EimiONMENTAl PROTECTION AGENCY
            OFFICE OF ENFORCEMENT
                 EPA-330/3-75-001
   An Application Of ERTS Technology
          To  The Evaluation  Of
    Coal Strip  Mining And Reclamation
       In The  Northern  Great Plains
NATIONAL FIELD INVESTIGATIONS CENTER-DENVER
        REGION
                    AND
DENVER.COLORADO
              FEBRUARY  1975
                                     Maw

-------
       ENVIRONMENTAL PROTECTION AGENCY

            OFFICE OF ENFORCEMENT
      AN APPLICATION OF ERTS TECHNOLOGY

            TO THE EVALUATION OF

      COAL STRIP MINING AND RECLAMATION

        IN THE NORTHERN GREAT PLAINS
National Field Investigations Center - Denver
                     and
                 Region VIII
              Denver, Colorado

                February 1975

-------
                         CONTENTS
  I.    INTRODUCTION 	     1
 II.    SUMMARY AND CONCLUSIONS	     5
III.    BACKGROUND	     9
       Climatology	     9
       Coal  Resources	    10
       Production	    12
 IV.    STUDY TECHNIQUES 	    15
       Satellite Data	    15
       Ground Truth 	    20
       Data Interpretation and Analysis 	    20
  V.    RESULTS AND EVALUATION OF
         THE DATA ANALYSIS	    39
       Wyoming Mines (Nos. 1-10)	    40
       Montana Mines (Nos. 11-19) 	    57
       North Dakota Mines (Nos. 20-30)	    77
                           m

-------
                             TABLES

   II-I     Summary of Mine  Classifications  	    6
   IV-1     Spectral  Bandwidths	   18
   IV-2     Standard Deviation Point  Intensities	   33
    V-l     Areal  Distribution of Land Use (Mine 4)  ....   49
    V-2     Areal  Distribution of Land Use (Mine 5)  .  .  .  .   53
    V-3     Areal  Distribution of Land Use (Near Mine  6).  .   54
    V-4     Areal  Distribution of Land Use (Mine 11).  ...   60
    V-5     Areal  Distribution of Land Use (Mine 12).  . -.  .   64
    V-6     Areal  Distribution of Land Use (Mine 13).  ...   67
    V-7     Areal  Distribution of Land Use (Mine 14).  ...   72
    V-8     Areal  Distribution of Land Use (Mine 15).  ...   75
    V-9     Areal  Distribution of Land Use (Mine 20).  ...   80
   V-10     Areal  Distribution of Land Use (Mine 21).  ...   84
   V-ll     Areal  Distribution of Land Use (Mine 22).  ...   89
   V-12     Areal  Distribution of Land Use (Mine 23).  ...   94
   V-13     Areal  Distribution of Land Use (Mine 26).  ...   99
   V-14     Areal  Distribution of Land Use (Mine 29).  ...   103
                             FIGURES

    V-l     Northern  Great  Plains Coal  Field	    2
   IV-1     Typical ERTS  Daily Ground Trace 	   16
IV-2, 3     Observatory Configurations; MSS Scanning.  ...   17
   I V-4     ERTS Coverage of  Continental US	   19
   IV-5     Red Band  Image  of Mine  No.  4	   22
   IV-6     Sample Symbolic Map of  Mine No. 4	   24
   IV-7     Sample Classification Map of Mine  No. 4  ....   26
   IV-8     Sample Interpretive Overlay  	   27
   IV-9     MSS-Geometric Corrections 	   30
  IV-10     Comparison  of Successive Classifications.  ...   34
  IV-11     Comparison  of Mine No.  4 Maps  	   37
    V-l     Locations of  Wyoming Coal Mines 	   105
    V-2     Satellite Image of Mine No. 4	   43
    V-3     Dave Johnston Mine Reclamation	   44
 V-4, 5     Topsoil Borrow  Areas	   45
 V-6, 7     Middle Area of  Mine	   46
    V-8     Classification  Map of Mine No. 4	   48
    V-9     Satellite Image of Mines No. 5 and 6	   51
   V-10     Classification  Map of Mine No. 5	   52
   V-ll     Classification  Map of Abandoned Mine
             Area Near Mine  No. 6	   55
   V-12     Locations of  Montana Coal Mines 	   106
   V-13     Satellite Image of Mine No. 11	   58
   V-14     Classification  Map of Mine No. 11	   59
   V-15     Satellite Image of Mine No. 12	   62
   V-16     Classification  Map of Mine No. 12	   63
   V-17     Satellite Image of Mines No. 13 and 14	   65
                                IV

-------
V-18    Classification Map of Mine  No.  13	   66
V-19    Classification Map of Southern  Section
           of Mine No. 14	   69
V-20    Classification Map of Northern  Section
           of Mine No. 14	   71
V-21    Classification Map of Mine  No.  15	   74
V-22    Locations of North Dakota Coal  Mines  	  107
V-23    Satellite Image of Mine No.  20	  .   78
V-24    Classification Map of Mine  No.  20	   79
V-25    Satellite Image of Mines No.  21,  22	   82
V-26    Classification Map of Mine  No.  21	   83
V-27    Classification Map of the South Mine  	   86
V-28    Classification Map of the North Mine	   88
V-29    Satellite Image of Mine No.  23	   91
V-30    Classification Map of Mine  No.  23	   92
V-31    Satellite Image of Mine No.  25	   93
V-32    Satellite Image of Mine No.  26	   97
V-33    Classification Map of Mine  No.  26	   98
V-34    Satellite Image of Mine No.  29	101
V-35    Classification Map of Mine  No.  29	  102

-------
                              CONVERSIONS
multiply                           by                  to obtain
Metric Unit.                                            English Unit
hectares*                        2.471                 acres
kilometers (km)                  0.621                 miles (mi)
meters (m)                       3.280                 feet (ft)
metric tons                      1.102                 tdns (short)

*Multiply pixels by:  0.486 to obtain hectares; or 1.199 to obtain acres.


                             ABBREVIATIONS
ERTS - earth resources technology satellite
LARS - Laboratory for the Application of Remote Sensing (Purdue University)
MSS  - multispectral scanner
SCS  - Soil Conservation Service
USGS - United States Geological Survey

                               GLOSSARY
borrow pit - an area of special excavation to provide material  for an
             embankment when there is insufficient excavation on or
             near the job site to form the embankment
embankment - an artificial ridge of earth and broken rocks, or  a fill
             with a top higher than the adjoining surface
high wall - the unexcavated face of exposed overburden and  coal  in an
            opencast mine, or the face or bank on the uphill  side  of a
            contour strip mine excavation
overburden - the material that overlies a deposit of coal,  especially as
             mined from the surface by open cuts
pixel -  the instantaneous field of view within the multispectral scanner
radiometric analysis  - an analysis based on the  intensity  of  the light  as
                       determined by  ERTS multispectral  scanner
spectral  signatures - a set of four radiant intensities  in the  ERTS  spectral
                      bands defining  a class of  material
spoil pile - the area where mine waste is disposed of  or piled
                                    VI

-------
                      I.   INTRODUCTION

     Historically, the production of coal  in  the  Northern  Great  Plains
has been minimal  by national  production standards,  although  reserves  are
very large.  The recent energy crisis,  however,  coupled with the ex-
tremely low sulfur content (averaging about 0.6  percent)'of  the  coal  have
produced increased interest in coal  mining in this  region.  But  the
potential adverse environmental impacts of surface  mining operations, as
demonstrated in other parts of the country, have offset this interest.
     The Northern Great Plains Resources Program, initiated  in 1973 by
the Department of the Interior, was to provide tools to quantify the
environmental impacts of actual and proposed development of  coal resources
in Wyoming, Montana, North and South Dakota [Fig. 1-1].  As  an integral
part of the program, the National Field Investigations Center-Denver
(NFIC-Denver), Office of Enforcement, EPA, was requested by  EPA  Region
VIII, Denver, Colo., to conduct a study of coal  mines in Montana, North
Dakota, and Wyoming.  The purpose of the study was  to document the size,
shape and location of the following:
     1.   The actively mined area within each coal  strip mine site
     2.   The untouched spoils piles within the mine site
     3.   The reclaimed or recontoured areas within the mine site
     4.   Newly vegetated recontoured areas within  the mine
     5.   Abandoned spoils piles that have been  naturally revegetated
          with native plants.

-------
Figure I —1.  Northern  Great  Plain*  Coal Field

-------
     The study was conducted with the use of remote sensing  data from
the Earth Resources Technology Satellite (ERTS},  recorded from late June
through late August 1973.   To verify mining data, a telephone survey of
companies operating active coal mines was conducted in.early 1974.
Other verification data included USGS and SCS aerial photographs of
mining areas, maps provided by mining companies,  and aerial  photographs
taken by NFIC-Denver personnel.
     Besides presenting a synoptic inventory of Northern Great Plains
coal strip mines, this report demonstrates how computer classification
of satellite data can be applied to monitoring land use.  Future ERTS
imagery of this area, when compared to the baseline information in this
report, can be used to detect various changes at the mine sites.

-------
                      II.   SUMMARY AND CONCLUSIONS

     Thirty active, inactive and proposed coal  mine sites  identified by
the Northern Great Plains  Resources Program in  the states  of Montana,
North Dakota and Wyoming were evaluated in this study.   Three under-
ground mines and three proposed mines were not  computer classified.
Computer classification was not attempted for eight additional mines be-
cause they were small, inactive, or lacked ERTS data.   Of the remaining
sixteen sites, computer classification was successful  at fourteen.
Attempts to classify the other two mines were unsuccessful because of
the small size of one mine and the masking effects of cultivated crop-
land at the other.
     Land use, or classification, at each mine evaluated was defined by
computer processing the ERTS data which was in the form of digital
magnetic tapes.  A multipectral classification algorithm was used after
accepted procedures for smoothing, filtering and line drop-out correc-
tions were applied to the digital data.
     In most cases, the computer classification techniques were success-
ful in separating mine areas into:  active mining areas, graded spoils
piles, ungraded spoils piles and revegetated spoils areas.  Other topo-
graphical and cultural features such as ponds,  roads, railroads and
cropland in the mine vicinities were defined.
     The results of the computer classifications are presented in Table
II-l.  Disturbed areas of all types for the mines classified totaled
       2                       2
45.2 km  (11,159 acres; 17.4 mi ).  In addition to this total, several

-------
                                                                 TABLE II-l.                                                                                          _
                                                       Summary of Mine Classifications
Mine*/
No
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
TOTALS
Spoils
Active Mine Ungraded Graded Revegetated Miscellaneous
Hectares Acres Hectares Acres Hectares Acres Hectares Acres Hectares Acres
Mine not computer classified
Mine not computer classified
Mine not computer classified
247^ 610-7 125 305 182 450
21 53 28 68 44 108
145 356 16 38 20 49
Proposed mine
Mine beginning operation
Proposed mine
Proposed mine
40 100 44 108 56 137
74-/ 182-/
42-' 104-7 25 61 23 56
61 151 508 1255 304 750 64 158
18 44 161-' 398-'
Mine not computer classified
Underground mine
Underground mine
Underground mine
36 90 81 199 6 14
135-7 334-' 162 401 136 337 138 341
129 320 536 ' 1324
362 892 164 405 32 78
Mine closed
Mine not classifiable
93-' 230-' 31 75
Mine not computer classified
Mine not computer classified
28 68 69 172 20 50
Mine not computer classified
896 1988 2055 5069 673 1661 348 856 434 1067
Total Disturbed-7
Hectares Acres
27
85
2
554
93
180




140
74
90
937
179




123
572
665
557

124


117

4519
66
210
4
1365
229
444




347
182
222
2315
442




304
1414
1644
1375

306


290

11,159
a/  Mine locations shown on Figures V-l, V-12, V-22 inside back cover.
b/  Area of mines not classified was determined from ERTS transparencies
c/  Includes ungraded spoils
d/  Includes roads and areas in preparation for mining

-------
square kilometers of land have been disturbed at the small  mines not
classified.  Ungraded spoils piles accounted for the largest portion
(45 percent) of the mine area.  Active mine areas including some ungraded
spoils accounted for another 20 percent of the total mine area.  Only
four mines had significant areas that had been revegetated, accounting
for 8 percent of the total area.  An additional 15 percent of the area
contained in 5 mines had been graded but not vegetated.  Some degree of
natural revegetation has occurred on both graded and ungraded spoils at
several mines.
     The application of ERTS technology to land use classification and
areal determination was most successful for the large mines containing
large areas with the same land use.  The minimum size object or land
area that can be defined by the ERTS sensors is about 0.5 hectare
(1.2 acre), about the size of a football field.  This limited spatial reso-
lution means that present ERTS technology can be effectively used only
for mines with homogeneous areas at least 200 m (660 ft) on a side and
preferably 340 m (1,100 ft) in minimum dimension.  ERTS technology to be
implemented in the next few years will substantially improve the spatial
resolution, allowing classification of small as well as large mines.
     The most efficient use of ERTS technology could be achieved by
collecting detailed ground truth or field data on one or two mines,
calibrating the computer procedures using this data, and then extrapo-
lating to other mines, thus completing the classification without addi-
tional data collection.  In practice, this did not prove feasible.  Sig-
nificant variations between mines, and even for the same mine,  were

-------
8
 observed  in  the  ERTS data for the same type of mine area such as ungraded
 spoils  piles.  This required that the classification algorithm be cali-
 brated  for each  mine site.  The accuracy of the classification at each
 mine was  thus  largely  dependent upon the availability of- ground truth
 such as low-level aerial photographs, maps, or field observation.
     Two  types of problems were encountered in distinguishing areas dis-
 turbed  by mining activities from adjacent lands not disturbed by mining.
 Due to  the semi-arid climate, native vegetation near mines  in Wyoming
 and most  of  Montana is relatively sparse.  If reclamation or natural
 revegetation of  spoils piles produced significant  vegetative cover, it
 was often difficult to distinguish  between such disturbed areas and the
 surrounding  lands because both types of terrain yielded similar ERTS
 imagery.  A  similar problem was encountered in distinguishing between
 nonvegetated mine areas and adjacent lands with eroded or cultivated
 soil.   These problems  also occurred for North Dakota mines  where adjacent
 lands often  consisted  of alternate  strips of bare  cropland  and growing
 crops.

