Evaluation of CMIP5 Gross Primary Productivity Using Atmospheric COS Seasonal Cycle

Disentangling respiration from photosynthetic absorption of carbon dioxide (CO2) by plants on a global scale remains a challenging task which is necessary for understanding the effects of atmospheric CO2 increase. A surrogate of CO2 for estimating carbon storage in the terrestrial vegetation has been found with carbonyl sulfide (COS), a trace gas which exhibits a striking similarity with CO2 in the diffusion pathway of leaves. Based on the biochemical similarity between COS and CO2, mechanistic as well as empirical models relating both gas concentrations have been developed recently in order to provide constraints on the estimation of the Gross Primary Productivity (GPP), the amount of carbon dioxide that is absorbed by ecosystems.

We have evaluated the joint COS and CO2 surface fluxes simulated by twenty-two Climate Model Intercomparison Project (CMIP5) models and calculated seasonal concentration amplitudes and phases in offline mode using the LMDZ transport model. These concentrations were compared to observations provided by the NOAA/ESRL global monitoring network.In this study we first estimated the photosynthetic COS fluxes with the help of known empirical, plant-specific parameterizations based on the models' GPP. The joint concentration and phase misfit evaluation (with regard to observations) of COS and CO2 serves as a tool to identify possible model biases. We then performed a sensitivity study of the amplitudes and phases of the joint concentrations of COS and CO2 to the leaf relative uptake (LRU), and show the key role of this parameter in the COS - CO2 relationship.