河南农业科学 ›› 2020, Vol. 49 ›› Issue (11): 172-180.DOI: 10.15933/j.cnki.1004-3268.2020.11.023

• 农业信息与工程·农产品加工 • 上一篇    

基于格网和模糊聚类的河南省冬小麦气象干旱风险区划与分析

黎世民1,2,张红利2,王来刚2,郑国清2,郭燕2,高建华1   

  1. (1.河南大学 环境与规划学院,河南开封 475004;2.河南省农业科学院 农业经济与信息研究所,河南郑州 450002)
  • 收稿日期:2020-08-10 出版日期:2020-11-15 发布日期:2020-11-15
  • 通讯作者: 郭燕(1983),女,副研究员,河南驻马店人,博士,主要从事农业遥感与信息技术研究。E-mail:guoyan8372@163.com 高建华(1964),男,河南临颍人,教授,主要从事区域发展与规划研究。E-mail:jhgao@henu.edu.cn
  • 作者简介:黎世民(1978-),男,河南许昌人,副研究员,在读博士研究生,研究方向:农业遥感、区域发展与规划。E-mail:57085397@qq.com
  • 基金资助:
    国家自然科学基金项目(41771142,41601213);河南省农业科学院科技创新创意项目(2021CX)

Winter Wheat Agrometeorological Drought Zoning and Analysis in Henan Province Based on Grid and Fuzzy Clustering Algorithm

LI Shimin1,2,ZHANG Hongli2,WANG Laigang2,ZHENG Guoqing2,GUO Yan2,GAO Jianhua1   

  1. (1.The College of Environment and Planning of Henan University,Kaifeng 475004,China;2.Institute of Agricultural Economics and Information,Henan Academy of Agricultural Sciences,Zhengzhou 450002,China)
  • Received:2020-08-10 Published:2020-11-15 Online:2020-11-15

摘要: 利用河南省110个县(市)地面观测点的降水距平百分率数据,基于格网尺度,采用模糊聚类方法进行冬小麦气象干旱区划和风险概率评估。空间相关分析指标——局域指标(LISA),Moran’I散点图(Moran scatter plots)和LISA聚集图(LISA cluster map)表明,济源、濮阳和商丘等17个县(市)落入“高-高”第一象限;信阳、驻马店和开封等地区的12个县(市)落入“低-低”第三象限,这些区域降水量较大,存在较强的空间正相关关系;信阳、驻马店等地区的6个县(市),落入“高-低”第四象限;焦作、杞县和信阳3个县落入“低-高”第二象限,这些区域存在较强的空间负相关关系。总体来看,县域降水空间变化的相关性较小。模糊分类结果表明,当模糊性能指数(FPI)、归一化分类熵(NCE)二者最小,区域划分为5类时,效果最佳。利用单因素方差分析(OneWay ANONA)进行差异显著性检验,不同的区划类别之间在0.05水平差异显著。格网尺度下的制图结果显示,干旱可能发生的区域具有空间分异规律,信阳和南阳的西南部属于湿润区域,发生旱灾的风险概率较低;三门峡以及濮阳、安阳和开封部分区域属于严重干旱区域,发生旱灾风险的概率较高;周口和漯河的大部分区域属于中旱区域。高风险区主要分布在三门峡、南阳等西部山区和许昌、郑州的边区,低风险区主要是在东北部和信阳等水稻种植区域。这与干旱的空间分布有一致性,但是部分区域存在差异,如濮阳的部分区域干旱程度高但是风险概率值却较低。

关键词: 干旱;格网;降水距平百分率;模糊聚类;区划;河南省, 冬小麦

Abstract: 110 ground observations of the precipitation distance percentage in Henan Province were applied to evaluate the agrometeorological drought zoning and risk probability analysis for winter wheat with fuzzy clustering algorithm based on grid. Indicators of spatial association indices-Moran scatter plots and LISA cluster map were adopted to conduct to the spatial exploratory data analysis.The results indicated that 17 counties(cities) of Jiyuan,Puyang and Shangqiu and so on fell into the first quadrant with “high high”;12 counties(cities) of Xinyang,Zhumadian and Kaifeng and other regions fell into the third quadrant with “ low-low”,where these regions existed strong positive correlations with more precipitation;6 counties(cities) of Xinyang,Zhumadian and Jiaozuo,Qixian,Huaiyang fell into the fourth quadrant with“high-low” and the second quadrant “low-high” respectively, where these regions existed strong negative correlations.It could be concluded that the correlations of spatial variation of precipitation among counties was small.Then,fuzzy k means was used to divide the precipitation distance percentage data into five classes,while the fuzzy performance index(FPI) and normalized entropy(NCE) were the smallest.There existed significance between different zones at the 0.05 level by One-Way ANONA test.According to the distribution maps,the spatial and differential laws of drought could happen in Henan.The south-west of Xinyang and Nanyang belonged to the humid area with lower drought risk probability;Sanmenxia,parts of Puyang,Anyang and Kaifeng counties belonged to the region of severe drought with higher drought risk probability;most areas of Zhoukou and Luohe belonged to the drought-stricken area.The high risk probability was mainly distributed in the western mountainous areas such as Sanmenxia,Nanyang and other western mountainous areas and Xuchang and Zhengzhou.The low risk probability value areas were mainly in the rice growing areas such as northeast and Xinyang.This was consistent with the spatial distribution of the drought, but there were also differences in some regions,such as the high level of drought in parts of
Puyang,but low risk probability value.

Key words: Drought;Grid;Precipitation distance percentage;Fuzzy clustering;Zoning, Henan province;Winter wheat

中图分类号: