A method for characterizing the spatial organization of deep convective cores in deep convective systems’ cloud shield
Abstract. Deep convective systems (DCSs) play a fundamental role in atmospheric dynamics, precipitation, cloud radiative effects, and large-scale circulations. Their associated deep convection exhibits complex spatial arrangements, commonly referred to as convective organization, which exerts an influence on the systems’ morphology that needs to be assessed. However, quantifying this organization remains challenging due to the lack of a general, robust, and consensual metric, in both observations and models. This study introduces a new method for characterizing the spatial arrangement of deep convective cores within the cloud shield of individual DCSs. The first step of this technique consists in decomposing the convective mask into elementary structures. Four key variables are then extracted to fully capture the organization of a scene. Two of these variables characterize the overall properties of the convective field, such as the size and convective fraction of convective cores. The remaining two variables are specifically designed to describe the spatial arrangement of deep convective cores: a characteristic convective scale using two-dimensionnal (2D) autocorrelation and an evaluation of the deviation from randomness by comparing it to a stochastic ensemble of synthetic convective fields. Two independent datasets, derived from satellite observations and kilometer-scale numerical simulations, each employing distinct convective core identification techniques are used to assess the generalization of the method. Finally, an unsupervised clustering algorithm identifies four distinct classes, revealing consistent and physically sound patterns of convective organization across both datasets. This demonstrates the method’s robustness and suitability to characterize the spatial organization of convective cores in convective systems’ cloud shield.