Drought risk analysis using remote sensing and GIS in the Oshikoto region, Namibia.

Abstract

Drought is a recurrent climatic process that occurs with uneven temporal and spatial characteristics over broad areas and over an extended period of time. Therefore, detecting drought onsets and ends as well as assessing drought severity using satellite-derived information is essential. This should be especially the case in an arid country like Namibia where drought is part of Namibia’s climatology. It is believed that proper planning and research using near real-time data can curb the devastating environmental and socio-economic impacts of drought. Weather data used currently are often from a very sparse meteorological network, incomplete and/or not always available in good time to enable delineating accurately and timely, regional- and local-scale droughts. Consequently, the detection and monitoring efforts are hampered to provide timely and unbiased information to decision makers for accurate drought relief allocation and for land reform purposes. Furthermore, even though, data obtained from satellite-based sensors such as the Advanced Very High Resolution Radiometer (AVHRR) have been studied as a tool for drought monitoring for many years and provides an extensive temporal record for comparison, its coarse spatial resolution limits its effectiveness at detecting local scale variability where severe droughts might go undetected due to these data constraints. The objective of this study was to evaluate satellite-based and meteorological drought indices for the spatial and temporal detection, assessment and monitoring of drought condition to accurately delineate drought characteristics of drought prone areas. The study computed the Vegetation Condition Index (VCI) and Normalized Difference Vegetation Index (NDVI) from the 250m resolution NDVI data obtained from the Moderate Resolution Imaging Spectroradiometer (MODIS) sensor and one- and three-months Standardized Precipitation Index (SPI) data from rainfall stations in the study area. Detailed analyses of spatial and temporal drought dynamics during three seasons (2005/6 - wet, 2006/7 - normal and 2007/8 - dry) have been carried out through index maps generated in a Geographic Information Systems (GIS) environment from the mentioned data. Analysis and interpretation of these maps, which give different drought scenarios, reveal that remotely-sensed drought indices can accurately detect and map the local and regional drought spatial occurrence. Moreover, statistical analysis found strong correlations between the regional crop production data and the remotely-sensed data. However, the results showed that the local and regional drought occurrences detected were not reflected in national crop production data, confirming the suspicion that important local spatial variations are only detected if higher spatial resolution data are used. The study concluded that fine spatial resolution satellite data should be used to aid decision makers in monitoring and detecting drought which will also aid the allocation of millions of dollars in drought relief funds.