Analysis of grids and surfaces - Spatial Data Analysis: An Introduction for GIS Users - page 168

A B

Val u e Val u e High : 56191.301

High : 13.212

Low : 0.000 Low : 0.000

Figure 10.16 Gradient (in degrees) with starting location superimposed (A)
and the cost surface from this location (B).

Figure 10.16 shows a gradient map (in degrees) with the starting location (A) and
the cost surface derived using these inputs (B).
Cost surfaces are computed in numerous dif erent contexts. In an application con-
cerned with identifying possible routes for the construction of a new railway, Cowen
et al. (2000) generated cost surfaces accounting for several dif erent factors, including
grade (dif erence between existing and desired elevation), road crossings, stream cross-
ings, and track cost. h e generation of alternative routes and comparisons between
them has been found to be very useful in many planning projects of this kind.

Case study

10.7

Gradient (in degrees) was computed using ArcGIS™ Spatial Analyst from the DEM
(with a spatial resolution of 1009.975 m) illustrated in Figure 10.17. h e DEM is
described by Dubois (2003). h e algorithm used is that detailed by Horn (1981) and
the output is shown in Figure 10.18.
h e southern part of Switzerland is dominated by the Alps, and this explains the
steep gradients in that region; in the north-west the Jura mountains are evident. In
terms of, for example, modelling the costs of transporting goods, this output would
suggest that it requires more ef ort to move goods over land in the south of Switzerland
than in the north. If a value can be assigned to a gradient derived from a DEM then
such data may provide the basis of a useful analysis.