为探究未来气候变化对流域生态需水量的影响，保障河流生态需水量，针对好溪流域进行生态需水量计算 及预测。基于好溪流域气象数据及下垫面条件建立流域生态需水模型，并根据 GF1-WFV 遥感影像数据订正后的 地表反射率和作物种植结构提升模型模拟精度。选择 CanESM2气候模式下的 RCP2.6、RCP4.5 和 RCP8.5 这 3 种 排放情景，建立气候变化背景下流域生态需水预测方法，计算现状年并预测未来年份的生态需水量及生态需水保 障程度。结果表明，基于光学遥感影像进行数据订正后，模型模拟精度有所提升，率定期的模型精度 R2从 0.80 提 升为 0.85，验证期的 R2从 0.75 提升至 0.78。应用提升精度后的模型进行生态需水预测，在 RCP2.6、RCP4.5 和 RCP8.5 情景下，2025—2100 年的年均生态需水分别增加了 0.27 亿、0.21 亿和 0.29 亿 m3，其中 RCP8.5 情景下生态 需水保障程度最高，RCP4.5 情景下生态需水保障程度最低。

Climate change has caused the redistribution of water resources by changing the global hydrological cycle. The main factors are changes in temperature and rainfall, and the variation of these climate factors directly led to the change in surface runoff. At present, people pay too much attention to economic development and ignore the protection of the ecological environment, which has led to the contradiction between the supply and demand of water resources in the watershed. At the same time, under the environment of global warming, the ecological water demand of the watershed has been more or less affected, and the ecological balance of the watershed has been seriously threatened. The previous research results were drawn up and Tennant, Penman-Monteith formula, and other methods were used to calculate the current annual ecological water demand of Haoxi Watershed. The results showed that the current ecological water demand of Haoxi Watershed was 157 million m3 from September to February of the next year, and 248 million m3 from March to August. Using CanESM2 meteorological prediction data under RCP2.6, RCP4.5, and RCP8.5 scenarios, the location of Haoxi Watershed was extracted and downscaled to obtain the climate prediction data. The grid data of 26 prediction factors were obtained by inverse distance weight interpolation NCEP reanalysis data. Multivariate stepwise linear regression was applied to establish the statistical relationship between the three prediction variables and the prediction factors, and finally, the prediction factors were determined. The daily precipitation, daily maximum temperature, and daily minimum temperature from 2025 to 2100 were selected as predictive variables to analyze climate change and the changing trend of ecological water demand of the watershed under each scenario in the next 75 years. The CanESM2 model data was standardized and verified in SDSM model to generate the long-term series data of the watershed in the future. The changes in precipitation, daily maximum mild temperature and daily minimum temperature in future years were also estimated. The obtained meteorological forecast data was put into the high-precision ecological water demand model established above to generate the long-term forecast value of ecological water demand to 2100. Based on the predicted value of ecological water demand, the degree of ecological water demand guarantee in the future years was calculated. Due to the lack of accuracy of land use data sources applied in the current research, the simulation accuracy of ecological water demand in Haoxi Watershed was low. Therefore, the remote sensing image data of GF1-WFV and ENVI software were used for remote sensing interpretation to obtain the surface albedo and high-precision land use data in the study area, and the interpreted data were put into the existing SWAT model to compare the improvement effect of data accuracy before and after substitution. In terms of the comparison of the simulation results of ecological water demand, the results of the original model were as follows: Periodic R2 is 0.75. The high-precision land use data and surface albedo obtained from the interpretation of multivariate remote sensing data were substituted into the simulation results obtained by the model. The periodic R2 value was 0.85 and the validation R2 is 0.8, and the verification period value was 0.78. Therefore, the simulation accuracy of the model was significantly improved. In the forecast changes of meteorological data in the next 75 years, the maximum temperature and minimum temperature under RCP2.6, RCP4.5, and RCP8.5 all had different amplitude increases. From the general trend, the temperature increase was the largest under RCP8.5 and the lowest under RCP2.6. Compared with the temperature, the rainfall showed no obvious change trend, and the overall fluctuation was large. In terms of the forecast of ecological water requirement, the ecological water requirement of the three different scenarios will increase to varying degrees, but the two scenarios of RCP2.6 and RCP4.5 will gradually stabilize after 2065, while the RCP8.5 will continue to increase. Therefore, the variation trend of ecological water demand guarantee degree can be obtained. RCP2.6 tends to be stable, RCP4.5 shows a downward trend, and RCP8.5 shows an upward trend. In terms of the comparison of the simulation results of ecological water demand, the simulation results of the original model were as follows: the rate periodic R2 is 0.8, and the verification period R2 is 0.75. The high-precision land use data and surface albedo obtained from the interpretation of multivariate remote sensing data were substituted into the simulation results obtained by the model. The periodic R2 value was 0.85 and the validation R2value was 0.78. Therefore, the simulation accuracy of the model was significantly improved. In the forecast changes of meteorological data in the next 75 years, the maximum temperature and minimum temperature under RCP2.6, RCP4.5 and RCP8.5 all had different amplitude increases. From the general trend, the temperature increase was the largest under RCP8.5 and the lowest under RCP2.6. Compared with the temperature, the rainfall showed no obvious change trend, and the overall fluctuation was large. In terms of the forecast of ecological water requirement, the ecological water requirement of the three different scenarios will increase to varying degrees, but the two scenarios of RCP2.6 and RCP4.5 will gradually stabilize after 2065, while the RCP8.5 will continue to increase. Therefore, the variation trend of ecological water demand guarantee degree can be obtained. RCP2.6 tends to be stable, RCP4.5 shows a downward trend, and RCP8.5 shows an upward trend. The results showed that the established ecological water demand model could be applied to the study of Haoxi Watershed, and the simulation accuracy improvement method based on the remote sensing simulation interpretation method could improve the accuracy of the model. In addition, the predicted ecological water demand and the degree of ecological water demand guarantee in the next 75 years and clarified the law of its change, which had important reference significance for guaranteeing the basic ecological functions of the river watershed.