Maximum-likelihood adjoint-state finite-element estimation of groundwater parameters under steady- and nonsteady-state conditions

摘要: A method is presented for estimating the hydraulic parameters of groundwater flow models under steady- and nonsteady-state conditions. The estimation problem posed in framework maximum-likelihood theory by means a log-likelihood criterion that includes prior estimates parameters. To allow an incomplete knowledge covariances head parameter errors, these are expressed terms few unknown statistical may be estimated jointly with Computational efficiency achieved evaluating gradient adjoint-state finite-element scheme using combination conjugate-gradient algorithms, coupled Newton's determining step size to taken at each iteration. Model structure identification criteria developed time-series literature (all which utilize concept) shown useful selecting best way parametrize region when number alternative schemes parametrization given. paper also demonstrates potential utility proposed optimum design space-time measurement networks. case study dealing three-dimensional multiaquifer system briefly discussed.