Uncoupling of microbial CO2 production and release in frozen soil and its implications for field studies of arctic C cycling

Knowledge of seasonal trends and controls of soil CO2 emissions to the atmosphere is important for simulating atmospheric CO2 concentrations and for understanding and predicting the global carbon cycle. This is particularly the case for high arctic soils subject to extreme fluctuating environmental conditions. Based on field measurements of soil CO2 efflux, temperature, water content, pore gas composition in soil and frozen cores as well as detailed temperature experiments performed in the laboratory, we evaluated seasonal controls of CO2 effluxes from a well-drained tundra heath site in NE-Greenland. During the growing season, near-surface temperatures correlated well with observed CO2 effluxes (r2>0.9). However, during intensive thawing of near-surface layers we observed up to 1.5-fold higher effluxes than expected due to temperature alone. These high rates were consistent with high CO2 concentrations in frozen soil (>10% CO2) and suggested a spring burst event during soil thawing and a corresponding trapping of produced CO2 during winter. Laboratory experiments revealed that microbial soil respiration continued down to a least -18 degrees C and that up to 80% of the produced CO2 was trapped in soil at temperatures between 0 and -9 degrees C. The trapping of CO2 in frozen soil was positively correlated with soil moisture (r2=0.85) and led to an abrupt change of the temperature sensitivity (Q10) observed for soil CO2 release at 0 degrees C with Q10 values below 0 degrees C being up to 100-fold higher than above 0 degrees C. The results of sub-zero CO2 production allowed us to predict the microbial soil respiration throughout the year and to evaluate to what extent burst events during thawing can be explained by the release of CO2 being produced and trapped during winter. Taking only the upper 20 cm of the soil into account, winter soil respiration accounted for about 40% of the annual soil respiration. At least 14% of the winter CO2 production was trapped during the winter 2000-2001 and observed to be released upon thawing. Thus, the site-specific winter soil respiration is an important part of the annual C cycle and CO2 trapping should be accounted for in future field and modelling studies of soil respiration dynamics in arctic ecosystems. In conclusion, we have discovered a soil moisture dependent uncoupling of CO2 production and release in frozen soils with important implications for future field studies of Arctic C cycling.