Modelling stratospheric composition for the Copernicus Atmosphere Monitoring Service: multi-species evaluation of IFS-COMPO Cy49

Abstract. The daily analyses and forecasts of atmospheric composition delivered by the Copernicus Atmosphere Monitoring Service (CAMS) are produced by the ECMWF Integrated Forecasting System configured for COMPOsition (IFS-COMPO). In 2023 this system was upgraded to Cy48 which solves explicitly for stratospheric chemistry through a module extracted from the Belgian Assimilation System for Chemical ObsErvations (BASCOE). In 2024 the system was further upgraded to Cy49 which improves the representation of stratospheric composition with an adjusted parameterization of Polar Stratospheric Clouds (PSC), updated chemical rates for heterogeneous chemistry, and the implementation of missing processes to simulate an accurate distribution of sulfate aerosols in the stratosphere. Here we report on these improvements and evaluate the resulting stratospheric composition in chemical forecast mode, where the model is constrained by assimilation of meteorological observations but not by assimilation of composition observations. These evaluations comprise 13 gas-phase species and sulfate aerosols in three case studies: a global-scale assessment during a quiescent period (July 2023 to May 2024) in the context of the operational upgrade of the CAMS system; the evolution of key tracers related to polar ozone depletion during the winter and spring seasons across several years; and the evolution of stratospheric aerosols over the three years following the June 1991 Mount Pinatubo eruption. The model captures the rapid increase of the sulfate burden after the Pinatubo eruption, with the peak of stratospheric sulfate burden timed correctly, gradual recovery, and expected vertical profiles for quiescent periods. A scorecard assessment of chemical forecasts in the stratosphere of IFS-COMPO Cy49 highlights good performance for O3, CH4, N2O, and H2O and adequate performance for HCl, ClO, BrO and BrONO2 in the polar lower stratosphere. The model performance is poorer for HNO3, N2O5, NO2 and ClONO2, highlighting the need to improve the representation of heterogeneous chemistry, particularly the interactivity between aerosols and gas-phase composition, and refine the parameterization of PSC to better capture their impact on gas-phase composition. Overestimations of CH4 and N2O in the upper stratosphere are potentially related to the Brewer–Dobson Circulation, and long-standing biases of NO2 and O3 in the upper stratosphere remain unresolved. Despite these points for further development, IFS-COMPO will be a useful tool for studies of the couplings between stratospheric aerosols and gas-phase chemistry. The current cycle paves the way for assimilating stratospheric composition observations beyond ozone.