Abstract. Carbon (C) turnover time is a key factor in determining C storage capacity in various plant and soil pools as well as terrestrial C sink in a changing climate. However, the effects of C turnover time on ecosystem C storage have not been well explored. In this study, we compared mean C turnover times (MTTs) of ecosystem and soil, examined their variability to climate, and then quantified the spatial variation in ecosystem C storage over time from changes in C turnover time and/or net primary production (NPP). Our results showed that mean ecosystem MTT based on gross primary production (GPP; MTT EC_GPP = C pool /GPP, 25.0 ± 2.7 years) was shorter than soil MTT (MTT soil = C soil /NPP, 35.5 ± 1.2 years) and NPP-based ecosystem MTT (MTT EC_NPP = C pool /NPP, 50.8 ± 3 years; C pool and C soil referred to ecosystem or soil C storage, respectively). On the biome scale, temperature is the best predictor for MTT EC (R 2 = 0.77, p < 0.001) and MTT soil (R 2 = 0.68, p < 0.001), while the inclusion of precipitation in the model did not improve the performance of MTT EC (R 2 = 0.76, p < 0.001). Ecosystem MTT decreased by approximately 4 years from 1901 to 2011 when only temperature was considered, resulting in a large C release from terrestrial ecosystems. The resultant terrestrial C release caused by the decrease in MTT only accounted for about 13.5 % of that due to the change in NPP uptake (159.3 ± 1.45 vs. 1215.4 ± 11.0 Pg C). However, the larger uncertainties in the spatial variation of MTT than temporal changes could lead to a greater impact on ecosystem C storage, which deserves further study in the future.

No abstract