  all of which may contribute spill  problems.
The nearby waterfront at Nichols contains a chemical plant with holding
ponds and storage piles only minimally protected (by levees) from the
river.
Avon (Figure 5)

The test area at Avon lies adjacent to Pacheco Creek at its confluence
with the Sacramento River.  This site contains one installation of signi-
ficance to the project, a major refining and storage area for petroleum
products.


Martinez (Figure 6)

The Martinez area also contains a major petroleum refining and storage
installation, although large parts of that installation are protected
and removed from the river by a railroad embankment.  At the south
anchorage of the highway bridge connecting Martinez with the north shore
of the Sacramento River is a large pile of discarded material from an
ore processing operation.  This material is not restrained from slough-
ing into the river and provides a continuous low-level source of dis-
colored runoff.  This out-wash has been visible regularly as a yellow-
orange stain in the river; it could also be inundated during periods
of high water.


Richmond (Figures 7 and 8)

The Richmond waterfront and refinery area on San Francisco Bay was chosen
as a test area because it presents numerous installations with potential
spill  sources.  Included in this area is another major petroleum refining
and storage complex; furthermore, there are several small boat harbors,
a whale-rendering plant, and a Navy fuel depot where fuel oil, stored
in tanks immediately adjacent to the shore, is transferred to and from
ships.  At the southern end of the Richmond test area is an extensive
scrap metal operation which frequently introduces hazardous substances
into the Bay.  The area also contains numerous small industrial instal-
lations with  concomitant potential as spill sources.


Oakland Estuary (Figure 9}

On the Oakland Estuary, there are several installations which present
                                   21

-------
c
3
O
k.

O

a  ~°
-  c
                         N011VAH3S3H  //


              /  &^    I     1VAVN
     FIGURE 4.  NICHOLS TEST AREA
                        22

-------

                       ^                \
                        ^ *-*. -fXy»fV' '. • '.          \
area coverage of 1/80,000
   scale photograph
area coverage of 1/10,000
   scale photograph
area coverage of 1/2500
   scale photograph
   FIGURE  5.   AVON TEST AREA
                     23

-------
                                       : ^.j: •  ;^ym -. £•

                                      ^xt4C:7^r'^*^*-'-" *« '•
                                                                   "
•
               FIGURE 6.   MARTINEZ TEST AREA
                                 24

-------
FIGURE 7.   RICHMOND TEST AREA
                  25

-------
  \'->.   w*  \ w
''";v"  .*'«-ii'il?'i
                                                         JJ
                                                       s: w
                                                       CD (n
                                                         **
                                                         c
                                                         b
                                                         Q.

                                                                       \



                                                                       \
                                                                          \
                                                          C
                                                          JS-
                                                          jn
                                                          tn

                                                          §
                                                          m

                                                              i
               FIGURE  8.   RICHMOND  TEST AREA
                                   26

-------
                                                    v       " 5?'
                                                      I  : ™$
                                              t>£± ' •< 1  	.iii.MJii.-A  njl
                                               J_J"_ '. U..J	I	^

                                                 -•-"-•••-™-'-1  •. , . »V . p -TTt III

                                                 :.   «»«&-..
FIGURE 9.  OAKLAND  ESTUARY  TEST AREA
                        27

-------
features of interest.  Included is another Urge scrap steel  processing
plant, a chemical plant of moderate size, numerous small  industries, and
docks for general ship commerce.


Hunters Point (Figures 10 and 11)

Hunters Point, in the City of San Francisco, contains several sewer out-
falls, numerous docks for general shipping, two fossil/fuel  power generat-
ing stations, and numerous small  industrial installations, including
some with holding facilities for  fuel oil or other petroleum products.
 Sensor and  Flight  Considerations

 Two major data output requirements must be
 surveillance  spill  prevention system.  The
 of the following:
                        met  for a  successful  aerial
                        system used  should be capable
              tions
              areas,
for potential
  Thus,  it is
        Detecting areas of industrial  activity near drainage features
        or waterways where a spill  of oil  or other hazardous substance
        would contribute to water pollution.

             This requirement can be satisfied by taking aerial
             photographs which will  permit a trained photointer-
             preter to detect and identify features and facilities
             and to assign each to a category in a suitable
             classification plan.  It should then be possible to
             concentrate efforts on further searches and delinea-
                                 spill  sources in the most likely
                                 possible  to assign a lower priority
             for further study to,  or to eliminate completely,
             those areas where little or no spill threat exists  due
             either to land being unused or being used for activities
             not hazardous to waterways (e.g., recreational areas,
             residences, grazing land,  and non-polluting industries).

     0  Studying in detail facilities classified as being potential
        sources of spills.

             The objective is to detect careless practices,
             inadequate protective measures, illegal waste
             storage and disposal practices, active discharge
             and disposal of materials  into waterways, and evi-
             dence of past spills as indicators of potential
             spills where corrective measures have not been
             taken.

A detailed study of facilities should be conducted only in those areas
identified as potential contributors to spill problems.  Data from the
detailed study should be adequate to define both the types of material
                                   28

-------
   i.
FIGURE 10.  HUNTERS  POINT TEST AREA
                     29

-------
                                     PotreroPt
FIGURE 11.  HUNTERS  POINT TEST AREA
                      30

-------
which may be Involved in a spill and the corrective measures which
should be undertaken.   In most cases a minimum of ground visits for
verification should be necessary.

In light of the two requirements listed previously, it is apparent that
a single photographic mission would not be the most efficient way to
obtain all necessary data, because of the requirement for a regional
analysis under the first item and a highly selective localized approach
under the second.  Rather, data from the first mission can be used to
select specific areas for coverage by a second mission.

Regional coverage should provide only enough detail to identify indus-
trial complexes.  Small area coverage should provide considerably more
detail for obtaining information on specific spill threats.

There are a wide variety of sensors, cameras, films, and filters avail-
able to meet the requirements of regional and local surveillance, as
well as vehicles to transport such a system to a target area.  It is
imperative that care be used in specifying the vehicle, camera system,
and flight parameters, in order to optimize image acquisition operations.

Among the factors necessary to optimize image acquisition specifications,
it is desirable to:

     0  Produce photographs at the smallest scale acceptable for the
        required interpretation accuracies

             Specification of an excessively large scale increases
             the cost of the photography (for example, costs per
             square mile are approximately 2.5 times more for
             1/5,000 scale than for 1/10,000 scale - see page 107),
             requires the interpretation of more photographs than
             necessary, and obscures some regional trends which
             may be obvious on smaller scale photographs covering
             larger areas.

     0  Specify a film/filter combination which provides optimim contrast
        between the target features and surrounding objects

             Use of the proper film/filter combination may reveal
             information which is undetectable because of spectral
             characteristics of the target.

     0  Expose photographs during periods when solar energy best
        illuminates the features of interest

Small relief features are rendered more visible by using short focal
length lenses which increase the stereoscopic parallax recorded at a
given photographic scale.  Long focal length lenses tend to decrease
vertical exaggeration or "flatten" ground features at the same scale.

Atmospheric haze can seriously degrade aerial photography, particularly
                                  31

-------
In the shorter wavelengths (ca. 380-500 nanometers).   Conventional  color
film is not generally used from altitudes above 20,000 feet, because
haze often renders the ground scene a low-contrast blue or green.
However, color infrared film, which is used with a minus-blue filter
(thus reducing haze effects], does not suffer to the same degree when
used from altitudes above 20,000 feet.  On clear days, conventional
color film may be used from higher altitudes and will produce good color
fidelity, but it should be noted that very clear days are atypical in
industrialized areas.  Therefore, color infrared film (Kodak f-j-[m type
8443 or 2443) exposed through a Wratten 12 filter is commonly specified
for photography taken at altitudes above 20,000 feet.  Under light to
medium haze conditions, the maximum altitude for conventional color
films should be decreased to 15,000 feet.  In conditions of extreme
haze, even color infrared film cannot be used with success, because of
the degrading effects of the haze.

Choice of appropriate large scale coverage is dependent upon more varia-
bles  than is the case with choice of small scale imagery.  Therefore,
no  a  priori decision was made; rather, a test of various scales  was
devised  in order to assess the merits of available choices of large
scales.  The scale test covered the range commonly available with con-
ventional aircraft and camera systems, and, for continuity and comparison,
extended from small scale  (1/80,000) through various increments  to very
large scale  (1/2,500).

At  1/2,500,  square to'circular features as small as 6 to 12 inches
across  can be seen on photography taken under fair to good atmospheric
conditions.  At 1/80,000,  features such as individual automobiles can
barely  be resolved under good atmospheric conditions.  Photography at
a scale  of 1/400,000, taken by NASA from U-2 aircraft over San Francisco
Bay,  was also evaluated to determine whether such photography might be
useful  for regional classification.  Many of the requirements set forth
for small scale coverage were not satisfied by such photography.

Four  identical Hasselblad  500EL cameras, modified for use with 100-foot
film  magazines and equipped with interchangeable lenses of 80mm  and
250mm, were  used in the test.  Shutters were electronically synchronized
to  provide identical coverage on each film/filter combination (Table 2),
using exposure parameters  as specified by the manufacturer.  Due to the
relatively small area to be surveyed, and the desirability of simultaneous
exposures when evaluating  several film/filter combinations, this 70mm
format  system was chosen as most compatible with the requirements of the
feasibility  demonstration.  A similar multispectral array of larger format
cameras  would have been impractical.

Specifications similar to  those for the photographic missions were pre-
pared for acquisition of thermal infrared imagery, based on practices
known to be  useful for imaging pertinent industrial components.  The
daytime  temperatures and associated thermal emission rates of most
surface  areas are quite different from those existing at night.  Usually
subtle industrial thermal  sources are best recorded during  the pre-dawn
                                   32

-------
              Table 2.  Film/Filter Combinations Used
    TTW
   Filter
           Remarks
Aerial Color
(Kodak SO-397)
     1A
   (Haze)
Color Infrared
(Kodak 8443 or
2443)
 Wratten 12
(Minus-Blue)
Panchromatic
(Kodak 2402)
Black-and-White
Infrared
(Kodak  2424)
 Wratten 47B
 (PIus-Blue)
 Wratten 89B
(Minus-Visible,
  Infrared
 Transmitting)
Records the ground scene in a
nearly true rendition of the
features as seen by the human
eye.  Provides maximum contrast
among features whose major
reflectance is in the visible
spectrum.  Provides very sharp
images and good exposure lati-
tude for various lighting
conditions.

Provides optimum haze penetra-
tion while recording the ground
scene in a variety of false
colors which provide contrast
among features of interest for
vegetation and environmental
studies.  Provides sharp images
and fair exposure latitude for
various lighting conditions.

Provides fair contrast between
oil slicks and background water.
Provides fair sharpness and
exposure latitude for various
light conditions.

Provides good shoreline delinea-
tion between water and land.
Provides good detectability of
vegetation conditions.
                                    33

-------
hours.  Since many industrial substances which might become spill  threats,
are at temperatures other than- ambient, thermal scanning provides  much
useful information eyen if one ts not primarily concerned wtth thermal
pollution.

In order to test the validity of these considerations, thermal infrared
imagery was acquired over all test areas at night at an altitude of
5,000 feet and over the Richmond test area at  two other times (mid-
afternoon and evening} and at three different  altitudes  (1,500, 5,000,
and 10,000 feet).  At 1,500 feet, thermal anomalies of as  low as one  to
two degrees centigrade and features as small as one to three  feet  in
diameter can usually be detected; at 10,000 feet, however, thermal
anomalies have to be at least three to five degrees centigrade and
features have to be at least five to ten feet  in diameter  to  be detected.
Features exhibiting extreme temperatures, such as vent stacks, pipelines,
and hot engines, can frequently be detected even if they are  smaller
than  these minimum averages.

The infrared scanner employed was a Texas Instruments Corporation  RS-14
 (operating in the 8-14 micrometer spectral region), chosen for its
excellent thermal and spatial resolution as well as its  dependability
and performance.
 Photographic  and Thermal  Data Acquisition

 Data acquisition began with  a reconnaissance  flight over the  entire  San
 Francisco Bay -  Sacramento  - San  Joaquin waterfront to delineate  candidate
 test areas for further surveillance.  The next activity involved  over-
 flight of the selected areas in order to obtain photography at  a  scale
 of 1/40,000 for  determination of  the exact sub-areas to be used during
 the remainder of the  study.  From the resultant series of photographs,
 five major test  areas were  selected for intensive study.  One area was
 selected for  the sun  angle  and photographic scale test and was  overflown
 several  times with  the entire four-camera multispectral array,  in order
 to obtain imagery at  representative sun angles and at representative
 scales,  as previously described.  From the results of these overflights,
 times and altitudes for  succeeding missions were selected to  give opti-
 mum coverage  of  the remaining test areas.

 All scale tests  were  scheduled to be flown at the same time under stable
 atmospheric conditions to minimize the effects of uncontrolled  atmos-
 pheric variables and  to  eliminate changes caused by target area fluctua-
 tions.  It was recognized that the findings from photographs  taken under
 any set  of atmospheric conditions may be different from those from photo-
 graphs taken  under  other conditions.  Nonetheless, major conclusions
 as to scale choice  will  be  independent of all but the most degraded
 atmospheric conditions.

