Quantifying the rainfall-water level fluctuation process in a geologically complex lake catchment.

Abstract

Simulating hydrologic processes in geologically complex environments is a difficult scientific task since it incorporates high level of uncertainty. Many studies have attempted to accurately quantify the rainfall-water level elevation relationship in freshwater bodies so as to predict flooding and drought events. For this purpose several types of models have been implemented including distributed, black box and conceptual models that often provide efficient results, depending on the availability of reliable data as well as on the level of understanding of the system. Nevertheless, in the particular effort, three different models have been used to describe the relationship between rainfall and water level elevation in Trichonis Lake during the period 1951-1997. A Transfer Function model, a Dynamic Linear Regression and a physically based model, consisting of the lake's water budget equation, its Digital Bathymetric Model and GIS algorithms. These models have been tested to assess their efficiency and applicability in a karstic environment and the aim of the study was to find the best modeling option for developing sustainable water management plans and establishing a flooding/drought warning system in the particular lake catchment. The results indicated that in areas with geologically complex conditions, simple, physically-based models operate better than mechanistic models which usually cannot describe adequately the complexity of the system.