Résumé

We combined published and new radiocarbon and ancillary data for uncultivated topsoils (typically 15 cm depth), to make two databases, one for the United Kingdom (133 sites), and one global (114 sites). Forest topsoils are significantly higher in radiocarbon than non-forest soils, indicating greater enrichment with "bomb carbon" and therefore faster C turnover, if steady-state conditions are assumed. Steady-state modelling, taking into account variations in atmospheric (CO2)-C-14, including the effects of 20th century nuclear weapons testing and radioactive decay, was used to quantify soil carbon turnover rates. Application of a model with variable slow (20 year mean residence time, MRT) and passive (1,000 year MRT) carbon pools partitioned the topsoil C approximately equally, on average, between the two pools when the entire data set was considered. However, the mean slow:passive ratio of 0.65:0.35 for forest soil was highly significantly different (p < 0.001) from the 0.40:0.60 ratio for non-forest soils. Values of the slow and passive fractions were normally distributed, but the non-forest fractions showed greater variation, with approximately twice the relative standard deviations of the forest values. Assuming a litter input of 500 g C m(-2) a(-1), average global C fluxes (g C m(-2) a(-1)) of forest soils are estimated to be 298 (through a fast pool of MRT 1 year), 200 (slow pool) and 2.0 (passive pool), while for non-forest soils, respective average fluxes of 347, 150 and 3.3 g C m(-2) a(-1) are obtained. The results highlight the widespread global phenomenon of topsoil C heterogeneity, and indicate key differences between forest and non-forest soils relevant for understanding and managing soil C.

Détails

Actions