CHANGE DETECTION AT MULTIPLE TEMPORAL SCALES - SEASONAL AND ANNUAL VARIATIONS IN LANDSCAPE VARIABLES

Landscape attributes vary at a hierarchy of temporal scales. Temporal aggregation of fine-scale observations is necessary before attempting to detect change in landscape attributes at broader time scales. This paper pursues three goals: (1) to demonstrate the limitations of the c lassic change detection approach in situations where landscape attribu tes are characterized by marked seasonal variations; (2) to explore th e effect of increasing the frequency of observation on the change dete ction performances, using a common data aggregation approach; and (3) to rest a method which monitors broad timescale changes in landscape a ttributes by faking explicitly into account the finer time-scale varia tions. These issues are explored with data on the landscape spatial pa ttern of three West African landscapes. The spatial structure of these landscapes is measured from AVHRR data, for a period covering two yea rs. it is shown that seasonal changes in landscape spatial pattern may be much greater than that caused by longterm changes. Therefore, comb ining information from multiple temporal scales of observation is esse ntial to quantify in a meaningful way the spatial dynamics of landscap es.