Evaporation Estimation from Climatic Factors

This study assessed the ability of two models, Local Linear Regression (LLR) and Artificial Neural Network (ANN) to estimate monthly potential evaporation from Pantagar, US Nagar (India) which falls under sub-humid and subtropical climatic zone. Observations of relative humidity, solar radiation, temperature, wind speed and evaporation have been used to train and test the developed models. A comparison was made between the estimates provided by the LLR model and ANN model. An extensive review of different methods applied for estimation of evapotranspiration in the sub-continent has also been presented in the study. Results have shown that the models were able to learn well about the events they were trained to recognize. For ANN model, the correlation coefficient for the training period was 0.9311 and for the testing period ended up to 0.9236. The value of Root Mean Square Error (RMSE) for the training period was 1.070 and for the testing period it reached 0.9863. In case of LLR model, the correlation coefficient for the training period was 0.9746 and for the testing period it was 0.9273. The value of RMSE for the training mode was 0.6121 and for the testing period it was 1.5301. The results provided a lot of confidence to carry out further research for different applications in more diverse range of climatic zones to cover the entire region.