CAJ | 학술논문

采用自组织映射图（ SOM）神经网络法对流域气象站点进行聚类分析，结合中国气候分区将流域划分为7个子分区。基于流域及其周边203个基本气象站点1961－2013年的降水资料，采用标准化降水指数（ SPI）、线性倾向估计和M－K趋势检验法分析了流域多时间尺度干旱的组合特征。结果表明：年尺度上，淮河下游地区的干旱频率最高，达37．74％。除了流域的北部和西部地区以外，其他地区的干旱均呈加重趋势；季节尺度上，流域的东北部和北部地区春旱发生频率最高，淮河下游地区夏旱和冬旱的发生频率最高，流域的北部和中部地区秋旱发生频率最高。流域小范围区域的春旱和夏旱呈加剧趋势，流域南部的大范围区域秋旱呈加剧趋势，全流域的冬旱呈减弱趋势。
채용자조직영사도（ SOM）신경망락법대류역기상참점진행취류분석，결합중국기후분구장류역화분위7개자분구。기우류역급기주변203개기본기상참점1961－2013년적강수자료，채용표준화강수지수（ SPI）、선성경향고계화M－K추세검험법분석료류역다시간척도간한적조합특정。결과표명：년척도상，회하하유지구적간한빈솔최고，체37．74％。제료류역적북부화서부지구이외，기타지구적간한균정가중추세；계절척도상，류역적동북부화북부지구춘한발생빈솔최고，회하하유지구하한화동한적발생빈솔최고，류역적북부화중부지구추한발생빈솔최고。류역소범위구역적춘한화하한정가극추세，류역남부적대범위구역추한정가극추세，전류역적동한정감약추세。
Combining with the Chinese climate divisions , the paper divided the basin into seven sub-re-gions by clustering analysis for meteorological stations and using self-organizing map (SOM ) neural net-work method .It analyzed the combination characteristics of multiple time scales of drought in Huanghua-ihai River basin by using standardized precipitation index ( SPI) ,linear trend estimation and M-K trend test based on the daily precipitation data of 203 meteorological stations from 1961 to 2013 .The results in-dicated that on annual scale , the downstream of Huaihe River has the highest drought frequency which reaches 37 .74%.Drought of other regions of the basin presents intensified trend except north and west regions of the basin .On seasonal scale , northeast and north regions of the basin have the highest spring drought frequency .The downstream of Huaihe River basin have the highest drought frequency in summer and winter .The northern and central region have the highest drought frequency in autumn .The drought in small area of the basin shows a tendency to aggravate in spring and summer .The drought in most parts of the basin has intensified trend in autumn .The drought in the whole basin shows a weakensing trend in winter.