80 
DISPLAY METHODS 
The display of multi-level data sets introduces a wide 
range of options and problems. We have experimented with 
many of these options in working up displays of various 
integrated data sets. Motion or more generally time vary 
ing real-time displays are very effective when they can be 
used. However, one is often limited to a hard-copy, map 
like product. We have combined color coding, line and 
point overlays, shading techniques, and stereo in our 
hard- copy products (Blair 1982). Specific examples will 
be discussed but a few general comments should be made 
first. 
Many map shading techniques have been discussed in the 
literature. Horn (1981) gives a good summary of shading 
methods using the reflectance map as a common reference. 
He points out that some of the simplest shading methods 
work very well. Our experience reinforces this fact as 
well. A slope histogram may be computed using the com 
ponent of slope in the direction facing away from the 
assumed light source. An effective grey scale shading 
lookup table can then be formed by integrating the slope 
histogram and normalizing the integral to the number of 
grey shades required. This is a technique frequently used 
in image processing to maximize contrast in an image. The 
method is computationally efficient, an average mini 
computer will shade a 3000 x 3000 pixel image in 15 minutes. 
Another shading technique we have found effective is a 
shaded contour format. This method lightens a contour on 
a positive slope and darkens it on a negative slope. The 
method will introduce relief shading without greatly chang 
ing the colors in a map. This requirement is important in 
maps such as geologic or land use where color shades have 
specific meaning. 
We have frequently used stereo in our map products. It is 
a familiar tool to geoscientists. A good quality stereo 
pair of a three-dimensional surface may be generated by 
shifting the picture elements in one image by an amount 
porportional to elevation. Here, elevation is whatever 
variable constitutes the vertical dimension of the surface. 
It may be magnetic field intensity, temperature, or almost 
any quantity. Our usual practice is to introduce parallax 
into only one image. Thus, one image remains unaffected 
but can be used with the shifted image for stereo viewing. 
EXAMPLE MAPS 
The methods and practices discussed above are illustrated 
in the following examples. Figure 2 is a print and stereo 
pair of a water- and land-use map prepared for the Pasco 
Basin in Washington. Portions of three one-degree square 
topographic files were registered for the elevation layer 
of the data base. Other layers consisted of Landsat clas 
sification maps, slope data, digitized canals, irrigation 
wells, streams, watershed boundaries and some political 
boundaries. This merged data base constituted a mixture 
of raster, line, and point data types. The main problem 
here was to combine the data sets in an output product