-------
                           III.   BACKGROUND

     The area known as the Northern Great Plains encompasses portions of
the states of Montana, North Dakota, South Dakota,  Nebraska and Wyoming.
At present, active coal mines are located only in Montana,  Wyoming and
North Dakota.  The area is sparsely settled and the principal activity
is ranching with a limited amount of farming.

Climatology
     The climate of the states included in the study area may be de-
scribed as a typical mid-latitude continental climate.  The Rocky Mountains,
located west of the study area, act as a barrier to the moist Pacific
air in the prevailing westerly flow.  There are no major geographic
features to modify air masses originating in the cold polar regions of
Canada, or the warm moist regions of the Gulf of Mexico before they
overflow the study area.
     Weather changes frequently occur over the area with a passage of
cold fronts and their associated pressure systems;  these generally
follow an easterly track.  Stagnant air masses which would adversely
affect atmospheric dispersion occur rather infrequently.  Warm windy
periods may occur in Wyoming, Montana and the Black Hills of South
Dakota during the winter as a result of warm Chinook winds reaching 40
to 80 km/hr (25 to 50 mph).  These periods may last several days and
tend to aid in the dispersion of any atmospheric pollutants.

-------
10
     Within the study area, average wind speeds range from 13 to
23 km/hr  (8 to 14 mph), while the average temperature varies between
6 and 9°C  (42 and 49°F) with annual average highs of 10 to 16°C (50 to
60°F) and  average lows of 1 to 2°C (30 to 35°F).  The mean annual relative
humidity  is 60 percent but averages 70 percent for the fall and
winter quarters.  Annual precipitation in the region varies from 25 to
50 cm (10  to 20 in.) per year.  Montana and Wyoming are generally more
arid than  the three eastern states in the region.  Usually 50 percent
or more of the rainfall occurs in spring while winter is the driest
time of the year.  The Dakotas receive about 91 cm (36 in.) of snow
annually,  except for the Black Hills area which receives 152 to 254 cm
(60 to 100 in.) per year.  Wyoming receives 91 to 152 cm (36 to 60 in.)
annually.
Coal Resources
     The Northern Great Plains Region contained about 46 percent of the
estimated  remaining coal resources of the United States as of January 1,
1972.  This reserve was contained in an area of 218,000 km2 (84,000 mi2),
representing only 18 percent of the total area of coal bearing rock in
the United States.   The presence of 46 percent of the tonnage in 18
percent of the area of coal  bearing rock provides convincing evidence of
the concentration of coal in the study area, a fact explained primarily
by the greater thickness of individual coalbeds in the region as
compared to the bed thickness in other parts of the United States.
     The coal  in the Northern Great Plains Region occurs in a sequence
of beds  460 to 910  m (1,500 to 3,000 ft)  thick of late Cretaceous to

-------
                                                                     11
Eocene Age.  This sequence is divided from bottom to top into the Hell
Creek, Tullock, Fort Union and Wasatch formations.  The coal  resources
are concentrated in the Fort Union formation of Paleocene Age, and only
subordinate amounts are present in the underlying and overlying
formations.  The rocks associated with the coal consist primarily of
poorly consolidated sandstone and siltstone with minor amounts of shale.
Many of the coal outcrops have been ignited by spontaneous combustion,
forest fires, lightning and acts of man.  Areas thus effected are
characterized by the red color of the baked rocks over the burnout coal.
These baked rocks are locally termed "clinker," but are more properly
identified as "baked shale."  Some of the burning took place thousands
of years ago, -ind some is recent.
     The coal bearing strata of the Northern Great Plains are nearly
horizontal.  Throughout most of the region, measurable dips are only 1
to 2°.  However, these low dips are locally reversed and give rise to
three broad regional structural features:  1) the Wiliston Basin,
centered in western North Dakota; 2) the Cedar Creek Anticine, near the
Montana-North Dakota line; and 3) the Powder River Basin in northeastern
Wyoming and southeastern Montana.  The steepest dips, commonly 5 to 10°
eastward, are along the western edge of the Powder River Basin in north-
central Wyoming.  Because of the gentle dips and the reversals of dips
associated with the three structural features mentioned above, the coal
bearing rocks are near the surface in most parts of the region.  In the
deepest part of the Powder River Basin in Wyoming, the coal bearing

-------
12
rocks are a little more than 610 m (2,000 ft) below the surface, but are
nearer the surface over most of the Wyoming Part of the basin.   In
eastern Montana, most of the coal is less than 460 m (1,500 ft) below
the surface.  In North Dakota all the coal is less than 370 m (1,200 ft)
below the surface, and 98 percent is less than 300 m (1,000 ft) below
the surface.  In South Dakota, all the coal is less than 300 m (1,000 ft)
below the surface, and about 80 percent is less than 150 m (500 ft)
below the surface.
     The Bull Mountain field, which lies mostly in Musselshell  and
Yellowstone Counties, Montana is a broad shallow basin with a northwest
trending synclinal access.  Throughout most of the basin, the coal
bearing seams dip 1 to 5° toward the center of the basin, but are nearly
flat to the east and south, and are steeper to the north.  In the
northwest part of the basin, the dip steepens to a maximum of 30°.  Most
of the past and current mining has been in the broad, central part of
the basin where dips are low.
Production
     Throughout the years, production of coal in the Northern Great
Plains has been minimal by national production standards.  In 1965, for
example, the region produced about 5 million tons, or 1 percent of the
national total of 512 million tons.  In 1972, the region produced about
20 million tons, or about 3.3 percent of the national total of 595
million tons.  This may be compared to 1972 production of 144 million
tons, or 24 percent of total national production, from the Illinois coal
basin in Illinois, Indiana and Western Kentucky.  The Illinois  basin is only

-------
                                                                    13
half the size of the Northern Great Plains coal  bearing region,  and it
contains only one-fourth as much coal.   The recent energy crisis,
and the extremely low sulfur content of coal  in  the Northern  Great Plains
(averaging about 0.6 percent), however, have significantly increased
interest in mining of coal  from this region.

-------
                                                                       15
                    IV.  STUDY TECHNIQUES

SATELLITE DATA
     The source of the remote sensing data for the Northern Great Plains
study was the Earth Resources Technology Satellite.  ERTS was launched
into earth orbit in July 1972.  Its orbit around the earth is inclined
9° with respect to the true north and south poles.  The altitude of the
satellite above the surface of the earth varies from 900 to 950 km
(560 to 590 mi) due to the slight flatness (oblateness) of the earth
and to the perturbing gravitational forces induced by the sun, moon and
earth.  ERTS completes an orbit every 103 min, making fourteen complete
orbits per day.  A typical ground trace for the daylight passes shown in
Figure IV-1.  A particular target on the earth's surface is covered by
the satellite every 18 days with respect to the previous over-pass.
Likewise, it views the entire earth every 18 days.  For every 18-day
period, the satellite ground trace repeats its earth coverage at the
same local time.  The overlap registration of a current ERTS frame
(rectangular format for the recorded data) is within 37 km (23 mi) of
the frame recorded 18 days earlier.
     ERTS is commonly referred to as an observatory containing various
sensors and communications systems [Fig. IV-2].  The sensor data used
for this study was obtained from the multispectral scanner (MSS).   The
MSS is a line scanning device which uses an oscillating mirror to
continuously scan perpendicular to the spacecraft path of travel

-------
16
   165  150 135 120 105  90 75  60  45  30 15   0  15   30  45  60  75  00  105 120  135 150 165
           DAY1
           (REPEATS EVERY 18 DAYS!
 180  165  150 135 120  105  00  75 60  45  30  15  0  15  30   45  60  75  90  105 120 135 150  165 18C
           Figure IV—1.  Typical  ERTS  Daily Ground  Trace

                           (Daylight  Passes  Only)

-------
   SOLAR ARRAY
  ORBIT ADJUST TANK
DATA COLLECTION ANTENNA
                                                                           17
                                      ATTITUDE CONTROL SUBSYSTEM
                       RETURN BEAM
                    VIDICON CAMERAS (3)
                                             WIDEBAND
                                             RECORDER
                                             ELECTRONICS
            WIDEBAND ANTENNA



  Mgf*-	ATTITUDE MEASUREMENT SENSOR

d     "^-MULTISPECTRAL  SCANNER

   S-BAND ANTENNA
X,
    Figure  IV —2.  Observatory  Configuration
              6 DETECTORS
              PER BAND:
              2» TOTAL.
              ->2FCn BANDS
              (EflTSBI
       SCAN MIRROR
       IDSCILLiTfS
       NOMINALLY
    Figure IV—3. MSS  Scanning  Arrangement

-------
18
 [Fig.  IV-3].  Six lines, with the same bandpass, are scanned simultaneously
 in each of four spectral bands for each mirror sweep.  Spacecraft motion
 provides the along-track progression of the six scanning lines.  Optical
 energy reflected from the earth's surface is sensed simultaneously by an
 array  of detectors in the four spectral bands with the bandwidths shown
 below:
                              Table IV-1
                          Spectral Bandwidths
Channel Number
1
2
3
4
Color Designation
Green
Red
Infrared I
Infrared II
Bandwidth (microns)
0.5 to 0.6
0.6 to 0.7
0.7 to 0.8
0.8 to 1.1
At the ground station, the continuous strip of imagery from ERTS is
transformed to framed images with a 10 percent overlap of consecutive
frames measuring 185 x 185 km (100 x 100 nautical mi).
     The complete orbital coverage of ERTS for the Continental United
States is shown in Figure IV-4.  In this ground trace pattern, every
location within this country is imaged every 18 days at the same re-
spective local time.  With a particular target in mind, one can check
the date of satellite coverage [Fig. IV-4] for orbit number and geo-
graphical coordinates.  With this information, a search of the NASA
data base is instituted to find the exact satellite frame(s) (imagery)
desired.  The imagery can be obtained in three forms:

-------
                                                            291 5"
                                     tT  \  V
                                      •  \  '  L-A-^
                                                           71"  7-r1
Figure  IV — 4. ERTS Coverage of Continental  United States

-------
20
     1.   Digital  computer compatible tapes
     2.   22.9 x 22.9 cm (9 x 9 in.)  optical  transparencies,
          scale 1:1,000,000
     3.   22.9 x 22.9 cm (9 x 9 in.)  black and white prints,
          scale 1:1,000,000.
GROUND TRUTH
     The ground truth for this study consisted of the following items:
     1.   Large scale low-level aerial photographs of various mines
          recorded by Soil Conservation Service and U. S.  Geological
          Survey from 1969 to summer 1973
     2.   Large scale detail maps of certain mines defining active,
          recontoured and revegetated areas
     3.   Aerial photography of certain mines recorded by NFIC-Denver
          personnel during the summer of 1974.
     The information contained in these types of ground truth data
was used during the computer analysis of the ERTS data.
DATA INTERPRETATION AND ANALYSIS
Data Source
     The ERTS data was obtained from the NASA data base for the thirty
areas  reported to be sites of coal strip mines.  The data consisted of
nine sets of 22.9 x 22.9 cm (9 x 9 in.) transparencies and the respec-
tive sets of digital computer compatible tapes.  Nine ERTS frames were
required to completely cover the Northern Great Plains area.   The
selectivity of the data was based upon the time interval of late summer
1973,  near-zero cloud cover over the respective mine locations, and
freedom from telemetry degradations and errors.

