Monitoring, Reporting and Verification systems for Carbon in Soils and Vegetation in African, Caribbean and Pacific countries

Abstract

Hyperspectral techniques are increasingly being used to determine chemical properties from the diffuse reflectance of visible and near infrared light from soil. In particular the applications from airborne sensors (imaging spectroscopy) provide information on the plough layer of cropland soils at very high spatial resolution. We illustrate the application of imaging spectroscopy to determine soil organic carbon in freshly ploughed croplands for two case studies: i) soil carbon content at the regional scale and ii) soil carbon stocks for a single field. At the regional scale the best results (accuracy of c. 3 g C kg -1 ) were obtained when local calibrations were applied for each agri- geological region. Both broad regional patterns and detailed within field patterns of SOC emerged. The former reflect gradients in geology and climate, while the latter could be related to erosion and sedimentation as well as previous land use. Apart from SOC content, bulk density and rock fragment cont are required to calculate the SOC stock of the plough layer. The proximal sensing platform of the Digisoil project was used to provide this complementary data for a field within the flight line of the regional case study. Geo-electric surveys provided spatial patterns in bulk density and rock fragment content. Although the accuracy of the SOC stocks (c. 10 Mg C ha -1 ) can still be improved, the confidence limits of the SOC stocks (48.6 ± 3 Mg C ha -1 ) allow detecting the effects of improved agricultural techniques (e.g . zero tillage or ploughing crop residues c. 0.2-0.5 Mg C ha -1 y -1 ) within 5-10 years.