Spectral Determination of Soil Properties under Vegetation
Remotely sensed and laboratory reflectance spectra have been used for many years to predict a range of soil properties including clay content and species, carbonate, iron oxide, salinity (EC) and Munsell soil colour. However, remote sensing of soil properties generally relies on substantial soil exposure so that the sensor can directly measure the soil reflectance without interference from vegetation. Modern agricultural practices such as minimum tillage and direct drilling maintain a cover of crop residue on the soil surface for the entire year, limiting the ability to remotely sense soil spectral properties. Despite this there have been few studies to determine the level of soil exposure required to predict soil properties.
This study used a controlled experiment to determine the influence of green and dry plant cover on the detection of specific soil spectral features. Reflectance spectra were measured for differing mixtures of four soils and three cover types. Soil exposure was manipulated by incrementally passing the chosen cover type over the soil within the field of view of the spectrometer, thus creating a ‘mixed-pixel’. Results show that soil chromophores dominate the reflectance spectra even with substantial vegetative cover. However there is a threshold at which the soil chromophores become completely obscured and information about the related soil properties is lost. This preliminary study indicates that there is scope for using this experimental method to examine the determination of soil properties from mixed-pixels.