Abstract
In this paper, we propose a new methodology for parameter identification of groundwater systems. Observations of either groundwater heads, pumping rates or hydraulic conductivities estimated from permeability tests, can be used to identify the distributed parameter systrem in a unconfined aquifer. Based on the extended Kalman filter algorithm, a finite element analysis is used to predict the groundwater heads and the pumping rates at the next time step, both of which are treated as the state variables in addition to the hydraulic conductivities. Finally, the proposed methodology is verified with the case using simulated data in a model of groundwater system, and is found to be reasonably accurate.
