东北地区大豆干旱指标构建及时空特征分析

CONSTRUCTION AND SPATIO-TEMPORAL CHARACTERISTICS OF SOYBEAN DROUGHT INDEX IN NORTHEAST CHINA

关键词:  大豆  东北地区  水分盈亏指数  干旱等级指标  时空特征

Key words:  s oybean  Northeast China  water budget index  drought grade index  spatiotemporal characteristics

摘要:

目的 东北地区大豆干旱严重，构建基于灾变过程的东北地区大豆干旱等级指标，分析研究区域大豆干旱时空分布特征，对大豆干旱灾害进行实时监测预警、区域防旱减灾具有重要意义。方法 文章利用1961—2020年东北地区101个气象站点逐日气象资料、45个农业气象站的大豆发育期资料以及大豆历史灾情资料，通过历史灾情资料反演，基于水分盈亏指数并进行累计概率分布转换获得大豆干旱样本序列，采用K-S分布拟合检验拟合样本序列概率分布，通过t分布区间估计法确定指标等级阈值，并进行验证。在此基础上，分析大豆干旱时空变化特征。结果 （1）将大豆水分盈亏指数CD50,i=0.56作为大豆干旱过程初始判别值，以CD50,i≥0.56干旱过程持续日数（D）为主导因子，以CD50,i≥0.56逐日累加值（SDI）为辅助因子，构建了基于日尺度气象资料的东北地区大豆干旱等级指标，实现了干旱的动态化判识，指标验证结果显示与灾情记录等级完全符合的样本占全部样本的82.4%。（2）东北地区大豆每站平均总干旱频次最高的年份为1989年，达到1.51次/站，其次是2007年和1977年。干旱频次整体呈由西北向东南增加的趋势，高值区集中在辽宁省北部、吉林省中西部以及黑龙江省西北部和中部地区。（3）轻度干旱高频区集中在黑龙江西北部、吉林西部和辽宁北部地区；中度干旱频次高频区集中在整个东北地区的东部地区，干旱频次最高的站点为吉林蛟河；重度干旱频次高频区集中在整个东北地区的中部地区，发生频次最高的站点为黑龙江北安；特旱干旱频次整体呈由西北向东南减少趋势。结论 通过构建动态化大豆干旱指标，实现大豆干旱等级动态判识，可以实时有效地对东北地区大豆干旱进行监测预警。在研究时间段内，轻旱频次最高，特旱频次最低。在20世纪60—80年代干旱高发区范围逐渐缩小，90年代之后逐渐增大，21世纪10年代又有所减小。

Abstract:

Soybean drought is severe in Northeast China, therefore, it is of great significance to construct a soybean drought index based on disaster processes and study the spatiotemporal characteristics of soybean drought in Northeast China for real-time monitoring and early warning of soybean drought disasters and regional drought prevention and mitigation. The daily meteorological data of 101 meteorological stations, and the soybean growth period data and the soybean historical disaster records of 45 agricultural meteorological stations in Northeast China from 1961 to 2020 were used in this study. Through the inversion of historical disaster data, the soybean drought sample series were obtained based on water budget index and cumulative probability distribution transformation. The probability distribution of the sample series was fitted by using the K-S distribution fitting test, and the thresholds of index grade were determined and verified by the t-distribution interval estimation method. On this basis, the spatiotemporal characteristics of soybean drought were analyzed. The results were showed as follows. (1) The soybean water budget index CD50,i=0.56 was the threshold for discriminating the soybean drought process, and the number of drought process days (D) with CD50,i≥0.56 and the surplus deficit intensity (SDI) with CD50,i≥0.56 were calculated respectively. The soybean drought grade index constructed by these two elements could reflect the occurrence of soybean drought in Northeast China well. The verification results showed that 82.4% of the samples were in complete agreement with the disaster records. (2) The highest average total drought frequency per station for soybean in Northeast China was occurred in 1989 (1.51 times/station), followed by 2007 and 1977. The overall trend of drought frequency increased from northwest to southeast, and the high values were mainly occurred in northern Liaoning province, central and western Jilin province, and northwestern and central Heilongjiang province. (3) The high-frequency areas of mild drought were mainly distributed in northwestern Heilongjiang, western Jilin and northern Liaoning; the high-frequency area of moderate drought was mainly distributed in the eastern part of Northeast China, and the station with the highest frequency of moderate drought was Jiaohe in Jilin; the high-frequency area of severe drought was mainly distributed in the central part of Northeast China, and the station with the highest frequency of severe drought was Bei'an in Heilongjiang; and the overall frequency of extreme drought decreased from northwest to southeast. In summary, by constructing a dynamic soybean drought index, the dynamic identification of soybean drought grade can be realized. Hence, the real-time monitoring and early warning of soybean drought in Northeast China can be conducted effectively. During the study period, the frequency of mild drought was the highest, and the frequency of extreme drought was the lowest. The extent of drought-prone areas decreased gradually during the 1960s to 1980s, and increased gradually after the 1990s, and then decreased in the 2010s.