 Sun angle tests  were  conducted to identify illumination angles  which
                                   34

-------
might tend to enhance or degrade the detectability of features of interest.
The photographic flights for sun angle testing were scheduled on the same
day as flights for scale testing to provide direct comparison between the
two tests and to minimize target area changes which might be encountered
by photographing the area on two different days.  Flights were made during
periods when the sun angle was approximately 10, 40, and 60 degrees above
the horizon.

From the results of these tests, times and altitudes for succeeding
missions were selected.  On vertical photography, sun angle proved not
to be a critical factor as long as the sun was at least 30 degrees above
the horizon,  rt was decided that the large scale photography should be
obtained at a scale of 1/10,000.

The various missions flown and the nature of the missions is shown in
Table 3.  Interpretation of the imagery from any given mission occurred
before subsequent missions were flown, thus mission planning could be
done on the basis of previously acquired data.  Large scale photography
was interpreted prior to interpretation of thermal imagery to provide
a reasonable data base against which to judge the thermal data.
Ground Truth Rationale

In photointerpretation operations for spill prevention, the goal is to
perform the necessary data collection functions exclusively by reference
to the photographs provided, without undertaking ground visits.  In many
cases this goal can be achieved because the features of interest are
easily recognized, counted, and measured on the photographs, and the
photointerpreter is confident that his interpretations are accurate.
In other cases, the interpreter cannot complete his data collection
functions working with the photographs alone, either because the imagery
is unsuitable  (spectral band, scale, or quality not optimum), the fea-
ture does not  lend itself to remote identification (such as identifica-
tion of most chemical components), or the interpreter lacks the necessary
experience to  derive the needed data even though the features may be
clearly identifiable by an experienced interpreter.

In many data acquisition operations, it is known at the outset that a
combination of photointerpretation activities and ground checking is
needed.  Costs for ground visits and commensurate time factors can become
quite high if  care in designing the data collection methods is not exer-
cised.  It is  therefore desirable to minimize ground checking requirements
in order to control costs associated with data collection activities.
Ground checking requirements for spill surveillance can be minimized  in
a number of ways:

     0  Photographic acquisition procedures and systems can be
        specified to maximize the amount of data obtainable from
        photointerpretation.
                                    35

-------
                  Table 3.   Remote Data Acquisition
    Mission

Visual  Reconnaissance
and Oblique Photography

     Small Scale
     Photography

     Scale Test
    Area

San Francisco
  Test Site

     All
    Avon
    Date

  19 July


  27 July


  20 August
     Sun Angle
       Test
    Avon
  20 August
     Large Scale
     Photography
     All
8-9 September
     Thermal IR
       Test
  Richmond
   9 August
 2030-2200 PDT
                              All Others
                         1400-1430 PDT

                          15 August
                         2300-0100 PDT
                                  36

-------
                             Table 3.  (cont.)
  Scale                      Film/Ft Her                Sun Angle

Oblique                       Color/lA                    50°-60°
                            Panchromatic/12
1/40,000                      Color IR/12                  50°-60
1/2,500;                      Color/lA;                   40°-60°
1/5,000;                     Color IR/12;
1/10,000;                  Panchromatic/478;
1/20,000;                     B&W IR/89B
1/40,000;
1/80,000

1/10,000                      Color/lA;                 10°, 40°, 60°
                             Color IR/12;
                           Panchromatic/476;
                              B&W IR/89B

1/10,000                      Color/lA;                   30°-55°
                             Color IR/12;
                           Panchromatic/47B;
                              B&W IR/89B

(Altitude:                      	                      	
1,500; 5,000;
10,000 feet)

(5,000 feet)                    	                      	

(5,000 feet)                    	                      	
                                  37

-------
     0  phQtotnterpretatton personnel  can be trained In specific
        tasks and disciplines that Increase the amount of data
        obtainable from photographs alone.

     0  Reference documents such, as photointerpretation keys can
        be prepared, providing valuable aids to photolnterpreters
        in identifying and comparing features seen on aerial images.

     0  Previous photographic records and interpretations can be
        consulted to provide additional data.

Ground truth can take several forms and can be obtained by a variety of
methods.  For spill  surveillance, most ground truth information would
provide additional information on potential spills of hazardous materials
Because of the extensive areas that are involved, ground data would
usually be obtained after the aerial photography has been acquired and
interpreted and only in those places selected by photointerpretation.
Ground data should support information from photointerpretation, either
to provide specific identifications, enumerations, or delineations, or
to determine the significance of features as a potential spill  threat.
When  evidence is needed to support enforcement actions, it is essential
that  interpretations made on aerial photographs be verified by an addi-
tional data source,  such as low altitude telephotography or on-site
ground observations, to provide complete confidence in data developed.

Additional sources of ground truth data include consultation with indus-
trial personnel responsible for the facilities in question and reference
to photographic interpretations performed for similar purposes.
Ground Truth Acquisition

In order to identify features which could not be positively identified
on aerial photographs, at least one ground visit was made to each test
area.  At this time, verification of imagery interpretation was made, and
documentary photographs, on 35mm and 70mm format, were taken of impor-
tant target features.

Verification was accomplished in one of several ways:

     0  Where project personnel were familiar with the facilities
        in question, their ground observations provided ground
        documentation.

     0  Where project personnel were not sufficiently familiar
        with a feature, consultation was arranged with manage-
        ment personnel at the facility in question.

     0  Where both ground access and management consultation were
        denied, appropriate state or local government officials
        familiar with, the facility were consulted.
                                   38

-------
In that all major facilities to be surveyed were represented by two or
more examples, extrapolation could be made from facilities where coopera-
tion was extended to those where cooperation was denied.

Thus, officials in several industrial operations were interviewed in order
to compile information on prevailing conditions and activities and on
the relationship of various components to potentially spillable materials.
For example, at one refinery, management personnel provided information
on storage tank capacities and the volumes and temperatures of tank con-
tents for comparison with data obtained through image interpretation.
Image Interpretation and Data Correlation

Small scale multiband photographs were interpreted to delineate areas of
interest for assessment of spill threats and to classify facilities and
land use practices.  Areas which would require large scale color photo-
graphic coverage to obtain detailed information on actual spill threats
or active sources of discharge were delineated on small scale coverage,
as were areas where low altitude oblique photographs were desired.  Large
scale photographs were interpreted to define the specific location of
spill threats and active outfalls.   Many of the areas studied in the
initial phase were eliminated from more detailed study, because they were
recognized as unlikely areas in which to find objects or conditions of
interest.  In most cases, unwanted features were merely enumerated but
not considered further.  Features which fell within the area or subject
of interest were then further interpreted.

All interpretation was performed by stereoscopic viewing wherever possible.
Stereoscopic viewing was found to be superior to monoscopy early in the
project, as it facilitates the interpretation of topographic features,
natural or man-made, which are important to a spill prevention study.
Monoscopy is best suited for gross identification and the rapid scan of
small scale imagery when delineating areas for detailed study.

Interpretation of all test site imagery and detailed interpretation of
the petroleum refinery in the Avon area provided the basis for evaluating
several aspects of the study.  Interpretation provided data for evaluating:

     0  Suitability of vertical photography in classifying land
        use activities

     0  Role of low altitude oblique photography as a supplement
        to vertical photography and as a substitute for ground
        photography

     0  Relative suitability of the four film/filter combinations

     0  Effects of sun angle variations on image interpretability

     0  Accuracy and limitation of mensuration using simple equipment
                                   39

-------
     0  Effects of different lens focal lengths and associated
        parallax factors on interpretability

All interpretation was done in a manner which, would permit it to be
related to an evaluation of aerial imagery in facility analysis for past,
present, and potential releases of hazardous materials.

Interpretations of color imgery were made on original positive trans-
parencies and on paper prints of the black-and-white photography.
Although the usual amount of degradation of image quality occurred in
the process of making black-and-white opaque prints, such degradation
was somewhat offset by the inherently easier interpretability of positive
prints as compared to negative transparencies.  Positive transparencies
were interpreted with both a simple stereoscope and a zoom stereoscope
(Bausch and Lomb Zoom 240); paper prints were interpreted with a simple
stereoscope only.  The unique perspective view afforded by low angle
oblique photographs combines aspects of both vertical aerial  photography
and ground views.  In this study, the oblique photographs proved to be
valuable for detailed interpretation, especially interpretation of objects
having height but relatively little areal extent, such as fences and
barriers.

As scale denominator numerically diminishes, the extent of topographic
relief is obscured.  This effect is greatest on small (70mm format)
imagery due to the small area coverage.  Topographic relief is important
in analysis of potential spill hazards from sites on hills or ridges, as
slope may render  ineffective a revetment which, if on level ground, would
adequately retain the contents of its enclosed tank.  In addition, one
must  consider the effects of relief on the drainage of surface water into
catch basins and  holding ponds, possibly resulting in overflow.  Since
70mm  format covers only 6% of the area of 9-inch format at comparable
scale, 1/10,000 scale 9-inch format photography should provide ample
realization of local relief in most instances, whereas 70mm format does
not.

If desired, a topographic study could be performed on a site such as the
ridge along Pacheco Creek in the Avon test area (Figures 5 and 12), which
contains a major part of the tank farm area.  The tanks are situated on
the sides and top of the ridge on excavated benches and are provided with
banked earth revetments and drains to catchment basins.  A detailed analy-
sis of the adequacy of earthworks during catastrophic  spills  and  of  the
fate of residue in catchment basins during periods of heavy rainfall is
possible only with extensive photogrammetric contouring of these instal-
lations and their environs, an operation probably not justified  inmost
cases.

Interpretation for identification and analysis of potential hazardous
spills combines industrial classification with interpretation for specific
hazard and protection features.  The unique analysis desired necessitates
interpretive aids, both to assist the trained interpreter and to provide
instructional  material for novices.  Such keys for spill prevention would
be constructed around a classification scheme which leads from gross
                                  40

-------
activity identification to specific identification of hazards, protective
measures, and evidences of previous problems.

A classification plan prepared for a specific discipline would contain
many categories of features that are not of interest, as well as a com-
prehensive listing of those that are of interest.  It was therefore
necessary to compile a classification scheme that would segregate areas
near waterways where further detailed study would be needed [and that
would permit a determination of priority for spill threat analysis) from
those recognizable as being of no consequence as potential spill sources.
A further requirement was that the categories in the plan represent
classes of features which could be recognized consistently in a logical
sequence by image interpreters on imagery specified for spill prevention.

Spills of oil and other hazardous substances are most likely to originate
from facilities wherein such products are stored, transported, or used.
The initial two divisions in the classification scheme, detailed in Table
4, were related to the stability of such facilities:

        0  Static
               An installation which is a permanent feature

        0  Transient
               An installation or facility which is not permanent,
               even over a short period, such as ships, trains,
               and trucks used for transporting hazardous materials

The next major division in the classification scheme involved the types
of materials which nright contribute to the spill of oil or other hazard-
ous substances and, obviously, involved classification as to solid, liquid,
and gaseous materials.  The classification under materials was not more
detailed because of the difficulty of absolute classification or identifi-
cation of a substance from aerial imagery; rather, that particular piece
of information, while necessary, would almost always have to be provided
by ground observation and perhaps by laboratory analysis of specimens
collected during ground observation.

Another division involved types of activities which could lead to a spill.
Such activities include mining, transportation, storage, refining and
processing, land development, and others.  Means were therefore developed
for identifying such activities and their potential, in any given setting,
for contributing to spill problems on or near water bodies.

All of the facilities and activities described, which are generally asso-
ciated with the industrial activity and its material handling facilities,
contained various protective measures designed to control spills.
Therefore, the classification scheme included a category for protective
features which would tend to retard or retain a spill, the integrity of
which are therefore important to a spill prevention program.  Such fea-
tures include barricades, fences, revetments, dikes, moats, cooling ponds,
and waste lagoons.
                                   41

-------
                    Table 4.  Classification Scheme
Static Facilities

Tank farms
Raw material, by-product, or
   waste piles
Waste lagoons and cooling ponds
Terminals
Pumping stations
Refining and processing
   installations
Power plants
Pipelines
Slaughterhouses
Livestock pens
Junk yards
Garbage dumps

Transient Facilities

Railroad cars
Trucks
Grading equipment
Ships

Materials*

Solid
Liquid
Gaseous

Activities

Mining
Transport
Storage
Manufacturing
Refining and processing
Waste treatment and disposal
Land development
Forestry
Agriculture
Recycling
Protective Features

Dikes, levees, dams
Barricades and railings
Revetments
Moats and trenches
Automatic control devices
By-pass storage
Cooling ponds and waste lagoons
Derailing prevention devices
Channel markers
Floating barriers

Evidence of Previous Problems

     Immediate Past
Vegetative alterations
Stains
Turbidity
Slicks
Fish kills
Thermal anomalies

     Remote Past
Drainage patterns
Vegetative alterations
Stains
Residues and deposits
Thermal anomalies
  Photointerpretation will yield only these categories plus  identifica-
  tion of material as raw input, output, or by-product.  Ground truth
  will provide specific identities, such as chemical composition, radio-
  activity, etc.
                                  42

-------
Finally, Indications of past spill problems would be good clues to areas
where potential spill problems may exist.  Therefore, categories were
constructed for ground features which might indicate recent or past spill
problems, including vegetative alterations, stains, thermal anomalies,
drainage patterns, and turbidity.

