Application of Artificial Neural Networks to Project Reference Evapotranspiration Under Climate Change Scenarios

Evapotranspiration is sensitive to climate change. The main objective of this study was examine the response reference evapotranspiration (ET0) under various change scenarios using artificial neural networks and Canadian Earth System Model Second Generation (CanESM2). Hargreaves method used calculate ET0 for western, central, eastern parts Prince Edward Island their two input parameters: daily maximum temperature (Tmax), minimum (Tmin). Tmax Tmin were downscaled with help statistical downscaling model (SDSM) three future periods 2020s (2011-2040), 2050s (2041-2070), 2080s (2071-2100) representative concentration pathways (RCP’s) including RCP 2.6, P4.5, 8.5. Temporally, there major changes in Tmax, Tmin, RCP8.5. temporal variations all RCPs matched reports literature other similar locations. For RCP8.5, it ranged from 1.63 (2020s) 2.29 mm/day (2080s). As a next step, one-dimensional convolutional network (1D-CNN), long-short term memory (LSTM), multilayer perceptron (MLP) estimating ET0. High coefficient correlation (r > 0.95) values both calibration validation showed potential estimation. results will decision makers water resource managers quantification availability island optimize use resources on sustainable basis.