设为首页  |   加入收藏
引用本文:
【打印本页】   【下载PDF全文】   查看/发表评论  【EndNote】   【RefMan】   【BibTex】
←前一篇|后一篇→ 过刊浏览    高级检索
本文已被:浏览 382次   下载 170 本文二维码信息
码上扫一扫!
分享到: 微信 更多
农田土壤有效态微量元素空间预测方法及影响因子定量分析
乔依娜1, 刘洪斌2
1.西南大学地理科学学院;2.西南大学资源环境学院
摘要:
为深入研究土壤有效态微量元素的影响因素,以重庆市江津区农田土壤为研究对象,利用1 265个样点数据,加入种植制度和母岩类型数据,构建土壤微量元素的虚拟变量回归预测模型,采用随机森林算法,定量分析了分类变量对土壤有效态微量元素影响的相对重要性。结果表明:加入种植制度和母岩的回归模型的拟合度高于普通线性回归模型,对土壤有效铁、锰、铜和锌变异的解释度分别提高了9.20%、38.99%、20.75% 和29.96%,并且提高了对土壤有效铁的预测精度,但对提高土壤有效锰、铜和锌的预测精度作用不明显。土壤养分和种植制度是影响土壤有效态微量元素含量的重要因素,种植制度和母岩中,种植花椒、种植水稻和遂宁组发育的土壤对农田有效态微量元素含量的影响相对较大。
关键词:  有效态微量元素  虚拟变量  随机森林  影响因素
DOI:10.13758/j.cnki.tr.2019.02.026
分类号:S159
基金项目:
Spatial Prediction of Soil Available Microelement Contents and Quantitative Analysis of Influential Factors in Farmland
QIAO Yina1, LIU Hongbin2
1.School of Geographical Sciences, Southwest University;2.School of Resource and Environmental Science, Southwest University
Abstract:
To further explore the influential factors of soil available microelements, the farmland soils in in Jiangjin of Chongqing were taken as the study objects, the data of 1 265 sampling sites, together with cropping systems and parent rock types, were used to construct the regression model of soil microelements and random forest model (RF) was used to quantitatively analyze the relative importance of influential factors of soil available microelements. The results showed that the prediction model with the cropping system and the parent rock type was higher than the conventional regression model, the fitting degrees (R2) of available Fe, Mn, Cu and Zn were increased by 9.20%, 38.99%, 20.75% and 29.96%, respectively, compared with the conventional regression model. The prediction accuracy was improved for soil available Fe, but not for available Mn, Cu and Zn. Soil nutrients and cropping system were relatively important for the contents of available microelements in farmlands, and among of them, the effects of pepper and rice planting as well as soils derived from Suining Formation were relatively important. This study could provide theoretical and scientific bases for reasonable fertilization and management of Jiangjin farmlands.
Key words:  Soil available microelements  Dummy variable  Random forest model  Influential factors

您是第2407965位访问者
版权所有 © 《土壤》编辑部
本系统由北京勤云科技发展有限公司设计   京ICP备09084417号