Séquestration de carbone par une forêt claire au Bénin : Variabilité interannuelle des flux et d'un indice de végétation

Abstract

Forests are playing an important role in climate change mitigation. The carbon budget associated to African ecosystems as well as their reactions to modified climate conditions are poorly documented. This study proposes an estimation of the annual carbon sequestration of an open forest in North-Western Benin (West Africa) and identifies the driving variables causing sequestration capacity’s variability, at three different time scales. This forest is dominated by the woody species Isoberlinia doka and is submitted to a Sudanian climate. The study site is part of the international AMMA-CATCH observation program. CO2 fluxes measurements, collected over a period of six years by an eddy co-variance system, were used in combination with meteorological data and with the MODIS Enhanced Vegetation Index. u* correction was applied to filter stable night-time conditions. The annual carbon sequestration estimate was made by summing the Net Ecosystem Exchange (NEE) provided at half-hourly scale and by filling the gaps resulting from raw data treatment and data capture systems failures. Daytime variations of NEE were first associated with the photosynthetically active photon flux density (PPFD) using a non-linear model, in order to extract the three following parameters: the saturation assimilation (P1500), the dark respiration (Rd) and the quantum light efficiency (α). The seasonal and inter-annual variabilities have then been analysed combining these parameters to environmental factors and to the vegetation index taking in account calculated anomalies. The spatial heterogeneity was also assessed by following spatial divergence of the P1500 and the EVI around the flux tower. On average, the forest acts as a carbon sink of 490 gC m-2 year-1. The seasonal variation of P1500 was strongly correlated to the vegetation index (r² = 0.69) and both saturation assimilation and dark respiration respond to soil moisture increase. The sink strength is strongly dependent on total annual precipitations (p-value = 0.002, r² = 0.97) and is better explained by respiration changes than saturation assimilation variability. More precisely, following behaviours have been clearly detected in our data: wettest years have the smallest annual respiration rates (significant) and the highest saturation assimilation (although non-significant). Interestingly, driest years show the highest respiration rates. Higher saturation assimilations were confirmed by an improved vegetation activity (significant). NEE anomalies could be partially explained by P1500 anomalies that were also confirmed by the vegetation index (p-value < 0.01). Spatial heterogeneity couldn’t be confirmed with EVI data for both seasons. An increasing assimilation before the first rains has also been detected that would also deserve being deeper investigated.