H13A-1485
An approach based on localized ensemble Kalman filter to estimate the heterogeneous dispersivity field

Monday, 14 December 2015
Poster Hall (Moscone South)
Shaohua Cao, Jichun Wu and Xiankui Zeng, School of Earth Sciences and Engineering, Nanjing University, Nanjing, China, Nanjing, China
Abstract:
Ensemble Kalman Filter (EnKF) synthesizes observation data from multiple sources to estimate parameters approach to the real values. Dispersivity is crucial in groundwater numerical modeling of flow and non-reactive solute transport, as it deeply affects the predicted pollution plume when we use a numerical model to simulate the pollutant transport, however, the dispersivity used in the model are very difficult to obtain. In this study, transport data observed from a two-dimensional confined aquifer is assimilated via a localized Ensemble Kalman Filter system to calibrate the dispersivity field, and the result is satisfying. Some additional examples are illustrated to investigate the effect of different factors such as the number of realizations, initial assumed guesses, number and configuration of observations, and observation error on the efficiency of this method. The result indicate that the number of realizations greatly affects the efficiency of this method, excessive or insufficient is not good. A better estimation can be obtained if initial assumed guesses and observations is similar with the real field.