Evaluation of a simplified double-moment aerosol scheme (SOL/INSOL) using satellite, airborne and ground based observations over the UK

Abstract. Uncertainty in aerosol processes, particularly aerosol–cloud interactions (ACI), remains high, limiting the predictive skill of clouds and rainfall across timescales. Aerosols also contribute to visibility degradation, highlighting the need for accurate forecasts to support safety and reduce transport‑related economic impacts. Consequently, the development of consistent, scalable, and computationally affordable aerosol representations that are applicable across both numerical weather prediction and climate timescales is increasingly important. To this end, we introduce and evaluate a new double‑moment prognostic aerosol scheme, named SOL/INSOL ("SOLuble and INSOLuble"), implemented within the Met Office’s UKV local area model. SOL/INSOL is a simplified configuration of the United Kingdom Aerosol and Chemistry used for climate projections. A global driving model provides lateral boundary conditions to the UKV, where aerosol precursor fields are initialized using the Copernicus Atmosphere Monitoring Service (CAMS) reanalysis dataset. Emissions ancillaries are monthly means used in the Met Office operational air quality model. This development enables coupling of number concentration with the UKV cloud microphysics scheme, CASIM, which will strengthen the model’s treatment of ACI. This study presents an in‑depth evaluation of simulated SOL/INSOL PM2.5 using surface and airborne PM2.5 observations, alongside a comparison of modeled aerosol optical depth with satellite retrievals for a selected case date. We then run a suite of case studies amounting to 31 dates over the UK which are compared to both UK-wide ground sites, and regional airborne observations over Manchester, London and Birmingham. We also perform sensitivity simulations, and evaluate the short-term forecasting skill of the model. Model skill is quantified in terms of normalised mean bias (NMB). We show that SOL/INSOL predicts PM2.5 with fidelity, reproducing surface PM2.5 mass concentrations to within 13 % of that observed. We show that using CAMS reanalysis for initial conditions perform well, demonstrating how variations in the relative contributions of these gridded precursor aerosol components translate into systematic changes in modal particle size distributions and associated PM2.5 concentrations. Our simulations investigating sensitivity to lead time show that SOL/INSOL has good predictive skill up to T+37 hours. We show that the simulated number concentrations agree well with observations, providing confidence in their use within the UKV cloud microphysics scheme. We find that SOL/INSOL is currently limited in its ability to accurately reproduce discrete, localised enhanced aerosol events and discuss future requirements to address this shortfall. We conclude that the simple aerosol scheme shows excellent promise for further testing in operational forecasting at the UK Met Office.