SPATIOTEMPORAL INFORMATION-SYSTEMS IN SOIL AND ENVIRONMENTAL SCIENCES

This work is concerned with spatiotemporal information systems and the ir application in soil and environmental sciences. Issues investigated in this work include developments in the space/time modelling of natu ral variations, composite spatiotemporal mapping, and the incorporatio n of various sources of information into space/time analysis. Theoreti cal models, simulation examples, as well as real-world case studies ar e discussed. The models can process data available in a space/time con text, offer valuable physical insight and produce sequential regional maps of natural variables that are considerable improvements over pure ly temporal or purely spatial analysis. Spatiotemporal characterizatio n involves a multi-scale description of natural processes that reveals the effects of observation and mapping scales. The limited availabili ty of hard data ultimately affects space/time analysis and, hence, the incorporation of soft data into the study of natural processes can be a very useful approach. In this context, it is shown that Bayesian ma ximum entropy analysis offers significant improvements over traditiona l minimum mean squared error techniques. (C) 1998 Elsevier Science B.V . All rights reserved.