轮作模式在农耕区土壤有机质推测制图中的应用

发布时间：2018-06-28 16:36

  本文选题：数字土壤制图 + 遥感解译　； 参考：《土壤通报》2017年04期

【摘要】：人类活动近年成为数字土壤制图亟需考虑的要素。本文以农业活动中轮作模式为例,将轮作信息应用于数字土壤制图,探讨其对土壤空间变异刻画的有效性。以安徽宣城两个县市的耕地平区为研究区,通过野外调查获得近年三种主要轮作模式,基于监督分类对多期遥感影像解译得到轮作类型空间分布图,使用方差分析探讨轮作对土壤表层有机质空间变异是否有显著性影响,采用随机森林重要性指标对自然环境因子、轮作模式、土地利用方式和归一化植被指数进行重要性排序,并构建不同的环境因子组合,利用基于相似度的土壤推测模型和随机森林模型进行制图和交叉验证。结果表明,轮作模式对土壤表层有机质有显著性影响,其重要性排序为第二,引入轮作使得基于相似度的土壤推测模型和随机森林模型制图精度分别提高4.8%~65.9%和1.9%~2.7%。
[Abstract]:Human activities have become an important factor in digital soil mapping in recent years. In this paper, the rotation pattern in agricultural activities is taken as an example, and the rotation information is applied to digital soil mapping, and its validity to depict soil spatial variation is discussed. In this paper, three main rotation patterns were obtained by field investigation in the flat area of two counties and cities in Xuancheng, Anhui Province. Based on the supervised classification, the spatial distribution map of rotation type was obtained by interpreting multi-period remote sensing images. The variance analysis was used to investigate whether crop rotation had significant influence on spatial variability of soil surface organic matter, and the importance of natural environment factors, crop rotation patterns, land use patterns and normalized vegetation index was ranked by random forest importance index. Different combinations of environmental factors were constructed, and the similarity based soil conjecture model and stochastic forest model were used for mapping and cross-validation. The results showed that the rotation model had a significant effect on the soil surface organic matter, and the importance order was the second. The precision of soil conjecture model based on similarity and random forest model was increased by 4.859% and 1.92.7respectively.
【作者单位】： 中国科学院地理科学与资源研究所资源与环境信息系统国家重点实验室;中国科学院大学资源与环境学院;南京大学地理与海洋科学学院;南京师范大学地理科学学院;虚拟地理环境教育部重点实验室(南京师范大学)江苏省地理环境演化国家重点实验室培育建设点江苏省地理信息资源开发与利用协同创新中心;Department
【基金】：国家自然科学基金项目(41471178,41530749,41431177)资助
【分类号】：S159.9;S344.1
，

本文编号：2078649