在全球典型的气候变化预警区——我国西南地区，以地面台站数据（CMA）为基准，评估地面网格化降水数据（GRID_pre）、气温数据（GRID_tmp）和卫星遥感降水产品（TRMM）、气温产品（MODIS）在研究区的时空探测能力，实现针对性数据融合，获取时空精度较高的空间化气象数据。基于该数据构建大尺度模式数据与区域网格气象数据间的统计降尺度模型，模拟研究区在RCP4.5和RCP8.5情景下的未来气候变化。进一步将未来气候数据驱动SWAT分布式水文模型，计算研究区典型流域不同情景下非点源氮磷输出负荷。在此基础上系统性分析研究区内历史时期、未来时期气温、降水因子的时空变化规律，以及氮磷流失的时空分布特征，有针对性识别其海拔依赖性及敏感性、突变性地区。主要得到以下结论： （1） 多源气象数据在西南地区的探测能力随海拔变化趋势 空间化气温数据在西南地区的模拟能力总体优于空间化数据降水数据；在不同时间尺度上，空间数据模拟能力由强到弱为：月＞年＞日。GRID_pre较TRMM更为稳定，但在月尺度上将降水的波动趋势刻画的更为平缓，且在高海拔区的探测能力明显不足，平均偏差（BIAS）达到69.5%；TRMM呈现出高值偏高，低值偏低的探测误差特征，且对高海拔站点的日值探测方面具有优势（BIAS≤8.7%）；GRID_tmp的探测能力优于MODIS，二者BIAS的空间分布差异不明显但前者的R值较高；MODIS对夏季气温的月尺度探测能力优于GRID_tmp，但MODIS数据的HKS评分随时间的变化波动频繁，总体数值较为零散，且空报现象较多。因此，有必要在研究区6000m以上的站点缺测较多且分布密度较低的地区进行TRMM与GRID_pre融合，得到优化的降水空间化数据。通过日数据训练得到的多元回归最佳融合方程为y=sin(0.8147*x1)-0.1893*x2+0.9936(R2=0.9706 , RMSE=1.133)。 （2） 西南地区历史气温、降水变化的海拔依赖性特征 研究区气温变化的海拔依赖性较强，而降水变化的海拔依赖性不显著。针对气温而言，高海拔区升温趋势最为显著；随着海拔升高，气温变化的突变时间提前。针对降水而言，研究区多年降水总体呈微弱上升趋势，低海拔区降雨变化平缓波动下降；低海拔区降水降低主要受到秋季降水大幅降低的影响；降水趋势变化不随海拔单调变化；周期分析结果表明，研究区降水总体周期为28年，但低海拔区降水具有12年的小周期。 （3） 西南地区未来气温、降水变化及其海拔依赖性特征 ASD统计尺度模型在西南地区对未来气温的模拟效果总体优于降水，从月际评估指标来看，统计降尺度模型在西南地区模拟效果较优，能够满足水文模型输入需求。总体而言，研究区未来升温趋势的海拔依赖性较强，而降水趋势的海拔依赖性相对较弱。RCP8.5情景下的研究区未来平均气温比历史时期增加3.65℃，比RCP4.5情景下增加1.09℃，且升温趋势较RCP4.5更为强烈，趋势较历史时期高出24.34%，而RCP4.5情景下则较历史时期降低46.29%；在不同情景下研究区未来年平均降水变化并不显著（分别增加0.013mm和0.048mm）。海拔＜2000m的地区未来升温趋势变化最剧烈，在海拔＞2000m的地区，升温趋势随着海拔升高逐渐放缓，未来升温的敏感区呈现进一步向低海拔地区转移的态势；随海拔升高，降水减少的趋势放缓，增加的趋势递增。在海拔3500~4500m范围内，降水变化受海拔调节程度较低。 （4） 流域氮磷流失对气候敏感度的海拔依赖性 在RCP4.5和RCP8.5两种未来气候情景下，澜沧江流域总氮流失分别减少了45.13%和47.53%，总磷流失则分别减少了33.13%和36.89%，研究区总体氮磷流失与气温呈反向变动，与降水呈正向变动，总氮流失对气候变化的响应更剧烈。在未来情景下，高海拔分区氮磷污染的减缓趋势较显著，中等海拔分区次之，而低海拔区氮磷流失只有微弱的降低趋势，由高海拔向低海拔分区流失负荷降低的递减率在5%~10%之间研究区内气候变化的海拔依赖性充分作用于非点源氮磷的流失，在不同的海拔区间进行非点源污染流失风险管控措施制定过程中，需充分考虑气候变化的作用。总氮流失对气候变化敏感度具有更强的海拔依赖特性，其突变条带区海拔为3000m左右。总磷流失对气温变化的敏感度较低，对降水变化的敏感度较高，其总体的海拔依赖特性不显著。总磷流失对降水变化敏感度的突变条带区海拔在4000m左右。

In the global climate change warning zone—the southwestern region of China, ground gridded precipitation data (GRID_pre), temperature data (GRID_tmp), and satellite remote sensing precipitation products (TRMM) are evaluated based on ground station data (CMA). The space-time detection capability of temperature products (MODIS) in the study area enables targeted data fusion to obtain spatialized meteorological data with high spatial and temporal precision. Based on this data, a statistical downscaling model between large-scale model data and regional grid meteorological data is constructed to simulate the future climate change of the study area under RCP4.5 and RCP8.5 scenarios. The future climate data drives the SWAT distributed hydrological model to calculate the non-point source nitrogen and phosphorus output load in different scenarios of the typical watershed in the study area. On this basis, systematically analyze the temporal and spatial variation of temperature and precipitation factors in the historical period and future period in the study area, as well as the temporal and spatial distribution characteristics of nitrogen and phosphorus loss, and identify the altitude dependence and sensitivity and mutation areas. Mainly got the following conclusions: (1) The detection ability of multi-source meteorological data in Southwest China varies with altitude The simulation ability of spatialized temperature data in the southwest region is generally better than the spatialized data precipitation data; on different time scales, the spatial data simulation capability is from strong to weak: month>year>day. GRID_pre is more stable than TRMM, but the trend of precipitation fluctuation is more gradual on the monthly scale, and the detection ability in high altitude area is obviously insufficient, the average deviation (BIAS) reaches 69.5%; TRMM shows high value , low-value detection error characteristics, and has advantages in daily value detection of high-altitude sites (BIAS≤8.7%); GRID_tmp detection capability is better than MODIS, the spatial distribution of BIAS is not obvious, but the former R The value is higher; MODIS is better than GRID_tmp for the monthly scale detection of summer temperature, but the HKS score of MODIS data fluctuates frequently with time, the overall value is scattered, and there are more empty reports. Therefore, it is necessary to integrate TRMM and GRID_pre in the area with more than 6000m in the study area and the distribution density is low, and