21_Acquisition_and_Use_of_Digital_Elevat
21_Acquisition_and_Use_of_Digital_Elevat
Classification
Danny M. Vaughn
Introduction
Digital elevation models (DEMs) were first created by the United States Geologic
Survey (USGS) through the National Mapping Program (NMP); and later sub-
contracted through the United States Forest Service (USFS). Many additional
public and private companies now create digital elevation models through aerial
and satellite surveys.
The DEM begins as a digital data file that stores elevations for a specific area of
the Earth’s surface. The file format is based upon a model (raster) which consists
of a grid developed at regularly spaced intervals. The grid spacing varies for
models, and is generally dictated by the total area covered by a DEM, scale (for a
discussion on scale, see the essay, Acquisition and Use of Topographic and
Orthophoto Maps in Imagery Classification), variations in terrain relief, and the
degree of detail desired. Elevations are assigned at the intersection of each grid
interval in either feet or meters above mean sea level.
Common scales include; 1:24,000 (covers 7.5 minutes by 7.5 minutes of the
Earth’s surface), 1:62,500 (15 by 15 minutes), 1:100,000 (1 by 1 degree), and
1:250,000 (one by two degrees). The 7.5-minute DEM is one of the most common
in use today. A small planimetric area (approximately 92 acres/37.2 hectares) of
the Earth’s surface is represented in one square inch on a map at this scale. DEMs
cover areas over most of the United States and its territories. More recent
developments in LiDaR (Light Detection and ranging) have improved elevation
data to a level of precision below one meter.
Unlike the USGS DEMs that are distributed by topographic quadrangle, the NED
(National Elevation Dataset) DEMs are distributed in pre-defined blocks. The
United States is partitioned in blocks of one degree latitude by one degree of
2
longitude. Figure 1 illustrates a portion of a block of a NED for the area centered
on Crawford County, Indiana. The spatial extent of the DEM covering the Taswell
Quadrangle in figure 2 is outlined in red on the NED model.
A NED model is a seamless, raster-based data set with elevation values expressed
in feet, and distributed on a grid at 30 meter intervals (spatial resolution). There is
a difference in the units of measure as elevation values are in feet, while the spatial
(grid spacing) resolution is in meters. This inconsistency is not really explainable
as the decision to use any unit of measure is made by the NED development team.
Figure 2 is a DEM of the Taswell, Indiana United States Geologic Survey 7.5
minute quadrangle. Elevation values are in meters above mean sea level, and the
grid spacing between elevations is 10 meters. Elevation ranges are classified by
colors and are listed in the accompanying legend. Although color does not
3
necessary reflect any specific land cover type, one can gain a rapid sense for terrain
by identifying the elevations and drainage patterns. For example, the stream
valleys (blue) exhibit a branching (dendritic) pattern providing a means for
determining drainage direction, ridge, and valley locations. The uplands are
colored brown, purple, grey, and salmon. Creating a colorized DEM classification
map enables an investigator to determine a basic vision of the topography which
can be used to associate elevation with potential class features to be interpreted
from high spatial resolution digital multispectral imagery.
Figure 1. Small subset of a National Elevation Dataset for Crawford County, Indiana.
4
Figure 2. Taswell, Indiana 7.5 minute Digital Elevation Model (DEM). Elevations are in
meters.
Vegetation species and the density of vegetation are often related to slope and
aspect. Study the vegetation types in figure 3, the aspect model in figure 4, and the
slopes model in figure 5. The red and brown-red colors in figure 3 represent
Gambel Oak trees, the orange color represents Mountain Maple trees, and the
yellow color represents Aspen trees. The homogeneous stand of dark green
features represents coniferous trees, principally Spruce and Fir. There are some
striking relationships between vegetation and aspect which can also be compared
to the slope and elevation models. The topography of the photograph is deceptive
which is common in high relief mountainous terrain. What appears to be a
topographically higher ridge trending across the lower third of the photograph in
figure 3 is actually a valley (figure 6). The application of spatial models can
5
Figure 3. Color aerial photograph of fall vegetation variances on the Wasatch Front, in
northern Utah.
6
Aspect of DEMs
Flat
North
Northeast
East
Southeast
South
Southwest
West
Northwest
Slope of DEMs
0-5
5 - 10
10 - 15
15 - 20
20 - 25
25 - 30
30 - 35
35 - 40
40 - 45
45 - 50
50 - 55
55 - 60
60 - 65
65 - 70
70 - 75
Figure 5. Slope (in degrees) model for the aerial photograph in Figure 3.
7
Figure 6. DEM model (in meters) for the aerial photograph in Figure 3.
Selected References