Ice Crystals in deep convective Clouds: interactions with Aerosols, Radiation and Electricity – ICCARE

The first part of the project is based on the exploitation of measurements from past airborne field campaigns (HAIC in French Guiana in 2015, and EXAEDRE in Corsica in 2018). Convolutional neural network (CNN) algorithms are applied to ice crystal image databases to automatically classify ice crystals and link their morphology to their growth patterns. In addition, measurements from space lidars (Calipso, Cats-ISS) will be used to provide information on the shape and orientation of ice crystals in the upper part of clouds. These analysis of the observations already available will make it possible to identify the crystal shapes that should be given priority in the numerical models, as well as the ice crystal growth modes in the different regions of the storm system, depending on its stage of development and its environment.
The second axis aims to develop a complete and coherent cloud model in order to explore numerically how the interactions between ice crystals and the physical processes linked to them control the development of thunderstorms and the associated damage. Building on the results of the first axis, the parameterizations of the morphological and microphysical properties of the crystals, the dominant growth regimes and the interactions with other physical processes are improved. Aerosol-microphysics-radiation-electricity interactions are considered for the first time in an atmospheric model. The first evaluation phase of this unique module will be carried out via sensitivity analyses on a wide range of academic thunderstorms (continental and maritime, mid-latitude and tropical, etc.).
Finally, the last axis is dedicated to three-dimensional simulations at hectometric resolution of deep convection systems observed during the HAIC and EXAEDRE measurement campaigns. The special case of a tropical cyclone observed during the ReNovRisk measurement campaign over the south-west Indian basin will also be addressed in order to analyze cloud-radiation interactions on a convective system with a lifetime of several days.

The study of the links between the morphology of ice crystals and the associated growth regimes (vapor diffusion, aggregation, icing) has begun. Data from the HAIC-Cayenne 2015 airborne campaign combined with satellite observations were used to identify the convective, stratiform and cirriform regions of the sampled clouds, and the number and mass fractions of the morphological classes were produced. Automated extraction tools for geostationary imagery (MSG) and sun-synchronous satellite measurements (A-Train: CALIPSO, MODIS) have been developed. For A-Train measurements, an additional tool has been developed to identify overpasses close to a region during a study period, and to locate the measurement files in the ICARE data center catalog. Relevant variables in the satellite measurements for comparison with Meso-NH simulation outputs were identified, extracted and collated for a thunderstorm with an anomalous electrical structure in Corsica, tropical cyclone Idai and a southern storm.
The development of the Meso-NH cloud physics module is progressing. Two secondary crystal formation processes (collisional ice breakup and raindrop shattering by freezing) have been integrated into the LIMA 2-moment microphysics scheme, and the ORILAM aerosol scheme has been coupled with LIMA by developing its initialisation (chemical composition, modes) using CAMS analysis. Different crystal shapes are being taken into account in LIMA: new prognostic variables (concentration in number of crystals of each shape) have been created, and mass-diameter and velocity-diameter relationships and a capacitance specific to each shape have been defined. The particle distribution shapes and ice crystal shapes assumed for the optical properties were made consistent between the ecRad radiative transfer code and those used in LIMA: new laws (mass/area/velocity-diameter) were defined in LIMA, compatible with the available databases of optical properties of icy clouds. Finally, the CELLS electrical scheme has been coupled to LIMA, and the effect of the electric field on the falling speed of hydrometeors and on the aggregation of ice crystals has been implemented. A review of the brightness temperature errors simulated by RTTOV in ice clouds as a function of ice content and cloud height was also carried out.
At the same time, simulations of real cases using Meso-NH have begun. The simulation of a thunderstorm over Corsica demonstrated the impact of desert dust on cloud ice, in particular by limiting the presence of supercooled water, a necessary condition for the formation of an electrical tripole structure. Lastly, the microphysical composition of cyclone Idai resulting from a preliminary simulation is currently being compared with MODIS and CALIPSO data.

