Decision-Making Tool for Groundwater Level Spatial Distribution and Risk Assessment Using Geostatistics in R

The present research work applies ordinary kriging using an R programming script to map the free surface of an unconfined alluvial aquifer, calculating, at the same time, the standard deviation of the estimations. Moreover, indicator kriging is also applied to calculate the probability of the aquifer level to lie below a certain limit that could cause a significant impact on groundwater resources availability. Kriging efficiency depends on the estimation of the optimal spatial dependence of the measurements. Therefore, classical variogram functions and the Matern model are applied to determine the spatial correlation of the measurements. The power-law variogram application provided the most accurate results. Maps associated with the hydraulic head spatial variability and prediction uncertainty, as well as a probability map were produced. The source R script is available with this work for public use. The concept, tools, and results of this work can be a useful framework for stakeholders and local authorities to design and decide optimal water resources management and governance.