15-2: The types are many; just a few: Land Ownership;
Political Units (Wards, Townships, Zoning, etc.); Housing Density; Recreational
Facilities; Archaeological Sites; Mineral Resources; Energy Distribution Infrastructure;
Depth to Water Table. BACK
15-3: The general land use map will tell you what
kind of neighborhood your future home would be in. If you are close to an industrial
or business area, perhaps you may want to sell the property or trade for a residential
lot. The vegetation map will warn you that some expensive clearing might be
necessary. The slope map will indicate the type of house you should develop
- if there is a moderate slope, the excavation, which also be influenced by
the soil type, can be tricky. Slope also indicates whether there could be landslide
or runoff flooding problems. This simple case begins to illustrate what a Geographic
Information System can do: the various thematic map imput data are important
in solving a problem or making a decision as to what needs to be considered
in getting something done. BACK
15-4: Land ownership; Topographic relief; Proximity
to Railroads; Direction of Prevailing Winds; Slope Stability; Forest Cover;
Access to power lines; Bearing strength of soil; Existing buildings; Proximity
to Housing; EPA requirements; Cost of living; Present highway locations; Ease
of excavation; Land Costs; Tax rates; Zoning laws. BACK
15-5: This is not a trivial deal - in fact, a lot
of work will be involved before you try to convince some "venture capitalists"
to underwrite your opening mining operations. First, you need to tap some data
sources: Land ownership is critical. Is the land public or private? Records
in the county seat where the possible ore body is a start. If it turns out to
be Federal land, then the Bureau of Land Management, the Bureau of Reclamation,
and the National Park Service are good sources. Similar agencies manage state
lands. If it is private land, the problem compounds. You need to track down
ownership but buying the land may involve some ethical problems. Let's assume
that the land turns out to be federal. Now you need some data elements that
will tell you if anyone has already found "suspicious" types of information
that suggest potential mineralization. The U.S. Geological Survey (or maybe
the State Geological Survey) may have produced geological or mineral maps that
include, and perhaps even detail, the land you are speculating on. See if any
mining companies have run geological mapping or geophysical surveying over the
area. If not, you are going to have to hire some company to map the area and
conduct appropriate surveys. This will include mineral assays (check the state
or county assayer's office to see if any are on file). The cap to this must
be drilling to get at the ore body. At this point you should have gotten "seed"
money from your bankrollers to pay for this. The prime output map will be a
three-dimensional display of the extent of the body and the percentage variability
of the silver it contains. Now, you need the "big cash" to develop it. So, other
interpretive maps may be needed to present your case to your backers. These
include maps that define the conditions that involve setting up the mining operations,
including environmental protection guarantees and haulage routes to smelters.
This answer is rather generalized and leaves out some essentials, but you get
the idea that becoming a millionaire this way isn't easy. And you can appreciate
the role of maps and data accumulation critical to decision-making. GIS has
a key part to play in this, as you will see. BACK
15-6: The first M is Measurement - observing, identifying,
and quantifying the parameters needed in a GIS. The second is Mapping - portraying
the characteristics of the Earth's surface as a series of themes. The third
is Monitoring - updating existing maps, discovering changes, and, in some applications,
getting essential near real time information. The fourth is Modeling - defining
how things work (processes) or relations (actions) vital to the decision-making
interact. BACK
15-7: Clearly, the ultimate factor that determines
how a GIS will be involved in Data Management is the user requirements. These
govern what inputs are to be sought out and integrated. One can also argue that
a second driver is the specific outputs to be developed, usually a combination
of statistics, graphical (including maps) displays, and written reports that
should include interpretations and bases for the decisions that are the final
goal. BACK
15-8: These five methods of encoding are common alternatives:
1) The dominant soil type within the grid cell can be selected for the cell,
namely the type that occupies the largest amount of land in that cell; 2) The
percentage of each soil type within the cell, which achieves a high degree of
detail but at a higher cost; 3) The presence or absence of each soil within
the cell can be encoded; 4) The soil type found at the centroid of the cell
is encoded to represent the value for the entire cell; 5) The Corner designation
method, in which data values are recorded at each corner of the grid cell, or
systematically at some one corner. BACK
15-9: The gray, purple, and dark green colors are
there on each side of the river. They are so close-spaced that they are barely
visible. This thinning along the river is just the fact that the river has very
steep banks (perhaps it is somewhat incised). BACK
15-10: The areas in dark green, purple, and gray
in the elevation map are the lower parts in the scene. When the river floods,
it spills its banks and carries across the medium green strip into these lower
areas but doesn't invade the other areas that are medium green and higher. BACK
15-11: The floodwaters are the key factor. Where
they cover the clays, and over a period of time saturated these fine-grained
soils, they produce optimum soil moisture contents that are higher the water
amounts held in the other soils. Of course, there can be years when flooding
doesn't occur but hopefully some moisture will be concentrated preferentially
in the clay soils anyway. BACK
15-12: Location and identification of major crops;
Distribution and identification of forests; Location and status of lakes; Major
categories of land use; Broad patterns of urban development; Characteristics
and interrelations of landforms; Indications of offshore sediment concentrations;
Wildlife habitat; Range land conditions; Disaster assessment; Mineral/Petroleum
exploration; Meteorological conditions. BACK
15-13: The vector method is definitely more accurate.
