A combined observational and modelling approach to evaluate aerosol–cirrus interactions at high and mid-latitudes

Abstract

Aerosol–cirrus interactions remain a major source of uncertainty in climate models due to the complex interplay of aerosol properties, ice nucleation pathways, and atmospheric conditions. In this study, we investigate the drivers of observed differences in cirrus microphysical properties between high and mid-latitudes from the CIRRUS-HL campaign by combining observations with simulations from a global aerosol-climate model. While mid-latitude cirrus exhibit median ice crystal number concentrations (Nice) one order of magnitude higher than those at high latitudes, aerosol concentrations (Naer) integrated across several sizes ranges are similar at cirrus altitudes in both regions. By coupling the model output with backward trajectories, we attribute the differences in Nice to diverse influences of specific ice-nucleating particle (INP) types with distinct freezing efficiencies rather than to total aerosol or INP number concentrations. Mineral dust plays a dominant role in cirrus formation at mid-latitudes, while aviation-emitted black carbon may contribute to high-latitude cirrus assuming it acts as an efficient INP. The model reproduces aerosol observations reasonably well but underestimates at high latitudes near 300 hPa. At mid-latitudes, it overestimates Nice at temperatures above 220 K, primarily due to an overestimation of the concentration of ice crystals detrained from convective clouds. Incorporating a size parametrization for convective ice crystals derived from CIRRUS-HL measurements significantly reduces this bias, which represents a fundamental improvement to the cloud scheme. These findings highlight the value of integrating observations with model simulations to interpret field measurements and improve the representations of cirrus clouds in global models.