Large differences in global and regional soil carbon stocks estimated by different products: intercomparison and evaluation with field measurements.

Abstract:

Soils are the major component of the terrestrial ecosystem and the largest organic carbon reservoir on Earth. However, it's a non-renewable natural resource and quite reactive to human disturbance and climate change. Even small modifications, of less than a percent, of the huge amount of carbon contained in soils may lead to sources or sinks of greenhouse gases that could be significant relative to those released by fossil fuel combustion. Consequently, soil carbon dynamic is an important source of uncertainties for future climate predictions. To tackle this difficulty, there is a growing need for global and specific information on the soils to better understand the mechanisms controlling soil carbon dynamic and better constraints the Earth system models. Several global information systems on soil parameters already exist but these products are not always consistent and lack of spatial details.

The aim of our work is to compare the soil organic carbon stock given by different products : The World Soil Information (ISRIC), the Harmonized World Soil Database (HWSD) and The Northern Circumpolar Soil Carbon Database (NCSCD) and better understand what are the main differences that may explain the inconstencies between the products.

We calculated global and regional soil carbon stocks with the three products at different depths and we observed that they may largely differ in particular in boreal regions. We also observed bigger differences in the stocks estimated for surface soils compare to deep soils. Differences in boreal regions may be due to high disparities in soil organic carbon concentration whereas differences in other regions may be more likely due to different bulk densities. Finally, we compared the three products with field data available within the International Soil Carbon Network and we observed that each products present important drawbacks in representing the spatial variability. The estimation of the global soil carbon stocks is still quite uncertain and improved geostatistical methods are urgently need to reduce the confidence interval of the most important organic carbon stock.