-------
                                                                      21

     The transparencies were used to visually locate the general  area of
each mine.  The precise geographical location of each mine was used in
cueing the computer for selection of the area(s) within the computer
tape to be radiometrically analyzed.  The detailed analytical work was
subsequently performed using a general purpose digital computer (IBM
360-44) and the digital computer compatible tapes which provided an
accurate radiometric analysis.
     The four separate spectral bands of the ERTS imagery afford a
multispectral identification and classification of a particular target
in question within the limits of the spatial resolution of the satellite
                                              2
scanner.  Its resolution capability is 4,860 m  (1.2 acres) which by
definition is 1 pixel, the instantaneous field of view within the multi-
spectral scanner.  A sample ERTS image constructed from the computer
compatible tapes is shown in Figure IV-5, a red band image of the Dave
Johnston Mine (Mine 4).  The tick marks on the perimeter of the image
are guides to a coordinate system used in selecting training sights as
discussed below.
     A multispectral statistical computer classification technique was
used to analyze the area in and around each coal mine.  This program
was an adaptation of one developed at the Laboratory for the Application
of Remote Sensing (LARS), Purdue University, for the classification of
agricultural (farm) land from digital aircraft and satellite data.  The
LARS program operates by reading the optical intensity (in this case
ground reflectance of sunlight) of each pixel in the four ERTS spectral
bands (tape) and then determines the statistical classification of that
element in relation to all other pixels in the target area or scene.

-------
22
          -N-
                  J
                   :T?ETCH
                                                      MINE NO. 4
                               nUG !'?« 1974 124052 JPL-IPl
        Figure  IV-5.  Red  Band  Image of  Mine  No.  4

-------
                                                                      23
Data Categories
     The satellite data for each coal mine were classified into the
following categories:
     1.   The actively mined area within each strip coal mine site
     2.   The untouched spoils piles within the area of the mine site
     3.   The reclaimed or recontoured areas within the area of the
          mine site
     4.   Newly vegetated recontoured areas within the mine
     5.   Abandoned  spoils piles that have been naturally revegetated
          with native plants.

Training Areas
     During the computer analysis-of satellite data, training areas
(areas of known target character such as active mine locations derived
from ground truth) were used to define the spectral characteristics of
the target classes into which  the total scene  is to be  classified.  The
LARS program  classifies each pixel  of the target area into one of  the
previously defined training classes and then computes a number represent-
ing the  statistical  confidence with which the  pixel falls into that
respective class.  By examining the confidence numbers, the  program
analyst  can determine the effectiveness of the classification routine.
Data Forms
     The output of data from the computer is presented  in several  forms.
A  symbolic display map of the  pixel  intensities in each of the four
spectral bands is used initially to identify and select training areas.

-------
   24
                   t eMTFP.it
              4 »==_-pFg.
                  = *€   =«
                                   i i£ «EE
                                8 F 6 * » « 4 4 . = « fg £= i • .
                                                       t=   e=== * «4 = =
                    Peg t=t=CeFc
                                                               4444=  4** 4 1* 44 168BSGB18
                                                              444* = 4=  .  4S*44(*t 4*eet>-
st   =- = 
-------
                                                                      25
A symbolic map for Mine 4 is shown in Figure IV-6.   Using maps or aerial
photographs as a reference, training sites are established on the symbolic
display and the coordinates are entered into the computer.  The statistics
for the four bands are computed and displayed in the form of histograms
for the training areas that have been selected.  After the selection of
the respective training areas has been finalized, a classification of
the image is made and a computer map is printed out in which the number
of the particular class is printed in place of each pixel [Fig. IV-7].
     By inspection of the histogram for confidence levels assigned to
all the pixels or elements within the target area, a lower limit of
confidence may be selected which permits a second classification map to
be printed, on which confidence levels lower than the selected limit
are replaced by blank spaces.  Lty observing this map, the analyst can
tell at a glance  in which regions of the image his choice of classi-
fication categories is deficient.  He can then combine similar classes,
determine new classes or alter the boundaries of training sites if
necessary.  Then  the process described above is repeated  until an
acceptable confidence level is reached for the particular classifi-
cation scheme being employed.  An interpretive overlay*  [Fig.  IV-8] is
drawn to illustrate the results of the final classification and anno-
tated to identify significant features.
ERTS Measurement  Limitations
     Limitations  can be placed on the accuracy or uncertainty of the
 *   Some  scale distortion  is  inherent  in producing an overlay map  (on the
    computer  line  printer) which can be avoided only by additional processing.
    In  scaling the overlay map  to a USGS topographic map, the longitudinal
    axis  of the  former must be  reduced to 80% of its nominal value.

-------
  26-


  1  73UM162M 11 1M1 11 113*13331 1 111 1133311111111
  2  22*12*6661 111 1331 112 HI 331 131 13111 1 1 3311 1 1331-"
  1  6? I 11221 1122013112263 3131111313)341 I 111 16U1
  4  1111 12211 71133333311113)3311111111 133311 It 11 1 1
  5  «>212->2 11226311 331331/3113111 111 11 11311221 11 11
  *  AMl221l?111313313im31l.lJllUllllllimilll
  7  IM1 I22l62oll 11 123131131111111111111111111111
  B  111116616211111221333111111111111111111111666  ~
  f)  61 11 1221 12662S&2?5131 1113111 111 11 11 11 11 1 1 1111
10  21 11 12522645 54?hf./S 1331 11 1211 11 11 11 116611 11 111
11  2112112*55453136*111 111 1221 112121111 lllll 1111
12  6211 1232&4554331261 lllll 111112126111 lllll 1111
13  •><.22211212A'.241312:'3313U 1111111 11 11 llUlllll
-I*  S',622U216655631<.46221 11 11 111 111 1 llllllllllll —
15  62&?222. "! 266 35434 ',4o66ll 11 II 11 11 11 116111111111
16  661122?22o333l33 1333 Jl llll 31
?fi  111 I 1 1 &f,64 2 1 1 33 3 311 > 5 -16 1 1 1 1 3 .J 3 3 I 2 1 1 3 1 1 1 1 1 1 1 .1 3
?«»  mimii:22mil2226Mllin331121111U33?331    10   15   20   25   30    35    40
30  11111131 111 11 131 1 112521 11 1143324 1131 J333111 11
31  611llim31ll?3333111?5?lll3431ll333H33 311111 33 11111311111131133. 111113311133113
12  '.1 11 1 I II 111 12621111 I 116T6122'. 331. '3333331 111 311111111111 1111 Ull 11111331111333111
31  AV,l 11 Ul 111221 11 111 11 12512333313331)311 311 1.13331131 111 11 111) 1111331111111111331
32?l 1 1111 112621 16323331 13333373311 11 11 lllll 1111 1111 113311 311 111 13331
35  1111116111126611111111112222133333113331 1 1 1 I I 1 11 \ 1 1 1 1 11 3 1 1 1 11 1 1 11 1 1 1 I 1 1 1 11313331
If.  Ill 1 11 1 16 11 2M 311 11 1111 11 1261 133333333 U 331 11 111 11 11111111133111 1111 I 11133311131
3/  1311 1111 11 11211 13111 131 1 1161 211 13113 3 Ul II 33! Ill 1121 lllll II 13 31 133J1 1111 11111331
in  lllllUlli;?'!'. ?113!llllll!12?6i24tJH1111l311116<>'i2i lllll ! 3331112J311111LU11111
l'»  1.UIM114I»'>111.I>1IM lJMlllll">M4'ilU l-H mi !255'.t4'ill 1111 331 11.11111 133)11 Jlllll
An  1111 u ii?22iii2ii u i J33Hiii22iim3 nil i:miJ22'."<,2mi limn iimm-'iimu
*!  ll?21 I 11 11221 11133 11333 11 1111222226 3 3 111 , n 2^«4643J2? 1 1 H 1 1 I I I 1 1 1 1 1 3 11 1 I 1 1 1 11 1 1
*?  11 11 1311 I 1111111333111 H31I 11 162t?5V,n3i2, *,,<,,. 1? rJ,l2mlll,, 333, 3m
V,  M11
«!>  1
*6  1
47  1
',b  I
49  I  11 1211 111 1131131316662111 11331333311
50  111 HI II 111 111131131161131131 13313111131 , 1 1 1 J l',-',',^^ 32 11 1 16','o63 1 1 1 1 1 3333! 1 1 11
                                             1 1 11 I 2 2 '. -3 <• «. ^ 22 (.?55/l^/t132llA',«3^4ab', I 111 111 12M31 III
  11112M13313333H211 lllll 111 11
                                            ,
mil2&tl21U113»31?2lllJjmiM3ll22522i,f^,.j33?21lli:2M/.S511111333llllllll
                                        i333 UJ3 31^31 12'.644^56t- 1 1333.' 11 3733)1 111
                                        1333J3»33)3331U'.'-336ft61 2 J 1 311J3 »H 13
                                        333313UJ53J313333*.'. 313
                                        1113311131111113 MI3J3141 366 111113331111
                                        llJltll'limn33i311U23M2H333333llll
                                        11 1 3 I 1 13 1 1 H 1 1 3 14 3 I 1 1 ?f.<. 3 352 <. 1 1 1 1 1 1 1 26 1 1
                                        1131 13112313121 121 1 lf>624 311 I 112621 112211
                                        31111 1)1 111)11131 1111 116 JlbM 111 11666161
                                        I M 1 11U t 1 I HI! 111 1 ?/?'•<> 4 3J 4 hi Iftrf 6AA611
                                        lllll 111 I 11 11 11 II lbS6/.«.? 244ft U? 122261111
  Figure  IV — 7. Sample  Classification  Map  -  Mine  4

-------
                                                                 27
                _3PRING 1973
FALL 1 B7.O
                                            SECOND HIGH WALL CUT
                                            AND SPOILS PILES
                                                    HIGH WALL DEFINITION FROM
                                                    RATIO IMAGE RED BAND
                                                          INFRARED 1 if A N D
       LEGEND
     ACTIVE MINE AND UNGRADED SPOILS
     FAUI972. SPUING 1973

     REVEGETATION, SPUING 1972
     HEVEGETATION , 196?, SPRING 1971
     TOPSOIL BORRO
         Figure  IV — 8.  Sample  Interpretive Overlay
               of a  Classification Map  -  Mine 4

-------
28 .
physical measurements carried out on the ERTS (frame) transparencies and
the computer compatible tapes.  Any measurements for linear distance and
surface area in the transparencies were made with scaling instruments
and light table microscopes.
     The uncertainty for linear distance (ALO) is
          ADL = ± 5 x 10   x  image scale (m).
     The image scale of the transparencies is 1:1,000,000.  The value
for ALD is
          ALD = ± 5 x 10"4 x  106 m = ±  500 m  (1,640  ft)

A distance  X, measured physically on the ERTS transparency  is accurate
to within ±500 m  (1,640 ft).
     The uncertainty  for  the  surface area  (ASA)  is  (rectangular)
          ASA  = ALD  (±X±Y) m2.
 For  the ERTS imagery
          ASA  = 500  (±X±Y) m2.
 For  example, a  rectangular  area with dimensions  of  X ±500 m and
 Y ±500 m would  have  the value
          Area  =  [XY  + 500  (±X±Y)]  m2.

