Close Window

Effects of seasonality and timing of image acquisition on remote observation of regenerating vegetation in the seasonally-dry tropical forests in Rondonia, Brazil

Stephen C Hagen, University of New Hampshire, steve.hagen@unh.edu (Presenting)
William A Salas, Applied Geosolutions, LLC, wsalas@agsemail.com
Mark J Ducey, University of New Hampshire, mjducey@cisunix.unh.edu
Stephen Frolking, University of New Hampshire, steve.frolking@unh.edu
Bobby H Braswell, University of New Hampshire, rob.braswell@unh.edu

This study examines how the timing of the remote observations of the land surface affects the resultant land cover characterizations. Specifically, we document the spectral properties of the land surface in a dynamic area of the Brazilian tropical rainforest at the beginning and end of the dry season. Over the course of a season, the biophysical properties of the vegetation and, correspondingly, the spectral signature of the vegetation can change for many reasons. These reasons include human directed land use change, as well as growth, mortality, and stress of vegetation. Studies conducted on the ground, and more recently with high temporal resolution (e.g. MODIS) data, have documented phenological changes in tropical vegetation from the wet to dry season. The analysis is separated into two parts. In the first part, using atmospherically corrected Landsat reflectance data acquired over Rondonia, Brazil in May and August 2003, we analyze the spectral dynamics of several land cover features, with a focus on regenerating vegetation. In the second part, we extend this analysis by exploring the consequences that seasonal differences in spectral signatures have on thematic land cover mapping. With a better understanding of the seasonal dynamics of the tropical land surface, specifically regenerating vegetation, we can improve our ability to monitor changes in this ecologically important region.

Science Theme:  LC (Land Use and Land Cover Change)

Presentation Type:  Poster

Abstract ID: 63

Close Window