In terms of in situ data, a statistical study of the dominant morphological classes and associated growth modes will be produced for selected HAIC and EXAEDRE flights. In addition, a database of depolarisation ratios and morphological classification indexes will be produced for the regions and periods of interest for the field campaigns and case studies selected on the basis of CALIPSO and CATS-ISS data.
The development of the Meso-NH cloud physics module will continue, paying particular attention to consistency between the different parameterizations. An initial assessment of the roles of the various components of this module and their interactions will be carried out on several academic storms.
Finally, the impact of different crystal shapes and ice formation mechanisms on the organization of deep convective systems and associated precipitation will be assessed on a case from the HAIC-Cayenne field campaign. The contribution of secondary crystal formation processes to cloud electrification and lightning flashes, and the feedback of the electric field on microphysics will be assessed on a case from the EXAEDRE field campaign in order to benefit from data from the SAETTA 3D lightning detection network. Finally, the microphysics-radiation interactions will be analyzed on a mesoscale convective system and on a tropical cyclone.

Conferences :
- Barthe, C., Modélisation des interactions aérosols-microphysique-rayonnement-électricité dans les systèmes de convection profonde. Journées scientifiques sur la convection profonde, 21 mars 2022, Bordeaux.
- de Sevin S., I. Vongpaseut, C. Barthe, S. Coquillat, et P. Tulet, Analyse du rôle des poussières désertiques sur des anomalies de structure de charge électrique. Ateliers de Modélisation de l’Atmosphère 2023, 9-11 mai 2023, Toulouse.
- Vongpaseut, I., et C. Barthe, Simulation des interactions électricité-microphysique-aérosols dans les orages en cas idéalisés. Ateliers de Modélisation de l’Atmosphère 2023, 9-11 mai 2023, Toulouse.

Dissemination:
Barthe, C., ICCARE : un projet pour comprendre le rôle des cristaux de glace dans le développement des orages. La Météorologie, 119, doi.org/10.37053/lameteorologie-2022-0076, 2022.

Today, the formation, growth and precipitation of ice crystals are misrepresented in numerical models, impacting the forecasting of the structure, evolution and impact of deep convective clouds (e.g., precipitation, hail, radiative impact, lightning, wind gusts). The ICCARE project aims at further understanding how ice crystals interacts with aerosol, microphysics, electricity, radiation and dynamics, and how these interactions control the evolution of deep convective clouds.
ICCARE is divided into 4 tasks. Task 0 concerns the coordination of the project (management, communication, datasets and numerical codes provisioning). Task 1 is dedicated to exploring new ways to describe ice crystal shape and type through airborne observations from past field campaigns (HAIC, EXAEDRE) and from data from space-borne lidars (CALIPSO, CATS). Task 2 focuses on the development of a coupled model with coherent parameterizations to model exhaustively aerosol-microphysics-radiation-electricity interactions. We will improve or develop the treatment of secondary ice formation mechanisms and the competition between homogeneous and heterogeneous nucleation of ice crystals, the description of ice crystal habits, the coherent coupling between microphysics and radiation, and microphysics – atmospheric electricity interactions. Finally, Task 3 relies on the innovative modeling system developed in Task 2 in synergy with observation products from Task 1 to analyze for the first time the aerosol-microphysics-radiation-electricity inside convective clouds observed during past field campaigns. For all case studies, after cloud contextualization and calibration of the numerical simulations, sensitivity analysis will be conducted. Interactions that take place in the different regions of the system will be analyzed depending on their developing stage, the role of cloud physical processes will be hierarchized, and their impact on the macrophysical cloud parameters will be evaluated.
A comprehensive numerical module of cloud physics will be developed in synergy with the exploitation of observations from past field campaigns and from satellites measurements. It will enable to better characterize the uncertainties in the cloud resolving models due to ice crystals and their associated physical processes. In the more or less long term, the impact and benefits are numerous: better forecasting of deep convective systems and associated hazards, identification of key processes to be integrated as a priority in large-scale models, improvement of the representation of the radiative impact of clouds in models… ICCARE will enable developing innovative numerical models and tools that will be made available to the whole scientific community (through the AERIS portal for observations, and through its website for Meso-NH) to advance our understanding in multiple topics for which scientific research is extremely active (e.g. aerosol-cloud and cloud-radiation interactions, risks associated with extreme meteorological events).
ICCARE relies on a consortium made up of 3 partners: LAERO and LATMOS are associated inside the same partner, CNRM and LaMP. The scientists involved in the project are expert in all the domains of cloud physics (dynamics, aerosol, microphysics, radiation, atmospheric electricity). They are actively involved in the development of the community model Meso-NH and of the radiative transfer model RTTOV, and in the organization and data processing of the HAIC, EXAEDRE and ReNovRisk field campaigns which are at the heart of the project.