When a classification scheme such as this is illustrated with ground and
aerial views and structured to provide dichotomous choices in classifi-
cation through each level, it becomes an image interpreter's key.
According to the proposed scheme, a complete key would allow the inter-
preter to go beyond general identification of activity [e.g., refining
and processing] to specific classification (e.g., oil refinery) of the
facility or of sub-areas in the facility (e.g., tank farm) and still
further, to include types of hazardous material probably present (e.g.,
highly volatile vs. less volatile petroleum products).  Once the hazard
had been so classified, the image interpretation keys would lead the
interpreter into an analysis of associated hazards (leakage, rupture,
etc.), protective measures (revetments, catch basins), and indications
of past occurrences  [stains, vegetative alteration) in order that he
might assess the magnitude of potential or actual problems.

For purposes of the image interpretation tests in the present study, and
as a demonstration of a method of classifying and analyzing spill sources,
an abbreviated key to gross identification of facilities was constructed.
The key is based on classification of a target site by determination of
the type(s) of activity occurring.

In an operational situation, image interpretation keys would be extensive,
due to the variety of activities and processes which offer potential
hazards.  In addition to extent of coverage, the keys would need to be
periodically updated as industrial practices change.

While various objective measurements are commonly used to evaluate the
characteristics of the image structure on film, there is very little
correlation between such measurements and the amount of information that
can be extracted by trained image interpreters.  The human interpreter
is a very powerful information processor in whom many complex visual and
mental processes take place during the analysis of a photographic image;
these processes can govern the amount of useful information which he
derives from that image.  It is difficult to relate a photographic
quality measurement obtained at several points on a photograph to the
information an interpreter can perceive by viewing that same photograph
in its entirety.  Therefore, the information obtained by a group of
qualified interpreters from the various film records, rather than various
quantitative measurements of image parameters, was used as the basis for
rating the sets of photographs taken with different films and filters
and at various scales.

From preliminary study of the aerial photographs tentative conclusions
were drawn regarding the interpretability of each film-filter-scale
combination.  Image interpreters, all of whom were unfamiliar with  the
                                   43

-------
target areas, were asked to make judgements relating to the interpret-
ability of particular target features on these various combinations.
The responses to these questions were scored by comparing them with
"ground truth" data.  The resulting information was used to define
film-filter-scale specifications for an aerial surveillance spill  pre-
vention system.

Correlating the responses of a group of photointerpreters requires a
careful analysis of the types of information desired (identifications,
measurements, enumerations, comparisons), the quality of photographs
[sharpness, contrast, color or tonal fidelity, stereoscopic parallax),
the atmospheric and illumination conditions (haze, clouds, sun angle),
and the experience of the image interpreters.  Because the interpreters'
responses are subjective, there will usually be a spread in answers
which complicates the scoring system.

A questionnaire and a response sheet were prepared, on which each inter-
preter submitted his judgements.  Results of the image interpretation
tests are presented in section V.
                                 44

-------
                              SECTION V

                               RESULTS

The imagery in this section has been compiled to illustrate significant
findings and to demonstrate the usefulness of multispectral remote
sensing techniques for spill prevention.

The color photographs in the scale test should be used for familiariza-
tion with the increase in detail and the decrease in area covered as
scale denomination increases numerically.  The black-and-white photo-
graphs are useful for comparison of film/filter combinations and for
evaluation of the appearance of various easily recognized features
(such as waterways, roads, storage tanks, trees, and buildings) on
various images, including the oblique views.

Having made these comparisons, the usefulness of remote sensing for
classifying ground features, detecting potential sources of spills,
and locating actual discharges of oil and other hazardous materials
will be more obvious.  The difficulties in actual identification of
certain activities and materials is evident when one examines the
imagery.

In order to visualize the above relationships, many of the photographs
contained in this section are mounted as stereograms and should therefore
be viewed with a stereoscope for three-dimensional interpretation.  A
large number of features of interest to this study can be seen on single
photographs; however, their interaction with other man-made objects and
with the natural environment, as well as with topographic features
which influence spill threats, are best visualized by stereoscopic
viewing.

It may be noted that some photographs in the 70mm stereograms are out
of linear alignment due to drift and heading change of the aircraft as
these photographs were exposed.  In addition, the amount of overlap
between the two members of a stereogram may vary somewhat.  Despite
the minor annoyance which such problems may cause, it was considered
that the photographs should be presented in this report in the form in
which they came from the photographic mission in order to realistically
illustrate imagery from operational systems.  The advantage to viewing
such material in this direct manner stems from avoidance of the image
degradation inherent in any corrective procedures.

Small scale photographs such as those in Figure 12, taken with color
infrared film and a Wratten 12 filter (color IR/12), are useful in the
classification of areas into industrial activity categories.  The photo-
graph on the right side of the triplet at the top of the page (scale
1/80,000) shows an oil  refinery located at the confluence of Pacheco
Creek and the Sacramento River in the Avon test area.  On the right is
a large bio-oxidation pond, commonly associated with oil refineries.
At the bottom center of the right hand photograph, a coke pile (residue
from oil refining) can be seen.  Surface drainage features should be
                                  45

-------
           Tank Farm and Oil  Refinery, 1/80,000
           Oil Refinery, 1/40,000
            Tank Farm,  1/40,000
FIGURE 12.   AVON TEST AREA.   Color  IR/12,  August  20, 1971
                                 46

-------
noted and compared with Figure 20.  The left hand photograph of the
triplet shows a tank farm (clustered storage tanks) where petroleum
products are held for processing and shipment.  The center photograph
provides 60% graphic overlap for three-dimensional viewing of the imaged
scene.

Many of the features necessary to spill threat analysis cannot be dis-
cerned on these 1/80,000 scale photographs.  Such features include
pipeline routes, revetments around tanks, and transportation terminals.

The lower two stereograms are of the same area, but at a larger scale
(1/40,000).  Comparison of the features; observed on the upper photo-
graphs with those on the lower photographs demonstrates that it becomes
possible at this scale to discern some large protective features such
as dikes and levees; others, such as barricades and fences,  do not
image.  Note that storage tanks with white tops can be resolved, while
tanks with dark tops tend to blend together.

At this time of year (summer) in this area, most vegetation  does not
appear red on color infrared/12 photography.  The predominate vegetation
is annual grasses which have dried and are therefore not reflective of
solar infrared energy.  Wetland vegetation does appear red,  however,
because it is physiologically active and therefore has a high solar
infrared reflectance.  This figure should be compared with Figure 13.

The 1/80,000 scale color photographs taken with Kodak SO-397 film and
using a Wratten 1A filter (color/1 A)(at the top of Figure 13) are of
unusually high quality due to the absence  of  significant amounts of
atmospheric haze at the time of exposure.  The features seen on Figure
12 (tank farm, refinery, coke pile, bio-oxidation pond, etc.) should
be compared with the same features on these photographs.  At the
1/40,000 scale, it is possible to follow some pipeline routes and, at
the 1/20,000 scale (Figure 13), the pipeline routes are easily visible,
as are some supporting structures for elevated pipelines.

The stereogram at the top of Figure 13 should be compared with the
black-and-white photographs in Figure 19.  Although most features are
visible on the black-and-white photographs, qualitative evaluations
based on them are more difficult to make.  For example, on the black-
and-white photographs, one might mistakenly identify the tonal changes
seen on the outer sides of the oval-shaped reservoir in the center of
the 1/20,000 scale example as green vegetation, indicating seepage and,
thus, a point of potential levee failure.  On the corresponding color
photographs, the features can correctly be identified as seasonal,
brown vegetation and therefore not indicative of a point of active
seepage.  Green vegetation can be seen along the creek banks in the
upper half of the photograph.  Vegetation evaluation is more easily
accomplished using color IR/12 photography, as seen in Figure 22.
Algae floating at the border of the reservoir are also visible -

Seasonal changes relating to water level and vegetative conditions may
be important to spill prevention surveillance.  Such changes can affect
                                  47

-------
                1/80,000
                1/40,000
                 1/20,000




FIGURE 13.   SCALE COMPARISON, AVON TEST AREA.  Color/lA, August 20, 1971
                                48

-------
the magnitude of the threat where heavy rainfall weakens structures,
increases the volume of hazardous material which might spill into a
waterway, or causes flooding, inundating inadequately protected waste
storage areas.  Seasonal differences in vegetation can also mask impor-
tant features or clog important drainage pathways and holding areas.

The stereograms in Figure 14 were taken on color/lA at a scale of
1/10,000.  The upper four photographs were taken with an 80mm (normal)
lens and the lower three photographs with a 250mm (telephoto) lens.
Topographic relief is more easily visualized on the photographs taken
with the shorter focal length lens because of the larger lens angle
and the lower altitude required to achieve the same photographic scale.
When viewed stereoscopically, levees, tanks, and buildings appear to be
taller in the upper series of photographs than in the lower series.
This effect makes it possible to photogrammetrically estimate heights
with greater accuracy on photographs taken with a shorter focal length
lens and also to determine whether levees and revetments are incorporated
at each tank site.

An oil spill stain can be seen on the ground at "A".  An unrevetted tank
is located on the creek bank at "B".  Railroad tank cars can be seen at
the loading point at "C" and a fossil-fuel-power generating plant is
at "D".

The 1/5,000 scale color/lA photographs in Figure 15 provide more detail
than the previous photographs, but they cover less area.  At this scale,
it is possible to determine the presence or absence of such features
as elevated pipelines and supporting structures ("A"), protective fences
along pipeline routes ("B"), oil leaks at pipe valves ("C"), pipelines
running through levees  ("D"), pumps  ("E"), and dark stains on the ground
from past oil spills.  Levee structures can be seen around storage tanks
in the lower photograph, as can individual pipelines.  In some areas,
incipient levee erosion is apparent  ("F").

The photographs in Figure 16 are considered to be of unnecessarily
large scale (1/2,500) for the requirements of a spill prevention system.
No additional significant information for detecting spill sources was
gained by use of this large scale photography over that which was
obtained at smaller scales.  Perhaps in areas where high concentrations
of industrial activities are present, one would want to specify 1/2,500
scale photography, but then only in very limited areas because of the
time required for interpretation.  The oil spill at "A" on Figure 14
appears at "A" on this figure.  It is also possible to see fences and
gates on this photography ("B").  Interpretability of specific features
seen at this scale can be compared with interpretability at other scales
by returning to previous figures.  As a photointerpreter gains experience
in searching for potential spill sources, and as his efficiency increases,
he will be able to use smaller scales for interpretation.

The reddish catch basin, seen in this series of stereograms  (Figure 17)
at scales of 1/10,000, 1/5,000 and 1/2,500, contains diesel fuel which
                                  49

-------
01
o
O «-i
o i—

o c:

^L3
—• 73
J> 3»
>•  —I
  i—i
—' O
— -z.

o o
"  TI
o
o «c
o m

•  5
GO i—I
tD O
-a J>
r+ t—

§m
cr ><
0) 3>
-S CD
  CD
VD m
       1
                         80mm (Normal Lens)
       m
       oo
               W^'^i
                         250mm (Telephoto Lens)

-------

FIGURE 15.   LARGE SCALE COVERAGE,  AVON  TEST AREA.
            August 20,  1971
Color/lA,  1/5,000,
                                  51

-------

                                                     CO
FIGURE 16.   VERY LARGE SCALE COVERAGE,  AVON TEST AREA.   Col or/1 A,
            1/2,500,  August 20,  1971
                                  52

-------
                  1/10,000
                         ^
                                                I
i
                   1/5,000
                   1/2,500
FIGURE 17.   OIL DISCHARGE, AVON TEST AREA.  Color/1 A, August 20, 1971
                                 53

-------
was inadvertantly drained from the tank at the right of the basin.  The
same reddish liquid is present at "A" in the ditch along the roadway,
indicating further drainage of a quantity of the same type of material.
The presence of this material was brought to the attention of refinery
personnel, who subsequently reclaimed about 1,000 barrels of diesel
fuel which was floating on the water in the catch basin.  This oil
spill was unknown to refinery personnel, who indicated that much of the
product would have been lost to evaporation and seepage had it not been
detected by photointerpretation.

A tank has been removed at "B"; the potential spill  threat from tanks
on this hill is aggravated by the fact that the slope leading to the
waterway is quite abrupt in some places.  Although most detail is visible
at the largest scale, area coverage is so reduced as to make topographic
judgements nearly impossible.  On a 9-inch format, 17 times more area
would be covered at comparable scale, thus alleviating topographic
obscurity.

These photographs should be compared with those in Figure 18, in which
the upper row of photographs are aerial oblique views of the same scene.
The stereo pair of oblique photographs at the top is helpful in evalua-
tion of the spill threat from tanks on the hill, although some areas
cannot be seen as clearly as in the vertical view because of the topo-
graphy.   If stereographic vertical coverage cannot be obtained, the
photographic team can take stereo oblique photographs for evaluating
 .ndustrial activities by three-dimensional viewing.   The black-and-
white oblique photograph at left center does not reveal the reddish
color of the catch basin.  At right center is a ground view.

