Covariance Correction For Estimating Groundwater Level Using Deterministic Ensemble Kalman Filter

Abstract

The main problem in developing a groundwater model is to determine model parameters, particularly hydrogeologic coefficients, in a precise way. In this research, Deterministic Ensemble Kalman Filter (DEnKF) is described as a modern sequential method for data assimilation and a localization scheme within the framework of DEnKF is applied. Najafabad aquifer (in Iran) with area of 1150 km2, is modeled in the time window of Oct. 2000 to Sept. 2007 to obtain water table level data when its values of hydrogeologic coefficients calibrated and verified. DEnKF assimilated 45 observations of true run into the model with 2, 5, and 10 times of calibrated values of hydraulic conductivity and specific yield. This filter has been run both with and without use of localization. Results show easily-implemented localized DEnKF is favorably robust in groundwater flow modeling.