Geographical research requires the management and analysis of spatial
data at multiple scales. As part of the U.S. Geological Survey's globa
l change research program a software system has been developed that re
ads raster data (such as an image or digital elevation model) and prod
uces a pyramid of aggregated lattices as well as various measurements
of spatial complexity. For a given raster dataset the system uses the
pyramid to report: (1) mean, (2) variance, (3) a spatial autocorrelati
on parameter based on multiscale analysis of variance, and (4) a monof
ractal scaling parameter based on the analysis of isoline lengths. The
system is applied to 1-km digital elevation model (DEM) data for a 25
6-km(2) region of central California, as well as to 64 partitions of t
he region. PYRAMID, which offers robust descriptions of data complexit
y, also is used to describe the behavior of topographic aspect with sc
ale.