基于1971—2015年陕西省无缺测情况、分布较均匀的58个气象站点的逐日降水数据,采用最大1日、3日、5日和7日的降水量代表极端降水,应用地区线性矩法研究区域极端降水的时空特征。结果表明：（1） 陕西省可划分为6个水文气象一致区,其中广义极值分布（GEV）在各一致区的模拟效果最好,各一致区最优频率估计值与同频率的实测值较为吻合。（2） 地区分析法计算的极端降水频率估计值较单站分析法具有更好的稳健性和准确性,尤其是在计算较长时段的极端降水情况下更显著。（3） 重现期为2 a一遇时,陕南的地区增长因子大于陕北;重现期达到5 a一遇时则相反,且随着重现期的增长,地区增长因子及其在陕南、陕北的差异也在增大。（4） 在100 a和50 a一遇重现期下,极端降水在陕西省南部较大,东部居中,中部的咸阳至商洛地区、西部的延安西北和榆林西部较小;极端降水的分布特征与陕西省独特的地理特征有关,尤其是东西向分布的秦岭阻挡了水汽向北输送,造成了极端降水的南北差异。

Extreme precipitation can cause severe disasters in arid and semi-arid regions, such as in the Shaanxi Province of Northern China. To investigate the spatiotemporal characteristics of extreme precipitation in the Shaanxi Province, this study adopted the daily precipitation data of 58 meteorological stations from 1971 to 2015 with no missing observations and relatively uniform distribution and used the maximum precipitation on the 1st, 3rd, 5th, and 7th day to represent the extreme precipitation. The regional L-moments method was further applied to study the temporal and spatial characteristics of the extreme regional precipitation, which involves the screening and processing of precipitation data, identification of homogeneous regions, goodness-of-fit test, quantile estimation for each region, comparison between at-site and regional estimation, deriving regional growth factors, and mapping of the spatial patterns of extreme precipitation. The results and conclusions of the study were: (1) Shaanxi Province can be divided into six hydrometeorological homogeneous regions, among which GEV distribution in each homogeneous region has the best simulation effect, and the estimated optimal quantiles of each homogeneous region are in good agreement with the measured value of the same frequency. (2) The estimated extreme precipitation quantiles calculated by the regional analysis method have better robustness and accuracy compared with the single-station analysis method, especially more significant in calculating extreme precipitation over a long period. (3) When the return period is once every 2 years, the regional growth factor of southern Shaanxi is greater than that of northern Shaanxi; when the return period is once every 5 years, the opposite is true, and with the increase of the return period, the regional growth factor and the difference between southern Shaanxi and northern Shaanxi also increase. (4) In the 100-and 50-year return periods, extreme precipitation is large in the south, centered in the east, the Xianyang-Shangluo region in the middle, the northwest corner of Yan’an in the west, and the west of Yulin are small. The distribution characteristics of extreme precipitation are related to the unique geographical characteristics of Shaanxi Province, especially the east-west Qinling Mountains, which block the water vapor transmission to the north, causing differences in extreme precipitation between the north and south.