Seasonal foliage changes in the eastern Amazon basin detected from Landsat Thematic Mapper satellite images

Seasonal changes in tropical forests are difficult to measure from the ground, especially in areas of high species diversity and low phenological synchrony. Satellite images, which integrate individual tree canopies and cover a large spatial extent, facilitate rests for stand-level canopy phenology. Variability in near-infrared radiance (TM bands 4 and 5) of several distinct vegetation types was used to detect seasonal changes in a series of three Landsat Thematic Mapper (TM) images from the wet season co the dry season in Maraba, Brazil (eastern Amazon basin). Despite different atmospheric and instrumental conditions among the images, spectral changes were distinguishable. A phenological process (leaf aging, leaf drop, water stress) was determined from the spectral changes for each vegetation type. Changes in the spectral properties suggest that during the dry season, upland terra firme forest increased the race of leaf exchange and some riparian vegetation was deciduous. Terra firme forest that had been altered by penetration of fires from nearby pastures increased in leaf biomass over a 14-month period. This study shows that a time series of images can provide information on temporal changes in primary vegetation and guide field studies to investigate seasonal changes that may not be detectable from the ground.