基于RS和GIS的石亭江中下游土壤铅和镍含量预测建模研究  

Modeling to Predict Lead and Nickel Contents in Soil of the Mid-and Lower Reaches of Shiting River Using RS and GIS

作　　者：姚苹[1,2] 张东 张世熔[1,2] 徐小逊 李婷[1,2] 

机构地区：[1]四川农业大学资源环境学院,成都611130 [2]四川省土壤环境保护重点实验室,成都611130 [3]德阳市农业局,四川德阳618000

出　　处：《农业环境科学学报》2014年第1期95-102,共8页Journal of Agro-Environment Science

基　　金：国家科技支撑计划(2012BAD14B18-02)

摘　　要：为了快速高效地获取区域土壤重金属含量数据,利用石亭江流域中下游Landsat 7 ETM+遥感影像及70个样点土壤表层(0~20 cm)重金属铅镍含量和地面数据建立预测模型并进行了空间反演。结果表明,仅用波段像元灰度值建立的土壤铅镍含量预测模型均达极显著水平(P=0.000),表明遥感图像的波段光谱信息能用于土壤铅镍含量的预测建模。在分别引入成土母质、海拔高度或pH等地面辅助因子后,铅镍含量预测模型确定系数R2明显增大(P=0.000),铅预测模型R2从0.276分别提高到0.571和0.606,镍预测模型R2从0.304分别提高到0.513和0.551,表明地面辅助因子能有效改善模型精度。与实测值分布图比较,最优模型预测反演图能较好地表现区域土壤铅镍含量分布的基本格局,但对于个别特殊值区域的反演效果仍有待进一步提高。Predicting soil heavy metal contents is critical for soil pollution assessment and early warming management. In this work, the re-mote sensing spectral data from Landsat7 ETM+, soil Pb and Ni contents(70 samples from 0~20 cm soil layer)and the related ground parameters in the mid- and lower reaches of Shiting River were integrated to construct a model for predicting soil Pb and Ni contents in this area. The space inversion was employed to check the model reliability. Results indicated that the high prediction accuracy for Pb and Ni contents could be achieved by the constructed model using remote sensing spectral data only(P=0.000), implying its reliability to predict soil heavy metal contents. When taking ground parameters such as soil parent materials and elevation or pH into consideration, the R2values of the model were significantly increased(P=0.000), with R2values for Pb being increased from 0.276 to 0.571 and 0.606, and R2values for Ni from 0.304 to 0.513 and 0.551, indicating the involvement of prediction accuracy by including ground parameters. The predicted values were in good agreement with the observed Pb and Ni contents in most cases.

关 键 词：铅 镍 RS 预测建模 GIS 空间特征 

分 类 号：X833[环境科学与工程—环境工程]