      The uncertainty in  the  computer analysis of the computer compatible
                                                                       2
 tape data  is limited  to one  pixel  which measures approximately 4,860 m
  (1.2 acres).  Due to the  digital character of the recorded data  the
 computer can resolve two  targets only  if they are separated by a dis-
 tance of at least 80 m (260 ft).  In  calculating the surface area of a

-------
                                                                      29
target, the computer counts the number of pixels present in the image
which is significantly more precise than through visual measurements.
The integrity of the satellite data is greatly enhanced by fourteen
distinct geometric corrections [Fig. IV-9] carried out immediately on
the telemetered data received at the various ground stations. The data
is further improved through precise geometric and contrast corrections
before being analyzed.
     A full technical description of the accuracy achieved in the anal-
ysis would require a mathematical treatment of the problems of separating
overlapping Gaussian distributions  in a multidimensional spectral space.
Some of the practical problems affecting accuracy will be discussed.
     A problem exists in obtaining  registration between the satellite
image and  the aerial photograph or  map used as ground  truth.  Some
features such as the location of a  pond or high wall  in the infrared
bands are  quite distinct,  but others, such as the exact boundary of  a
revegetated area, are not.  The latter can result in  the inclusion of an
erroneous  area in a training site or in extending a training site to a
large enough size to avoid an area  that is questionable.
     Lack  of adequate spatial resolution is the most  significant problem
in using ERTS imagery.  The individual picture elements are approximately
the size of a football field.  For  a large mine this  is not significant
but for smaller mines, the size of  the picture elements severely limits
the amount of detail that  can be seen.  This results  in the blending of

-------
      30
                          GEOMETRIC FOOTPRINT
  MACmiUDE OF CORRECTION -
   MEIERS ON THE GROU1D
                                                                                                      GCOMETRIC FOOTPRINT
                                                    VACKITUD! OF CORRECTION -
                                                      METERS ON THE GROUnO
SCAIE ADJUSTMENT -
ALIITUOE VARIATIONS
MTHINAFIIAUI

B CORRECTIONS
WHIN THE
FBA.ME
AlIIIUBi VARIATIONS
KITHIN A FRAME
(CORRECTIONS
MIHINAFRA.ME
SCORRECriONS
(1IIHIN A FRA.V
•CORRECTIONS
AITHI.1 A FRAME
IMAGE SKEW
 AUSED BY EARTH
ROTATION

(FUNCTION OF
LATITUDE!
AVERAGE VELOCITY
CHANJE FROVI
NOMINAL
CAUSED BY FINITE
SCAN Tl •:
ix-inxio* h,

(AT EDGE OF FRAME)
                           	IT
                        -H  t— J»
JX-IIIX lO'o
                                                  (AT IDF AND MTTW
                                                  OF FRAME)
                                                  lU-IISMEIEHS
SPECTRVMANO OFFSE TS
BAND 1 |
BAND I J
BAND ]

BAND!

1 .


                                                  JX- 1I7METERS
                                                                            SCAN MIRROR

                                                                            VELOCITY CORRECTION
                          PERSPECTIVE CORRECTIONS

                          SWEEP TO
                          FRAME IN DIRECTION
                          OF SCAN
                                                                                                                              JXMAX-lliMETERS
                                                                            FRAME CORRECTION
                                                                            1 DIRECTION OF
                                                                            SPACECRAFT TRAVEL -
                                                                            IEBR CORRECTION)
ALIGNMENT
CORRECTIONS -RELATIVE
AIIG1MENTOFMSS
TO AMS

ROLL
                                                                                                                iX— t  1—
                           Figure   IV —9.  MSS  Geometric   Corrections

-------
                                                                      31
 picture  elements  that straddle the border between two different classes
 of  terrain.  The  resulting spectral signature is a blend of the two
 types  of material viewed.  These effects are noticeable at the border
 between  two distinctly different classes of material where one or two
 pixels on the border will drop into some intermediate spectral class.
 For smaller mines, these can represent a significant portion of the
 area.
     The low ERTS spatial resolution also causes a statistical problem
 in  selecting a training site commonly composed of only 16 to 25 picture
 elements.  A much larger number of representative samples of a class is
 desirable to adequately define the mean and standard deviation of that
 class  relative to other classes against which the classification algorithm
 is  trying to discriminate for the size mines that are being used in this
 analysis.
     Since the multispectral scanner detects only the optical properties
 of  the surface observed, there are circumstances where uncertainty can
 arise.   One of the most common is the extensive areas around a mine from
which  the vegetation has been stripped by trucks and other equipment.
These are particularly noticeable in the area about to be mined in front
of the high wall.   The exposed soil is difficult to separate from spoils
where the earth has been disturbed.  Although graded and ungraded spoils
can be separated,  in general, exposed dirt has about the same spectral
signature regardless of its  form or physical  state.   Other sources of

-------
32
 misclassification  arise when  spoils piles have naturally overgrown with
 vegetation,  in which  case  they  resemble  the  spectral characteristics of
 the natural  landscape.   In areas  that are being revegetated  the  spectral
 difference  between grass planted  at different times seems  to be  due to
 the ratio of exposed  soil  to  vegetation.
      The accuracy  and extent  to which a  mine can  be classified also
 depends  upon the level of  detail  supplied in the  ground  truth available.
 Three computer runs usually suffice to overcome the mathematical  data
 analysis problems  involved in multispectral  classification.   The re-
 sulting  accuracy is then determined by the detail  provided in the ground
 truth and the ability to secure registration of the satellite image
 to the ground truth.
      As  an  example of the  changes in the classification  map  that result
 from small  changes in definition  of the  training  sites,  Figure  IV-10
 displays two overlays of the  northwest corner of  the Dave  Johnston Mine
 that resulted from the second and third  attempts  at classification.
 While small  deviations in  the revegetated areas defined  as Classes 5  and
 6 are noticed, the general features remain consistent.   Class 6  has  two
 arms extending northward that enclose the Class 5 area on  the east and
 west.  Two  sections of revegetated area  extend northward with the Class
 6 area to the west of Class 5.  A small  finger of Class  6  extends south-
 ward on  the east side of yet  another class of reclaimed  spoils.

-------
                                                                      33
     The differences in the classification maps were caused by expanding
the size of the training areas by one additional column to include
additional representative samples of the different classes.  To illus-
trate how small the differences between the classes are in the second
and third classification runs, the numerical values for the intensities
can be examined.  The light intensities in each spectral band are ex-
pressed on an arbitrary scale from 0 to 256.  The intensities of the plus
and minus one standard deviation points about the mean value are given
below.
                              Table IV-2
                 Standard Deviation Point Intensities
Class
5
6
Run
2nd
3rd
2nd
3rd
Green
Band
102
104
90
88
- 114
- 112
- 98
- 98
Red
Band
114 -
118 -
96 -
94 -
126
130
106
106
Infrared
Band 1
118 -
122 -
106 -
108 -
128
132
116
116
Infra-ed
Band 2
116 -
120 -
106 -
106 -
124
126
114
114

-------
                                REVECETATED SPRING 1973
                                        TOPSOIL BORROW AREA
                                             REVECETATEO SPRING 1972
                                                                                  HIGH WALl
2ND RUN
                                                                                             3RD RUN
           Figure IV —10. Comparison of Successive Classifications

-------
                                                                        35
     To determine the accuracy of the computer classification,  an overlay
drawing was prepared from a 1974 aerial  photograph outlining the features
of the mine at the same scale as the computer line printer overlay.   An
exact comparison could not be made because the horizontal  and vertical
dimensions of the line printer output differ 20 percent in scale.  The
two maps of Mine 4 were brought into optimum registration  and over-
printed in Figure IV-11.
     Comparison of the two maps shows that in the northwest section of
the mine roads were included as a part of the revegetated  areas in the
computer classification.  A small strip of land between the mine haulage
road and the automobile road to the west was included as part of the
mine.  The northwesternmost tip of the mine seems to have  been revegetated
recently where the computer classified a high wall area.  In the center
of the mine a revegetated area just east of the road leading to the
second cut but unmarked on the company supplied map is discerned.  A
road and some unreclaimed spoils cutting across the mine to the high
wall was classified as background.  Revegetated areas planted in the
spring and fall of 1972 were classified as background; examination of
photographs showed that the revegetation that had grown in these areas
was more dense than in other revegetated plantings and more closely
resembled the natural vegetative cover of the background classification.
The 1974 high wall extends further to the northeast than the positions
determined by the computer analysis in 1973; this is the area currently
being mined.

-------
 36
     A numerical comparison of the areas allotted to high wall and
spoils shows that the computer recorded 86.3 percent of the acreage
shown in the aerial photo one year later.  The computer counted
105.7 percent of the acreage devoted to revegetated land and topsoil
borrow.  The overall area of the mine as determined by the computer
was 96.9 percent of the area recorded in the aerial photo one year
later.

-------
\
Aerlol  Photo Mop of Min« 4 (1974)

Satollila  Mop of Mine  4 (I973|
                               0        0.5        1.0       15       20       1.1
                                                 KIlOMITIti
                      Figur«  IV —11. A Companion  of Mine 4  Clarification  Mop»
   LEGEND
    A.— ACTIVt MINI
    §.— HlveOf TATID'SfOUS
    C.— TOPSOIl 1OIHOW
    0.— EXPOSED  SOU
    I.— IOADS
                                                                                                                                                     CO

-------
                                                                     39
            V.  RESULTS AND EVALUATION OF THE DATA ANALYSIS

     This section presents the results of the computer analysis of the
 ERTS satellite data for coal strip mines in the Northern Great Plains.
 Of  the thirty mines identified via satellite, fourteen were successfully
 classified using the techniques described in Section IV.  Eight mines
 were not classified because of their small size or lack of ERTS data.
 Two of the mines were not classified because the necessary computer
 compatible tape was not received from the NASA data base; classification
 results will be reported when available.  Three mines were not yet in
 operation and three underground mines were not classified.
     Satellite images are presented for each mine classified.  A classi-
 fication map is also presented showing the shape, location and relative
 size of the active mine areas, graded and ungraded spoils piles, and
 areas in various stages of revegetation.  Where appropriate, adjacent
 topographical and cultural features such as stream beds, cropland, roads
 and railroads are shown.  The area of each type of land use is tabu-
 lated.
     The problems encountered in applying ERTS technology to the classi-
 fication of strip mines are also discussed.  Such problems include
masking of disturbed areas by revegetation and similarities between
cropland and mine areas.
     Coal  production figures given for the mines for 1972, 1975 and 1980
were obtained from the mining companies for the Northern Great Plains
Research Committee.

-------
40
                             WYOMING MINES

MINE 1 - WYODAK RESOURCES DEVELOPMENT CO.. UYODAK MINE
     The Wyodak mine, about 10 km (6 mi) east of Gillette [Fig. V-l,
inside back cover], produced 0.56 million metric tons (0.-62 million
tons) of coal in 1972.   Coal production for 1975 and 1980 is projected
to be 0.9 and 1.4 million metric tons (1.0 and 1.5 million tons),
respectively.  The thickness of the coal deposit averages about 27 m
(90 ft) within the mine.
     Based on a satellite image of this area, the surface area of the
mine was about 27 hectares  (66 acres) in 1973.  The computer tape for
this mine was not received  from the NASA data base; however, when
available, a classification map will be addended to this report.
MINE 2 - AMAX COAL CO.,  BELLE AYR MINE
     About 22 km (14 mi) south of Gillette [Fig. V-l], this mine yielded
only 0.027 million metric tons (0.03 million tons) of coal in  1972.  A
major increase in production is expected in the future with projected
levels of 1.6 and 8.6 million metric tons (1.75 and 9.5 million tons)
for 1975 and 1980, respectively.  The coal deposit averages 21 m (70 ft)
thick.
     The surface area of this mine, as determined from a satellite
transparency recorded in July 1973, was about 85 hectares  (210 acres).
The computer tape for this  mine was not received from the  NASA data
base; however, when available, a classification map will be addended
to this report.

-------
                                                                       41
MINE 3 - BEST COAL CO.. EAST ANTELOPE MINE
     This mine is about 80 km (50 mi) north of Douglas in the Thunder
Basin National Grassland [Fig. V-l].  The mine is small with a 1972
production of only 910 metric tons (1,000 tons).   No increase in pro-
duction is expected through 1980.  The average thickness of the coal
deposit is about 12 m (40 ft).
     The surface area of the mine, measured from a USGS aerial photo-
graph was about 1.5 hectares (3.7 acres) in 1969.  The land surrounding
the mine is characterized by essentially bare soil with little vegetation.
The optical characteristics (spectral signature)  of this land as re-
corded in the ERTS data were nearly identical to  those of the mine area
precluding any meaningful  computer classification of land use within the
mi ne.

