REFERENCE EVAPOTRANSPIRATION PREDICTION USING VARIOUS HEURISTIC AND STATISTICAL APPROACHES
Loading...
Date
2024-02-01
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
G. B. Pant University of Agriculture & Technology, Pantnagar-263145
Abstract
The accurate estimation of reference evapotranspiration (ET0) has paramount importance and is crucial in irrigation planning and scheduling, watershed hydrology studies, drought forecasting and monitoring, water resource management and planning, etc. In the present study from Guwahati station (Assam), the standard FAO-56 based Penman-Monteith (PM) method was utilized to estimate daily ET0 which was considered an output to assess the models. The different soft computing and statistical techniques such as ANN, wavelet based ANN (WANN), ANFIS and MNLR models were used for the prediction of daily reference evapotranspiration in the study area. Gamma test (GT) was used to determine and select the best input combination of climatic parameters (i.e., mean temperature, mean relative humidity, wind speed and solar radiation) having the least gamma and V-ratio values. The qualitative and quantitative performance evaluation criteria were done by visual inspection and using statistical and hydrological indices such as coefficient of determination (R2), root mean square error (RMSE), coefficient of efficiency (CE) and Willmott index (WI) respectively, which were used for assessing the prediction accuracy of the developed models. Based on the comparison of the models, the results revealed that the WANN-11 model performed the best as compared to ANN-8, ANFIS-02 (trap-2) and MNLR models for prediction of reference evapotranspiration of the study area.The sensitivity analysis was also carried out for the best developed model to detect the most sensitive input parameter based on the performance of the model. It was found that mean relative humidity was the most sensitive input parameterfor daily reference evapotranspiration prediction of the study area. (
Description
Keywords
Citation
Theses of M. Tech