The raster cell method is based on which feature occupies more of a cell's area
- this causes loss of information about the lesser occupants of cells containing
two or more features (or categories or classes). BACK
15-14: Despite its greater inaccuracy (less precise
location of individual map categories), the raster method is much easier to
process, assuming the same cell size is maintained for each data element or
layer. When vectors are used, they will almost invariably have different outlines
from one theme map to any other (each such map will have its own vector patterns).
These are difficult to overlay but computer programs do exist to allow analysis
based on point arrays or other decision layouts. BACK
15-15: Any GIS product - including a series of thematic
maps - that PP&L undertook to construct would be subject to "coarse" resolution
(23 acres or about 300 meters on a side), that is, the map categories would
have to be those which tend to extend more or less continuously over large areas,
so that essential details for some purposes would not be discernible. The features
mapped would be restricted to those of the Level 1 categories in the Anderson
Land Use/Land Cover Classification. Particularly limited would be expressions
of transportation (road networks) .and energy flow (power line) infrastructures,
since these linear features tend to be narrow (less than a square acre's width).
Landsat opened up the possibility of a much more detailed set of end products
with many more categories specifiable at its resolution (even better with TM).
Realizing this, PP&L in 1980 approached the ERRSAC group at NASA Goddard with
the (accepted) offer to conduct a joint study on the applicability of remote
sensing imagery to improving the GIS data base and maintaining currency on the
new sets of categories chosen. This led to the Harrisburg and Berwick projects,
the latter of which involved using Landsat MSS data to develop, analyze, and
produce GIS data elements and maps of a nuclear power plant on the Susquehanna
River near Berwick, PA at a site about 20 miles east (and normally downwind)
from the writer's residence in Bloomsburg. Unfortunately, the data tapes that
have the interesting results were unreadable in 1999 so that the demonstration,
which proved quite successful, cannot be included in this Tutorial. BACK
15-16: A photointerpreter, examining Landsat images,
could have been able to produce maps of A., Landforms, and C., Stream Order
(probably only as far as the 3rd or 4th orders) without resort to additional
data sources. Map B., could be constructed if DEM data or a digitized topographic
map data set (done by the PP&L/Goddard team) were available. Map D requires
ground measurements and Landsat would offer little help. Landsat, or other space
observations, would be of some aid in putting together the other maps (D = Flood
Prone Areas; E = Agricultural Potential) only if there were auxiliary information
and, for D, 20 or more years of observations. BACK
15-17: The broad land cover pattern evident in the
PP&L map is effectively duplicated in the Landsat classification. The big differences
are in details, level of classification, and accuracy - Landsat being much more
specific for some classes. Thus, for convenience, the PP&L map combines several
urban categories that are subdivided into three plausible classes in the Landsat
map. However, the PP&L map contains several classes, such as forested and non-forested
wetlands, that were chosen from ground surveys for specialized purposes not
considered germane in the Landsat classification (in principle, they probably
could have been singled out in Landsat but weren't). One big reason, as already
mentioned, for differences between the two maps is the "coarser" mapping level
adopted for the PP&L map: 23 acres compared with the 1.1 acre for Landsat. Differences
also result from arbitrary class selections and intended use. The PP&L choices
tended to lump together some diverse natural features into general categories;
the Landsat map is more of an indication of real spectral classes that properly
correspond to bonafide land cover classes. BACK
15-18: As it appears, for me (NMS) this map raises
some concerns and doubts. It was made by others nearly 20 years ago - individuals
no longer available to ask about details. While the red pattern in the white
square represents a set of conditions that are said to meet all major criteria,
and its selection as located seems based on ruling out red areas outside because
they don't satisfy some unspecified criteria, there arises the question of why
the large areas in the scene that are shown as black may not have also had some
favorable localities. Sometime one cannot easily accept at face value a reported
result; this Tutorial contains several instances but better examples just weren't
available. BACK
Collaborators: Code
935 NASA GSFC, GST,
USAF Academy,
Webmaster: Bill Dickinson Jr.
Primary Author: Nicholas M. Short, Sr.
email: nmshort@epix.net
Contributor Information
Last Updated: July '99
Site Curator: Nannette Fekete
Please direct any comments to the Site
Curator eerstweb@gst.com.