-------
42

MINE 4 - PACIFIC POWER AND LIGHT CO.. DAVE JOHNSTON MINE

Background
     Situated about 48 km (30 mi) northeast of Casper and 26 km (16 mi)
north of Glenrock [Fig. V-l], the Dave Johnston Mine produced 2.4 million
metric tons (2.6 million tons) of coal in 1972.  Production is expected
to increase only slightly as reflected by projections of 2.4 and 2.5
million metric tons (2.7 and 2.8 million tons) for 1975 and 1980, re-
spectively.
     As shown in the satellite image [Fig. V-2], the mine has a long,
narrow configuration.  The 8 km (5 mi) long high wall lies in a south-
east direction.  Excavation is advancing to the northeast into an arid
landscape thinly covered by sagebrush and prairie grass.  Spoils piles
and revegetated areas lie to the southwest of the high wall.
Revegetation was begun in 1969 and closely follows the excavation activi-
ties.  The area of unreclaimed spoils piles is thus relatively small.
The status of reclamation as of Spring 1973 is shown in Figure V-3
provided by the mining company.
     For verification, aerial photographs of various areas of the mine
were taken by NFIC-D personnel, four of which are shown in Figures V-4
through V-7.

Classification of Mine Areas
     As shown in Figure V-3, about 55 percent of the disturbed area of
the mine has been revegetated.  The computer classification defined the
disturbed areas but difficulty was encountered in separating older

-------
                                                       43
-N-
         •i^E_i_sa
        I ((i! i •. I M 1111; \ i i i •! ' i' ' i 11' f •   • • •• : •' '• •', i

        - "*•-».   ' *''"'**             '"  "
       •- ,
                                           MINE  NO. 4
         'H.fi.P.1
         c ;'4::-i:-iii05-4;:  : :
                  JBn
Figure V-2.  Satellite  Image  of  Mine  No.  4

-------
                                                                                >>v
                                                                            //• .—/^v"
                                                                          :- f J§* ^V
                                                                            «^v-n

^
Figure V — 3. Revegetation of Dave Johnston Mine



 (Map Provided by Pacific Power and Light Co.)

-------
                    Figure  V-4.                                         Figure  V-5.



       Topsoil  Borrow  Area in  Foreground.         View  of Southeast  End of -Mine Looking  North,



Revegetated  Area  Between  Road  and  Spoils  Pile.     Note Topsoil  Borrow Area North  of Mine.
                                                                                                     en

-------


                Figure  V-6.                                    Figure  V-7.



Middle  Area of  Mine  Showing Second Cut.    Middle Area of Mine Showing Discontinuity



                                                            in High  Wall.

-------
                                                                     47
revegetated areas from undisturbed native vegetation.  This separation
was achieved using Figure V-3.  Difficulty was also encountered in
defining the high wall location in the northwestern area of the mine
during the initial classification.  The high wall was then defined by
preparing a ratio image of the green band to the infrared 1 band.  A
composite map [Fig. V-8]* of the classifications of various mine areas
was prepared using both the ratio image and the multispectral classi-
fication.
     In the northwestern corner of the mine, land revegetated in Spring
1973 was discernible from grass in an adjacent area planted one year
earlier.  A topsoil borrow area on the west side was also identified.
Proceeding to the southeast, revegetated areas planted in 1969, 1970 and
1971 as designated on the map produced spectral signatures the same as
undisturbed native vegetation thus indicating a comparable density of
vegetation had been achieved.
     The southeastern portion of the mine contains much revegetated land
and many topsoil borrow areas.  In the borrow areas, a scraper has
removed strips of topsoil leaving native vegetation in alternating
strips as shown in Figure V-4.  Some natural revegetation has taken
place in the disturbed strips.  The revegetated land and the borrow
areas were spectrally indistinguishable from one another as both consist
of areas of vegetation mingled with bare soil.
*  Some scale distortion is inherent in producing an overlay map (on
   the computer line printer) which can be avoided only by additional
   processing.  In scaling overlay maps to USGS topographic maps, the
   longitidinal axis of the former must be reduced to 80% of its nominal
   value.

-------
48
                 Figur* V —3. Classification  Mop  of Mine 4 —  Pacific Power  and Light Company,
                                    Dave  Johnson Mine,  Glenrock,  Wyoming

-------
                                                                     49
     The land north of the high wall  in the southeastern part of the
mine had a signature characteristic of revegetation or topsoil  borrow.
No activity in this area was indicated on the company map [Fig.  V-3].
An NFIC-D aerial photograph [Fig. V-5] verified, however, that this was
a topsoil borrow area.
     Characteristics of the most extensive spoils piles near the second
cut in the center of the mine are shown in Figures V-6 and V-7.
     By counting pixels, the computer keeps track of the area in each
classification.  The area of each classification for the Dave Johnston
Mine is presented in Table V-l.   Of the 554 hectares (1,365 acres) of
disturbed area, about 55 percent has been revegetated.
                               Table V-l
                    Areal Distribution of Land Use
                                Mine 4
Land Use

High wall and spoils
Northwest Portion of Mine
Spring 1973 Revegetation
Topsoil Borrow
Spring 1972 Planting
Spring 1971 Planting
Fall 1970 Planting
Fall 1969 Planting

Southeast Portion of Mine
Spring 1972 Planting
Fall 1972 Planting
Fall 1972 Planting and Topsoil
Borrow in Southwesternmost
Corner

Total Mine Area

Pixels
508

26
7
67
17
33
46
196

30
36


368
434
1,138
Area
Hectares
247

13
3
33
8
16
22.
95

15
18


179
212
554

Acres
610

31
8
80
20
40
55
234

36
43


442
521
1,365

-------
 50
MINE 5 - BIG HORN COAL CO.. BIG HORN MINE

Background
     This mine, at Acme about 11 km (7 mi) north of Sheridan,  produced
nearly 0.9 million metric tons (1.0 million tons) of coal  in 1972
[Fig. V-l].  Production is expected to remain the same through 1980.   The
mine lies in a valley at the confluence of Goose Creek and the Tongue
River.  Mine 6, which is discussed in the next section, is a few kilo-
meters to the west.
Classification of Mine Areas
     The satellite imagery [Fig. V-9] revealed a complex setting for
this mine.  The mine is bound on the north by the Tongue River and on
the west by a small reservoir on Goose Creek.  Graded spoils lie between
the reservoir and the active area of the mine [Fig. V-10].  On the east
is a bluff with a surface of eroded soil and brush that gives  the same
spectral signature as the graded spoils.  The mine is being extended
eastward into the bluff.
     To the west of the mine, a gravel pit and a coal loading  area are
distinguishable.  Across the Tongue River Valley to the north  of the
mine, an area of abandoned spoils piles and a pond from previous mining
activity were defined.  The area disturbed by mining activities in this
vicinity is shown in Table V-2.

-------
                                                                                     51
ABANDONED  MINE
 SLATER  CREEK
                  -N-
                           11 lltjItttj'lUjll II JIHI JHIIJIHU 11 t I j I I I I j I I ! I (III!

                                                       ----
                            1330-17232-L-5  UED   17WJG73    E-14
                            
-------
52
                                                                       Tongul Rivtr Vollty-
              Tongu*  River  Voll.y
                          ACTIVE MINE

                          SPOILS PIUS NORTH  OF RIVER

                          GRADED SPOILS
                   I   I   POND
                              Figurt  V —10. Claisificotion Mop  of Mine 5— Big Horn  Cool Company,
                                                 Big Horn Mint, Acmt  Wyoming

-------
                                                                        53
                               Table V-2
                    Area! Distribution of Land Use
                                Mine 5
Land Use

Active mine
Graded spoils
Total, Big Horn Mine
Abandoned spoils north of
Tongue River
Total, Acme, Wyoming

Pixels
44
90.
134
5!
191
Area
Hectares
21
44_
65
28
93

Acres
53
108
161
68
229
MINE 6 - WELCH COAL CO.
Background Information
     Welch Coal Company operates a small strip mine about 16 km (10 mi)
northwest of Sheridan anc! a few kilometers west of the Big Horn Mine
[Fig. V-l].  Production in 1972 was only 18,100 metric tons (20,000 tons)
with no projected production through 1980.
     The active mine is very small.  Computer classification was thus
difficult and of little value.  However, extensive spoil piles of an
abandoned mine in the area north of the Tongue River and the town of
Kleenburn [Fig. V-9] were classified and are discussed below.
Classification of Abandoned Mine Areas
     Four areas of spoils piles with varying degrees of re-vegetation
are apparent [Fig. V-ll].  Several small ponds were defined in the three
major spoils areas.  Exposed spoils in the three areas yielded differing
spectral  signatures.

-------
 54
     Spoils in the northwestern area were overgrown with vegetation.   In
the northeastern area, the spoils were uniformly graded and gave the
highest intensity signature because they were oriented in an optimum
direction (east-west) for reflecting sunlight.
     Several ponds and areas of vegetation break up the structure of  the
middle spoils pile.  By carefully locating the training areas, the spectral
signature of the spoils alone was determined.
     Much of the southeastern area is covered by ungraded spoils.  Some
of the spoils are partially overgrown with vegetation.
     To the west of the middle spoils is an area where much light-colored
soil is eroded and exposed.  Such an area is frequently classified into
the same categories as the spoils piles because of similar spectral
signatures.
     The areal distribution of each classification of the abandoned spoils
piles is summarized in Table V-3.

                               Table V-3
                    Areal  Distribution of Land Use
                     Abandoned Spoils Near Mine 6
Land Use

Overgrown spoils
Lakes (total)
Northern section spoils
Middle section spoils
Southern section spoils
Total Area

Pixels
32
41
67
113
117
370
Area
Hectares
16
20
33
55
57
180

Acres
38
49
80
136
140
444

-------
                                                                                                        55
       LEGEND

       •EVEGETATED  SPOILS

       IRREGULAR SPOILS
       REGULAR SPOILS

B§?^   STREAM  BED

'    '   POND
                 EXPOSED SOIL
               SLATER CREEK
           Figur« V —11. Cloilif icotion Mop of Abandoned Spoilt  Pil»« North o( Min«  6-
                             W«lch Coal  Company, Kleonburn, Wyoming

-------
56
MINE 7 - CARTER OIL CO.
     Carter Oil Company has proposed a mine about 12 km (7 mi)  north
of Gillette [Fig. V-l].  It is doubtful that mining will  begin  before
1985.

MINE 8 - ATLANTIC RICHFIELD
     Situated 68 km (42 mi) south of Gillette, this mine  is just begin-
ning operations with 1975 production estimated to be 0.18 million metric
tons (0.2 million tons) [Fig. V-l].   Production is expected to  increase
sharply, reaching 6.4 million metric tons (7.0 million tons) annually by
1980.  No indication of the mine was visible on the ERTS  imagery.

MINE 9 - REYNOLDS METALS, INC.
     A mine has been proposed 32 km (20 mi) south of Sheridan [Fig.  V-l].
Production plans are uncertain.
MINE 10 - KERR-McGEE
     No 1975 production is planned at this proposed mine 63 km (39 mi)
south of Gillette and a few kilometers north of Mine 8 [Fig.  V-l].
Coal production of 6.4 million metric tons (7.0 million tons) is planned
for 1980.

-------
                                                                     57
                             MONTANA MINES

MINE 11 - DECKER COAL CO.. DECKER MINE
Background
     West of the Tongue River Irrigation Reservoir 30 km (19 mi)  north
of Sheridan, Wyoming, this  mine is about 24 km (15 mi) northeast of
Mines 5 and 6 [Fig. V-12, inside back cover].  Coal production was only
0.7 million metric tons (0.8 million tons) in 1972; however, a major
increase in production to 5.4 and 11.8 million metric tons (6 and 13
million tons) is expected for 1975 and 1980, respectively, making this
the second largest mine in the Northern Great Plains.
     The satellite imagery [Fig. V-13] shows that the disturbed area of
the mine is characterized by a circular pattern.  The loading area is in
the center with radial roads extending to the active mine area.  The
mine is advancing along a circular arc into the hillside to the west and
south.

Classification of Mine Areas
     In addition to the high wall, two types of disturbed areas were
present in the mine area.  The area to the west and south of the high
wall was being prepared for mining and the vegetation had been removed
[Fig. V-14].  However, the surface had not been disturbed to any great
extent.  The land to the east and north of the high wall contained
spoils piles.  Both types of disturbed areas displayed the same spectral
signature precluding separation by computer analysis.