The 1/5,000 scale stereoscopic photographs at the bottom of Figure 18
show a catch basin with an overflow pipeline ("A") connecting the catch
basin with a drainage ditch.  Oil can be seen in the basin, which is
nearly filled to capacity with oil and water, thus restricting its
usefulness for receiving liquid from nearby tanks if a spill should
occur.

The photographs in Figure 19 were used in testing the interpretability
of the various scales and film/filter combinations employed in this
study.  These photographs should be compared with Figures 12-16 to
assess the interpretability of black-and-white photographs versus color
photographs; notice that it is difficult to delineate natural surface
drainage networks on this black-and-white photography.  A comparison of
the black-and-white photographs in Figure 20 with the color photographs
in Figures 12-16 indicates the interpretability of features such as
vegetation, oil stains, fences, pipelines, and storage tanks.

The delineation of water boundaries and drainage networks is facilitated
by the use of black-and-white  infrared photography taken with a Wratten
89B filter (B&W IR/89B), as shown in Figure 21.  Because water absorbs
almost all energy in the near-infrared region of the spectrum, water
will image black on a positive infrared print.  This dark tone contrasts
                                  54

-------
                   Color/lA, October 5, 1971


     Black-and-White IR/89B,
        October 5, 1971
Color/lA, September 23,  1971
                               \   X
           1/10,000,  Color/lA,  August  20,  1971


FIGURE 18.   OIL DISCHARGE,  AVON TEST AREA.
                                 55

-------
            1/80,000
            1/40,000
            1/20,000




FIGURE 19.   AVON TEST  AREA.   Pan/47B,  August  20,  1971
                                 56

-------
                  1/10,000
                  1/5,000
FIGURE 20.   SCALE COMPARISON,  AVON TEST  AREA.   Pan/47B, August 20, 1971
                                 57

-------
                   1/80,000
                   1/10,000
                   1/2,500


FIGURE 21.   WATER BOUNDARY  DELINEATION,  AVON  TEST  AREA.   Black-and-White
            IR/89B,  August  20,  1971
                                 58

-------
with most shoreline features (such as sand, vegetation, and rocks),
which image light in tone.  For evaluation of industrial facilities,
B&W IR/89B has some limitations.  Image sharpness is somewhat limited,
tonal contrast between structures may be reduced, and shadow areas are
very dark, all of which may obscure important detail.  The photographs
in this figure should be compared with those in Figures 12-16, 19, and
20.

The use of color IR/12 photography for vegetation studies has received
wide acceptance in the fields of agriculture, forestry, rangeland
management, and aquatic studies.  This "false color" film displays
features with a high solar infrared (700-900 nanometers) reflectance
as various shades of pink or red, depending on the intensity of the
reflectance.  Healthy vegetation has a high infrared reflectance, thu:.
its red coloration.  Dead vegetation does not appear red on such photo-
graphs because of its low infrared reflectance.  In aquatic studies,
submerged vegetation will not appear red because of the absorption of
infrared energy by overlying water.

In Figure 22, living vegetation along the creek banks and at various
other points  appears red.  Dead vegetation, such as on levee banks and
in open fields, appears blue or tan.  This should be compared with pre-
vious photographs to evaluate the appearance of various features of
interest.

The thermal infrared (8-14 micrometers) images in Figures 23A and 23B
were taken over the refinery at the Avon test area from 5,000 feet at
about 2300 PDT.  Warmer water (light toned) in Pacheco Creek, starting
at "A" on Figure 23A, has been imaged, as have cold areas (dark toned)
such as the covered reservoir at "B" and the wet meadow at "C".  Most
significantly, the overflow pond seen in Figures 17 and 18 is at  "D"
and exhibits  a high emissivity  (light in tone), characteristic of oil
on water.

The refinery  facilities previously detailed can be seen in Figure 23B.
The oil separator pond is visible at "E", heated water in the canal at
"F", and heated water being discharged at  "G".  These images should be
compared with previous figures for analysis of features seen in this
area.

The black-and-white stereogram in Figure 24 is at a scale of 1/80,000
and is typical of the type of photography available from existing sources,
in this case  the USGS.  The Martinez refinery is at "A", the Benecia -
Martinez bridge at "B", the Avon refinery at "C", and the chemical plant
at Nichols at "D".

This stereogram should be compared with the 1/80,000 scale photographs
in Figures 12, 13, and 19.  It will be noted that many features of
interest previously described can be seen in this figure, but since
color photographs provide information not available from black-and-white
photographs (colors of water bodies, vegetation, soils, buildings, etc.),
                                   59

-------
                 1/20,000
                  1/10,000
                  1/5,000

FIGURE 22.   VEGETATION ANALYSIS,  AVON  TEST AREA.   Color  IR/12, August
            20,  1971
                                 60

-------
FIGURE 23A.   THERMOGRAM, AVON TEST AREA.  2306 PDT, August 9, 1971 at
             5,000 feet
                                 61

-------
FIGURE 23B.   THERMOGRAM,  AVON  TEST AREA.   2306  PDT,  August  9,  1971  at
             5,000 feet
                                  62

-------
FIGURE 24.   NICHOLS-AVON-MARTINEZ.   1/80,000,  Panchromatic,  May 14,  1970
            by USGS
                                 63

-------
the use of the latter allows fewer judgements regarding potential  spill
sources.  When comparing other figures of the same area with  this  one,
it should be recognized that some of the variations are due to differ-
ences in dates of photography.

The upper stereogram in Figure 25 at 1/40,000 scale shows a refinery
at Richmond, with waste water holding ponds,  refinery facilities at "A",
and storage tanks at the top of the strip.  Waste materials are being
discharged into the San Francisco Bay at "B".  A large scale stereogram
of the waterfront at "B" is shown in the lower right and in the oblique
view at lower left.

The waste discharge at "B" was detected on the small scale (1/40,000)
photographs before the flight line for large scale coverage was selected.
This is a typical example of the way in which a spill prevention survey
should be conducted.  The preferred sequence is to obtain small scale
coverage, interpret that photography making the necessary judgements
regarding areas of interest for further study, locate large scale  flight
lines, and acquire large scale coverage.  Interpretations can then be
made relative to the potential spill threat by a careful evaluation of
the  levees, drainage ditches, holding facilities, waste piles, and
adjacent waterfronts.  The storage pond at "C" is contained in Figures
26 and 30-32.  An oil separation facility can be seen at "D" in the
lower stereogram and a patch of floating oil  at "E".

The  image in Figure 26 is from a thermal infrared scanner operating in
the  8-14 micrometer range.  As before, the tonal values represent  various
surface temperatures, light-toned features being hot and dark-toned
features being cold.  This line was flown at 10,000 feet at 0100 PDT
over the Richmond refinery area.  Water ponds at higher than ambient
temperature are visible at "A", and both hot and cold storage tanks are
clustered at "B".  The cooling pond at "C" is the one seen in Figure 25
at "C".  In this thermogram, a discharge of heated effluent is apparent
which was not visible on the corresponding photography.  This image
should be compared with the images in Figure 25 and 27-32.

Storage tanks exhibit various emission characteristics, depending  on
the  temperature of the fluid contained in the tank, the volume of the
fluid contained in the tank (and thus empty space above the fluid), and
the material from which the tank roof is fabricated or covered (painted
or bare metal, asphalt, wood, etc.).  It is therefore misleading to
conclude that the tonal value of the tank top image reveals something
about the volume of fluid in the tank.  However, an oblique view,  in
which the side of a tank is visible, will give some indication of the
volume of material stored in the tank, provided a temperature difference
can be recorded by the thermal scanner.  The tanks located at  "A"  in
Figure 27 exemplify these features.  The waste discharge point at  "B"
in this figure may also be seen in Figure 25, also at "B".  Individual
vents of cooling towers are visible at "C", and an oil separator is
visible at "D".  This thermal infrared image was taken from 1,500 feet
at about 0100 PDT over the Richmond refinery area.
                                   64

-------
        ro
                                 1/40,000, Color  IR/12, July  27,  1971
cn
O
n:


I
D

-H
m
oo
                   Color/1 A,  November  5,  1971
                                                      1/10,000, Color/lA, September 9, 1971

-------
FIGURE 26.   THERMOGRAM,  RICHMOND TEST AREA.   0100  PDT,  August 15,  1971
            at 10,000 feet
                                 66

-------
FIGURE 27.   THERMOGRAM,  RICHMOND TEST AREA.   0048  PDT,  August  15,  1971
            at 1,500 feet
                                 67

-------
Figure 28 is a thermal  infrared image of a refinery at the Richmond test
area, taken at about midnight from 5,000 feet.   The main discharge point
from the holding pond at "A" can be seen, as well  as seepage of warm
water along the levee.   A dispersion of warm water into San Francisco
Bay is visible at "B".   Thermal imagery taken at about midnight or some-
what later usually provides very useful information on surface tempera-
tures and emission characteristics.

The thermal infrared image in Figure 29 was taken  from 1,500 feet at
about 0100 PDT, the subject being the same holding pond in the preced-
ing thermal image.  At  this scale, it is possible  to see clearly the
main discharge from the pond, as well as several heated water seepage
points along the levee.  This should be compared with thermal  images
taken at 10,000 and 5,000 feet in Figures 26 and 28, respectively, as
well as with Figure 32, which shows color aerial photographs of the
holding pond.

Figure 30 is a thermal  infrared image of the Richmond refinery area
from 5,000 feet taken at about 2200 PDT.  Some recognizable features
can be seen on this image, even though imagery acquired in early night-
time is not the most desirable.  Some surface emission from daytime
solar heating remains during pre-midnight hours and usually results in
undesirable imagery; post-midnight hours are therefore recommended for
thermal infrared flights.  Storage tanks are located.at "A", a cooling
point at  "B", warm water drainage from a refinery at "C", and a pleasure
boat harbor at "D".

Figure 31 is a daytime thermal infrared image of the Richmond area,
taken at 5,000 feet at about 1600 PDT.  In order to record detail on the
land areas, it was necessary to adjust the scanner to record the water
areas as black (cold).   In reality, the water temperature difference
between night and day was slight, while the land temperature difference
between night and day was as much as 30°F.  The holding pond at "A" is
seen to exhibit the same tonal value as the bay water; thus, the seepage
detected at night on thermal imagery is not detectable here (however,
tidal differences were probably a factor as well).  One might also be
tempted to conclude that both the bay and the pond were at the same
temperature, which is most certainly not the case.  This ambiguity arose
because the thermal scanner was deliberately "tuned" to record land
emissivity data and not water data.  The temperature difference between
land and water may be a few degrees late at night, while, in the daytime,
it may be 20 degrees or more.  Because the land cools so rapidly after
sunset, thermal imagery taken late at night can usually record emissivity
information over both land and water areas without losing detail in
either one.  This can only rarely be accomplished during daylight or
early evening hours.  Figures 31, 30, and 29 should be compared, noting
the contrasts from daylight to midnight.

By comparing the thermal infrared images in these figures with the color
photographs of the same area, one can evaluate the utility of thermal
infrared images in spill prevention surveillance.   It is suggested that
thermal infrared imagery is useful only where features of interest are
                                  68

-------
FIGURE 28.   THERMOGRAM, RICHMOND TEST AREA.   0012 PDT,  August  15,  1971
            at 5,000 feet
                                  69

-------

FIGURE 29.   THERMOGRAM,  RICHMOND TEST AREA.  0048 PDT, August 15, 1971
            at  1,500  feet
                                  70

-------
FIGURE 30.   THERMOGRAM, RICHMOND TEST AREA.   2207 PDT,  August 9,  1971
            at 5,000 feet
                                  71

-------
FIGURE 31.   THERMOGRAM,  RICHMOND TEST AREA.   1557  PDT,  August 9,  1971
            at 5,000 feet
                                  72

-------
visible as a function of their temperature or thermal emission difference.
The imagery presented demonstrates that, for purposes of spill prevention,
unique information rarely occurs on thermal imagery.

The stereo photographs at the top of Figure 32 were taken to permit
detailed evaluation of the levee condition and the drainage seen on the
thermal images.  On the low altitude oblique photograph at lower left,
a drain pipe can be seen passing through the levee at "A".

The color photographs in the lower right part of this figure were taken
of a color television display produced by an image enhancement device
(International Imaging Systems, Digicol),  With this device it is possi-
ble to "contour" density difference by assigning a unique color to each
density, thus making a multicolor  display from a black-and-white image.
A technique called "density slicing" permits the operator to measure the
relative areas of the various densities automatically by adjusting the
gain setting.  The relative areas portrayed on the imagery for a parti-
cular water temperature can be determined by this device, which aids in
monitoring the dispersion of heated effluent into a body of water.  The
holding pond seen in the previous imagery is displayed as a yellow or
greenish-yellow feature in these images, which were created electronic-
ally from the images in Figures 26 and 29.

