Anciens doctorants

Matieu HENRY

Carbon stocks and dynamics in Sub Saharan Africa.

Thèse soutenue le 07-06-2010


Fonctions actuelles


Projet de thèse 

Carbon stocks and dynamics in Sub Saharan Africa


Climate change became the most important global environmental crisis and it consideration revealed the emergence of a new political regime. Forests play a key role because they significantly contribute to the carbon cycle and they meet the consensus among the different groups of interests. Represented by the United Nations Convention on Climate Change, the climate community tends to establish an equitable repartition of the efforts of mitigate climate changes and adapt to the modifications of our environment. However, sub?Saharan African countries are mostly in margin of the process mostly because of the lack of knowledge about the quantitative contribution these countries can make to the global efforts. Quantification of carbon stocks becomes crucial to increase their participation to the climate discussions and develop actions. Better understanding the carbon cycle also allows a better comprehension of our interaction between the global climate and our activities. The aim of this thesis is to better understand the dynamics of carbon stocks in sub?Saharan Africa. The objectives were (1) to understand the most important factors influencing the estimation of carbon stocks and carbon stock changes, (2) to inventory the state of knowledge at continental scale and (3) to fulfill some of the gaps while identifying the necessary further researches. Three field experiments where implemented in Ghana to measure the variability of biomass within plots, between plots and between wet evergreen, moist evergreen and moist semi?deciduous forests. The wet evergreen forest was selected to analyze the variation of biomass within and between trees based on an original method to build tree allometric equation. The three forests were logged and one of the forests was selected to measure the impact of logging activities on carbon stocks. In parallel, an intensive data collection in various libraries of the world allowed the development of the first tree volume and biomass allometric database and the first biomass and carbon stock database at continental scale. The data were organized and reviewed to allow the comparison between equations and carbon stocks data.
About two third of the continental carbon stock for Africa is found in the soil. By analyzing the various soil databases and soil maps for Africa, I found that variations of soil organic carbon estimates can be up to 30%. Using the tree allometric database, I found that the choice of the tree allometric equation to estimate the tree biomass can induce a variation of aboveground carbon stocks and stock changes of about 40% in tropical rainforest. Using default biomass values from the literature, I estimated that aboveground carbon stocks ranges 10 ?105 Pg C while improving the data collection and harmonizing the data allow reducing the variability of carbon stocks ranging 58?76 Pg C. Based on the field measurements from the three forests in Ghana I found important variability of carbon stocks between forests, within forests and within measured plot. Aboveground biomass ranges 348?364 ha?1 between ecological zones, 218?485 Mg ha?1 between plots and 0 ? 4681Mg ha?1 within plots of 1 ha. The biomass of a tree was mainly influenced by the diameter of the trunk, the crown diameter and the wood density in a tropical rainforest of Ghana. However, various factors influence the ecosystem biomass variability.
At plant scale, the tree biomass is influenced by the tree species, the plant functional type, and the growth strategies. At ecosystem scale, the biomass is influenced by the effect of topography, plant distribution, slope, soils, history of perturbations and forest management. At biome scale, the biomass is influenced by climate, ecological zone, age, structure, and management. At continental scale, the biomass estimation is mainly influenced by biome classification, available databases, quality of data and descriptors.
When considering the impact of anthropic activities on carbon stocks, I estimated that deforestation and forest exploitation emissions ranged 0.06?0.5 and 0.03 – 0.08PgC Yr?1 respectively. While deforestation was already studied in previous research, I focused on forest degradation caused by selective logging. In Ghanaian rainforests, selective logging impact 12 Mg C ha?1. Based on the relation between harvested volume and the impact forest carbon stocks, I estimated that 88 Tg C yr?1 are emitted at continental scale which mean that selective logging would contribute to 20 – 25% of the continental C balance. The potential for decreasing GHG emissions from selective logging exist but is limited by the accuracy and the cost of field measurements. Forest restoration was identified as the highest potential activity for emission reduction and C sequestration in Ghana. While this activity is not recognized by the climate convention, enhancement of forest C stock can sequester significant amount of C. It was estimated that intensification of already existing cocoa farms was the most feasible option to sequester important amount of C in Ghana while increasing the farming income of about 19%. However, C sequestration and financial compensation are among many other aspects to better consider when implementing developing activities. The land and forest tenure issues have to be particularly addressed in Ghana to allow effective implementations. The exclusion of some group of the society would lead to massive forest destructions which already happened in the past. Sub?Saharan Africa faces important gaps related to the understanding of the contribution of the African ecosystems to the C cycle. Only one percent of the necessary tree allometric equations are currently available. The decomposition of wood products in tropical rainforests is poorly known and would influence the GHG balance of selective logging and certainly also deforestation. The impact of selective logging on soil dynamics is not known. The dynamics of forest regeneration in Africa after perturbations such as selective logging are poorly known. While C stocks are assumed to recover after a 40 years period, the structure of the forest and the biodiversity are still strongly affected. The use of remote sensing is often presented as a panacea and the solution to monitor natural resources such as forest biomass. I however point out that the estimation of biomass using available satellite imageries of Ghana is not an accurate option while the use of high resolution imageries are too costly for countries that have low human and technical capacities to achieve their forest inventories using these techniques. This study highlights the issue of the language used to describe the environment. The land description used in global land analysis is often not in relation with the language used in the field to describe the vegetation in term of structure, floristic composition, position in the landscape and management. The identification of the degraded forest is strongly limited by the poor definitions and the lack of data.
While several attempts to harmonize the vegetation descriptors were developed based on classification systems, they however face the complexity of wording, defining and translate our perception of environment. At last, very few socio?economic studies considered the impact of carbon services to local farmer livelihood. They are however considered as part of the main process of deforestation. There is an urgent need to identify the potential for improving farmer livelihood while sequestering carbon stocks and reducing emissions from deforestation and forest degradation. This would be achieved by increasing the consideration of agricultural activities in developing countries in the climate change convention. But still, the agricultural sector will be considered if a carbon quantification system can be implemented for farming activities in developing countries.


Thèse soutenue le 7 juin 2010 devant le jury suivant :

  • Prof. Riccardo Valentini, University of Tuscia, directeur de thèse
  • Dr. Martial Bernoux, IRD, directeur de thèse
  • Prof. Marco Marchetti, University of Molise, Examinateur
  • Dr. Raphaël Manlay, AgroParisTech-ENGREF, Examinateur
  • Prof. Davide Travaglini, University of Firenze, Examinateur
  • Dr. Frederic Achard, Joint Research Centre, Examinateur


UFR G-ENV — AgroParisTech

Montpellier campus
648 rue Jean-François Breton
BP 44494
34093 Montpellier Cedex 5