Identifying lake longer-term dynamics via fuzzy relations and non-linear Kalman filters

A novel simulation modelling technique that may be used to reconstruct or predict longer-term water level dynamics of lakes and reservoirs has been elaborated. On the grounds of the conservation law and a kind of self-regulation property of these lakes, by using Kalman filters for the unmeasurable part of hydrologic-cycle dynamical path and output observation equation in traditional and in fuzzy-relational a simulation modelling technique was developed that is believed to have wide applicability. It has been applied to reconstruct successfully the behaviour of Ohridean and Prespanean Lakes during various time periods, despite the available records are interrupted, uncertain and unreliable due to historical circumstances.