A geostatistical approach to the inverse problem in groundwater modeling (steady state) and one-dimensional simulations

摘要: The problem of estimating Hydrogeologic parameters, in particular, permeability, from input-output measurements is reexamined a geostatistical framework. field the unknown parameters represented as ‘random field’ and estimation procedure consists two main steps. First, structure parameter identified, i.e., mathematical representations variogram trend are selected their established by using all available information, including hydraulic head permeability. Second, linear theory applied to provide minimum variance unbiased point estimates hydrogeologic (‘kriging’). Structure identification achieved iteratively three substeps: selection, maximum likelihood estimation, model validation diagnostic checking. methodology was extensively tested through simulations on simple one-dimensional case. results remarkably stable well behaved. estimated smooth, while small-scale variability statistically described. As quality improves, reproduces more features original field. also shown be rather insensitive deviations assumptions about