Figure 33 illustrates a waterfront with several potential spill sources.
The upper photographs were taken at a scale of 1/40,000 for a regional
survey of the Richmond area, using color IR/12 photography.  An oil
catch basin holding heavy oil waste from tanker bilges can be seen at
"A".  The basin is seen on color film, at a scale of 1/10,000, in the
center strip of photographs.  It is also possible to see dark stains
along the shoreline of the pond and containers adjacent to the pond.
An oil separation facility is located at "B"; this feature is difficult
to identify on the small scale coverage.  At "C" are stored materials
awaiting transport.  The two photographs in the lower row, one in color
and one in black-and-white, provide a means for comparing the usefulness
of each type of oblique photography.  The two photographs are mounted
for stereoscopic viewing to facilitate comparison.  By alternately
blinking his eyes while viewing in stereo, the observer can make meaning-
ful comparison.

A steel salvage operation at Richmond is seen in Figure 34 at "A".  A
cluster of tanks can be seen at "B" and at several other locations in
this area.  A manufacturing plant with some shoreline residue visible
is at  "C".  Disposal practices near the shoreline are visible in the
center and lower photographs.  The vessel in the pond behind the earthen
dam is surrounded by an oil slick.  The black-and-white photograph is
included for comparison of information obtainable from both color and
black-and-white photographs.  Oblique photographs, such as those in the
lower part of this figure which show scrap steel and associated materials,
are very useful for evaluating potential spill hazards.   Use of low
altitude oblique photographs may eliminate the need for a time-consuming
ground visit.
                                   73

-------
CO
ro
1C
o

o
I—I
•z.
CD



I
o
o

o
m
oo
                         1/10,000, Color/lA,  August  20,  1971
                                                          0012 PDT,  August 15,  1971 at 5,000 feet
             Color/lA, October  5,  1971
                                                          0048 PDT, August 15, 1971 at 1,500 feet

-------
          1/40,000,  Color  IR/12, July 27, 1971
          1/10,000,  Color/lA, September 9, 1971
                                      R.,-'.
           1  .C'tt



       Col or/1 A,  November  5,  1971





FIGURE 33.   OIL WASTE POND,  RICHMOND TEST AREA.
                                 75

-------
           1/40,000,  Color IR/12, July  27, 1971
                               •••Kfe^milfre
           1/10,000,  Color/lA, September 9, 1971
       Pan/47B, September 17, 1971    Col or/1 A, September 17, 1971
FIGURE  34.  STEEL SALVAGE, RICHMOND TEST AREA.
                               76

-------
Figure 35 shows  industrial  and mining  activities.  The upper photographs
were taken of the manufacturing plant  producing prefabricated concrete
products, seen at "C" in Figure 34.  Residual material is piled along
the shoreline.  The center  photograph  shows a rock quarry at the
Richmond test area at "A".  Storage tanks are clustered on the hill.
Trucks and earth-moving equipment can  be seen in the parking area in
the lower stereogram at "B".

Figure 36, of the Oakland Estuary area, illustrates an area with high
shoreline utilization, where both polluting and non-polluting operations
are collocated.  The color  IR/12 stereogram was used for an initial
survey to select features of interest.  A steel salvage operation is
visible at "A", with residue piles and storage tanks at the water's
edge.  There is a containerized cargo  transshipment point at "B", ship
loading cranes are also visible.  Natural gas storage tanks are located
at "C" and a railroad yard  at "D".

The color/lA stereogram in  the center  shows, at larger scale, the steel
salvage operation visible at "A" on the small scale color IR/12 photo-
graphs.  Adjacent to the salvage operations is a small chemical plant
at "E".  The tanks at "F" are abandoned (this fact was determined by
ground checking and could not be determined by photographic interpreta-
tion).  A residue pile extending into  the water is visible at "6".

The oblique photographs at  the lower left of Figure 36 are useful for
detailed interpretation of  industrial  activities.  The ground view shows
the residue pile at "G" and adjacent abandoned tanks.

The sequence of photographs in Figure  37 shows a copper ore residue
(roasted pyrite) pile, first observed  and photographed on a reconnais-
sance flight.  The small scale color IR/12 photographs at upper right
reveal the residue pile and a color discharge into the water at "A".
An oil refinery can be seen at upper right, with ship-loading docks at
"B".

The lower photographs were  taken at later dates and show removal of the
residue pile from the shoreline.  The  lower right photograph is a ground
view of the residue pile at the waterfront, taken after removal had
begun.  This type of photographic presentation has been used to illus-
trate the sequence of corrective measures following enforcement actions.
In this case, corrective measures followed issuance of a Cease and Desist
Order by the California Water Resource Control Board.

Figure 38 illustrates a chemical plant and associated waste disposal
facilities ('"A") for manufacture of alum, various acids, ammonium hydrox-
ide,  and other products, as determined by interview with plant person-
nel.  The waste material is slurried lime, used to neutralize the by-
products.  The oblique photographs in  the center show the residue
holding areas and dikes.

The large scale photographs at the bottom show the area in more detail
and reveal a point where the levee is  very narrow ("B"), a condition
                                   77

-------
            1/10,000,  Color/lA,  September 9, 1971
            1/40,000,  Color IR/12,  July  27,  1971
            1/10,000,  Color/lA,  September  9,  1971




FIGURE 35.   INDUSTRIAL AND MINING OPERATIONS,  RICHMOND TEST AREA.
                                 78

-------
               1/40,000, Color IR/12, July 27, 1971
               1/10,000,  Color/lA, September 9, 1971
                                                     Ektachrome X,
                                                     September 15, 1971
Color/lA, September 13,  1971    Color/lA, July 19, 1971
FIGURE 36.   HEAVY INDUSTRY,  OAKLAND  ESTUARY TEST AREA.
                                  79

-------
         CO
         73
         m
         GO
         i—i
         O

         m

         -a
         i—i
         I—
         m
00
o
        2
                      Color/lA, July 19, 1971
                       1/10,000,  Color/lA,

                        September 9,  1971
         1/40,000, Color IR/12, July 27, 1971
Color/lA, September 13, 1971    Color/lA, September 23, 1971

-------
            1/40,000,  Color  IR/12,  July  27,  1971
           1/10,000, Col or/1 A,  September 9,  1971




FIGURE 38.   CHEMICAL PLANT,  NICHOLS TEST AREA.
                                 81

-------
which might lead to failure and a spill  of waste materials.   The circu-
lar canal at "C" is used to hold cooling water from the chemical manu-
facturing processes.

Figure 39 presents the recognition characteristics of a fossil  fuel
power generating plant ("A").  The upper photographs were taken on
color IR/12 at a scale of 1/40,000 and demonstrate a problem caused by
weather; low clouds over the Hunters  Point  test  area  obscure  part  of
the ground scene.  An underwater sewage discharge is located at "B".

A power generating plant can usually be recognized by the presence of
tall vent stacks, nearby fuel storage tanks, cooling water intake and
discharge facilities, power transformers, transmission facilities, and
the absence of raw material and/or waste piles,  large employee parking
lots, and finished products stored for transport.

The installation seen in the center and lower photographs of Figure 39
fits these criteria and was correctly identified by photointerpretation.
On the center photograph, the wall-like revetments around the four tanks
to the rear of the power plant appear to be as tall as the tanks; that
this is not the case is difficult to determine on a single photograph,
but is strikingly simple in three-dimensional viewing.

The lower stereogram is composed of low altitude oblique photographs.
An intake channel with debris from a nearby land fill operation is
visible, as is a turbulent cooling water outfall channel.

Using the same criteria for power plant identification described in
Figure 39 and in the photointerpretation key, the feature seen in
Figure 40 was correctly identified by photointerpretation as a fossil
fuel power generating plant.  This facility, at Pittsburg on the
Sacramento River, can be seen at "A" on the 1/40,000 scale color IR/12
photographs and on the 1/10,000 scale color/lA photographs.   An oil
separator pond and two adjacent empty waste water evaporation ponds are
apparent at "B", as are water intake pumps on the shoreline at "C", and
power transformers at "D".  The dark streak in the river at "E" is a
discharge from a sewage treatment plant at "F".   The appearance of the
sewage treatment facility and outfall on the aerial oblique views
should be compared with the vertical views.

The capacities of storage tanks and associated revetments was estimated
by photogrammetry from these 1/10,000 scale photographs  (Table 5) to
evaluate the value of simple photogrammetric instruments in mensuration
of spill prevention photography.  No attempt was made to measure earthen
revetments or irregularly shaped objects; the major effort was given to
the determination of storage tank capacities and the capability of
simple vertical wall revetments, such as those at the power plant site,
to retain the contents of the enclosed tanks.  Errors in photogrammetric
measurements were the result of inability to determine accurately the
exact scale of individual frames, parallax in the tube magnifier, coarse
graduation of the measuring reticle, and most importantly, human error.
                                  82

-------
         1/40,000, Color IR/12, July 27, 1971
          1/10,000, Color/lA, September 9, 1971

                         Color/lA,  September  13, 1971
FIGURE 39.   POWER PLANT,  HUNTERS  POINT  TEST  AREA.
                                 83

-------
 CD
-a
o
GO
CO
              1/40,000,  Color IR/12,  July 27,  1971
                                                    Color/lA, July 17,  1971
m
GO
m
}
   ^.M$*I':^LmL Z=Z.  r*L
                           1/10,000,  Color/lA, September 9, 1971

-------
           Table 5.  Photogrammetric Measurements, Avon-Pittsburg
Tank Number         Photo Height        Photo Diameter     Actual  Height
 (Avon)                (Feet)              (Feet)             (Feet)

80-235                  33.1                 113               41.8
80-231                  36.5                 113               41.7
Reservoir                --                  515
55-267                  36                   112               30
104-217                 40                   120               54
104-26                  51                   120               54
15-696                  46.2                  46.5             47.2
80-517                  37                   110               41.5
80-2                    37                   110               42
40-642                  44.2                  77               48
25-428                  28.5                  77               29.3
56-209                  36.7                  97               40
10-214                  22.3                  54               24
15-391                  27                    56.5             29.3
7-502                   40                    34.7             40.4
55-32                   33                   109               30
25-509                  26.6                  75               29.5
80-317                  34.7                 114               41.8
20-133                  26.9                  65.5             30
315-690                 54.5                 199               56
66-142                  33.4                 111               35
PG&E Pittsburg
  Tank                  45.3                 150               48
  Revetment             23.6                 285               21.5
film:   Color  (SO-397), roll #13
scale:  1/10,000
date:   September 8-9, 1971
                                  85

-------
                           Table 5.  (cont.)
Tank Number
 (Avon)

80-235
80-231
Reservoir
55-267
104-217
104-26
15-696
80-517
80-2
40-642
25-428
56-209
10-214
15-391
7-502
55-32
25-509
80-317
20-133
315-690
66-142
Actual  Diameter
    CFeet)
     117.4
     117.4
     515
     115
     120
     120
      47.5
     120
     115
      80
      78
     100
      56
      60
      36
     115
      78
     117.4
      75.4
     200
     116.3
Photo Volume
   CBbls.)

   60,000
   65,000

   65,000
   81,000
  103,000
   14,000
   65,000
   65,000
   37,000
   22,700
   47,000
    9,000
   11,500
    7,000
   55,000
   20,000
   66,000
   15,000
  300,000
   58,000
Actual Volume
   CBbls.)

   80,000
   80,000
1,000,000
   55,000
  104,000
  104,000
   15,000
   80,000
   80,000
   40,000
   25,000
   56,000
   10,000
   15,000
    7,000
   55,000
   25,000
   80,000
   20,000
  315,000
   66,000
 % Error
Q/olume)

   -25
   -19

   +18
   -22
   -1
   -7
   -19
   -19
   -7.5
   -9
   -16
   -10
   -23
    0
    0
   -20
   -18
   -25
   -5
   -12
PG&E Pittsburg
  Tank
  Revetment
     160
     299.3
  140,000
  270,000
  168,000
  270,000
   -17
    0
                                 86

-------
The use of larger format Imagery would greatly reduce such errors; this
is especially true because scales are only approximately known from
altitude and lens focal length and are most precisely computed where
individual frames cover a reasonably large area.

Figure 41 shows a steel mill and nearby waste material holding areas at
Pittsburg ("A").  Waste water holding ponds are visible at "B" and "C"
on the small scale, 1/40,000, color IR/12 photographs and can also be
seen on the larger scale, 1/10,000, color IA photographs at center and
lower left.  The oblique photograph and stereogram in the lower row show
the pond seen at "C" in the center left photographs.  Note the chemical
plant at "D" in the 1/10,000 scale stereogram.  One should be able to
see waste water ponds on the photographs in this figure.  It should also
be seen that very little care is exercised in maintaining the integrity
of these waste water basins to prevent uncontrolled spills.

Figure 42 shows a variety of transportation and loading facilities for
oil and other petroleum products.  The upper photographs show the dock
("A") at the oil refinery at Martinez; pipes are visible on the pier.
These photographs should be compared with the pier at "B" on Figure 37.
A ship refueling area at Richmond is seen at "B" on the center photo-
graphs.  Unprotected pipelines connect tanks on the hill with the dock
area.  The dark materials in the water on the inner side of the pier at
"F" could be a small oil spill.

The railcar loading facility for pressurized gas at Avon is seen at "C"
on the photograph at the center of this figure.  A truck loading facility
is at "D", and railcars for oil transport can be seen at "E".

