| Summary: | Arctic permafrost soils are estimated to contain 1700 Gt of carbon and a warming climate signals more rapid release of CO2 from these environments, further exacerbating the warming effect of greenhouse gases. CO2 emissions from Arctic soils during the cold season were previously thought to be negligible; this is not the case and their contribution to the overall annual carbon budget is expected to increase in a warming world. Few observational and modelling studies have been directed at Arctic cold season carbon cycling processes, ultimately making any future projections of CO2 release unreliable. This unreliability is compounded by uncertainty in (1) the land component of Earth System Models regarding the parameterizations of biogeophysical and biogeochemical processes driving CO2 release during winter and (2) projections of future climate in the Arctic. Here, we use the CLM5.0 model in point mode for Trail Valley Creek near Inuvik NT, forced by ERA5 reanalysis to evaluate the performance of CLM5.0 for present-day conditions using flux tower measurements of net ecosystem CO2 exchange as model constraints. We then apply a bias-corrected NA-CORDEX ensemble of RCP4.5 and RCP8.5 scenarios to quantify the variability in future winter CO2 emissions due to (1) model uncertainty in future climate conditions and (2) parameter uncertainty in CLM5.0 relating to the temperature sensitivity of soil respiration (Q10), the moisture threshold for soil decomposition and the thermal conductivity scheme for the snow layer. The latter is a strong control on the thermal evolution of soil temperature, and therefore soil respiration during the winter.
|
|---|