Estimation of evaporation from the water surface using the norm operator

Abstract

Due to the lack of precipitation in recent years, some regions of Turkey are in danger of drought. This situation increases the importance of planning water resources and makes it necessary to develop water budget calculations. One of the important steps in water budget calculations is the correct estimation of the amount of evaporation. For this reason, a different method has been developed for evaporation estimation and the applicability of this developed method has been tested with the meteorological parameters of Lake Eğirdir, one of most important freshwater resources of Turkey. Eğirdir Lake is located within the borders of Isparta province in the Mediterranean Region, Turkey. Firstly, evaporation estimation models were developed by artificial neural networks (ANN) method using 490 days of air temperature, water temperature, sunshine duration, and solar radiation parameters of Lake Eğirdir. After the meteorological parameters were transformed into a dimensionless form through normalization, the norm function was applied to these parameters as a part of the modeling process. The values obtained by the function are used as input parameters in the N-ANN method. In both cases, the pan evaporation values obtained with different network structures were compared and it was seen that the N-ANN models developed with the norm operator in general gave more appropriate results.