Computational uncertainty analysis of groundwater recharge in catchment

In this paper, a computational environinformatics (environmental informatics) operation for mapping the groundwater climatological recharge in regional sub-basin is presented. It is based on a soil–water balance (SWB) and spatial statistics integrated in a GIS environment. Mediterranean is a region with large demands for groundwater supplies. However, water catchment data are affected by large uncertainty, arising from sampling and modelling, which makes predicting groundwater recharge difficult. Geostatistic tools for GIS are able to incorporate imput data (coverages, shape files, raster, grids) in water data processing, allowing for modeling spatial patterns, prediction at unsampled locations, and assessment of the prediction uncertainty in a meaningful way that can provide a more suitable interpretation. An issue model of linear kriging, termed as lognormal kriging in form of a probability map (LKpm), is emphasized in this study because a soft description of the recharge in terms of probability is consistent to mitigate the uncertainty of the SWB estimates. The approach was applied to a test site in the Tammaro agricultural basin (South Italy) for the incorporation of change of support in water recharge downscaling modeling. So, the estimate of uncertainty at unsampled locations, via LKpm, was used to explain the probability of exceeding a value range of the water recharge samples' distribution. In this way, the probability of exceeding the median recharge (215 mm year− 1) is low in the southeastern portion (48%) of the basin area and high in the northwestern remaining portion (52%).