Kibambe Lubamba, Jean-Paul
[UCL]
(eng)
The Democratic Republic of the Congo (DRC) hosts 50% of Congo Basin dense forests, the world’s second largest tropical forest which plays a major role in regulating the CO2 fluxes between the Earth surface and the atmosphere. Satellite Earth observation technology is the most efficient way to monitor deforestation and forest degradation taking place in DRC.
Pixel- and object-based satellite image processing methods have, however, their own assumptions leading to discrepancies that have to be consolidated to ensure the credibility and the transparency of forest cover change estimates from remote sensing.
This thesis simulated future forest cover losses for the whole of the DRC from 2005 to 2035 at 1-km spatial resolution using neural network-based modeling fed by spatially-explicit variables. These included human population distribution and multimodal accessibility to major towns and big villages. The former variable was obtained by disaggregating coarse scale population data available at the third administrative-level. The latter was modeled using a functional approach of the transportation network instead of a single proximity to
roads and rivers such as previously modeled by most land-cover change models. The developed forest cover loss model was calibrated using land-cover maps from years 2000-2005 and validated with a land-cover map from the year 2010.
The consolidated national annual rate of gross deforestation between 2000 and 2005 corresponded to 0.29%±0.05%. About 70% of these changes were found at locations lying within a daily journey from and to the nearest major town or big village, i.e., 7 hours. Simulations under a proposed forest conservation scenario indicated that the DRC will likely remain a high forested and low deforested country if mitigation actions are undertaken. In turn, a business-as-usual scenario demonstrated that the DRC’s human population will at least double in 2035, leading to a total area of forest cover loss corresponding to13.5% of the forest area in 2005 and an annual rate of forest cover change of about 0.48% between 2005 and 2035.
Findings from this thesis highlighted the need for long-term and strategic land-use planning policies in DRC to conciliate economic and agriculture development, infrastructure building, and population growth with a sustainable use of the forest resources, by especially taking
into account the regional context of Congo Basin countries.
Bibliographic reference |
Kibambe Lubamba, Jean-Paul. Modeling the forest cover loss in Democratic Republic of the Congo. Prom. : Defourny, Pierre |
Permanent URL |
http://hdl.handle.net/2078.1/133430 |