摘要
【目的】在三峡库区王家沟小流域,研究不同土地利用方式对SOM含量反演精度的影响。【方法】利用来自园地、旱地、水田3种土地利用类型土壤样本的有机质含量及高光谱数据,构建了不同土地利用方式下的土壤有机质含量反演模型。【结果】不同土地利用方式下的土壤有机质反演模型精度并不相同:水田土壤有机质含量反演模型精度最高(总相关系数为0.838,标定集和验证集的相关系数分别为0.865和0.776);旱地有机质含量反演精度居中,总相关系数为0.781;园地土壤标定集和验证集中土壤有机质预测值和实测值之间也具有显著的相关性,但其精度最低(总相关系数为0.700)。混合来自园地和旱地两种土地利用方式的土壤样本建立的紫色土土壤有机质含量反演模型精度(总相关系数为0.528)低于单一土地利用方式的土壤有机质含量反演精度。【结论】与土壤类型相比,有机质含量的反演精度对土地利用方式更为敏感。考虑土地利用方式的影响,有助于提高土壤有机质含量反演精度。
【Objective】In the Wangjiagou watershed of the Three Gorges Reservoir area,the influence of different land use types on the inver-sion accuracy of SOM content was studied.【Method】Based on analysis of soil organic matter content and soil reflective spectrum,the inver-sion models were built for soil organic matter content in different land use types:garden land,dry land and paddy field.【Result】The inver-sion models of soil organic matter contents in three land use types had different inversion accuracy:the accuracy of soil organic matter content in paddy field was the highest(the total correlation coefficient is 0.838,and correlation coefficients calibration sets and validation sets are 0.865 and 0.776,respectively).The accuracy of Inversion model in dry land was in the second place(its total correlation coefficients is 0.781).The predicted values and measured values of soil organic matter content in garden land had significant correlation,although the model had the worst accuracy(its total correlation coefficients is 0.700).The inversion accuracy of soil organic matter content in purple soil based on mixed garden and dry land samples is lower than that of single land use.【Conclusion】Compared with soil types,the inversion accuracy of soil organic matter content is more sensitive to land use types.Therefore,considering the impact of land use types,it is helpful to improve the accuracy of inversion model of soil organic matter content.
作者
徐丽华
谢德体
XU Li-hua;XIE De-ti(College of Resources and Environment,Southwest University,Chongqing 400716,China)
出处
《西南农业学报》
CSCD
北大核心
2020年第2期369-373,共5页
Southwest China Journal of Agricultural Sciences
基金
中央高校基本科研业务费专项资金(XDJK2016 C083)
重庆市基础科学与前沿技术研究一般项目(cstc2016jcyjA0184)
国家自然科学基金项目(41671291)。
关键词
土地利用
SOM
高光谱数据
偏最小二乘回归
Land use
Soil organic matter
Hyperspectral data
Partial least squares regression
