陈鹏翔,江志红,彭冬梅. 2017. 基于BP-CCA统计降尺度的中亚春季降水的多模式集合模拟与预估[J]. 气象学报, ():-, doi:10.11676/qxxb2017.017
基于BP-CCA统计降尺度的中亚春季降水的多模式集合模拟与预估
Projection of spring Precipitation in Central Asia Using Multimodel Statistical Downscaling Based on Canonical Correlation Analysis
投稿时间:2016-08-10  最后修改时间:2016-09-22
DOI:10.11676/qxxb2017.017
中文关键词:  春季降水  BP-CCA  统计降尺度  中亚
英文关键词:spring  precipitation, BP-CCA, statistical  downscaling, Central  Asia
基金项目:国家自然科学基金
作者单位E-mail
陈鹏翔 新疆维吾尔自治区气候中心 cpx1860@163.com 
江志红 南京信息工程大学 zhjiang@nuist.edu.cn 
彭冬梅 新疆兴农网信息中心 pengdongmei126@126.com 
摘要点击次数: 468
全文下载次数: 242
中文摘要:
      利用中亚地区30个观测台站逐月降水资料及同期ERA40再分析资料,结合8个CMIP5全球气候模式模拟与未来预估大尺度环流场,使用基于变形典型相关分析的统计降尺度方法(BP-CCA)建立降尺度模型,评估多个气候模式对当前气候下中亚地区春季降水的降尺度模拟能力,并对春季降水进行降尺度集合未来预估。结果表明建立的降尺度模型能够很好地模拟出交叉检验期内春季降水的时间变化和空间结构:降尺度春季降水与相应观测序列的平均时间相关系数为0.35,最高为0.62,平均空间相关系数为0.87。气候模式对中亚春季降水的模拟能力通过降尺度方法得到了显著提高:8个模式降尺度后模拟的降水气候平均态相对误差绝对值降至0.2%-8%,相比降尺度前减小了10%-60%,模拟的降水量场与相应观测场的空间相关均超过0.77;对比降尺度前多模式集合结果,多模式降尺度集合模拟的相对误差绝对值由64%减小至4%,空间相关系数由0.47增加到0.81,标准化均方根误差降至0.59,且多模式降尺度集合结果优于大部分单个模式降尺度结果。多模式降尺度集合预估结果表明,在RCP4.5排放情景下,21世纪前期(2016-2035年)、中期(2046-2065年)和末期(2081-2100年)的全区平均降水变化率分别为-5.3%、3.0%和17.4%。21世纪前期中亚大部分地区降水呈减少趋势,降水呈增加趋势的站点主要分布在南部。21世纪中期整体降水变化率由减少变为增加趋势,21世纪末期中亚大部分台站降水增加较为明显。21世纪初期和末期可信度高的台站均主要位于中亚西部地区。
英文摘要:
      By using 30 meteorological station precipitation data in Central Asia, the European Centre for Medium-Range Weather Forecasts 40-year reanalysis dataset (ERA-40) and eight CMIP5 (Coupled Model Intercomparison Project Phase 5) climate models, statistical downscaling models are constructed based on BP-CCA (the combination of empirical orthogonal function and canonical correlation analysis) . Evaluation of the multiple models to the spring precipitation downscaling simulation ability and project the future changes of precipitation. The results show that the average time coefficient of Downscaling spring precipitation and corresponding observation is 0.35, the highest of 0.62, The spatial correlations are all improved to more than 0.87. The absolute values of domain-averaged precipitation relative errors of most models are reduced to 0.2%-8% after statistical downscaling. As a result of statistical downscaling multimodel ensemble (SDMME), the relative error is improved from 64% to 4%, the spatial correlation increases significantly from 0.47 to 0.81, and the RMSE reduced to 0.59. These demonstrate that the simulation skill of SDMME is relatively better than that of multimodel ensemble (MME) and most individual models’ downscaling. The projections of SDMME reveal that under the RCP4.5 (Representative Concentration Pathway 4.5) scenario, the projected domain-averaged precipitation changes for the early (2016-2035), middle (2046-2065) and end (2081-2100) of the twenty-first century is -5.3%, 3.0% and 17.4%, respectively. For the early of the century, the precipitation of most areas shows a decreasing trend and the increasing trends in the south region of Central Asia. The stations with significant increasing trends are in both the middle and end periods, with larger magnitude for the latter. With the passage of time, the SDMME forecast gradually enhance the credibility of the results.
查看全文   查看/发表评论  下载PDF阅读器
分享按钮