-------
  58
MINE NO.  11
                   ItmmMimi iiiimmiun
                  """" * -^**-*fe^fi«i4»«* •• •• MMttMHBBMIMi
                                  , -pTv, ,- -•.taw
                                  II litiltttitill 11M 111111111
                   1390-17S3M-5  FED    17AUG73    £-14
                   (N.G.P.)
                   C N44-37/HIGS-37  C M44-34/U106-24  HM 151
                   •SUM EL51 «135    5433
                    APPLIED :KDJ =   £.74 DEGREE:.  ft^PECT =
                   1.392930
                   »* CTPP MINE HO,  11 »«
                   KTRTCH2
                   STPETCH 5 32-112

                               JDA AiJG 19»  1974 124416 JPL/IPL
                                                                  TONGUE RIVER

                                                                RRIGATION  RESERVOIR
                                                                          -N-
               Figure  V-13.   Satellite   Image  of  Mine  No.  11

-------
                                                                                                         59
                                           LEGEND
                                      ACTIVE  MINE AREA (HIGH  WALL)






                                      UNGRADED SPOILS
                              li-UJ   STREAM BOTTOM





                              II   SHALLOW WATER AND PONDS






                              ^H   DEEP  WATER








                                              Tongut  Siv.r Irrigalio
                                                  Mint  Offlc«l
Figure V—14.  Clai iificati on  Mop of Mine 11— Decker  Cool Company,  D e e k e r, Monte

-------
60
     Because the active area is a narrow cut, the picture elements blend
the active area with the disturbed areas on both sides.  The pure
spectral signature of the active area was present for the north-south
cut, but the majority of the picture elements covering the active area
fell into a different class because of the blending.  This effect was
particularly noticeable in the southern portion of the mine.  Small ponds of
water were defined in the active area.
     The upstream end of the Tongue River Irrigation Reservoir is east
of the mine [Fig. V-13].  Two distinctly different spectral signatures were
observed in this portion of the reservoir.  Either the end of the reservoir
has been shoaled by silt deposition giving the light image area, or more
probably, turbid river inflow had discolored the southern end of the reservoir.
The v&lley of the Tongue River south of the reservoir was clearly
distinguishable.
     The areal  distribution of land use at this mine is summarized in
Table V-4.
                               Table V-4
                    Areal  Distribution of Land Use
                                Mine 11
Land Use

Active mine
Ponds in mine
Spoils
Area in preparation
for mining
Total Area

Pixels
83
9
90
107
289
Area
Hectares
40
4
44
52
140

Acres
100
11
108
128
347

-------
                                                                      61
MINE 12 - WESTMORELAND RESOURCES. SARPY CREEK MINE
Background
     Westmoreland Resources operates this strip mine about 40 km (25 mi)
northeast of Hardin in the upper Sarpy Creek drainage area [Fig. V-12].
The mine began operation in 1973.  Projected annual production for 1975
and 1980 is 3.6 million metric tons (4 million tons).
     A small tributary of Sarpy Creek bisects the mine site.  A substan-
tial area of cropland (light-colored areas) is present in the vicinity
of the mine as shown in the satellite image [Fig. V-15].
Classification of Mine Areas
     This mine had just begun operation when the ERTS image was taken
during the summer of 1973.  A large area in the vicinity of the mine was
distinguishable as disturbed area, different from the native rangeland
[Fig. V-16].  Separation of this area into mined areas and cropland was
difficult.
     Those areas classified as mined areas exhibited spectral signatures
similar to the high wall and spoils areas of Mine 14 about 32 km (20 mi)
to the east on the same ERTS frame.  Close examination of a 1971 US6S
aerial photograph of the mine vicinity revealed that part of the area
south of the tributary stream, which showed the same spectral signature
in 1973 as the mined area, was actually cropland in 1971.  The area to
the north of the stream classified as  mined area does not appear to be
cropland in the 1971  aerial  photograph.   Whether this area is being
mined or has been placed in cultivation since 1971  is unknown.
     A summary of the land classification is shown  in Table V-5.

-------
  62
SARPY  CREEK
                          1373-1729Q-R-5 IO   31.JUL73   .£-9
                          (H.G.F'.)
                          C N45-58/U107-2e  S H45-5S4iiO?-17 Hffi 1
                          SUM EL54 A2132    5201
                           APPLIED SKEW =   £.67 DEGKEC'. ASPECT =

                          • SIMP HUE NO. 12 »»
                          iriSEPT
                          A1TPTCH2
                          ITRETCH I 42- 99
                                                                       TRIBUTARY
                                                                        MINE  NO.  12
                                                                               -N-

                                      JDA AU6 19t117i_123SOS JPL/IPL
            Figure   V-15.  Satellite  Image  of  Mine  No.   12

-------
                                                                                               63
                                                                     LEGEND
                                                                       ACTIVE MINE

                                                                       BACKGROUND VEGETATION

                                                                I    I   IACKGIOUND VEGETATION

                                                                       STREAM IED
Figur. V-16. Cloilllicalion Map ol Min.  12—  W.itmor.  Land Re1ourc«l.
                  Sorpy Cf««k Min«,  Hardin,  Wyoming

-------
 64
                                Table  V-5
                     Areal  Distribution  of  Land Use
                                 Mine  12
Land Use


Mined area north of tributary
Mined area south of tributary
Total area


Pixels

99
53
152
Area

Hectares

48
26
74


Acres

119
64
182
 MINE  13  -  PEABODY  COAL CO..  BIG SKY MINE
 Background
      Peabody Coal  Co. operates the Big Sky Mine about 8 km (5 mi)
 southwest  of Colstrip [Fig.  V-12].  Annual coal production is estimated
 to  be 3.6  million  metric tons (4 million tons) during the 1975 to 1980
 time  period, a significant increase over 1972 production of 1.4 million
 metric tons (1.6 million tons).
      The Big Sky Mine and the large Rosebud Mine at Colstrip are visible
 in  the satellite image [Fig. V-17].  Colstrip is in the headwaters of
 Armells Creek, a small tributary of the Yellowstone River.

 C1 assi fi cati on of Mi ne Areas
      Comparison of the classified image with a USGS aerial  photograph indv
 cated that excellent discrimination was achieved between graded and
 ungraded spoils [Fig. V-18].  Disturbed ground in the railroad car
 loading area and an area where turning trucks had eroded the  grass from
 the soil  yielded the same spectral  signatures as the spoils.
     Two classes were identified  for surrounding undisturbed  areas.   One
was grass covered,  the other grassy areas  interspersed  with patches  of

-------
                                                                                  65
ARMELLS  CREEK
              -N-
   MINE  NO.  13
                         1373-17290-R-5 FED   31JUL73    E-?
                         (H.G.P.)
                         C fM5-53.-M107-28  £ N45-5&/M07-17  MM 1
                         iUM £154 6Z132    5^01
                          APPLIED SKEW =   t.V PEGPiE:. AIPECT =
                         i.'3^059
                         •» iTRIP !1IflE: MQ.  12U4 **
                         IfCEPT
                         (WTPTCH2
                         :TPETCH  45-11^

                                     JDA MJG T?» W74 123530 JPL
MINE  NO. 14
                                                                        RAILROAD
     Figure  V-17.  Satellite  Image   of  Mines   No.  13  and  14

-------
66
                -N-
  Truck Turning Area
                                                   Railroad Loading Area
                                                  ACTIVE MINE AND
                                                  UNGRADED SPOILS

                                                  GRADED  SPOILS
 Figure  V—18. Classification Map of Mine  13—  Peabody Coal Company,
                     Big Sky  Mine, Colstrip, Montana

-------
                                                                     67
evergreens.  Comparison with the aerial  photograph  indicated  that these
types of vegetative cover were successfully separated.
     The area of disturbed land is summarized in Table  V-6.

                               Table V-6
                    Areal Distribution of Land Use
                                Mine 13
1 anrl Ifco


Railroad loading and turnaround
Graded spoils
Active mine and ungraded spoils
Total area

Pixels
47
51
87
185
Area
Hectares
23
25
42
90

Acres
56
61
104
222

-------
 68
 MINE 14 - WESTERN ENERGY CO.,  ROSEBUD MINE
 Background
      The Rosebud  Mine  at Colstrip  is the  largest coal strip mine in the
 Northern Great Plains  [Fig.  V-12].   It produced 5 million metric tons
 (5.5 million  tons)  of  coal  in  1972 with the  same value projected for
 1975.   By 1980 the  yield of the mine is projected to  increase nearly
 four-fold to  17.5 million metric tons (19.3  million tons) per year.
      Because  of its size, the  mine is an  excellent target for satellite
 imagery and does  not pose the  problems of small size  training areas and
 blending of spectral signatures that are  associated with smaller mines.
 The  mine is clearly visible  in the satellite image [Fig. V-17].  Geo-
 metrically the mine is complicated, consisting of three arms of spoils
 piles  in the  northern  section, a northwest-southeast  central arm, a
 goose  neck and the  southern  active mining area.
 Classification of Mine Areas
     The mine  was classified in two steps since the initial aerial
 photograph obtained  from  the Company covered only the southern section.
 Classification  of the southern section defined the high wall, the spoils
 piles, an  older revegetated area, and a more recent attempt at revegetation
 [Fig. V-19].   Also  identified were the areas in preparation for mining
 south of  the high wall  and a railroad and highway corridor skirting the
 east side of the mine.   Distinguishing the mined area from eroded soil
west of the southern section and the  goose neck proved difficult.

-------
KZI
LEGEND


  ACTIVE MINE

  •(VEGETATED

  GtADED SPOILS

  UNGRADED SPOILS

  NATUHAUY REVEGETATiD

  MISCELLANEOUS
                                                EXPOSED HOCK OUTCIOPSING
                                                                                                                           HAILIOAD AND HIGHWAY CO««IDO«
                        Figure  V —19. Classification Map of the Southern Section  of Mine 14 — Western  Energy Company,
                                                          Rosebud Mine, Colstrip,  Montana
                                                                                                                                                            cn

-------
70
     When the spectral signatures derived from training sites in the
southern part of the mine were applied to the northern part, the results
were poor due to the changing character of the spoils.  New training
sites were thus established in the northern part of the mine.  Exam-
ination of aerial photographs covering the northern section shows that
the spoils piles occur in well-ordered rows that are oriented along the
direction of the arms of the mine.  The orientation of the spoils piles
to the incident solar radiation thus changes with location, also changing
the spectral signature.
     Figure V-20 illustrates the classification derived from training
sites on spoils piles in the northern section of the mine, nlus a few of
the original training sites from the southern section of the mine.  Some
overlapping occurred since several different classes of spoils piles
were used that closely resemble each other spectrally.  It was possible,
however, to distinguish between graded, irregular ungraded, regular
(uniform) graded and partially overgrown spoils.
     The high wall areas in the northern section of the mine did not
classify separately because they were too narrow.  In the one successful
attempt at classifying the northern section highwall, only a few pixels
one column wide were detected.  Due to natural revegetation, the western
part of the spoils (shown by the dashed line) near the landing strip on
the southwestern arm classified as background.  The area northeast of
the drainage ditch has naturally revegetated.  The full extent of the
area mined was thus not classified as disturbed area.

-------
                                                                                                                   71
                                                                                                       DIAINAOf OUCH
Figure V —20. Cloilifieotion Mop of  the Northern Section  of Mint U — W«it«rn En«rgy Company,
                                    Roi*bud Mine, Colttrip, Montana

-------
72
     In the diagonal  middle part of the mine,  the drainage ditches  and

roads that cut across the graded spoils [Fig.  V-20]  were identifiable.

     Table V-7 summarizes the classification of mine areas.


                               Table V-7

                    Areal Distribution of Land Use
                                Mine 14
Land Use

Northern section
High wall
Ungraded spoils
Partially overgrown spoils
Subtotal
Middle section
Graded spoils
Ungraded spoils
Roads and drainage ditches
Subtotal
Southern section
High wall
Graded and seeded
Ungraded
Graded for special studies
Subtotal
Total high wall
Total graded
Total ungraded
Road and ditches
TOTAL AREA

Pixels


26
479
530

299
335
132
766

100
85
207
241
633
126
625
1046
132
1,929
Area
Hectares


13
233
257

145
163
64
372

49
41
101
117
308
61
304
508
64
937

Acres


31
575
636

359
402
158
919

120
102
248
289
759
151
750
1255
158
2,315

-------
                                                                     73
MINE 15 - KNIFE RIVER COAL CO.. SAVAGE MINE
Background
     The small Savage Mine is about 13 km (8 mi) northwest of Savage
near the North Dakota border [Fig. V-12].  Annual coal production from
1972 to 1980 is expected to average 0.29 million metric tons
(0.32 million tons).
Classification of Mine Areas
     Four east-west cuts and a smaller cut to the south comprise the
mine [Fig. V-21].  In the north area, the two cuts have advanced toward
each other, meeting in the high wall area shown.  A similar mining
pattern was present in the middle section.  The small  southern cut and
its highwall were separately discernible.  Much of the disturbed area
may not have been excavated, but may represent roads and areas stripped
of vegetation in preparation for mining.   The classification of mine
areas is summarized in Table V-8.

