Nonlinear Kalman filter in simulation modelling of lake water dynamics

Results of a thorough study on modelling the hydrological cycle and the level water dynamics of a natural lake basin for the purpose of developing and identifying a simulation model of medium and longer-term dynamics of water level are presented. It has been used to model and simulate water level dynamics of Prespanean and Ohridean Lakes. The actual simulation modelling technique makes use of available recorded data and observed phenomena, as a-priori knowledge, involving the whole cycle from precipitation to evaporation and evapotranspiration in the basin. The resulting simulation model is represented by a set of discrete-time state equation, derived on the grounds of the conceptual model of interconnected multiple tanks for the hydrological cycle, and of discrete-time observation (output) equation. An applicability and feasibility discussion is given for the nonlinear case, although dynamic structures of Kalman filter for both linear and nonlinear modelling case have been derived