Vertical Velocity Deduced From Ice Change in MesoNHforC2OMODO RASTA‐km, a Data Set of Deep Convection Simulations Assessed Against Observations

Abstract Measuring vertical velocity is crucial for improving our understanding of deep convection. The Convective Core Observations through MicrOwave Derivatives in the trOpics (C2OMODO) mission aims to retrieve vertical velocity using ice‐sensitive microwave measurements acquired at two closely spaced time intervals. In preparation for the mission, it is essential to better understand the relationship between vertical velocity and satellite‐observed ice‐sensitive microwave brightness temperatures. However, vertical velocity within convective cores is rarely measured, and no comprehensive data set currently exists. To address this gap, we developed MesoNHforC2OMODO RASTA‐km (RAdar SysTem Airborne), a data set of kilometer‐scale Meso‐NH simulations created for the C2OMODO project. It is based on five airborne field campaigns during which the RASTA radar was deployed: CADDIWA, EXAEDRE, HAIC–Cayenne, HAIC–Darwin, and MAESTRO. The data set comprises approximately 1.2 billion atmospheric columns and 2.1 million updrafts, of which 717 000 are in the growth phase, derived from more than 50 simulations. It has been validated against RASTA radar and satellite observations. Using this data set, we confirm results from previous idealized studies. First, temporal variations in C2OMODO brightness temperatures are linearly related to temporal variations in ice water path. Second, the vertical velocity of growing convective cells can be inferred from C2OMODO measurements. We further demonstrate that vertical velocity can be inferred in deep convection from C2OMODO brightness temperatures with a gradient boosting method regressor, an ensemble‐based machine learning model.