| Summary: | Surface water is a critical resource in semiarid West-African regions that are frequently exposed to droughts. Natural and artificial wetlands are of high importance for different livelihoods, particularly during the dry season, from October/November until May. However, wetlands largely go unmonitored. In this work, remote sensing is used to monitor wetlands in semiarid Burkina Faso over large areal extents along a gradient of different rainfall and land use characteristics. Time series of data from the Moderate Resolution Imaging Spectrometer (MODIS) from 2000 to 2012 is used for near-infrared (NIR)-based water monitoring using a latitudinal threshold gradient approach. The occurrence of 21 new water bodies with a size larger than 0.5 km<sup>2</sup> over the 13-year analysis period results from a postclassification change detection. Yearly cumulative spatiotemporal analysis shows lower water extents in the drought seasons of 2000-2001, 2004-2005, and 2011-2012. Multiple wetlands indicate a positive trend toward a larger yearly maximum area, but a negative trend toward shorter flooding duration. Such a negative trend is observed particularly for natural wetlands. The temporal behavior of five selected case studies demonstrates that monthly negative anomalies of watercovered areas coincide with the occurrence of drought seasons. The successful application of remote sensing time series as a tool to monitor wetlands in semiarid regions is presented, and the potential of novel early warning indicators of drought from remote sensing is demonstrated.
|
|---|
