数据处理方法对中国区域平均降水序列精度的影响

利用地面观测站点资料研究大范围地区的降水长期变化规律,选取不同的指标会使分析结果出现显著差异。利用中国大陆区域内2139个国家站的逐日降水资料,比较不同数据处理方法得到全国和中国西部、中国东部地区的降水量、降水日数和降水强度的区域平均时间序列,探讨对其变化趋势估算的偏差。研究表明,1951—1957年估算的中国区域平均降水量原始值出现虚假的偏高,使趋势估算出现较大误差;1951—2016年中国西部地区平均降水量距平百分率时间序列的波动幅度显著偏大;区域平均降水量、降水日数和降水强度的距平和标准化距平序列较为可信。全国平均降水量、降水日数的原始值和距平值序列都基本反映了中国东部湿润地区降水的变化,降水量距平百分率的变化主要由西部干燥区域的降水变化构成,降水量标准化距平则可综合反映湿润和干燥地区的降水变化。

The influence of data processing on constructing regional average precipitation time series

The selection of different anomaly indicators will make a significant difference to study the long-term variation of precipitation in a wide range of areas using surface observation data. Daily precipitation data of 2139 national stations in the Chinese mainland were used to construct the regional average time series of precipitation, precipitation days and precipitation intensity, obtained from different indicators. These series were compared to explore the deviation of the long-term trend estimation in precipitation for western China, eastern China, and the entire China. The results shows that from 1951 to 1957, the regional average raw values of precipitation, precipitation days, and precipitation intensities in China had spuriously high biases due to the lack of data from stations in western China, which caused large deviations in the linear trend estimates from 1951 to 2016. In western China, the time series of the regional average precipitation anomaly percentages had excessive fluctuations. The changes in series of anomalies and normalized anomalies were reasonable in every region. The regional average time series of the original values and anomaly values of precipitation amount and days of the entire China basically reflected the precipitation changes in the humid regions in eastern China, while the precipitation anomaly percentages were mainly composed of the precipitation changes in the arid regions. Normalized anomaly could comprehensively reflect the precipitation changes in humid regions and arid regions.