Title Pan Evaporation Estimation Using Neural Networks and Neuro-Fuzzy Methods
Authors Özgür KIŞI, Selcan AFŞAR
Abstract Evaporation, as a major component of the hydrologic cycle, is important in water resources development and management. The application of artificial neural networks and fuzzy logic to evaporation modeling are limited in the literature. Therefore, the artificial neural network and neurofuzzy models for estimation of pan evaporation using climatic variables were investigated in the study. The daily mean air temperature, minimum temperature, maximum temperature and mean humidity and pan evaporation data of four weather stations in Kayseri, Kırşehir, Nevşehir and Yozgat were used. Various input combinations of weather data were used as inputs to the multi-layer perceptron (MLP), radial basis neural networks (RBNN), generalized regression neural networks (GRNN), neuro-fuzzy (ANFIS) and multiple linear regression (MLR) so as to evaluate degree of effect of each of these data on evaporation and to compare the models with each other. Root mean square error (RMSE), mean absolute error (MAE) and correlation coefficient (R) were used as comparing criteria. According to the RMSE and R statistics, the RBNN was found to be superior to the ANFIS, MLP, GRNN and MLR for 3 stations and MLR was found to be slightly better than the RBNN for 1 station. According to the MAE statistic, the RBNN performed better than the others for 2 stations and the ANFIS and MLR were found to be better than the RBNN for the other 2 stations.
Keywords Evaporation, artificial neural networks, fuzzy logic, estimation
Journal TABAD - Research Journal of Agricultural Sciences
Issue Issue 1
Page 45-51
Year 2010
Download Download this article.!!!
This article has been downloaded 2175 times.

© 2016 www.nobel.gen.tr. All rights reserved | Design by Nobel Science and Research Center