Below-cloud scavenging of aerosol by rain: a review of numerical modelling approaches and sensitivity simulations with mineral dust in the Met Office's Unified Model

Abstract

Abstract. Theoretical models of the below-cloud scavenging (BCS) of aerosol by rain yield scavenging rates that are 1–2 orders of magnitude smaller than observations and associated empirical schemes for submicron-sized aerosol. Even when augmented with processes which may explain this disparity, such as phoresis and rear capture in the raindrop wake, the theoretical BCS rates remain an order of magnitude less than observations. Despite this disparity, both theoretical and empirical BCS schemes remain in wide use within numerical aerosol models. BCS is an important sink for atmospheric aerosol, in particular for insoluble aerosol such as mineral dust, which is less likely to be scavenged by in-cloud processes than purely soluble aerosol. In this paper, various widely used theoretical and empirical BCS models are detailed and then applied to mineral dust in climate simulations with the Met Office's Unified Model in order the gauge the sensitivity of aerosol removal to the choice of BCS scheme. We show that the simulated accumulation-mode dust lifetime ranges from 5.4 d in using an empirical BCS scheme based on observations to 43.8 d using a theoretical scheme, while the coarse-mode dust lifetime ranges from 0.9 to 4 d, which highlights the high sensitivity of dust concentrations to BCS scheme. We also show that neglecting the processes of rear capture and phoresis may overestimate submicron-sized dust burdens by 83 %, while accounting for modal widths and mode merging in modal aerosol models alongside BCS is important for accurately reproducing observed aerosol size distributions and burdens. This study provides a new parameterisation for the rear capture of aerosol by rain and is the first to explicitly incorporate the rear-capture mechanism in climate model simulations. Additionally, we answer many outstanding questions pertaining to the numerical modelling of BCS of aerosol by rain and provide a computationally inexpensive BCS algorithm that can be readily incorporated into other aerosol models.