论文部分内容阅读
收集了4种不同数据来源的再分析降水数据:欧洲中期天气预报中心大气再分析数据(ERA-intreim)、英国East Anglia大学气候研究中心数据(CRU TS3.21)、全球降水气候学项目数据(GPCP V2.2)、TRMM多卫星降水产品——TRMM 3B43(V7)和中国气象科学数据共享服务网提供的降水实测数据(CMD),用均值、偏差和偏差百分比及标准差等统计指标对再分析降水数据在洞庭湖流域的模拟精度进行评估。结果表明:(1)4种再分析数据年、季节降水量的年际变化与CMD总体相似,其中秋冬季节与CMD的一致性好于春夏季节。(2)ERA-interim数据对年均降水量的高估比例为20%以上,对季节平均降水量的高估比例为10.13%~33.65%,年、季降水量偏差百分比变化的年际变化较大;标准差的年内变化范围为12.40%~70.13%,数据集的离散程度最大。GPCP和CRU两种数据的年季降水量偏差在-6.19%~5.71%之间变化,均值与CMD数据最为接近;CRU数据年季降水量偏差百分比的年际变化和标准差的年内变化均较GPCP的大。TRMM数据对年季降水量的高估比例为3.67%~6.52%,降水量偏差百分比的年际变化稳定;标准差的年内变化范围为4.58%~14.03%,数据集的离散程度最小。(3)ERA-interim、CRU、GPCP和TRMM数据年降水量的偏差范围分别为1 000.21~-649.51、333.89~-643.82、292.55~-686.85、256.20~-561.27mm。再分析降水数据对降水的低估区域均出现在以南岳站为中心的衡阳盆地,而高估区域分布则有明显差异:ERA-interim的高估极值出现在地势较高的西部,CRU和GPCP的高估极值出现在南部山地,TRMM无明显高估区域。ERA-interim数据离散程度的空间分异明显,CRU和GPCP居中,TRMM分异较不明显。所有再分析数据在西部和南部山地区域的离散程度最大。
Reanalysis precipitation data were collected from four different sources: ERA-intreim at the European Center for Medium-Range Weather Forecasts, CRU TS3.21 at the University of East Anglia in the UK, and Global Precipitation Climatology Project Data ( (GPCP V2.2), TRMM multi-satellite precipitation products - TRMM 3B43 (V7) and precipitation data (CMD) provided by the China Meteorological Science Data Sharing Service Network. The data of average, deviation and deviation percentage and standard deviation Analyze the simulation accuracy of precipitation data in Dongting Lake Basin. The results show that: (1) The annual variation of annual and seasonal precipitation of four kinds of reanalysis data is similar to that of CMD. The consistency of autumn and winter seasons with that of CMD is better than that of spring and summer. (2) The ERA-interim data over-estimated the annual precipitation over 20% and the over-estimated seasonal precipitation from 10.13% to 33.65% The standard deviation ranged from 12.40% to 70.13% in the year, with the largest data set dispersion. The annual precipitation variation of GPCP and CRU data varied from -6.19% to 5.71%, with the mean close to that of CMD data. The annual variation of annual precipitation percentage deviation and annual variation of standard deviation of CRU data were all lower than those of GPCP Big. The over-estimated percentage of TRMM data for annual precipitation is 3.67% ~ 6.52%, and the annual variation of percentage of precipitation deviation is stable. The variance of standard deviation ranges from 4.58% to 14.03% and the data set has the smallest degree of dispersion. (3) The annual precipitation deviations range of ERA-interim, CRU, GPCP and TRMM are 1 000.21 ~ -649.51,333.89 ~ -643.82,292.55 ~ -686.85,256.20 ~ -561.27mm, respectively. Reanalysis precipitation data underestimated areas of precipitation are found in the Hengyang basin centering on the Nanyueyue station, while there is a significant difference in the overestimated regional distribution: the extreme value of ERA-interim overestimation appears in the western part of the area with higher CRU and GPCP The overvalued extreme appears in the southern hills, with no noticeable overestimation of TRMM. The discrepancy of spatial distribution of ERA-interim data is obvious, CRU and GPCP are the center, and TRMM differentiation is less obvious. All reanalysis data have the highest degree of dispersion in the western and southern mountain regions.