Fate of soil organic carbon stocks after deforestation for agriculture: a diachronic approach in the Amazon forest
Thesis defended on Thursday, 27 November 2014 at Montpellier SupAgro
Soil organic carbon (SOC) plays a major role in maintaining soil properties and is a major sink of carbon susceptible to anthropogenic disruption such as changes in land use and land management. In the Amazon forest, deforestation releases greenhouse gases (GHG) by burning the vegetation but the SOC stocks are also likely to change and contribute to GHG emissions. This study showed that these changes in stocks at biome scale are not well quantified owing to the frequent use of chronosequence approaches with their many uncertainties and the lack of data on the management of the agrosystems set up after deforestation.
The study was based on a diachronic analysis of an agricultural site in Guyana, that had been cleared forest leaving the residues on the ground without burning. Three cropping systems were established: a meadow and two maize/soybean systems, one ploughed and one no-till. The aim was to determine the fate of forest carbon and the SOC in agrosystems.
Changes in carbon stocks were measured up to 5 years after deforestation in the top 30 cm horizon and a carbon balance was estimated for the top 100 cm horizon after 5 years. The decomposition of the wood debris, left on the ground after deforestation, was determined by measuring the loss of mass and characterised using Rock-Eval pyrolysis. The particle size distribution of SOC was measured 4 years after deforestation. Isotopic analysis (δ13C) was used in the meadow soil to distinguish the carbon originating from the meadow.
The carbon resulting from deforestation increased SOC stocks but only in the short term as the wood debris decomposed rapidly and did not lead to sustainable SOC storage. Five years after deforestation, the SOC stocks in the meadow were similar to those in the forest, owing to significant input of carbon from the roots, whereas in the cropped fields the stocks dropped by about 18%, with no discernable effect from ploughing. The drop in forest carbon, which applied to all the soil particle size fractions, was compensated by the increase in carbon in the meadow, which was not the case for annual crops. The RothC model was validated in these conditions although it slightly overestimated the stocks in the cropped fields. These results, applied to the conditions in the Amazon forest, show that the decrease in SOC observed here is lower than in tropical wet zones as a whole, probably owing to the optimum management of the site and the short observation time.