Constraining the budget of atmospheric carbonyl sulfide using a 3-D chemical transport model

Carbonyl sulfide (OCS) has emerged as a valuable proxy for photosynthetic uptake of carbon dioxide (CO2) and is known to be important in the formation of aerosols in the stratosphere. However, uncertainties in the global OCS budget remain large. This is mainly due to the following three flux terms: vegetation uptake, soil uptake and oceanic emissions. Bottom-up estimates do not yield a closed budget, which is thought to be due to tropical emissions of OCS that are not accounted for. Here we present a simulation of atmospheric OCS over the period 2004-2018 using the TOMCAT 3-D chemical transport model that is aimed at better constraining some terms in the OCS budget. Vegetative uptake of OCS is estimated by scaling gross primary productivity (GPP) output from the Joint UK Land Environment Simulator (JULES) using the leaf relative uptake (LRU) approach. The remaining surface budget terms are taken from available literature flux inventories and adequately scaled to bring the budget into balance. The model is compared with limb-sounding satellite observations made by the Atmospheric Chemistry Experiment - Fourier Transform Spectrometer (ACE-FTS) and surface flask measurements from 14 National Oceanic and Atmospheric Administration - Earth System Research Laboratory (NOAA-ESRL) sites worldwide. We find that calculating vegetative uptake using the LRU underestimates the surface seasonal cycle amplitude (SCA) in the Northern Hemisphere (NH) mid-latitudes and high latitudes by approximately 37ppt (35%). The inclusion of a large tropical source is able to balance the global budget, but further improvement to the SCA and phasing would likely require a flux inversion scheme. Compared to co-located ACE-FTS OCS profiles between 5 and 30km, TOMCAT remains within 25ppt (approximately 5% of mean tropospheric concentration) of the measurements throughout the majority of this region and lies within the standard deviation of these measurements. This provides confidence in the representation of atmospheric loss and surface fluxes of OCS in the model. Atmospheric sinks account for 154GgS of the annual budget, which is 10%-50% larger than previous studies. Comparing the surface monthly anomalies from the NOAA-ESRL flask data to the model simulations shows a root-mean-square error range of 3.3-25.8ppt. We estimate the total biosphere uptake to be 951GgS, which is in the range of recent inversion studies (893-1053GgS), but our terrestrial vegetation flux accounts for 629GgS of the annual budget, which is lower than other recent studies (657-756GgS). However, to close the budget, we compensate for this with a large annual oceanic emission term of 689GgS focused over the tropics, which is much larger than bottom-up estimates (285GgS). Hence, we agree with recent findings that missing OCS sources likely originate from the tropical region. This work shows that satellite OCS profiles offer a good constraint on atmospheric sinks of OCS through the troposphere and stratosphere and are therefore useful for helping to improve surface budget terms. This work also shows that the LRU approach is an adequate representation of the OCS vegetative uptake, but this method could be improved by various means, such as using a higher-resolution GPP product or plant-functional-type-dependent LRU. Future work will utilise TOMCAT in a formal inversion scheme to better quantify the OCS budget.