In order to tackle climate change and improve the spatial-temporal collocation of water resources, the localization management was performed for the parameters of the Hargreaves model with reference evapotranspiration (ET0), which was calculated with the Penman-Monteith model using the daily meteorological data from 88 weather stations in Huang-Huai-Hai Plain during 1961-2010. The long series of daily temperature data from 2010 to 2100 in representative concentration paths RCP4.5 and RCP8.5 scenarios were simulated with the statistical downscaling model (SDSM4.2) and the second generation of Canadian Earth System Model (CanESM2). The long series of daily temperature data were used to calculate ET0 in the 21st century with the calibrated Hargreaves formula. The temporal change trend of ET0 during 1961-2100 was analyzed by the Mann-Kendall trend test and relative anomaly method, and the spatial distribution of ET0 during 2010-2100 was analyzed by the Kriging method in this study. Results indicated that the standard error of the corresponding parameter in the calibrated Hargreaves formula was very small, and the fluctuation range of correlation index between the calibrated Hargreaves model and the Penman-Monteith model was 0.65-0.85, with a mean value of 0.80, which indicated the high accuracy of the calibrated Hargreaves model. The determination coefficient of the maximum temperature and the minimum temperature between the observed and simulated values in the calibration period (1961-1990) and the verification period (1991-2005) was greater than 0.95, and the root mean square error was very small, indicating the high accuracy of the temperature simulated by the SDSM4.2. ET0 in the Huang-Huai-Hai Plain in 21st century under the 2 climate scenarios all had a trend of increasing in the future. The average values of ET0 under the RCP4.5 scenario in 2020 s, 2050 s and 2080 s were 940.71, 949.49 and 955.39 mm, respectively, 5.14%, 6.12% and 6.78% greater than the benchmark value (an average of 894.76 mm from 1961 to 2010). The average values of ET0 under the RCP8.5 scenario were 940.12 mm in 2020s, 966.72 mm in 2050 s and 986.97 mm in 2080s, respectively, 5.07%, 8.04% and 10.3% greater than the benchmark. The increase under RCP8.5 was slighter than that under RCP4.5, which was maybe caused by higher temperature and solar radiation under RCP8.5. From an overall perspective, the change trend of relative anomaly of mean temperature and ET0 was similar: most years before 2055 had negative anomaly, while after 2055 had positive anomaly. However, the value of positive and negative anomaly in ET0 had bigger fluctuations than that in mean temperature. The relative anomaly of mean temperature and ET0 increased gradually as time went on, and the increase of them under RCP8.5 was higher than that under RCP4.5. ET0 in the Huang-Huai-Hai Plain under RCP4.5 decreased gradually in a spoon-like curve from the border of Hebei, Shandong and Henan to the surrounding areas, and the minimum was observed in Tangshan and Laoting in Hebei, Dongtai in Jiangsu, and Zhumadian in Henan. The spatial distribution of ET0 under RCP8.5 in 2020 s was very similar to that in 2050 s, ET0 decreased gradually from the border of Hebei, Shandong and Henan to the surrounding areas, and the minimum was observed in Tangshan and Laoting in Hebei, Dongtai in Jiangsu, and Zhumadian in Henan. But there was a great difference in the spatial distribution of ET0 in 2080 s, especially for the spatial distribution of the greatest values. The maximum values were distributed in these areas in the Huang-Huai-Hai Plain: Huimin in Shandong, Xinxiang in Henan, Bengbu in Anhui and Xuyi in Jiangsu. The increase of ET0 would further aggravate the shortage of water resources in Huang-Huai-Hai Plain. This research provides valuable information for irrigation schedule. © 2016, Editorial Department of the Transactions of the Chinese Society of Agricultural Engineering. All right reserved.