-------
74 ,
   -N-
                                                    NORTHERN
                                                    TWO CUTS
     LEGEND

      ACTIVE  MINE
      (HIGH WALL)

      UNGRADED  SPOILS
                                                           ROAD
                                                             MIDDLE
                                                           TWO CUTS
                                             SOUTHERN SECTION
               Figure V — 21. Classification Map of Mine  15 —
               Knife River Coal Company, Savage, Montana

-------
                                                 75
           Table V-8

Area! Distribution of Land Use
            Mine 15
Land Use

Northern two cuts
Spoils, roads and areas in
preparation for mining
High wall
Subtotal
Middle two cuts
Spoils, roads and areas in
preparation for mining
High wall
Subtotal
Southern section
Spoils, roads and areas in
preparation for mining
High wall
Subtotal
TOTAL AREA

Pixels

145
15
160
154
18
172

33
3
36
368
Area
Hectares

70
7
78
75
9
84

16
2
18
179

Acres

174
18
192
185
22
206

40
4
43
442

-------
 76
MINE  16  -  RELIABLE COAL CO.. RELIABLE MINE
      Located 87 km (54 mi) northeast of Billings [Fig. V-12], this mine
produced less than 910 metric tons (1,000 tons) of coal in 1972; 1975 and
1980  projections were not available.  Because of its small size this
mine  was not classified.

MINE  17  -  DIVIDE COAL CO., STORM KING MINE
      This  small mine is 48 km (30 mi) north of Billings.  Previously
an underground mine, surface operations were begun in late 1972, producing
less  than  0.11 million metric tons (0.12 million tons) of coal.  No
estimates  of future production were available.  Because of its small
size  this  mine was not classified.

MINE  18 -  WESTERN COAL CO.. WESTERN MINE
      This  small underground mine, 64 km (40 mi) north of Billings,
produced less than 910 metric tons (1,000 tons) of coal in 1972;
projections for 1975 and 1980 were not available.   Since this is an
underground mine, ERTS technology does not apply and the mine was not
classified.

MINE  19 - NIES COAL CO., NIES COAL MINE
     Also an underground operation, this small mine 74 km (46 mi) north
of Billings produced less than 0.11 million metric tons (0.12 million
tons)  of coal  in 1972;  projections were not available.   The mine was not
classified.

-------
                                                                    77
                          NORTH DAKOTA MINES

MINE 20 - BAUKOL NOONAN INC.. CENTER MINE

Background
     This mine is about 5 km (3 mi) east of Center [Fig. V-22, inside
back cover] and 2 km west of an impoundment on Square Butte Creek.  Coal
production in 1972 was 1.3 million metric tons (1.4 million tons) and
will be similar (1.4 million metric tons) in 1975.  By 1980 production
is expected to increase to 4.1 million metric tons (4.5 million tons).
     The mine is in two sections with only the north one active.  These
areas are clearly visible in the satellite image [Fig. V-23].
Classification of Mine Areas
     A thermal electric generating plant adjacent to the impoundment
[Fig. V-24] was defined.  An adjacent coal pile was also defined and had
the same spectral signature as dark areas of the mine.  The east end of
the road connecting the plant and mine was identified.  A parking area
surrounding a large building is alongside this road.
     The northern section of the mine is the smaller but is rapidly
expanding.  Mining is proceeding to the north and west.  The southern
mine section is in the shape of a long arc curving to the southwest.
Some revegetation may have occurred to the east of the high wall and
spoils piles in the southern section.   Without ground observations, it
was impossible to distinguish this revegetated land from the surrounding
cropland.
     The areal  distribution of land use is summarized in Table V-9.

-------
78.
                SQUARE  BUTTE CREEK.
    MINE  NO.  20
                             I Mil I IIII111111111 11 II1111111111 [ 1111 [ III111111 j IIII
                                          I I I I I I I I II I tl I I ! I I III Mill I II I I I I I I
                             13-^-17054-1-5 RED   2?Jit?:    M.j.F. E-7
                             C rW7-££/U101-l2  '. M'-SmCO-?? HDQ 1?£
                             iUH EL54 n:i;J    !:i!
                             Irr.ECT
                             •• :TPI? ni?,r ••:. ^; .»
                             nff.::: ;•:.. =   z.w :ES=E-::. -:=E:T =
                             i.3»l5Q
                             ftiWcna
                             :*:"•> ? 44- 25

                                        Jft WG K.' 1?74 033421 jR/IPl."
                                                                          -N-
         Figure  V-23.  Satellite  Image  of  Mine  No.   20

-------
                                                                                                                                  79
       LEGEND

(ffi&l   ACTIVE MINE

[£53   UNGRADED SPOILS
K53   IOADS AND POWER PLANT

ti&ft   STREAM IED

1^81   RESERVOIR
                                  Figure V — 24. Clauificotion  Mop of Mint 24— Boukol Noonan Inc.
                                                  C»nl«r Mine, C*nl«r, North Dakota

-------
80
                               Table  V-9
                    Areal  Distribution  of  Land  Use
                                Mine  20
Land Use

Northern section
High wall
Spoils
Subtotal
Southern section
High wall
Spoils
Road
Subtotal
TOTAL AREA

Pixels

19
66
85

56
100
12
168
253
Area
Hectares

9
32
41

27
49
6
82
123

Acres

23
79
102

67
120
14
202
304

-------
                                                                       81
MINE 21 - NORTH AMERICAN COAL. INDIANHEAD MINE

Background
     Coal production from this mine at Beulah [Fig.  V-22] was 0.9 million
    metric tons (1.0 million tons) in 1972.  Expected production for 1975
will be similar (1.1 million metric tons) but may increase to 4.1 million
metric tons (4.5 million tons) by 1980.
     The Knife River a few kilometers to the southeast is between the
Indianhead Mine and the Knife River Coal Mining Co.  operation (Mine 22)
[Fig. V-25].  A substantial area in the vicinity of these mines is
cropland.

Classification of Mine Areas
     Several examples of the problems involved in the classification of
strip mines using satellite multispectral data were encountered at this
site.  In the easternmost toe of the mine [Fig. V-26], the spectral
classification of the graded and seeded area changed in the northern
part due to a dark discoloration of the soil.  In the middle of the
mine, an old spoils pile that was overgrown with vegetation had a spectral
signature indistinguishable from vegetation outside the mine.  It was
difficult to obtain a representative spoils pile signature anywhere in
the eastern portion of the mine.  Close examination of Soil Conservation
Service aerial photographs indicated that there was little consistency
in the size and angular orientation of the spoils piles.  Consequently,
the spectral signature varied with location in the mine.  The effect
became more noticeable at the northern end of the mine at the site of
current activity.   There the spoils were piled up in unusually high and

-------
   82
 MINE NO.  21
KNIFE RIVER
                                                                   NORTH  MINE  NO.  22
                                                                           -N-
                        1352T7114-R-5 RED   10JUL73-    E-30
                      ;  C N47-£54JiO£-34  S H47-25/U102-24  HDG 192
                        SUH EL57 AZ131    4903
                         APPLIED SKEW =   £.59 DEGREES. ASPECT =   ,
                        1.391117
                        ** STF'IP MINE fffiS. 21i22 »«
                        ASTRTCH2
                        STRETCH  36- 83
                                    t T ' f F B
                       •^^^-	 -.;, .•JW.^&r
                                                                  SOUTH  MINE  NO.  22
      Figure  V-25.  Satellite   Image  of  Mines  No.   21  and  22

-------
                                                                                 LEGEND
Figure V — 26.  Classification Map of Mine 21 —  North American  Coal Company,


                           Indian Head,  North Dakota
00
OJ

-------
84
large mounds at varying orientations to the sun,  producing an extremely
large statistical  variation in the spectral signature.   The area was
very difficult to classify and was outlined from  pixels  that did not
fall into any of the other classifications.
     The orientation of the spoils in the western cut of the mine was
relatively uniform and produced a consistent terrain that classified
easily into graded and ungraded spoils.
     The classification of mine areas is summarized in Table V-10.

                              Table V-10
                    Areal Distribution of Land Use
                                Mine 21
Land Use

Western arm
Ungraded spoils
Graded spoils
Subtotal
Eastern section
Graded and seeded
Graded and seeded (dark soil)
Graded only
Ungraded spoils
Active mine and large spoils
piles (Northern part)
Subtotal
TOTAL AREA

Pixels

155
62
217

170
114
219
180
278
961
1,178
Area
Hectares

75
30
105

83
55
106
87
135
447
572

Acres

186
74
260

204
137
263
215
334
1153
1,414

-------
                                                                     85
 MINE 22 - KNIFE  RIVER  COAL  MINING  CO..  BEULAH MINE
 Background
      This Company operates  two  mines  near  Beulah on  the  Knife River east
 of Mine 21 [Fig.  V-22].   One  mine  consists of two long and narrow cuts
 about 8 km (5  mi) south  of  Beulah  clearly  visible in the  satellite image
 [Fig.  V-25].   The second mine is a similar distance  northeast of Beulah
 and north of the Knife River.   Coal production  through 1980  is expected
 to remain at the 1972  level of  1.4 million metric tons (1.5  million
 tons).
 Classification of Mine Areas
      The south mine  consists  of two long,  parallel east-west cuts with
 small  north-south cuts on the western edge.  The spoils have not been
 revegetated or graded.   The mine was  classified into high wall (dark
 material)  and  ungraded spoils (light material)  [Fig.  V-27].
     At the time  of  the  initial computer analysis, the existence of the
 north mine was unknown.   Therefore, no  training sites were selected from
 the  north  mine.   Upon  realizing that the north mine  existed, a computer
 classification was performed  using  the  statistics from the training
 sites in  the south mine.  Comparison of the resulting classification map
 with a  1971 aerial photograph showed that  the areas on the northern and
 eastern  portions  of the mine were fairly well defined but the spoils
 piles on  the southwestern corner were largely missed.  The explanation
 for this  is not known,  but experience with Mine  21  suggests that  if the
 spoils piles were overgrown  with vegetation,  they  would give  the  same
classification  as the surrounding  agricultural  land.   Questions

-------
86
                                                   Figur*  V-27.  Clot i f i c otio n Map of Min« 22 —
                                          South  Mint of Knife •:»•-  Coal  Company, louloh.  North  Dakota

-------
                                                                     87
such as this can only be resolved by more extensive ground observations.
     Mine 21 is on the same small element of the image extracted from
the Earth Resources Technology Satellite frame and would be expected to
have spectral characteristics similar to the north mine.  The north mine
was classified using the statistics from training sites at Mine 21.
Comparison of the resulting classification map [Fig. V-28] with the
similar map derived using training sites from the south mine showed that
the use of statistics from Mine 21 resulted in systematically larger
areas classified as part of the mine.  However, the southwest corner of
the mine was still largely classified as agricultural land.  The classi-
fication of the north mine based on Mine 21 training sites was selected
as the most representative.  Because of the difficulty in extrapolating
classifications from mine to mine, no further sub-classification of the
disturbed areas of the north mine was attempted.  Table V-ll summarizes
the classification of disturbed areas for both mines.

-------
                                               -N-
Figure V — 28.  Classification Map  of  North  Mine  of



Knife  River Coal Mining  Co.  Beulah,  North  Dakota

-------
                                                                    89
                             Table V-ll
                    Area!  Distribution of Land Use
                               Mine 22
          Land Use
                                     Pixels        Hectares       Acres
South Mine
 North arm
     High wall                       116             56            139

     Spoils                          138             67            166

                                     254            123            305

 South arm
     High wall                       151             73            181

     Spoils                          202             98            242

                                     353            171            424

     Subtotal south mine                 607           295            728

North mine

     Using training sites
        from Mine 21                     763           371            916
     Using training sites
        from southern part of
        Mine 22                      488            237            586


     TOTAL AREA                        1,370           665          1,644

-------
90.
MINE 23 - CONSOLIDATION COAL CO.,  GLEN HAROLD  MINE

Background
     This mine 8 km (5 mi) southeast of Stanton and  64  km (40  mi)
northwest of Bismark [Fig. V-22] produced 1.3  million metric tons  (1.4
million tons) of coal  in 1972.   An annual yield of  3.6  million metric
tons (4 million tons)  is forecast for the 1975 to 1980  period.
     The mine is in very hilly terrain in the  bluffs south of  the
Missouri River.  The disturbed area is large,  as shown  in the  satellite
image [Fig. V-29].  The physical configuration is heavily influenced by
the terrain that has divided the mine-into a north  section and two south
sections separated by a northeast-southwest ridge.

