Abstract:
Given the large uncertainty of the terrestrial carbon cycle in projecting future climate, it is important to understand the degree to which the variability of the atmospheric CO2 affects the variability of land carbon fluxes. Here we quantify this sensitivity by analyzing the spatial and temporal variability of the monthly Gross Primary Productivity (GPP) and Net Ecosystem Exchange (NEE), from a series of carefully-designed experiments using a standalone version of the NASA GEOS-5 CatchmentCN terrestrial biosphere model. Our control case imposes a maximum level of CO2 variability which forces the model with 3-hourly global fields of CO2 concentration extracted from the NOAA CarbonTracker for the period of 2001 to 2014. We then perform a series of experiments in which different levels of CO2 variability are successively removed to force the model: (i) we average the 3-hourly CO2 fields into daily fields; (ii) we average the daily fields into monthly fields; (iii) we average the monthly fields over the 14 years to get a climatological seasonal cycle of CO2; (iv) we average the climatological cycle into an annual mean field; and (v) we spatially average the annual mean field into a single value of the global mean CO2 concentration. Our preliminary results indicate that as the level of temporal variability of the CO2 fields increases, the monthly global GPP decreases. The spatial variability of the CO2 fields such as the North-South gradient tends to increase the monthly global GPP, but the temporal variability of the CO2 overall offsets the increase. When the climatological cycle of the CO2 is removed (e.g., case (iv)), the seasonal cycle of the GPP is amplified, showing the largest differences in the Northern Hemisphere summer (July-August).