我国农业干旱灾害时空格局分析

ANALYSIS OF SPATIAL AND TEMPORAL PATTERNS OF AGRICULTURAL DROUGHT DISASTER IN CHINA

关键词:  农业干旱  旱灾面积  旱灾占比  空间特征  时间特征  聚类分析  旱灾分区

Key words:  agricultural drought  drought area  relative drought percentage  spatio-temporal pattern  clustering analysis  drought zoning

摘要:

目的 分析国家尺度上农业干旱灾害时空分布规律。方法 文章收集了1998—2017年的农业干旱受灾面积统计数据，采用多年移动平均、变异系数和聚类分析等方法对国家尺度上农业干旱空间分布及动态变化进行研究。结果 表明国家尺度上，农业干旱灾害在空间和时间分布上均表现出一定的规律性，在空间分布上，干旱受灾面积较大的省份主要分布在华北和东北，这些区域也是我国农业生产的主产区。旱灾占比较大的省份则主要集中在西北和东北区域，表明该区域旱灾发生较为严重；在时间动态变化上，基于旱灾面积和旱灾占比的评价结果表明，20年来我国旱灾总体呈现出波动下降的趋势。另外，国家尺度上旱灾占比的变异系数仅为0.49，远小于省级尺度上的变异系数均值。这主要由于国家尺度上同一年中不同省份旱灾面积的增减相抵，从而降低了国家层面旱灾占比的波动性。根据农业旱灾的发生特点，以地理区域作为分类单元，以旱灾面积、旱灾占比和旱灾占比变异系数三类参数进行Z值标准化和聚类分析，最终将我国划分为3个干旱区域。旱灾Ⅰ区包括西北，该区域常年受旱灾影响较为严重。旱灾Ⅱ区包括东北和华北，该区域旱灾发生较为严重，但旱灾波动性较大。旱灾Ⅲ区包括西南、华中、华南和华东，该区域总体旱灾发生较轻，旱灾年际波动幅度小，即常年旱灾发生较轻。结论 基于历史统计数据，可以阐明区域尺度上农业旱灾的时空变化规律，可以为农业生产者及管理者提供参考。

Abstract:

In order to investigate the spatio-temporal pattern of drought disasters in agriculture at the Chinese national scale, authors assembled the historical statistical data on drought affected areas in 31 provinces or Autonomous Administration Regions from 1998 to 2017. The evolution of spatial distribution and temporal dynamic of drought events during these two decades, was analysed by methods of multi-year moving average, coefficient of variation as well as clustering analysis. The results indicated that some regularity in both spatial and temporal aspects of drought disaster pattern. In terms of absolute acreage affected by drought events, provinces in North China and Northeast China with, large agricultural productions were mostly impacted. Provinces with higher percentage on agricultural areas affected by drought were mainly concentrated in Northwest China and Northeast China, where the relative severity of drought were revealed. As to the temporal dynamic change, the evaluation results based on drought affected areas and relative drought percentage displayed a drifting down trend during the past 20 years on the national level. In addition, the coefficient of variation for relative drought percentage at the national level amounted 0.49, which was far less than the mean value of the coefficients of variation for 31 provinces. This was resulted by an offset of fluctuation for drought affected areas in different provinces in the same year, leading to a reduction of relative drought percentage on the national scale. Z score normalization and cluster analysis were applied at the geographic level to 3 variables: absolute drought affected area, relative drought percentage and coefficient of variation for the relative drought percentage. Finally, the country was classified into 3 drought affected zones according to the characteristics of drought disasters. The Drought-Region Ⅰ consisted of provinces in Northwest China, where drought disaster occurred steadily and constantly. North China and Northeast China belonged to the Drought-Region Ⅱ, where drought disaster occurred stiffly in general but fluctuates from year to year. The Drought-Region Ⅲ included Southwest China, Centre China, South China and East China, where drought disaster occurred moderately in general with a minor fluctuation. In summary, the spatial and temporal dynamics of agricultural drought disaster can be illustrated based on historical statistical data. This kind of analysis can be considered for supporting agricultural management and decision making.