摘要
有机质含量是土壤肥力评价的重要依据,快速监测南疆大面积分布的荒漠土壤有机质含量,可为后备耕地资源的合理开发与利用提供重要数字依据。通过野外采样与室内光谱及有机质含量测定,获取了240个荒漠土样的有机质含量与可见光近红外光谱数据。在此基础上,分析了不同有机质含量土样的光谱特征、有机质与光谱数据的相关性,并采用多种数据处理方式构建了反演模型。研究结果表明,不同有机质含量的土样在反射率和曲线形态上均存在一定的差异,土壤有机质在400~842 nm波段与反射率具有较高的相关性,相关系数最大值-0.32位于588 nm波段,反射率经连续统去除处理后相关性得到了明显提升,达极显著相关水平波段数量明显增加,尤其在797~1330 nm、1852~1872 nm、2155~2338 nm波段的改善效果显著,连续统去除的最大相关系数0.55位于86 4 nm波段。不同建模方法的精度对比结果表明偏最小二乘法(PLSR)具有最高的建模精度,在7种不同数据处理方式的PLSR模型中,以SG平滑+峰值归一化+正交信号校正(SGS+MAN+OSC)处理模式下的建模集的决定系数(R^2)最高、均方根误差(RMSE)最低,分别为0.81和0.98;该模型预测集的R^2、RMSE、残余预测偏差(RPD)分别为0.76、0.99和2.01,表明该模型具有较好的预测能力,可推荐为南疆荒漠土壤有机质的光谱定量反演模型。
Soil organic matter（SOM） is an important factor for soil fertility evaluation. High-speed monitoring the distribution of SOM content in desert soil in southern Xinjiang would provide important digital basis for the rational exploitation and utilization of land resources. This study determined the contents of SOM and their visible-near infrared spectroscopy data of 240 desert soil samples. On the basis of these, the spectral characteristics of soil samples with different SOM content and the correlation between them were analyzed, the inversion models were constructed by using a variety of data processing methods. There were some differences in the reflectance and curve morphology of soil samples with different SOM contents. SOM in the 400-842 nm band had a high correlation with the reflectance,and the maximum correlation coefficient in the 588 nm band was-0.32. The correlation of the reflectance was significantly improved after continuum removal, and the number of significant correlation bands was significantly increased, especially in the 797-1330, 1852-1872, 2155-2338 nm bands, and the maximum correlation coefficient of continuum removal was 0.55 in the 864 nm band. The partial least squares method（PLSR） had the highest modeling accuracy. The SGS ＋ MAN ＋ OSC processing model had the highest R^2 value of 0.81 and the lowest RMSE value of 0.98 among the seven different data processing methods in the PLSR model. The R^2, RMSE and RPD of the model predictive dataset were 0.76, 0.99 and 2.01, respectively, which indicated that the PLSR model has a good prediction to SOM content and can be recommended as a quantitative inversion model for SOM content of desert soil in Southern Xinjiang.
作者
李阳
刘新路
彭杰
李祥
吴家林
LI Yang;LIU Xin-lu;PENG Jie;LI Xiang;WU Jia-lin(Plant Science College of Tarim University,Alaer 843300,China)
出处
《土壤通报》
CAS
北大核心
2018年第4期767-772,共6页
Chinese Journal of Soil Science
基金
国家自然科学基金项目(41361048
41561049)资助
关键词
高光谱
有机质
荒漠土壤
南疆
反演模型
Hyperspectral
Organic matter
Desert soil
Southern Xinjiang
Inversion model
作者简介
李阳（1991-），男，陕西渭南人，硕士，主要从事干旱区土壤属性的光谱快速诊断与评价研究。E-mail：TarimSam@163．com;通讯作者：E-mail：pjzky@163．com
