Climate change and its impact on reference evapotranspiration at Rasht City, Iran

There are various factors of uncertainty regarding the impact of climate change on reference evapotranspiration (ETo). The accuracy of the results is strictly related to these factors and ignoring each of them reduces the precision of the results, thereby affecting their applications. In this study, the uncertainty related to two methods of calculating ETo, the Hargreaves-Samani (HGS) and Artificial Neural Network (ANN); the climate change models: Atmosphere-Ocean General Circulation Model (AOGCM); and downscaling method under the climate change scenario (A2) for the period 2010 to 2039 was evaluated. Meteorological data for the Rasht station located in the northern part of Iran collected for 1961-1990 was used to evaluate the climatic data and calculate ETo. Since there were no lysimeter installed in the area, the FAO Penman-Monteith (PM, 1998) method was adopted as the reference ETo method and the ETo of the period (1961-1990) produced by HGS and ANN methods was evaluated using performance functions including mean absolute error (MAE) and regression coefficient (R2). Next, the Hadley Centre Coupled Model, version 3 (HadCM3) climatic model and the Canadian Global Climate Model, version 3 (CGCM3) climatic model and the Statistical Downscaling Model (SDSM) were applied to generate maximum and minimum temperatures for use in simulating ETo using the HGS and ANN methods for 2010 to 2039. Results obtained showed average temperature increases of 0.95 oC with the HadCM3 model and average temperature increases of 1.13 oC with the CGCM3 model, relative to observed temperatures for 2010 to 2039. Accordingly, the predictions showed average increases of ETo ranging from 0.48 to 0.6 (mm/day) for the period of 2010 to 2039. The study also revealed that uncertainty with the AOGCMs is more than the ETo models applied in this study