obtain optimized spatial data of precipitation. The multivariate regression optimal fusion equation obtained by daily data training is y=sin(0.8147*x1)-0.1893*x2+0.9936 (R2=0.9706, RMSE=1.133). (2) Altitude-dependent characteristics of historical temperature and precipitation changes in Southwest China The altitude dependence of temperature changes in the study area is strong, while the altitude dependence of precipitation changes is not significant. For the temperature, the warming trend of the high altitude area is the most significant; as the altitude rises, the sudden change of the temperature change is advanced. For precipitation, the precipitation in the study area showed a slight upward trend, and the rainfall in the low-altitude area fluctuated smoothly. The decrease of precipitation in the low-altitude area was mainly affected by the sharp decrease of precipitation in autumn. The change of precipitation trend did not change monotonously with altitude; the cycle analysis showed that The total precipitation period in the study area is 28 years, but the precipitation in low-altitude areas has a small period of 12 years. (3) Future temperature and precipitation changes and their altitude dependence in Southwest China The simulation effect of ASD statistical scale model on future temperature in Southwest China is better than precipitation. From the monthly evaluation index, the statistical downscaling model has better simulation effect in Southwest China and can meet the input requirements of hydrological model. In general, the altitude dependence of the future warming trend of the study area is strong, while the altitude dependence of the precipitation trend is relatively weak. The average future temperature of the study area under the RCP8.5 scenario increased by 3.65 °C compared with the historical period, which was 1.09 °C higher than that of the RCP4.5 scenario, and the warming trend was stronger than RCP4.5, and the trend was 24.34% higher than the historical period, while RCP4 Under the .5 scenario, it decreased by 46.29% compared with the historical period; in the different scenarios, the average annual precipitation change in the study area was not significant (increased by 0.013mm and 0.048mm, respectively). The temperature rise trend of the region with altitude <2000m is the most intense region. In the region with altitude >2000m, the warming trend will gradually slow down with the elevation, and the sensitive area in the future will shift to the lower altitude. With the elevation, The trend of reduced precipitation has slowed down and the trend of increase has increased. In the range of 3500~4500m above sea level, the change of precipitation is less regulated by altitude. (4) Altitude dependence of climate and phosphorus sensitivity in watershed nitrogen and phosphorus loss Under the two future climate scenarios of RCP4.5 and RCP8.5, total nitrogen loss in the Lancang River Basin decreased by 45.13% and 47.53%, respectively, and total phosphorus loss decreased by 33.13% and 36.89%, respectively. Total nitrogen and phosphorus loss and temperature in the study area. In the reverse direction, the precipitation is positively changing, and the total nitrogen loss is more responsive to climate change. In the future scenario, the mitigation trend of nitrogen and phosphorus pollution in the high-altitude zone is more significant, followed by the medium-altitude zone, while the nitrogen-phosphorus loss in the low-altitude zone has only a slight decrease trend, and the decline rate of the loss from the high-altitude to the low-altitude zone is Between 5% and 10%, the altitude dependence of climate change in the study area is sufficient for the loss of non-point source nitrogen and phosphorus. In the process of formulating control measures for non-point source pollution loss at different altitudes, climate change should be fully considered. The role. Total nitrogen loss has a stronger altitude-dependent characteristic for climate change sensitivity, and its abrupt strip zone has an elevation of about 3000m. Total phosphorus loss is less sensitive to temperature changes, more sensitive to precipitation changes, and its overall altitude-dependent characteristics are not significant. The abrupt zone of the total phosphorus loss sensitivity to precipitation changes is about 4000m above sea level.