The influence of spatial variability of rainfall to hydrological model fitting - Application to Spercheios river basin

The subject of spatial variability of rainfall has always been a field of study. Its representation and impact in the adaptation of hydrological models has been and is being investigated by scientists. In this study homogenous and isotropic vector fields of rainfall for the region of Sperheios basin were originally created by using the program ARC-GIS 9.3 for the geostatistical method Co-Kriging. This was done by introducing data points taken from 14 gauges from the study area during a 50-year period. The data emerged after the primary sample processing, through expansion and completion techniques. The goal was the most adequate depiction of the spatial variability of rainfall and the extraction of the average monthly rainfall time series for the above period. The hydrological simulation was done using the program HYDROGEIOS supported by MapWindow. The model represents both the surface and the underground system. The rainfall data was obtained by applying the classical method of Thiessen polygons and the above method of Co-Kriging. A semi-automatic calibration of the model was performed through global optimization for both data sets and two scenarios were developed. The calibration was done in steps each of which required many tests. All the parameters were optimized in selected groups in each test. In this way, firstly were detected which parameters the model is sensitive to and secondly the range of feasible region for which each parameter has a physical interpretation. The aim was to identify that group of parameters which lead to a better simulation in reference to the observed hydrograph. The goodness-of-fit criteria of the model in both calibration and validation were the efficiency, the average bias and the error of zero discharges. Finally, the two scenarios were compared for both the calibration and the validation period. Particular emphasis was placed on the comparison of the parameters in relation to their physical interpretation and on the average annual water balance in reference to its realistic representation.