The aim of this study was to determine the amplitude and the driving factors of the spatial variability in soil CO2 efflux in a young European beech forest. Soil CO2 efflux was measured in 2003 and 2004 in seven beech plots differing in terms of soil type and leaf area index. After eliminating temporal fluctuations due to soil temperature and soil water content, standardized soil CO2 efflux varied significantly among plots over a large range given the homogeneity of the land cover type. Correlation analyses revealed that this spatial variability could not be explained by root biomass, litter C content, soil C contents, stand basal area or stem density. Conversely, very significant correlations were found with topsoil bulk density, superficial soil C/N ratio, and leaf area index. Multiple regression analysis led to a model relating standardized soil CO2 efflux to C/N ratio and topsoil bulk density, thus explaining 87% of observed interplot spatial variability. This study highlighted the need to consider spatially varying soil factors such as C/N ratio and bulk density when experimental schemes are elaborated to estimate mean soil CO2 efflux at forest scale.
