| 基于统计降尺度模型的江淮流域极端气候的模拟与预估 |
| |
| 引用本文： | 陈威霖,江志红,黄强.基于统计降尺度模型的江淮流域极端气候的模拟与预估[J].南京气象学院学报,2012(5):578-590. |
| |
| 作者姓名： | 陈威霖 江志红 黄强 |
| |
| 作者单位： | [1]南京信息工程大学气象灾害省部共建教育部重点实验室,江苏南京210044 [2]南京信息工程大学大气科学学院,江苏南京210044 [3]浙江省海宁市气象局,浙江海宁314400 |
| |
| 基金项目： | 国家科技支撑计划项目（2009BAC51801）;国家自然科学基金资助项目（40875058）;江苏高校优势学科建设工程资助项目（PAPD）;江苏省研究生培养创新工程项目（CX09B_229Z） |
| |
| 摘 要： | 利用江淮流域29个代表站点1961--2000年逐日最高温度、最低温度和逐日降水资料，以及NCEP逐日大尺度环流场资料，引入基于多元线性回归与随机天气发生器相结合的统计降尺度模型SDSM（statistical downscalingmodel），通过对每个站点建模，确立SDSM参数，并将该模型应用于SRESA2排放情景下HadCM3和cGcM3模式，得到了江淮流域各代表台站21世纪的逐日最高、最低温度和降水序列以及热浪、霜冻、强降水等极端气候指数。结果表明，当前气候下，统计降尺度方法模拟的极端温度指数与观测值有很好的一致性，能有效纠正耦合模式的“冷偏差”，如SDSM对江淮平均的冬季最高、最低温度的模拟偏差较CGCM3模式分别减少3℃和4．5℃。对于极端降水则能显著纠正耦合模式模拟的降水强度偏低的问题，如CGCM3对江淮流域夏季降水强度的模拟偏差为-60．6％，但降尺度后SDSM—CGCM3的偏差仅为-6％，说明降尺度模型SDSM的确有“增加值”的作用。21世纪末期在未来SRESA2情景下，对于极端温度，无论Had．CM3还是CGCM3模式驱动统计模型，江淮流域所有代表台站，各个季节的最高、最低温度都显著增加，且以夏季最为显著，增幅在2—4℃；与之相应霜冻天数将大幅减少，热浪天数大幅增多，各站点冬季霜冻天数减少幅度为5—25d，夏季热浪天数增加幅度为4～14d；对于极端降水指数，在两个不同耦合模式HadCM3和CGCM3驱动下的变化尤其是变化幅度的一致性比温度差，但大部分站点各个季节极端强降水事件将增多，强度增强，SDSM—HadCM3和SDSM-CGCM3预估的夏季极端降水贡献率将分别增加26％和27％。
|
| 关 键 词： | 统计降尺度模型 江淮流域 极端气候指数 |
Projection and simulation of climate extremes over the Yangtze and Huaihe River Basins based on a Statistical Downscaling Model |
| |
| CHEN Wei-lin,JIANG Zhi-hong,HUANG Qiang.Projection and simulation of climate extremes over the Yangtze and Huaihe River Basins based on a Statistical Downscaling Model[J].Journal of Nanjing Institute of Meteorology,2012(5):578-590. |
| |
| Authors: | CHEN Wei-lin JIANG Zhi-hong HUANG Qiang |
| |
| Institution: | 1. Key Laboratory of Meteorological Disaster of Minisry of Education, NUIST, Nanjing 210044, China; 2. School of Atmospheric Sciences, NUIST, Nanjing 210044, China; 3. Haining Meteorological Bureau of Zhejiang Province, Haining 314400, China) |
| |
| Abstract: | Based on the observed daily maximum, minimum temperatures and daily precipitation data at the 29 meteorological stations which locate in the Yangtze and Huaihe River Basins as well as the daily NCEP reanalysis data, the statistical downscaling model （SDSM）, which is a combination of multiple linear regression and stochastic weather generator, has been used to calibrate the parameters of SDSM at each station. Subsequently the statistical downscaling model is applied to construct scenarios of the cli- mate extremes during the end of the 21st century by using predictor sets simulated by two GCMs （i. e. HadCM3 and CGCM3, hereafter referred to as SDSM-HadCM3 and SDSM-CGCM3, respectively ） forced by the special report on emission scenarios （SRES） A2. Future scenarios of the climate extremes such as heat waves and intensity precipitation in the 21st century at the 29 meteorological stations are constructed. The evaluation of simulated extreme indices of temperature and precipitation for the current climate shows that the downscaled temperature-related indices match the observations well, and SDSM can modify the systematic cold biases of the AOGCMs. For example, compared with the raw CGCM3, SDSM can reduce the ＂cold bias＂ of the winter maximum and minimum temperature by 3 ℃ and 4.5 ℃, respectively. For indices of precipitation extremes, most AOGCMs tend to underestimate the intensi- ty, but SDSM improves this significantly. The bias of simple daily intensity index （ISDI） in summer for SDSM is -6% ,while that of CGCM3 is as high as -60. 6%. Overall,compared with the AOGCMs, the downscaling model really has ＂added values＂. Scenario results using A2 emissions show that in all seasons there is a significantly increase of mean daily-maximum and minimum temperature at the 29 meteorological stations, accompanied by a decrease of the number of frost days and an increase of the heat wave duration. For instance, the frost days in winter are projected to a significantly decrease of 5-- 25 days at the 29 stations while the heat wave duration in summer increases by about 4--14 days. There is good agreement in the direction and magnitude of the projected changes of temperature-related indi- ces between the two driving AOGCMs. A warming environment will also give rise to changes in ex- treme precipitation events such as the maximum 5-day precipitation and the heavy rainfall proportion （R95t ）. Precipitation extremes are projected to increase at most of the 29 meteorological stations in a quite consistent manner between the two global scenarios. For example, the R95t in summer is projected to increases by 26% and 27% by SDSM-HadCM3 and SDSM-CGCM3 ,respectively. |
| |
| Keywords: | statistical downscaling model the Yangtze and Huaihe River Basins climate extremes indices |
| 本文献已被 维普 等数据库收录！ |
|
