基于距离聚类与K-means动态聚类的棉田土壤养分评价研究  

Soil nutrient evaluation of cotton field based on distance clustering and K-means dynamic clustering

作　　者：范向龙 吕新[1,2] 张泽 高攀[2,3] 张强 印彩霞 易翔[1,2] FAN Xianglong;LYU Xin;ZHANG Ze;GAO Pan;ZHANG Qiang;YIN Caixia;YI Xiang(Key Laboratory of Oasis Eco-Agriculture,Xinjiang Production and Construction Group,Agriculture College of Shihezi Universit,Shihezi 832003,Xinjiang,China;National and Local Joint Engineering Research Center for Agricultural Big Data of Xinjiang Production and Construction Corps,Shihezi 832003,Xinjiang;School of Information Science and Technology,Shihezi University,Shihezi 832003,China)

机构地区：[1]石河子大学农学院,新疆生产建设兵团绿洲生态农业重点实验室,新疆石河子832003 [2]新疆兵团农业大数据国家地方联合工程研究中心,新疆石河子832003 [3]石河子大学信息科学与技术学院,新疆石河子832003

出　　处：《干旱区研究》2021年第4期980-989,共10页Arid Zone Research

基　　金：兵团重点领域创新团队项目(2018CB004);兵团国际合作计划项目(2018BC009);国家博士后面上项目(2017M623282);石河子大学创新发展专项(CXFZ201903);新疆兵团棉花生产大数据关键技术及农业大数据平台研发应用(2018AA004)。

摘　　要：为了解棉田土壤养分状况及确定养分丰缺程度,研究采用主成分结合距离聚类、K-means动态聚类以及土壤养分综合评价方法对新疆棉田土壤进行分析评价。结果表明:(1)在主成分分析中,有效铜、有效锰和碱解氮起主要作用,有效铜含量为1.82 mg·kg^(-1),丰缺评价状况属于高等水平,有效锰含量为11.36 mg·kg^(-1),属于较低水平,碱解氮含量为122.07 mg·kg^(-1),属于高等水平,土壤速效磷、速效钾、有效锌和有效铁含量较低,土壤养分含量分布不均匀。在距离聚类和K-means动态聚类中,有机质、碱解氮、有效锰含量较低,其余养分含量较高。在距离聚类中,土壤各类养分可表示为:第Ⅰ类>第Ⅴ类>第Ⅳ类>第Ⅱ类>第Ⅲ类,而在K-means动态聚类中可以表示为:第Ⅲ类>第Ⅰ类>第Ⅴ类>第Ⅱ类>第Ⅳ类。(2)在土壤综合肥力指数评价值中(IFI),1连和16连的等级高;2连、3连、4连、6连、15连、19连和二监区的等级较高;8连、9连、10连、11连、12连、17连、18连和20连在中等水平。5连、7连、一监区和农市站的等级较低。K-means动态聚类比距离聚类分类效果好,可以更加科学合理、准确有效地对土壤养分进行综合评价。This study used principal component analysis combined with distance clustering, K-means dynamic clustering, and comprehensive soil nutrient evaluation methods to analyze and evaluate the soil nutrient status of cotton field soil in Xinjiang, China. The results showed that the available copper, manganese, and alkali hydrolyzable nitrogen played a major role principal component analysis: The available copper content was high(1.82 mg·kg^(-1));the available manganese content was low(11.36 mg·kg^(-1));and the alkali hydrolyzable nitrogen content was 122.07 mg·kg^(-1). The available phosphorus, potassium, zinc and iron contents were low. Thus, the distribution of soil nutrients was uneven. In the distance clustering and K-means dynamic clustering, the organic matter, available nitrogen, and available manganese contents were lower, whereas the other nutrient contents were higher. In the distance cluster, the soil nutrients could be expressed as: class Ⅰ > class Ⅴ > class Ⅳ > class Ⅱ >class Ⅲ, whereas the soil nutrients could be expressed in the K-means dynamic cluster as: class Ⅲ > class Ⅰ >class Ⅴ > class Ⅱ > class Ⅳ. In the comprehensive evaluation of soil nutrients(IFI), the grades of 1 and 16 are higher. The second company, third company, fourth company, sixth company, 15 th company, 19 th company, and the second prison area had higher levels, whereas the 8 th, 9 th, 10 th, 11 th, 12 th, 17 th, 18 th and 20 th companies were in the middle level. The grades of companies 5, 7, and 1 supervision district and agricultural city station were lower. Thus, K-means dynamic clustering is better than distance clustering, as it can be more scientific, reasonable,accurate, and effective for the comprehensive evaluation of soil nutrients.

关 键 词：土壤养分 距离聚类 K-means动态聚类 综合评价值 

分 类 号：S562[农业科学—作物学]