Delineating spatial variations in soil carbon using remotely sensed data: A New England case study

Abstract

Remotely sensed imagery is used in many land management applications to provide rapid, cost-effective and up-to-date information on land condition. This paper examined whether a combination of remote and proximal sensing technologies can be used to identify the spatial variability in shallow (0-30 cm) soil organic carbon (SOC) content of soils in an agricultural land system. A demonstrator section of farming land in the northern tablelands region of NSW was used to test the hypothesis that "typical" land management agency datasets associated with landform, geology and land cover, coupled with precision agriculture sensing technologies for measuring apparent soil electrical conductivity and photosynthetically active biomass, can be used to delineate zones of %SOC. SOC explained 47% of the variance when all landscape variables were combined but when an upper limit of only four variables was applied this dropped to 33 %. When the landscape was stratified into vegetation and topography zones however, %SOC was found to explain up to 98% of a combination of proximal plant reflectance (red band), normalised difference vegetation index (NDVI) measurements of the pasture species, and the elevation.