In recognition of the need for a system for classifying land use activi-
ties which may present current or potential hazards, a classification
plan for spill prevention studies was designed as previously stated.
The next logical step was to develop, from the classification plan, an
aid to identification by photointerpreters at each level.  When struc-
tured as a series of pairs of mutually exclusive statements and illus-
trated with aerial imagery, the plan becomes an image interpreter's key,
to be used by persons not necessarily familiar with the varied activities
and processes encountered in spill prevention.  A representative key to
first order classification (Appendix C) was devised to demonstrate the
format and for use with the photointerpretation test later described.
Development of the key beyond the first order of classification and
development of illustrative and collateral material for such a compre-
hensive key were beyond the scope of this project.

The results presented for Figures 12 through 42 are illustrative of the
photointerpretation performed on the imagery acquired during this project.
This interpretation included evaluation of the utility of the various
film/filter combinations at each of the various scales imaged,  for the
purpose of detecting the features of interest in spill prevention sur-
veillance; Table 6 graphically presents these results.  Highlighted
entries represent combinations of scale and film/filter, wherein a
particular feature can be readily identified and classified.  In the
                                   87

-------
           1/40,000,  Color  IR/12, July  27, 1971
           1/10,000,  Color/lA,  September 9, 1971
Color/lA,  July 17,  1971
1/10,000,  Color/lA,  September 8,  1971
FIGURE 41.   STEEL MILL  AND WASTE,  PITTSBURG TEST AREA.

-------
        1/10,000, Martinez Test Area, Col or/1 A, September 9, 1971
         1/10,000,  Richmond  Test  Area,  Col or/1 A,  September 9,  1971
          1/5,000,  Avon Test Area,  Color/lA,  August  8,  1971
FIGURE 42.   OIL SHIPMENT FACILITIES.
                                  89

-------
Table 6(a).  Results of Photo-interpretation

X - Detectable
0 - Undetectable

N - Not recommended

E - No example
1. BULK
STORAGE




DIKES







DRAINAGE


PIPING




2. INDUSTRIAL
WASTE



number of tanks

material stored
in tanks
structural con-
dition of tanks
leaks or seepage
around tank area
condition of
dikes
volume behind
dikes
potential failure
points
compacted or
loose soil banks
drainage piping
thru dikes
valve condition
above or below
ground
condition of
pipes
trash and debris
in dikes

type of material
volume of lay-
over or pile
detention time
inflow rate
COLOR IR/12

o
o
o
"
o
oo
r-H
X-


0
0
0
n

o

n

0

n

0
0

0
n


0
n

0
o
o
o
«
o
^r
r-{
' X


0
0
X-
n

0

0

0

n

0
n

0
n


0
n

0
o
o
o
r>
o

-------
Table 6(a).  (cont.)
COLOfy 1A
o
o
o
f\
o
co
iH
X-
0
0
X-
0
0
0
0
0
0
0
0
0
0
0
0
0
o
o
n
o
•• •>
Jt-
•. *W^
0
0
''.'.. '" •
0
0
0
0
X-
0
0
0
X-
0
0
0
o
0
o
w
o
fsl
tH
.-.' '
XV
0
0
fc.
- 'J&-
X-
0
0
0
^
fs $' f
. ? .•".
0
X-
0
*;;
X-
0
0
o
0
o
*>
o
iH
.H
*
0
0
s •. s •"
''%{<
%
*: ~, ,
:%3
,'..?< f
''a£
' ' v s:
X-
'--, '''.
0
fy','
4%>
'(' ,..tf
0
'X,
%t
ii
0
o
0
o
•V
U1
rH
<.
,a?"
0
0
i>
.,'"•„
"«J" !
^Ai,,
"*?''<
** '
#!
¥<:
\"$
,-. ?
^X;,/,
«.' X
--J1X.! ( .^
0
'i«
0
,!/•
&'
*/- x;
j.'.. '.'»i.4/».
0
o
0
in
cs
i-H
INFRARED/89B
o
o
o^
o
oo
r-H
x " o
0
X-
''A,' :
/A :
-a,;
£-
1?;
!,; *
ttg '
,!&'•;
!>f '•'•:
,<* -'
:x-'
-*,
'-X'5-"'
"".-. ^ ^
0
*/'
f/>f
X-
3t
» '
''A'll.
0
0
0
X-
0
0
0
0
0
0
0
0
0
0
0
0
o
o
o
**
o

0
0
,&
0
:V
0
ir
0
:/lK
0
PAN/47B
c5
O
0
«s
o
oo
^H
N
N
N
N
N
N
N
N
N
N
N
N
.
N
N
N
o
o
o
#1
0
•
-------
Table 6[b).  Results  of  Pliotoiinterpretation

X - Detectable
0 - Undetectable
N - Not recommended
E - No example



DIKE
CONDITION



EFFLUENT
LOCATION
3. OPEN
STORAGE
PILES




4. PIPELINES
OVER
WATERWAYS


5 . RAILROADS
AND TANK
CARS
6 . MARINE
TERMINALS

source of waste
product
compacted or
loose soil
potential failure
point
leaks
discharge point
to river
volume of
material

type of material
drainage pattern
observed drain-
age in waterway
condition of
lines
condition of
holding structure
leaks

type of material
spillage in
transfer
transfer
operations
oil in harbor
COLOR IR/12

o
0
0
o
CO
•H
X-

0
0

0
0

0

0
X-

0
0
n

0

0
0

0
E
0
o
o
0
-3-
tH
X :

0
o

X-
X-

u

0
X

&
0
n

0

0
0

&
X-
0
o
o
o
C-J
.H
X

0
o

X-
X

u

0
X

X
0
0

X-

0
X-

X
X
0
o
o
S
t-i
X.

X-
X"

X
X

X-

0
'*

X
X-
X-

X

0
X

X
X

o
o
o
LO
tH
X

X
X

X
X

X

X-
X

X
X-
X-

x

0
X

X
X

o
o
CS
tH
X

X
*

X
X

X

X-
X

X
X-
X-

*<

0
1

X
X
                92

-------
Table 6(b).  (cont.)
COLOR/ 1A
o
o
0
0
00
1-1
X-

0
0
0

0
0
0
0
X-
o
o
o
o
.H
•;&*

0
0
*x;'

X-
0
X-
X-
;'fx:;
' *.'.',..
0
0
X-

0
0


X-
E
0
0
v •• '->.-!
;* '''f
0
X-

fV ' ; £
/?"•* *^i
E
o
o
o
0
CM
iH
«'

0
0
.^
;
;X ',
0
X-
*'.
*.
'"''.•.„'
0
0
v<
..„. ...'.J:
0
:-i;
'•n?i' ttv-
;' ^.-
'^",
E
o
o
0
0
tH
^
£
,
'*'
i;
,*-

X' '
':*'
*,
'*'
:.*'
• -
0
0
-X

0
2*v
' "" s^
^
$'f
i';.
o
o
o
LO
1-1
X !

? '.
i:'
& '
v" .,
X
*''
'/f /
Jt'
X
.'L.'.'f
X-
X-
'X '

0
.X
' "'••',

"%••
"'&;
O
o
«s
CM
1-1
X
INFRARED/ 89 B
o
o
o
A
O
00
1-i
0
,
1
*
r
, ,
;x
'x''
'%
•'*
:x
'V
X-
X-.
*
^
0
t
%

•!k
f¥
0
0
0

0
0
0
0
0

0
0
0

0
0


0
0
o
o
o
o
iH
0

0
0
0
*
0
0
0
IT
0

0
0
0

0
0


0
0
o
o
o
o
rH
X-

0
0
X-

0
0
0
x
0

0
0
0

0
X-


0
0
o
0
0^
0
i— i
i-l
v •
*v
^ *
0
0
'It

X-
X-
0
*
0

0
0
'X, '
' -
0
* .

,,
*\
0
0
0
0

•H
X "

0
0
x

X-
X-
0
•X
0

0
0
X ,
-A. , 	
0
X "
..
•'
X
0
o
o
U1
CM
T-H
X
'
0
0
x ;
,
X
X
0
X
0

0
0
X

0
X


X-

PAN/47B
o
o
o
o
00
^
N

N
N
N

N
N
N
N
N

N
N
N

N
N


N
N
o
o
o
•s
0
^
N

N
N
N

N
N
N
N
N

N
N
N

N
N


N
N
o
0
0^
0
CM
1-1 '
X-

0
0
u
0
0
iH
3£

0
0
x-l'V

X-
0
0
X-
X-

0
0
0

0
0


X-
V
X
0
0
X
X-

0
0
X-
0
0
o
c
1-1
X '

0
0

,
X
X-
0
%
X-

0
0
X
1- 	
0
0


X-
':*
0
X
•:.-.:

/x*:
'*
0
o
UO
CM
i—t
-X

0
0
X
11
,x
x"
0
x
X-

0
0
'k

0
'x.

1
;x
'X
          93

-------
Table 6(c).  Results of Photointerpretation
X - Detectable
0 - Undetectable
N - Not recommended
E - No example
7. REFINERIES
AND
PROCESSING
PLANTS
8- tt&W1-
SCHEME
9 . PROTECTIVE
FACILITIES
transfer
location of raw
materials - crude
type of raw
material
drainage pattern
condition of
storage tanks
condition of
pipelines
location of
effluents
water condition
near outfall
able to classify
fences along
lines
railroad car
barricades
dikes present
or absent
tracing pipeline
routes



COLOR IR/12
o
o
o
o
00
iH
0
0
0'
X-
0
0
X-
0
*
0
0
0
0



0
o
0
^
0

x :
X
, *
X
;
•X
X



               94

-------
Table 6(c).   (cont.)
COLOR/ 1A
0
o
o
0
CO
rH
0
0
0
0
0
0
X-
X-
"X
0
0
0
0



o
o
o
*
d>

m
T-H
X,.
0
0
>&
0
0
k
0
•3f
X-
?
X-
X



o
o
m
*v
CM
iH
X
0
0
X
0
0
X
0
*
X
*
X
r (
: x
X



PAN/47B
o
0
o
r>
o
oo
!— 1
N
N
N
N
N
N
N
N
Xf
^^ 1
N
N
N
N



o
o
o
*\
o
-3-
rH
N
N
N
N
N
N
N
X-
X
N
N
N
N



o
o
0_
0
cs
i-H '
0
0
0
X-
0
0
X—
X-
X
0
X-
0
X-



^)
0
0^
0
iH
iH
X-
0
0
1
0
0
X-
X-
X
X-
X
0
X—



o
0
0
m
rH
X *
0
0
X
0
0
X-
X-
X
X-
X
X-
3C



o
o
m
CN
iH
X
0
0
X
0
0
X
\r 	
X
V-
X
X
X



       95

-------
case of panchromatic film with Wratten 47B filter,  many applications
are not recommended, as this film/filter combination is extremely sensi-
tive to degradation of information content by intervening atmospheric
haze.

The features which are listed in Table 6 as undetectable (0) for all
film/filter combinations  and all scales had to be identified through a
ground visit to the site.   Those features which were marginally detect-
able  (X-) with a given film/filter combination at  a given scale required
supporting ground data for complete identification.  The following list
includes features which required a ground visit for proper identification:

               0  Material  stored in tanks
               0  Structural condition of tanks
               0  Condition of control  valves in pipelines
                  or tanks
               0  Condition of pipes and holding structures
                  of elevated pipelines
               0  Inflow  rate and detention time of indus-
                  trial waste
               0  Identity of material  in railroad  tank cars

Assistance in accomplishing these identifications was obtained by inter-
views with key personnel  at the power plants in Pittsburg and at Hunters
Point, the refineries at  Avon and Martinez, and the chemical plant in
Berkeley, among others.

Qualified photointerpreters not associated with the project were given
a test, using project imagery and the photointerpretation key in order
to arrive at an unbiased  comparison of the findings.   Results of the
test substantiated the conclusions previously drawn and are presented
in Appendix D.
                                  96

-------
                             SECTION VI

                          ACKNOWLEDGEMENTS

Earth Satellite Corporation wishes to acknowledge the support and guidance
provided by the EPA Project Officer, Mr. John Riley, without whose help
the successful completion of this study would have been considerably more
difficult.

Special thanks are due Mr. Benjamin Jones, Phillips Petroleum Company,
and Mr. Mark Samii, Pacific Gas and Electric Company, for the information
they provided concerning facilities, activities, and procedures in their
respective fields.  Employees of the California Water Resources Control
Board provided considerable information.

Project Manager for this study was Dr. Paul M. Maughan.  Dr. Robin I.
Welch served as Principal Investigator, and Dr. Allan D. Marmelstein
was Co-Tnvestigator.  Other key EarthSat participants included Dr- Robert
N. Colwell and Mr. 0. Ray Temple.
                                   97

-------
                              SECTION VII

                              REFERENCES

1.  Avery, T. Eugene, 1968, Interpretation of Aerial Photographs, Second
    Edition, Burgess Publishing Company, Minneapolis.

2.  Chandler, P. B., Dowdy, W. L., and Hodder, D. T., 1970, "Study to
    Evaluate the Utility of Aerial Surveillance Methods in Water Quality
    Monitoring:   Final  Report SD 70-565 State of California Water Resources
    Control Board Contract 9-2-42, North American Rockwell.

