Earth approach A dual isotope to isolate carbon pools of di ff erent turnover times

Abstract

Soils are globally significant sources and sinks of atmospheric CO 2 . Increasing the resolution of soil carbon turnover estimates is important for predicting the response of soil carbon cycling to environmental change. We show that soil carbon turnover times can be more finely resolved using a dual isotope label like the one provided by 5 elevated CO 2 experiments that use fossil CO 2 . We modeled each physical soil fraction as two pools with different turnover times, using the atmospheric 14 C bomb spike in combination with the label in 14 C and 13 C provided by an elevated CO 2 experiment in a California annual grassland. In sandstone and serpentine soils, the light-fraction carbon was 20–40 % fast cy-10 cling with 2–10 yr turnover and 60–80 % slow cycling with turnover slower than 100 yr. This validates model treatment of the light fraction as active and intermediate cycling carbon. The dense, mineral–associated fraction also had a very dynamic component, consisting of 5–10 % fast cycling carbon and 90–95 % very slow cycling carbon. Simi-larly, half the microbial biomass carbon in the sandstone soil was more than five years 15 old, and 40 % of the carbon respired by microbes had been fixed more than five years ago. Resolving each density fraction into two pools revealed that only a small component of total soil carbon is responsible for most CO 2 efflux from these soils. In the sandstone soil, 8–11 % of soil carbon contributes more than 85 % of the annual CO 2 efflux. The 20 fact that soil physical fractions, designed to isolate organic material of roughly homoge-neous physico-chemical state, contain material of dramatically different turnover times is consistent with recent observations of rapid isotope incorporation into seemingly sta-ble fractions, and with emerging evidence for hot spots of decomposition within the soil matrix. Predictions of soil response using a turnover time estimated with the assump-25 tion of a single pool per fraction would greatly overestimate near-term response to changes in productivity or decomposition rates. Therefore, these results suggest more 10190 BGD rapid, but more limited, potential for change in soil carbon storage due to environmental change than has been assumed by more simple mass-balance calculations.