Abstract:

 Soil reflectance spectroscopy within the visiblenearinfrared (VNIR) region (0.382.5 μm) contains much information of soil composition and their physical and chemical parameters. The determination of a soil reflectance spectrum is not only rapid, inexpensive but also nondestructive and environmental friendly, and moreover, the spatial distribution of soil reflectance information can be mapped by hyperspectral remote sensing. Based on the large quantities of soil samples acquired in the ecogeochemical survey of Jiangsu Province, we have related the soil ecogeochemical parameters, such as cation exchange capacity (CEC), organic matter, pH, ferric oxide, and contents of Al, Fe and Si, to the reflectance parameters of soil samples. The results indicate that the CEC of soil can be accurately predicted by first derivative (FD) values of soil reflectance spectra at 1400, 1900 and 2200 nm or by reflectance values directly at the shorter wavelengths (~400 nm). Soil organic matter, types of iron oxides, and contents of Al, Fe and Si can also be determined by their respective relevant wavelengths on the reflectance or FD spectra. Soil pH in Jiangsu was found to be highly correlated to FD values near 656 nm; however, this relationship showed a regional dependence, which may account for the nonlinear relationship between soil pH and reflectance parameters. The more complicated mechanisms relating to the relationship between soil pH and their reflectance parameters need to be further clarified.

Key words:

Key words: hyperspectral remote sensing; soil; reflectance spectrum; soil ecogeochemical parameters