3.  Chandler, Philip B., 1971, "Oil  Pollution Surveillance", AIAA Joint
    Conference on Sensing Environmental  Pollutants,  Palo Alto, California,
    No. 71-1073.

4.  Chisnell, Thomas C.  and Colt,  Gordon E., 1968, '"Industrial  Components'
    -- A Photo Interpretation Key  on Industry", Photogrammetric  Engineer-
    ing, 24(4),  pp.  590-602.

5.  Conomos, T.  J.,  et_ al_, 1971, "Drift of Surface and Near-Bottom Waters
    of the San Francisco Bay System, California", San Francisco  Bay
    Region Environment  and Resources Planning Study, U. S. Department
    of Housing and Urban Development, Research, and  Technology,  U. S.
    Department of Interior, U. S.  Geologic Survey, San Francisco,
    California.

6.  Conrod, Alfred C. and Rottweiler, Kurt A., 1971, "Water Quality
    Measurements with Airborne Multispectral Scanners", AIAA Joint
    Conference on Sensing of Environmental Pollutants, Palo Alto,
    California,  No.  71-1096.

7.  Department of the Army, 1967,  Aerial Photomath Workbook, SUPR 62048,
    U. S. Army Intelligence School,  Fort Holabird, Maryland.

8.  Department of Defense, 1967, Image-Interpretation Handbook,  Vol.  I,
    NAVAIR 10-35-685, Naval Reconnaissance and Technical  Support Center,
    Washington,  D. C.

9.  Edgerton, A. T., Meeks, D., and  Williams, D., 1971, "Microwave Emission
    Characteristics  of Oil Slicks",  AIAA Joint Conference on Sensing  of
    Environmental Pollutants, Palo Alto, California, No.  71-1071.

10.  Goolsby, A.  D.,  1971, "Water Pollution Detection by Reflectance
    Measurements", AIAA Joint Conference on Sensing  of Environmental
    Pollutants,  Palo Alto, California, No. 71-1069.

11.  Guinard, N.  W.,  1971, "Radar Detection of Oil Spills", AIAA  Joint
    Conference on Sensing of Environmental Pollutants, Palo Alto,
    California,  No.  71-1072.
                                  99

-------
12.  Harvey,  D.,  Kohler,  R.,  and Myskowski,  G.,  1961,  "Photographic
    Considerations for Aerospace Reconnaissance",  Technical  Note No.  4,
    Wright-Patterson AFB Contract No.  AF 33C616) 8126,  Itek  Laboratories,
    Lexington,  Massachusetts.

13.  Hines,  W.  G.,  and Palmer,  R.  H.,  1971,  "Municipal  and Industrial
    Wastewater Loading in the  San Francisco Bay, California,  1970",
    San Francisco  Bay Region Environment and Resources  Planning  Study,
    U.  S.  Department of the  Interior,  U.  S. Geological  Survey, and
    U.  S.  Department of Housing and  Urban Development,  Research,  and
    Technology,  San Francisco,  California.

14.  Klemas,  Vytautas, 1971,  "Detecting Oil  on Water:   A Comparison of
    Known  Techniques", AIAA  Joint Conference on Sensing of Environmental
    Pollutants,  Palo Alto, California, No.  71-1068.

15.  Mi Hard, J.  P. and Arvesen, J. C., 1971, "Effects  of Skylight
    Polarization,  Cloudiness,  and View Angle on the Detection of Oil
    on  Water",  AIAA Joint Conference  on Sensing of Environmental
    Pollutants,  Palo Alto, California, No.  71-1075.

16.  Welch,  R.  I.,  1971,  "The Use of  Multiband Aerial  Photography  for
    Water  Resource Management", Ph.D.  Dissertation, University of
    California  at  Berkeley.
                                  100

-------
                               SECTION VIII

                                 GLOSSARY

Collateral Data - All sources other than the imagery being interpreted
which contribute to the body of data on a given target.

Displacement, Radial - Dimensional changes in photography caused by lens
characteristics and relief differences between object location and the
point on the imagery corresponding to the ground position vertically
beneath the camera (nadir).

Format - The actual size of the negative or positive transparency on
which an image is produced.

Ground Truth - Actual conditions, identifications, sites, and other
information on objects of interest in image interpretation.

Imagery - Representations of objects produced electronically or optically
on film, electronic display devices, or other media.

Interpreter, Image - An individual trained in the process of detecting,
identifying, analyzing, and accurately locating, with respect to a known
reference, objects and activities portrayed on imagery, as well as
determining the implications of those objects and activities.  The term
is replacing the older designation "photointerpreter" or PI, as non-
photographic imaging systems come into widespread use.

Key, Image Interpretation - An illustrated reference material designed
to aid image interpreters in the rapid, accurate identification of an
object from the study of its image.

Magnifier, Photointerpreter's or Tube - Measuring device consisting of
a transparent tube, with a lens element at one end and a reticle at the
other, used to measure distances of up to about one inch on  imagery.

Micrometer - Unit of length equivalent to 10   meters.  Formerly called
micron.  Commonly used in discussion of the infrared portion of the
spectrum, wavelengths from 0.70 micrometers (700 nanometers) to 1,000
micrometers.

Mensuration - Measurement of images on film.

Monoscopy - Viewing or interpretation of a single image, therefore not
achieving a three-dimensional effect.

Mosaic -  An assemblage of overlapping aerial photographs which have been
matched to form a continuous photographic representation.

Nanometer - Unit of length equivalent to 10"9 meters and to 10~3 micro-
meters.  Formerly called millimicron.  Commonly used when working in the
visible portion of the spectrum; about 400-700 nanometers wavelength.
                                  101

-------
Oblique Aerial  Photograph - Photograph taken with the camera axis  inten-
tionally directed between the horizontal  and the vertical.   A photograph
is generally considered oblique if the camera is tilted more than  3° from
the vertical.   Commonly referred to as an oblique.

Parallax - The  apparent displacement of the  position of a body with res-
pect to a reference point or system, caused  by a shift in the point of
observation.

Photogrammetry  - The art of obtaining reliable ground distance measure-
ments from photography.

Powerpod (Zoom  Stereoscope) - Main body of a zoom stereoscope which
contains both the variable magnification  (zoom) mechanism and the  separate
optical paths  for stereoscopic viewing.  Resolution of the  instrument
is determined  by quality of the powerpod.

Principle Point - Point on a frame of imagery corresponding to the nadir,
or the ground  point vertically below the  aircraft at the instant of
exposure on vertical photography.

Rhomboid (Zoom  Stereoscope) - Opto-mechanical arms  which rotate in a
horizontal plane about the powerpod to provide image separation for
stereoscopic viewing.

Scale - The ratio of a measured distance on  a map,  photography, or mosaic
to the corresponding distance on the ground.  As the denominator of the
ratio increases, the scale is said to become smaller:
                        Large - 1/12,000 or less
                       Medium - 1/12,000 to  1/25,000
                        Small - over 1/25,000

Scale, Photointerpreter's - A rule, usually  six to ten inches long,
graduated in .001 feet and/or 0.1 millimeters, used for making long,
but precise, measurements on imagery.

Spectrum, Photographic - The segment of the  electromagnetic spectrum
between about  360 nanometers and 900 nanometers.

Stereogram - Two or more photographs with sufficient overlap and conse-
quent duplication of detail to make possible stereoscopic examination
of objects or areas common to adjacent frames.

Stereographic  Coverage - Photographic coverage with overlapping aerial
photographs to  provide a three-dimensional presentation of the picture;
a 53% overlap is generally regarded as minimum.

Stereoscope -  A binocular optical instrument for assisting the observer
to view two properly oriented photographs to obtain the mental impression
of a three-dimensional model.

Stereoscope, Mirror, or Prism - Viewing device which uses prisms or
                                  102

-------
diagonal mirrors to achieve greater image separation than simple stereo-
scopes for viewing large-format imagery.

Stereoscope, Simple or Pocket - Viewing device usually having only two
simple lenses and fixed magnification best suited for field work or non-
critical investigation.

Stereoscope, Zoom - Complex stereoscope for use with transparencies and
light tables, featuring continuously variable magnification, high
optical resolution, interchangeable lenses, and variable image separation,

Stereoscopy - Production and use of three-dimensional effects in inter-
pretation of imagery.

Wedge, Parallax - Measuring device for height determination from stereo-
scopic pairs of photographs.  It consists of two slightly graduated
lines printed on a transparent template,  which can be stereoscopically
fused into a single line for making parallax measurements.
                                  103

-------
                               SECTION IX

                               APPENDICES

                                                                 Page No

A.  Representative Costs for Aerial  Surveillance                   107

B.  Photointerpreter Selection and Interpretation                  "109
    Procedures

C.  First Order Photointerpretation Key                            111

D.  Photointerpretation Test                                       H3
                                  105

-------
                                APPENDIX A

Representative Costs for Aerial Surveillance

The costs associated with various remote sensing missions, when contracted
with commercial aerial  survey firms, vary depending upon scale, weather
contingencies, size of area covered, time constraints, distance to target
area from the base of operations, and maximum flying height required.

Color and black-and-white vertical photographic coverage can be flown
commercially for the approximate costs listed below:
Average Costs for Aerial Photographic Coverage
Scale
1/5,000
1/10,000
1/20,000
1/40,000
Black-and-White
cost per sq. mi .
(dollars)
18-20
6-8
2-3
.8-1.0
Color (positive transparency)
cost per sq. mi. (dollars)
20-25
8-10
3-4
1-2




These costs assume use of a 9-inch film format in an operational system.
Color positive transparencies in 70mm format, which will be required for
oblique, and possibly short-strip vertical photography, cost $3 to $4
per frame.

Any photo acquisition system must be accompanied by a well-organized and
viable image interpretation capability.  Suitable viewing, projection
and measuring devices, and comprehensive reference materials must also
be available.

Using the equipment described in section IV, and 1/40,000 scale color
infrared photographs, a photointerpreter can classify industrial facilities
and major spill threats at a rate of about 3 stereo pairs per hour depend-
ing on the density of industrial features and the number of potential
spill threats.  He can perform detailed interpretations on 1/5,000 scale
color photographs, listing and delineating specific data on spill threats,
such as their areas, distances from waterways, numbers  of storage tanks,
and associated environmental characteristics.  It is difficult, however,
to define an average output rate because of the variability in  features
of interest, need for mensuration, and the proximity to drainage features
                                  107

-------
leading to waterways.  However,  the rate involved is estimated to be
about two stereo pairs per hour.   These average hourly outputs can be
used to compute costs based on the salary level of the photointerpreter.
We feel a salary level of $10,000 to $12,000 per year (equivalent GS-11)
is representative.
                                  108

-------
                                APPENDIX B

Photo-interpretation Selection and Interpretation Procedures

A person selected for imagery interpretation training should possess the
capacity to extract bits of information from imagery and assemble those
bits to arrive at an identification, as well as being able to deduce
from image-derived and collateral data the significance of objects and
facilities.  The interpreter must be able to make judgements and identi-
fications with a high degree of accuracy and to his own satisfaction,
as he may have to defend his interpretation.  The imagery interpreter
is the only link between the raw data of limited direct use and a finished
intelligence product.  Furthermore, he should have good color vision and
be free of eye disease or uncorrectable vision defects, have the ability
to see three-dimensional images on stereoscopic pairs of photographs, and
have the mental attitude and capacity to perform repetitive tasks during
prolonged periods of time, as necessary, to develop useful information
from photographic images and supporting data.

Prospective interpreters must be trained in photography, image interpre-
tation in one or more disciplines, photogrammetry, systems for remote
sensing, and elementary physics.  The more extensive an interpreter's
background, the more readily he may adapt to a multidisciplinary field
such as spill prevention.

Field work is advantageous for interpreters in order that they become
familiar with the appearance and function of ground activity.  The collec-
tive body of collateral data or experience will directly influence the
depth and accuracy o-f analysis.  The photointerpreter should also be
involved in specifying mission parameters and in preparing flight maps
to optimize the factors under his control.  He should obtain reference
materials to guide his interpretation efforts, such as photointerpreta-
tion keys, ground photography, reports, and other descriptions of the
target conditions and associations which he will be asked to evaluate.

Arrangements for film processing and editing are generally made by the
service providing the aerial photography.  Ground checking activities
are usually planned ahead of aircraft flights, although, in the spill
prevention system defined here, the final step would be a ground check
of those areas where spill threats had been identified.  Thus, no prior
planning for ground checking would be needed except, perhaps, to define
the format of materials needed for ground checking efforts.

When small scale photographs are received, the photointerpretation team
immediately identifies the map area imaged by each photography.  Each
team member would then analyze photographs assigned to him.  He would
systematically scan each stereoscopic pair in a convenient fashion --
left to right working from top to bottom -- until all areas of the photo-
graph have been viewed.  He would then proceed to the next pair of photo-
graphs and so forth.
                                  109

-------
A team leader generally functions to coordinate interpretation tasks,
assigning individuals to work on jobs most suited to their abilities
and maintaining periodic checks on interpreter performance.