Classification of Mine Areas
     At the time of data collection, reclamation had progressed to the
point where most of the mine area could be separated into graded or
ungraded spoils.  About half of the west section had been graded with
the rest of the irregular or ungraded spoils [Fig.  V-30].  Most of the
east section is ungraded with small irregular  sections.  The highwall is
the south edge of these two sections but was not detected in the classifi-
cation.  Separation of spoils types was not clear in the north area.
Table V-12 summarizes  land use in the mine.

-------
                                                                              91
                                                                >
MISSOURI  RIVER
MINE NO. 23
              -N-
1&M7054-L-5  RED   27JUL73    H.G.P. £-7
C f!47-E£'m01-i£ S H47-2i4ttQ8-S9  HOG Itt
SUM EL54 S2133    5145
 flPPLlED SHU =  £.60 I'EGREt;.  ASPECT =
1.391150
** STRIP HM MQ. 23 »«
ASTPTCH2
STRETCH 0 35- %

           JUA AUG 26. 1974 033321 JPL/IPl
       Figure  V-29.  Satellite  Image  of  Mine  No.  23

-------
92
        LEGEND
       GRADED SPOILS
       UNGRADED SPOILS
       MIXED SPOILS, ROADS
                                                                                                         ROAD
                   Figure V —30. Clonificotion Mop of Mint 23— Ccniolidation Coal Company,
                                   GI»n/Harold Mine, Stanton,  North Dakota

-------
                                                                                93
MINE NO.  25
        -N-
                          1333-17104-R-5  REB   15flUG73    M.G.P. E-ll
                          C N48-50/U101-58  S N43-4S4I101-47  HUG 192
                          SUfl 0.49 A2140     5410
                           APPLIED SKEU =   2.52 DEGREE',.  ACPECT =
                          1.390208
                          M STRIP MINE 1C. 25 «»
                          ASTRTCH2
                          STRETCH 9 32- 67
                                      JM m 26»  1974 03i5£t JPL'IPL
           Figure   V-31.   Satellite  Image   of  Mine   No.   25

-------
94
                              Table V-12
                    Areal Distribution of Land Use
                                Mine 23
Land Use

North Section
Assorted spoils
West Section
Graded spoils
Ungraded spoils
East section
Graded spoils
Ungraded spoils
Roads
TOTAL AREA

Pixels
170
209
167
128
407
65
1,146
Area
Hectares
83
102
81
62
198
32
557

Acres
204
251
200
154
488
78
1,375
MINE 24 - BAUKOL NOONAN, INC.. NOONAN STRIP MINE

     This mine 5 km (3 mi) east of Columbus [Fig.  V-22] produced
0,4 million metric tons (0.5 million tons)  in 1972.   An aerial
photograph of the mine indicated that, its two large  areas of spoils
piles measure 5 km (3 mi)  by 1.3 km (0.8 mi) and 2,4 km (1.5 mi)
by 2.i km (1.3 mi).   As the mine is closed  with no future production
expected, no classification was performed.

-------
                                                                    95
MINE 25 - CONSOLIDATION COAL CO.. VELNA MINE

Background
     Located 40 km (24 mi) southeast of Minot [Fig. V-22], this mine
produced 0.43 million metric tons (0.47 million tons) of coal  in 1972.
Production for 1975 and 1980 is estimated to remain at this level.
Classification of Mine Areas
     The satellite image of this mine [Fig. V-31] was very difficult to
interpret visually.  But two sources assured that the mine is located
within this particular part of the ERTS frame: one was an accurate
examination of the ERTS frame at the known geographical coordinates of
the mine; another was the location of the mine at the northern edge of
an easily identifiable strip of land that represents the terminal moraine
of the continental ice sheet during the Glacial Age.  A careful comparison
of the satellite image with aerial photographs showed two bright areas
at the ends of the mines that can be identified, and dark vegetation is
growing in stream bottoms immediately southwest of the mine.
     All attempts at multispectral classification failed to separate the
middle area of the mine lying between the two bright areas at the ends
from the surrounding cropland.  In classifying other mines it was observed
that when the land surrounding the mines was planted in the rectangular
pattern common in midwestern agricultural practice, an effective camouflage
pattern was formed.  That factor made the mine very difficult to distinguish
with the multispectral  recognition algorithms currently used.

-------
96
MINE 26 - KNIFE RIVER COAL CO..  GASCOYNE MINE

Background
     Located 8 km (5 mi) northeast of Gascoyne [Fig.  V-22],  this  mine
produced only 0.15 million metric tons (0.16 million  tons)  of coal  in
1972.  A major expansion of mining activity is planned,  however,  with
annual production for the 1975 to 1980 period forecast at 2.7 million
metric tons (3 million tons).

Classification of Mine Areas
     A majority of the land in the vicinity of the mine is cultivated
farmland as shown in the satellite inage [Fig. V-32].  Discrimination
between mine areas and the adjacent striped cropland  proved to be diffi-
cult.  Only two classes of spoils (light and dark) could be separated
[Fig. V-33].  The mid-section of the mine was very irregular and  contain-
ed several small ponds.  The computer classification  map of this  area
was also irregular and difficult to interpret.  Several  small remotely
located pits to the east and south of the mine were picked up in  the
computerized classification.  The coal loading area to the south  of the
mine was distinct.  Extensive enlargement of the mine had occurred
compared to a 1957 aerial photograph.  The areal distribution of  land
use is summarized in Table V-13.

-------
                                                                         97
                      i I»hi n n i ti i iimn H Jimum j JIM
           II III I I) li i I Ml II Mill HIM I I I II Illllll [ I IIHl (I tilt I
         . 135£-171£0-L-5  PED   10JUL73     M.G.P,  E-i
         : C H46-014I1W-09  3 H4t-uO/U10i-5'? HM 192
          JUM EL53 AZ129     4903
          IfCECT -  ^FLIED 'KIU =   £.67 DEGREES.
          ft'PECT =  1.3920&
          *• STRIP HIIC HD. 26 ««
           STRETCH J 35- 93

                       JWt SER.27. 1974,084033 JPL/IRU
                                                           MINE  NO. 26
                                                                  -N-
Figure  V-32.  Satellite   Image  of  Mine  No.  26

-------
98

                                                                       LEGEND


                                                                    UNGRADED LIGHT SPOILS

                                                                    MIXED LIGHT AND  DARK SPOILS

                                                                    COAL LOADING AREA

                                                                    SPOILS WITH NUMEROUS SMALL PONDS
            Figure V—33.  Classification Map of Mine  26 —Knife  River  Coal Mining Company,
                                         Gascoyne, North  Dakota

-------
                                                                    99
                              Table V-13
                    Areal  Distribution  of Land Use
                               Mine  26

Land Use

Active mine and spoils
Small ponds
Railroad loading area
TOTAL AREA


Pixels
192
37
26
255
Area

Hectares
93
18
13
124


Acres
230
44
31
306
MINE 27 - BONSNESS COAL MINE
     Production in 1972 from this mine 5 km (3 mi)  south of Columbus
[Fig. V-22] was less than 0.11 million metric tons  (0.12 million tons).
The mine was closed by 1973 and it was therefore not classified.  An
                                                                       2
aerial photograph shows that spoils piles cover an  area of about 5.8  km
(2.3 mi2) in a strip 7.2 km (4.5 mi) by 0.8 km (0.5 mi).

MINE 28 - UNDERWOOD COAL CO.. INC.. UNDERWOOD MINE
     This company operates a mine 3 km (2 mi) southeast of Underwood
[Fig. V-22].  Production in 1972 was less than 0.11 million metric tons
(0.12 million tons).  Projections for 1975 and 1980 were not available.
     This mine was not located on the ERTS imagery  and, therefore, was
not classified.  On a Soil Conservation Service aerial  photograph, the
mine dimensions were roughly 400 m (1300 ft) by 270 m (900 ft).

-------
 100
MINE 29 - HUSKY BRIQUETTING INC.. LEHIGH STRIP MINE
Background
     This small mine 5 km (3 mi) east of Dickinson [Fig. V-22] produced
less than 0.11 million metric tons (0.12 million tons) in. 1972.
Production estimates for 1975 and 1980 were not available.
Classification of Mine Areas
     The small size of this mine compared to the spatial resolution of
the ERTS imagery [Fig. V-34] and the presence of cropland in the vicinity
limited the accuracy of the classification.  The mine is divided into
three sections by the Heart River and'the Northern Pacific Railroad
[Fig. V-35].
     Inspection of an aerial photograph indicated that spoils in the
north area are overgrown with vegetation.  Two of three ponds were
identified in the classification map.
     In the west section, the bright spectral signature of the spoils
was indicative of recent mining activity.  Two of three ponds in this
area were also identified.   A 17 hectare (7 acre) area that has been
revegetated was not identified in the computer classification map due to
its small  size.
     In the southern part of the mine,  the large pond and mine buildings
were dominant.  The spoils  piles in the northern part of the southern
segment were overgrown with vegetation  and did not classify uniquely.

-------
                                                                        101
           135M7114-L-5  PEB    10JUL73    E-30
           C N47-254J1Q2-34  S N47-25/II102-S4  HDG 192
           SUN EL57 ^2131    490S
           INSECT
           «* STRIP MItlE fO.  2? «
            ^PPLIEB SKEU =   £.59 I'EGPIE:,  ASPECT =
           1,391117
           ftSTRTCH2
           STRETCH t 33- 37

                       JDA SEP 27»  1574 084222 JPL/IPL
                                                         MINE NO.  29
                                                                  -N-
Figure  V-34.  Satellite   Image   of  Mine   No.  29

-------
102
                                                                                             -N-
                                                                ERODED SOIL
                            MINE BUILD INGS
                  LEGEND


               GRADED SPOILS

               LIGHT COLORED SPOILS AND  BARE  SOIL

               NATURALLY  REVEQETATED SPOILS

               LISHT COLORED QROUND

               STREAM BED

               PONDS AND  MINE BUILDINGS
                                                                        HIGH WALL
              Figure  V —35. Clauif ica tion  Map  of Mine 29— Huiky  Briquetting Inc.
                           Lthigh Strip Mint,  Dickinion, North Dakota

-------
                                                                 103
The two revegetated areas in this section are too small  to identify;
however, the bright spoils behind the high wall  showed up distinctly.
Spoils were being graded so close to the high wall that the spatial
resolution of the image was inadequate to define the high wall.   Table
V-14 summarizes the classification of land areas.
                              Table V-14
                    Areal Distribution of Land Use
                                Mine 29
Land Use

Northern section
Ponds
Overgrown spoils
Subtotal
Western section
Ponds
Bright spoils

Pixels

2
33
35

5
26
Area
Hectares

1
16
17

2
13

Acres

2
40
42

6
31
      Subtotal                        31             15           37
Southeastern section
      Pond and overgrown spoils       35             17           42
      Overgrown spoils and
         revegetated spoils          110             53          132
      Bright spoils                   31             15           37
      Subtotal                       176             85          211
      TOTAL AREA                     242            118          290

-------
 104
MINE 30 - ARROWHEAD COAL CO., ADAMS MINE
     Production in 1972 from this small mine was less than 0.11 million
metric tons (0.12 million tons).  Production estimates for 1975 and 1980
were not available.
     Overall mine dimensions on an aerial photograph were 4.6 km (2.8 mi)
by 340 m (1,100 ft).  The mine was too small to locate on the ERTS image
and therefore was not classified.

-------
                                                           105
w. 6H

^v
^SHERIDAN
\
%
\
"\\
^
°\B9
i
V
BUFFALO*^
\ \
\ ^^
\
?
|
1

xx
"" 	 WYOMING j£
I * O
II I*
41 .<
.. -N- IQ
i1
^
IS
-!S

/ i
7H /X j
1 ^-^
GILLETrEaJ ^-i I
^-\ 	 : 	 1^9°
— "^ \ 1
0 i j
i i
o> 1 NEWCASTLE* j
10 I
• ' •
J B
ii 8 !
   N



                                \


                                                           ,eo
                                                           lui
Figure  V — 1.   Locations  of  Wyoming Coal Mines

-------
                                                       106
               CANADA
       USA
       SHERIDAN*
Figure V —12.  Locations of  Montana Coal  Mines

-------
                                                                         107
i  NORTH DAKOTA
                      24,
                               CANADA
                         '•COLUMBUS
                          27

                      H H
JS
1





CO
00
j
"S
                                                              -N-
                                                       M/NOT
                                                             25
                                       SOUTH DAKOTA
                                                                  GP0857-646
      Fiqure V — 22.  Locations of North  Dakota Coal Mines

-------