洪水灾害是当今世界最危险的自然灾害之一，不可预测的洪水给人们带来了不同程度的生命和财产损失。通过对洪水灾害风险的研究，可以为防灾减灾工作提供技术支持，减少洪水带来的影响，同时保护生存、生态环境。
本文针对CMIP6数据进行降尺度研究，分析未来湟水流域气候变化情况。基于2010、 2015 和 2020年的土地利用数据预测2030年、2040年和2050年湟水流域土地利用分布情况。在此基础上，结合 RCP4.5 和 RCP8.5 两种气候背景下的气候模式模拟和分析了气候变化和土地利用变化对流域径流的共同影响。基于以上气候变化预估结果和径流预测结果，构建洪水人口风险评估模型。评估了研究区在RCP4.5和RCP8.5情景下2026-2055年洪水灾害人口风险的演化。本研究取得的主要结果如下：
（1）根据对NEX-GDDP 未来气候数据的分析，ACCESS-CM2、ACCESS-ESM1-5、CNRM-CM6-1、EC-Earth3、MIROC6、MIROC-ES2L、MPI-ESM1-2-HR、MPI-ESM1-2-LR、MRI-ESM2-0和NESM3这十种气候模型，在RCP8.5情景下的降雨变化波动较RCP4.5情景下的降雨变化波动更为剧烈，但RCP4.5和RCP8.5两种情景都预示未来降雨量预计将有所增加，但幅度并不明显。RCP4.5和RCP8.5两种情景表明年平均气温均呈现明显的上升趋势，未来30年将快速上升1.5°C。
（2）根据Logistic-CA-Markov模型对湟水流域未来30年的土地利用变化进行了预测：2030年至2050年建设用地明显增长，面积占比由2030年的5.59%上涨到2050年的10.24%，草地明显下降，面积占比由2030年的50.82%下降到2050年的42.26%。湟水流域整个区域内土地利用覆盖变化不会发生巨大的结构性变化。
（3）在十个气候模型数据以及其多模式集合平均(MME)下，在ACCESS-CM2、CNRM-CM6-1、MIROC-ES2L、MRI-ESM2-0、NESM3、MIROC6和MME气候模式下的不同重现期下未来极端径流在增加，在ACCESS-ESM1-5、EC-Earth3、MPI-ESM1-2-HR和MPI-ESM1-2-LR个气候模式的不同重现期下未来极端径流在减少。总体来看，2026年至2055年湟水流域洪水风险呈现上升趋势，并且RCP4.5情景下的洪水灾害发生的风险高于RCP8.5情景下洪水灾害发生的风险，出现这种变化的原因可能是在RCP8.5情景下，气温上升幅度较RCP4.5情景下更高，且RCP8.5的情景下，降雨量上升幅度明显小于RCP4.5情景。
（4）湟水流域洪水人口风险较高的区域主要集中在湟水流域中部人口密集的湟水谷地区域，湟水流域四周洪水人口风险较低，但在东南部流域出水口区域洪水人口风险较高；2025-2055年间，不论是RCP4.5还是RCP8.5情景下，湟水流域的总体洪水人口风险是呈现增加的趋势，主要集中在中部较为平缓且沿河流而分布的居民区，但是RCP8.5情景下洪水人口风险的变化相较于RCP4.5情景下的变化更为剧烈。

Flood disaster is one of the most dangerous natural disasters in the world today. Unpredictable floods have brought varying degrees of loss of life and property to people. Through research on flood disaster risks, we can provide technical support for disaster prevention and reduction work, reduce the impact of floods, and at the same time protect survival and ecological environment.
This paper conducts a downscaling study on CMIP6 data to analyze future climate change in the Huangshui River Basin. Based on the land use data in 2010, 2015 and 2020, the land use distribution in the Huangshui River Basin is predicted in 2030, 2040 and 2050. On this basis, the joint impacts of climate change and land use change on watershed runoff were simulated and analyzed by combining climate models under two climate backgrounds, RCP4.5 and RCP8.5. Based on the above climate change prediction results and runoff prediction results, a flood population risk assessment model was constructed. The evolution of the population risk of flood disasters in the study area from 2026 to 2055 under the RCP4.5 and RCP8.5 scenarios was evaluated. The main results obtained in this study are as follows:
(1) According to the analysis of NEX-GDDP future climate data, ACCESS-CM2, ACCESS-ESM1-5, CNRM-CM6-1, EC-Earth3, MIROC6, MIROC-ES2L, MPI-ESM1-2-HR, MPI- For the ten climate models ESM1-2-LR, MRI-ESM2-0 and NESM3, the fluctuations in rainfall changes under the RCP8.5 scenario are more severe than those under the RCP4.5 scenario, but RCP4.5 and RCP8.5 Both scenarios indicate that rainfall is expected to increase in the future, but the magnitude is not significant. Both scenarios, RCP4.5 and RCP8.5, indicate that the annual average temperature shows a clear upward trend and will rise rapidly by 1.5°C in the next 30 years.
(2) Based on the Logistic-CA-Markov model, the land use change in the Huangshui River Basin over the next 30 years was predicted: from 2030 to 2050, the construction land will increase significantly, with the area proportion rising from 5.59% in 2030 to 10.24% in 2050, while the grassland will decrease significantly, with the area proportion falling from 50.82% in 2030 to 42.26% in 2050. There will be no significant structural changes in land use and cover changes in the entire Huangshui River Basin.
(3) Different recurrences under ACCESS-CM2, CNRM-CM6-1, MIROC-ES2L, MRI-ESM2-0, NESM3, MIROC6 and MME climate models under ten climate model data and their multi-model ensemble mean (MME) The future extreme runoff will increase under this period, and will decrease under different return periods of ACCESS-ESM1-5, EC-Earth3, MPI-ESM1-2-HR and MPI-ESM1-2-LR climate models. Overall, the flood risk in the Huangshui Basin shows an upward trend from 2026 to 2055, and the risk of flood disasters under the RCP4.5 scenario is higher than the risk of flood disasters under the RCP8.5 scenario. The reasons for this change may be It is that under the RCP8.5 scenario, the temperature increase is higher than that under the RCP4.5 scenario, and under the RCP8.5 scenario, the rainfall increase is significantly smaller than the RCP4.5 scenario.
(4) The areas with higher flood population risk in the Huangshui River Basin are mainly concentrated in the densely populated Huangshui Valley area in the central part of the Huangshui River Basin. The flood population risk around the Huangshui River Basin is low, but the flood population risk is higher in the outlet area of the southeastern basin. High; between 2025 and 2055, under both the RCP4.5 and RCP8.5 scenarios, the overall flood population risk in the Huangshui Basin shows an increasing trend, mainly concentrated in the relatively gentle residential areas in the middle and distributed along the river. However, The changes in flood population risk under the RCP8.5 scenario are more dramatic than those under the RCP4.5 scenario.