This project focuses on the Okavango Delta of northern Botswana, one of the world's largest freshwater wetlands (approximately 20,000 km2). The Okavango Delta annually experiences two significant disturbance regimes, flooding and fire. A remote sensing time series database spanning the period 1989 - 2003 period is available to this project and consists of over 96 medium resolution images (30 m) and 4 high resolution images (4 m). The primary objective of this research is to improve understanding the impacts of these disturbances on vegetation processes by 1) extracting the vegetation patterns over both space and time as mediated by flooding and fire disturbance regimes from remotely sensed data and 2) examining the ecosystem processes related to the previously mentioned disturbance regimes at various spatial and temporal scales using ecological modeling. Information derived from remotely sensed imagery will be utilized to map the vegetative response due to disturbances (flooding and fire) that occur in the Delta. Frequency of disturbances will be extracted using a 1-D Fourier analysis and correlated to known oscillations of precipitation. Ecotonal shifts due to observed disturbances will be tracked for the 1989-2003 period using remotely sensed data. Additionally, leaf area index (a measure of plant vigor) will be computed using vegetation indices derived from Landsat TM, ETM+, and ALI satellite data and correlated to field derived measurements of leaf area index. Land cover information, leaf area index, and climatic data will be fed into ecosystem models to track net ecosystem productivity through the 15-year sequence of imagery.
Ecosystem processes on the Earth occur at various scales, and the work presented here will address the incorporation of vegetation characterization via remotely sensed data into ecosystem models. These results will provide not only better descriptions of critical landscape processes across scales but moreover refined information on how the scale of observation and modeling impacts broader understanding of global climate change. The ability to link ecosystem information collected at the field level with remotely sensed data is an ongoing research topic and has many implications for future directions of science. Remote sensing provides the means to make measurements over extended areas, yet the mechanisms of scaling surface parameters from the stand level to the level of observation are not known. Most biogeochemical processes do not scale linearly from one scale to another and are complicated by spatial heterogeneity across a landscape. This project will address many of the issues facing ecological scaling and the continued successful utilization of remotely sensed data.
