Reference evapotranspiration (ETo) is a key factor in the hydrologic cycle and quantifying ETo for the future periods is essential for an efficient management of water resources. The objective of this research was to project the possible changes in ETo in the future periods in Iran. For this purpose, observed climate data in 40 stations across Iran were collected from 1984 to 2014 and ETo was calculated using these data and the FAO56–PM method. A multi-model ensemble of 27 CMIP6 models under two scenarios, SSP1-2.6 and SSP5-8.5, for the periods 2031–2060 and 2061–2090, was used. A linear regression (LR) model was applied for downscaling and bias correction of CMIP6 temperature using the observed temperature data. The output of the LR model was entered into an artificial neural network (ANN) model with an optimized structure, trained by the observed temperature and the calculated ETo, to project the future ETo based on the downscaled and bias corrected CMIP6 future temperature. Results showed that ETo will increase in both 2031–2060 and 2061–2090 periods under SSP1–2.6 and SSP5–8.5 scenarios, although the rate of increment was higher under SSP5-8.5 in all stations. Considering all stations, the average of ETo changes was ~0.9 mm/day for 2061-2090 under SSP5-8.5. It was found that the ETo rising rate was higher in arid environments, and therefore, shifting to drier conditions with higher water demands will be expectable in such climates in the future. This could cause severe water crises in these environments.

Keywords