REGRESSION TREE ANALYSIS OF SATELLITE AND TERRAIN DATA TO GUIDE VEGETATION SAMPLING AND SURVEYS

Monitoring of regional vegetation and surface biophysical properties i s tightly constrained by both the quantity and quality of ground data. Stratified sampling is often used to increase sampling efficiency, bu t its effectiveness hinges on appropriate classification of the land s urface. A good classification must be sufficiently detailed to include the important sources of spatial variability, but at the same time it should be as parsimonious as possible to conserve scarce and expensiv e degrees of freedom in ground data. As part of the First ISLSCP (Inte rnational Satellite Land Surface Climatology Program) Field Experiment (FIFE), we used Regression Tree Analysis to derive an ecological clas sification of a tall grass prairie landscape. The classification is de rived from digital terrain, land use, and land cover data and is based on their association with spectral vegetation indices calculated from single-date and multi-temporal satellite imagery. The regression tree analysis produced a site stratification that is similar to the a prio ri scheme actually used in FIFE, but is simpler and considerably more effective in reducing sample variance in surface measurements of varia bles such as biomass, soil moisture and Bowen Ratio. More generally, r egression tree analysis is a useful technique for identifying and esti mating complex hierarchical relationships in multivariate data sets.