Two gridded precipitation datasets from the Climatic Research Unit (CRU) and the Global Precipitation Climatology Centre (GPCC) are evaluated using the station-observed precipitation over China for the period of 1901-2013 at the seasonal,annual,and interdecadal time scales,respectively.The main results are:Firstly,for the period of 1961-2013,both CRU and GPCC products are able to describe the temporal-spatial variation of precipitation in China,and both show a better performance in eastern China than in westem China and in summer than in winter.On the annual and seasonal time scales,comprehensive comparisons are also performed through analyzing the percentages of absolute deviation,the root-mean-square errors,and the correlation coefficients between two products with the station observations.It is found that the CRU data shows large biases in the Tibetan Plateau and some other areas of large mountains,and the annual precipitation trend derived from CRU data is also smaller than that from station observations over areas such as the Altun Shan Mai,the Loess Plateau,southeastern China,and the lower reaches of the Yangtze River basin,while GPCC data is relatively more consistent with the station observations in both precipitation amount and trend.Secondly,for the period of 1901-1961,in-situ observed precipitation at 65 stations are used to compare with CRU and GPCC products.It is found that CRU shows large deficiencies to the west of 110°E in the arid and simi-arid regions in China,while GPCC agrees well with observations.Finally,the two gridded datasets and station observations for the period of 1961-2013 are used to compute the standardized precipitation index (SPI),which is then used to describe the degree of aridity in China.The results show that the GPCC is closer to the observations than CRU in terms of the dry and wet events variations in China.For instance,GPCC can capture the severe drought in the summer of 1997,but CRU cannot.In summary,this study suggests that GPCC is a better choices compared to CRU for studying the long-term precipitation trend in China.%利用1901～2013年中国大陆地区的气象台站实测降水资料,对东英吉利(East Anglia)大学气候研究中心(Climatic Research Unit,CRU)和全球降水气候中心(Global Precipitation Climatology Centre,GPCC)的降水资料分别从季节、年际和年代际尺度上进行了评估.结果表明:1961～2013年CRU与GPCC降水资料均能较准确地描述中国大陆地区的降水特征,且在东部较西部地区、夏季较冬季与站点实测降水情况更为一致.将中国大陆划分为不同区域并在其季节、年际和年代际时间尺度上通过比较降水偏差绝对值的百分比、均方根误差和相关系数等统计量后发现:CRU在青藏高原和其它较大的山脉附近与站点实测降水的差别较大,且年均降水趋势在西北一带的阿尔金山脉、黄土高原、东南地区和长江下游地区,比实测降水的年均趋势小、甚至出现趋势相反的情况.此外,CRU降水的年代际变化趋势也偏小.而GPCC数据不论是降水量还是降水趋势都更接近实际情况.在1901～1961年,通过与65个长期气象观测站点的降水时间序列比较发现,CRU在110°E以西地区与站点观测的降水资料间的差别较大,而GPCC与站点观测资料的吻合较好.最后,利用1961～2013年两套降水资料和站点实测资料分别计算了标准化降水指数(SPI),简单分析了中国大陆地区的干旱变化,发现GPCC对旱涝的时空变化特征的描述比CRU更接近站点实际观测;并且CRU也没有反映出1997年夏季中国地区出现的严重干旱情况,而GPCC较为准确地反映出了这一干旱事件特征.因此,本文的研究结果认为,就中国大陆地区长时期降水资料而言,GPCC的适用性优于CRU.
展开▼
