Simulating Global Atmospheric CO2 and Local Atmospheric COS for a Continental Mixed Forest

Simulated hourly global atmospheric [CO2] for the year 2000 exhibits a systematic error in the seasonal cycle of simulated [CO2] in the Northern Hemisphere mid-latitudes, characterized by early spring drawdown of [CO2] relative to the observations. We have evaluated the simulation of carbonyl sulfide (COS) in SiB3, a land-surface model, at a continental mixed-forest site to separately evaluate seasonal variations in simulated photosynthesis and ecosystem respiration. Preliminary results of our simulation at WLEF in Wisconsin, US, show that the calculation of photosynthesis (rather than respiration) is the cause of the systematic error in the simulated seasonal cycle of atmospheric CO2. COS is consumed in plant tissues by a reaction catalyzed by carbonic anhydrase, and therefore behaves as a tracer of gross photosynthesis over land surfaces. We computed COS flux in the land surface model by assuming complete oxidation of intercellular COS, and compared the simulated flux to fluxes of [COS] estimated from the observed jump in [COS] between the atmospheric mixed layer and the free troposphere. Simulated surface exchanges of COS systematically led those derived from observations by several weeks in the spring, indicating that the simulated initiation of photosynthesis, rather than the timing of ecosystem respiration, is the cause of the systematic error in the simulated seasonal cycle of the flux of CO2. Employing COS has assisted in evaluating our systematic error of early drawdown of atmospheric CO2 by the biosphere in the spring and early recovery of atmospheric CO2 in the autumn, generally quite a difficult task due to the similar dependencies of photosynthesis and respiration on temperature and moisture. These results indicate that the ratio of COS uptake to CO2 uptake provides a sensitive indicator of the ratio of photosynthesis to respiration.