Noyaux glaçogènes d'origine naturelle et modélisation de leurs interactions avec les nuages arctiques

Ice nucleating particles (INPs) are aerosol particles capable of promoting the formation of ice crystals in clouds. In the Arctic-where global warming occurs about four times faster than the global average-clouds containing ice dominate the atmosphere. In particular, mixed-phase clouds (i.e. clouds composed of both liquid and solid phases) are common. When modelling the Arctic climate, errors in the representation of cloud phase partitioning strongly affect model accuracy, as cloud phase influences the radiative budget, precipitation, and cloud lifetime. All these properties depend on aerosol-cloud interactions, whose representation in models remains limited by the still scarce-although growing-knowledge of INPs. Key open questions include identifying the relevant INP species in the Arctic, their sources, their efficiency in ice nucleation, and ultimately their influence on clouds and the radiative budget-questions that continue to hinder improvements in Arctic climate modelling. In this thesis, I address in three steps the role of INPs in Arctic aerosol-cloud interactions and their impact on cloud properties. First, I develop, evaluate, and apply a new methodology for the identification of INP emission sources, based on the Lagrangian model FLEXPART-WRF combined with observational data. I then use this method to shed new light on the origins and species contributing to recently measured INP concentrations. Second, I implement new high-latitude aerosol sources in the regional model WRF-Chem, and use them together with ice-nucleation parameterisations to represent the dynamics of INP populations in the Arctic. Finally, after comparing the modelled INPs with observations, I couple them to the aerosol-aware microphysics scheme of WRF-Chem in order to investigate their effect on Arctic clouds during summer, when mixed-phase cloud occurence and INP concentrations are the highest. Results show that, in the central Arctic, INPs mainly originate from continental or marine sources near the Russian coasts of the Barents and Laptev Seas, as well as from coastal Greenland, where mineral dust emissions are known to occur. In addition, some results suggest biological sources from the ice-free Arctic Ocean for INPs active at temperatures above −15 ◦ C. The implementation of these high-latitude emissions, together with corresponding ice- nucleation parameterisations, enabled a realistic representation of Arctic INPs in WRF-Chem. When temperature- dependent activation was applied, the activated fraction of modelled INPs dropped by several orders of magnitude compared to those produced by purely temperature-based parameterisations. This had significant impacts on cloud microphysics, especially in the central Arctic, highlighting a potentially strong influence of INPs, and the need for updated aerosol-aware microphysics schemes.