Biomass assessment in african forest-savanna using ALOS PALSAR data

The purpose of this work is to assess biomass in forest savanna and to analyse changes in biomass from 2007 to 2010 in Cameroon, by using ALOS PALSAR data. The methodology includes an inter-calibration step, and multi-image filtering for noise reduction and preservation of the resolution. An existing direct model previously developed by CESBIO over various forests, linking γ_hv^o to biomass, is inversed using a probabilistic bayesian approach. There is a mean of 52 t.ha^-1 of biomass in 12 553 km² which represents 32.63 Mt of carbon not taken into account in existing carbon budget assessments. Besides, forest savanna transition and changes in AGB from 2007 to 2010 are analysed. Changes are mainly due to shifting cultivation, grazing fires, degradation and deforestation.