Assessing Water Availability in Mediterranean Regions Affected by Water Conflicts through MODIS Data Time Series Analysis

Water scarcity is a widespread problem in arid and semi-arid regions such as the western Mediterranean coastal areas. The irregularity of the precipitation generates frequent droughts that exacerbate the conflicts among agriculture, water supply and water demands for ecosystems maintenance. Besides, global climate models predict that climate change will cause Mediterranean arid and semi-arid regions to shift towards lower rainfall scenarios that may exacerbate water conflicts. The purpose of this study is to find a feasible methodology to assess current and monitor future water demands in order to better allocate limited water resources. The interdependency between a vegetation index (NDVI), land surface temperature (LST), precipitation (current and future), and surface water resources availability in two watersheds in southeastern Spain with serious difficulties in meeting water demands was investigated. MODIS (Moderate Resolution Imaging Spectroradiometer) NDVI and LST products (as proxy of drought), precipitation maps (generated from climate station records) and reservoir storage gauging information were used to compute times series anomalies from 2001 to 2014 and generate regression images and spatial regression models. The temporal relationship between reservoir storage and time series of satellite images allowed the detection of different and contrasting water management practices in the two watersheds. In addition, a comparison of current precipitation rates and future precipitation conditions obtained from global climate models suggests high precipitation reductions, especially in areas that have the potential to contribute significantly to groundwater storage and surface runoff, and are thus critical to reservoir storage. Finally, spatial regression models minimized spatial autocorrelation effects, and their results suggested the great potential of our methodology combining NDVI and LST time series to predict future scenarios of water scarcity.