Geostatistical Modeling

The geostatistical simulation will be employed for generating alternative realistic 2D or 3D numerical representations (realizations) of the spatial distribution of key hydrogeological and/or hydrogeochemical parameters to simulate flow and transport processes in application case studies. Stratified Monte Carlo simulation will be performed via conditional Latin hypercube (LH) simulation, which yields a more representative distribution of simulated attributes. Hence, fewer realizations will be required to capture the range of values for each simulation location, eventually alleviating the computational burden related to the repeated evaluations of the complex salinization models of hydrogeological flow and transport. In a second level, and in those cases afforded by data availability, geostatistical inverse modeling will be employed to render parameter realizations more realistic; this will be achieved by considering the information brought by hydrogeochemical measurements through flow and transport models linking the input parameters with those measurements. The set of alternative and realistic, parameter realizations constitute a model of uncertainty regarding the real but unobserved distribution of parameters in space and will be used for risk-conscious decision-making in a comprehensive water resources management context.