Artificial Neural Network for Estimating of Monthly Pan Evaporation in Qattinah Meteostation Based on only Temperature Data

Abstract

Evaporation losses from the water of Lake Qattinah form a large part of the losses of storage in this dam, which are greatly affected by climate change in the region, Besides that, the precise knowledge of these losses is very useful for the management of the water source which is important to irrigating of large area of agricultural land, therefore, this study investigates the potential of artificial neural networks (ANNs) in estimating of monthly evaporation in Qattinah meteostation using temperature data only. The models based on monthly air temperature data only as inputs, and the monthly measured pan evaporation data, used as outputs of the networks. This data was split into three datasets for training and validation and testing in the ratios 70:15:15 respectively. The network was trained and verified using a back-propagation algorithm with different learning methods, number of processing elements in the hidden layer(s), and the number of hidden layers. models were able to well estimating of monthly evaporation values in study meteostation with high reliability. Results shown good ability of (3-14-1) ANN to predict of monthly pan evaporation with correlation coefficient 96.41%, and the value of root mean square error 9.88 mm/month for the validation datasets. This study Recommend using artificial neural networks models in the forecasting of evaporation time series in Lake Qattinah, which have great importance in completing of the missing data.