Several methods for delineating features of interest and marking them
with an identifying code are commonly used.  One method is to write
directly on the film, either temporarily with grease pencil or permanently
with a marking pen using India ink or a similar lacquer type ink.   Another
method is to indicate on a frosted acetate overlay,  in either pen  or pen-
cil, the data of interest.  The method chosen depends largely on the data
reproduction and distribution requirements for the finished product.
The method used to mark a feature is usually to encircle (delineate) the
boundary, enclosing the feature or area, and to write an identifying sym-
bol or code adjacent to it in a way that does not obscure useful image
components.

Data sheets should be prepared to provide a common format for entries
pertaining to target data.  These forms can be developed to be compatible
with the standard practices of the data user.

The practicing photointerpreter should periodically  compare his  results
with those of other interpreters, and with reference materials,  to insure
consistency of his interpretation.  It is usually advisable to have more
than one interpreter analyze photos of vital  areas to minimize errors.
                                  110

-------
                               APPENDIX C

First Order PhotoTfiterpretation Key

A photointerpretation key for spill threat analysis would be unique,
since it would combine elements of municipal, industrial, and agricul-
tural activity analysis.   Although comprehensive keys to each of these
land uses have been developed by the military,  few seem to be readily
available for civilian application.  As far as  is known by the project
staff, nothing comparable to a spill prevention key exists other than
the preliminary key presented here.
        FIRST ORDER PHOTOINTERPRETATION KEY FOR FACILITIES
              ASSOCIATED WITH POTENTIAL SPILL  SOURCES

1.  Area contains railroads,  highways,  pipelines  and/or docks  with  assoc-
    iated pumps,  tanks,  and holding areas  	  Transport  facilities

1.  Area contains few,  if any,  of the  above features 	  See #2

    2.   Area contains  large open  pits,  large piles  of  bulk material, and
        a conspicuous  network of  roads  leading into  pit and storage pile
        areas 	  Open pit mining or borrow pit facilities

    2.   Area contains  few,  if any,  of  the  above features 	  See #3

3.  Area contains numerous  storage  tanks usually  close to buildings and
    structures with tall  vent stacks,  numerous pipelines, storage reser-
    voirs, slag and waste piles,  cooling towers and/or loading docks....
    	  Refinery and  processing  facilities

3.  Area contains few,  if any,  of the  above features 	  See #4

    4.   Area contains  large buildings  with numerous  vents and  stacks on
        rooftop,  storage yards  for  finished projects,  facilities  for
        unloading from trucks,  trains,  or  ships,  and large parking  lots
        for employees'  vehicles....Manufacturing  and assembly  facilities

    4.   Area contains  few,  if any,  of  the  above features 	  See #5

5.  Area contains a series  of rectangular  or circular  ponds with  stirring
    or  aerating mechanisms  visible; large  settling  basin(s) visible;
    piles, bales  or containers  of waste paper, scrap steel, wrecked autos,
    ships, rail cars,  or other salvageable material  also visible  	
    	 Waste treatment,  recycling,  and disposal  facilities

5.  Area contains few,  if any,  of the  above features 	  See #6

    6.   Area consists  of either cleared land,  with  leveled or  terraced
                                   111

-------
        hills and filled galleys,  or bare soil  and rocks  along shoreline
        or in spits with steeply sloping banks  	
        	 Land development and  filling facilities

    6.   Area contains few,  if any,  of the above features  	  See #7

7.  Area contains buildings with tall vent stacks, numerous  transformers
    and power!ines, liquid  holding  basins and/or fuel  storage tanks	
    	 Fossil  fuel  power generation facilities

7.  Area contains few,  if any,  of the above features  	  See #8

    8.   Area contains either even  rows of trees or plants with clearly
        defined boundaries  or a continuous cover of vegetation of  uniform
        density and color with  clearly defined  boundaries, a network  of
        access roads and/or irrigation facilities (canals, reservoirs,
        pumps, pipes, and sprinkler systems)	 Agriculture facilities

    8.   Area contains few,  if any,  of the above features  	  See #9

9.  Area contains large buildings  or other nonresidential  facilities
    under construction on cleared  or filled land.  Could  be  part of an
    existing industrial complex	 Construction facilities

9.  Area contains few,  if any,  of the above features  	  See #10

    10. Area consists of tree covered or recently cleared parcels  with
        road networks and log storage piles usually visible  	
        	  Forestry facilities

    10. Area contains few,  if any,  of the above features  	  See #11

11. Area contains large parcels that are devoted to stockpiles; buildings
    separated by driveways, loading docks and/or rail  sidings; or  large
    paved areas usually present, but no  evidence of processing or  manu-
    facturing 	 Storage  area facilities

11. Area contains few,  if any,  of the above features  	
    	 Non-polluting facilities and areas
                                  112

-------
                               APPENDIX D

Photointerpretation Test

The photointerpretation test was designed to provide a  basis  for a  com-
parison of findings of the project staff with regard to interpretability
of the various film-filter-scale combinations.   Twelve  interpreters not
familiar with the project or test areas were tested, and the  results  are
presented here.   Test results substantiated the findings reported in
Table 6.  The questionnaire concludes this appendix.
                    RESULTS OF PHOTOINTERPRETATION TEST

Dates Tests were Administered:  11/22/71 - 11/23/71

Number of Tests Administered:  12
Responses:                                    Number of     Percentage
                                               People          of
	Answering	Total

Question 1-(a)

Refining and Processing Installations:           11             92%
Waste lagoons and cooling ponds:                  1              8%
All other responses:                             _0_           	0%

                               Total             12            100%

Question 1-(b)

Storage tanks:                                   12            100%
All other responses:                             _0_             0%

                               Total             12            100%

Question l-(c)

Refining and processing installations:           11             92%
Storage tanks:                                    1              8%
All other responses:                              0            	0%

                               Total             12            100%

Question 1-(d)

Yes:                                             10             83%
                                  113

-------
Results of Photointerpretation Test (cont.)
Responses:                                    Number of     Percentage
                                               People          of
	Answering	Total

Question 1-(d) (cont.)

No:                                              _2            17%

                               Total              12           100%

      (Note:  For all  "No" answers,  the  following responses  were
              given regarding what subjects  felt the correct
              activity was:

                    1  Waste  Treatment and Recycling and Disposal
                    1  Refining and Processing)
Question 2-(a): (Pipeline Plotting)
     Film Type:  PAN 47-B
        1/80,000                                  0              0%
        1/40,000                                  1           8.33%
        1/20,000                                  6          50.0 %
        1/10,000                                  3          25.0 %
        1/5,000                                   2          16.67%
        1/2,500                                  _0         	0%

                               Total              12         100.  0%

     Film Type:  IR 89-B
        1/80,000                                  0              0%
        1/40,000                                  2          16.67%
        1/20,000                                  5          41.67%
        1/5,000                                   3          25.  0%
        1/2,500                                   1           8.33%
                                                 _L           8.33%

                               Total              12         100.  0%

     Fi 1m Type:  Color
        1/80,000 (this scale not given  on          0              0%
                  this film type)
        1/40,000 (this scale not given  on          0              0%
                  this film type)
        1/20,000                                  9           75.0%
        1/10,000                                  3           25.0%
        1/5,000                                   0              0%
        1/2,500                                   0              0%
                                  114

-------
Results of Photointerpretation Test (cont.)
Responses:
                             Number of
                              People
                             Answering
Percentage
   of
  Total
Question 2-(a) (cont.)
     Film Type:   Color IR
        1/80,000
        1/40,000
        1/20,000
        1/10,000
        1/5,000
        1/2,500
                               Total
                               Total
Question 2-(b):  (Guard Fence  Detection)
     Film Type:
        1/80,000
        1/40,000
        1/20,000
        1/10,000
        1/5,000
        1/2,500
     Film Type:
        1/80,000
        1/40,000
        1/20,000
        1/10,000
        1/5,000
        1/2,500
PAN 47-B
                               Total
IR 89-B
                               Total
     Film Type:   Color
        1/80,000 (this scale not given  on
                  this film type)
        1/40,000 (this scale not given  on
                  this film type)
        1/20,000
        1/10,000
        1/5,000
                                12
                                 3
                                 4
                                 3
                                 1
                               _0

                               12
                                0
                                1
                                2
                                4
                                1
                               _4

                               12
                                 1
                                 0
                                 1
                                 4
                                 1
12


 0

 0

 3
 6
 3
 100.0%
  8.33%
 25. 0%
 33.33%
 25
                0%
             8.33%
                0%
100. 0%
     0%
  8.33%
 16.67%
 33.33%
  8.33%
 33.33%

100.  0%
  8.33%
     0%
  8.33%
 33.33%
  8.33%
 41.67%

100.  0%
                                               0%

                                               0%

                                            25.0%
                                            50.0%
                                            25.0%
                                  115

-------
Results of Photointerpretation  Test  (cont.)
Responses:
Number of
 People
Answering
Percentage
   of
  Total
Question 2-(b)  (cont.)

        1/2,500
     Film Type:   Color IR
        1/80,000
        1/40,000
        1/20,000
        1/10,000
        1/5,000
        1/2,500
                               Tota'
                              Total
    _0

    12
     1
     1
     2
     2
     3
    12
      0%
 100.  0%
   8.33%
   8.33%
  16.67%
  16.67%
  25.  0%
  25.  0%

 100.  0%
                                 116

-------
           RESPONSE SHEET FOR EPA SPILL PREVENTION SYSTEM
      1.  Note on the 1/80,000 scale Pan 47B photos the installation
delineated on the frames.  Identify each half of the installation from
the following list.  (Check only one box in each column):
  a.
North
  b.
South
                Slag piles
                Waste lagoons and cooling ponds
                Terminals
                Storage tanks
                Pumping stations, pipelines and outfalls
                Refining and processing installations
                Slaughterhouses
                Livestock pens
                                   117

-------
c.  Identify the entire complex from the following list.  (Check one box)
    Mining
    Transport
    Storage
    Manufacturing  and  assembly
    Refining and Processing
    Waste treatment and  recycling  and  disposal
    Land development and  filling
    Forestry
    Agriculture
    Construction
    Power generation
       d.  Verify your interpretation on the color IR photos.   Is your
  interpretation the same?             Yes or No.  If not, what  is the
  activity  carried on at the complex?
                                118

-------
      2.  It is obvious that in a facility  such  as you are viewing a great
many pipelines exist interconnecting the  various  components.

          a.  At what scale can you  begin to  follow  pipeline routes?  Work
from 1/80,000, 1/40,000, 1/20,000, etc.,  for  each film before  proceeding.
             Pan 47B       IR 89B        Color        Color  IR


1/80,000




1/40,000



1/20,000




1/10,000



1/5,000




1/2,500
                                   119

-------
      b.  At what scale can  you  see  guard  rails  or fences  protecting  the



pipelines from vehicle traffic along main  roads?








             Pan 47B     IR  89B      Color      Color  IR





1/80,000








1/40,000








1/20,000








1/10,000








1/5,000








1/2,500
                                  120

-------
1

5

AccfNmoii Number
2

Subject Field &, Group
05B
SELECTED WATER RESOURCES ABSTRACTS
INPUT TRANSACTION FORM
Organization
Earth Satellite Corporation
Washington, D. C.
    Title
         A Feasibility Demonstration of An Aerial Surveillance  Spill  Prevention System
 10
Authors)
    Welch,  Robin  I.
    Marmelstein,  Allan D.
    Maughan,  Paul  M.
                                16
Project Designation

  EPA,  ORM Contract No. 68-01-0145
                                21
                                   Note
22
    Citation
23
    Descriptors (Starred First)
    *Water pollution sources, oil, chemicals, remote  sensing,  aerial  sensing,
     photography,  photogrammetry
25
    Identifiers (Starred First)
    *Multispectral  photography, thermal
      hazardous  materials
                                     imagery, color photography,  photointerpretation,
27
    Abstract
    Acquisition and interpretation of multispectral aerial  photography and thermal
    infrared  imagery was performed to evaluate remote  sensing  applications to spill
    prevention surveillance.  The San Francisco Bay area was used  as  a test site, with
    major sub-areas delineated which contained facilities and  activities  which might
    lead  to spills of oil and other hazardous substances into  waterways.

    Results demonstrated that high quality, small scale  (1/40,000  to  1/60,000) color
    infrared  photography can be used for regional surveillance leading to classifica-
    tion  of land use into areas where potential spill  sources  exist.   High quality,
    large scale (1/5,000 to 1/10,000) color aerial photography can be used for local-
    ized  delineation of potential spill sources.  Localized surveillance  should be
    supported by low angle oblique telephotography and limited ground surveillance.

    Recommendations are given for an operational  spill surveillance system using
    multi-scale aerial photography obtained on a  9-inch  film format.   Use of thermal
    infrared  imagery is not indicated at this time, as additional  information acquired
    is  minimal compared to resources required for its  acquisition.
A bsj.r3 c tor _ . . _ . •
Allan D. Marmel stein
Institution
Earth Satellite Corporation
WR;102 (REV. JULY 1969)
WRSI C
                                          SEND TO:  WATER RESOURCES SCIENTIFIC INFORMATION CENTER
                                                 U.S. DEPARTMENT OF THE INTERIOR
                                                 WASHINGTON. D. C. 20240


                                                          •&U.S. GOVERNMENT PRINTING OFFICE-.1972 O-464-291

-------