基于GEV干旱指数和DFA方法的苏北地区季节性干旱研究

为分析近年来江苏省苏北地区季节性干旱灾害特征,利用苏北五市1989~2013年的逐月降水量资料,建立月降水量的广义极值分布函数,通过分析广义极值干旱指数(GEVI),研究苏北五市近25a的季节性干旱分布和干旱等级的时空变化特点。在此基础上,利用去趋势波动分析方法(DFA)对苏北五市月平均降水量的周期性规律进行研究。结果表明,苏北五市的月降水量均服从广义极值Fréchet概率分布,且具有从西北(徐州)地区向东南沿海(盐城)地区递增的空间分布特征,整体上属于干旱灾害频发的地区,且以轻旱为主且季节性特征明显。苏北五市的月降水量存在幂律相关性,即降水量具有较强的正长程相关性,因此,未来苏北地区干旱的总体趋势将与过去特征正相关。

SEASONAL DROUGHT RESEARCH BASED ON GEV INDEX AND DFA METHOD IN THE NORTH OF JIANGSU PROVINCE

For the analysis of the characteristics of seasonal drought in the north of Jiangsu Province, a new drought index, which is based on the generalized extreme value (GEV) distribution, was introduced. The generalized extreme value distribution function is established using the mean monthly precipitation data from 1989 to 2013 of north Jiangsu Province. By computing the generalized extreme value index (GEVI) drought index, the seasonal drought distribution and the temporal and spatial variation characteristics are analyzed of north Jiangsu five cities in recent 25 years. From meteorology drought, the GEVI and standardized precipitation index (SPI) widely used in north region of Jiangsu are compared in this paper. The applicability of drought monitor between GEVI and SPI in north Jiangsu Province is discussed in greater detail. The results show that the two drought indexes are similar in computational principle. Based on the different distributional pattern of precipitation, they are given different drought grades in some drought events. It is more exhaustively and objectively to fit the precipitation distribution in generalized extreme value distribution. On this basis, using the detrended fluctuation analysis (DFA) method study the law of month average rainfall of North Jiangsu five cities. The results show that the five cities both obey Fréchet probability distribution, and from (Xuzhou City) northwest to the southeast coastal (Yancheng City) increases the space distribution features of region. The southeast part of North Jiangsu Province is the largest area of monthly precipitation, and the biggest change of the precipitation month. Five cities overall belong to the area of drought disaster and give priority to with the light drought and obvious seasonal characteristics. There is a powerlaw correlation in Northern Jiangsu Province five cities, which has a positive long range correlation. Therefore, the overall trend of the future will be associated with the past characteristics. The result indicate that GEVI and DFA method is an effective method for assessing drought conditions at different time scales and could be of great value for the dry season response.