Large convective storms, known as mesoscale convective systems (MCSs), are the main drivers of extreme rainfall and severe weather. Accurately representing these storms in Earth system models is essential for predicting their variations and changes.
Feng et al. [2025] apply ten different feature tracking methods to assess MCSs in an ensemble of next-generation global kilometer-scale or storm-resolving simulations. Although different tracking methods produced somewhat different estimates of storm frequency and rainfall in observations, consistent patterns emerged when comparing model simulations with observations.
While the models generally capture storm frequency well, they tend to underestimate the rainfall amount from these storms and their contribution to total precipitation, particularly over oceans. Most models predicted heavier MCS rainfall for a given amount of atmospheric water vapor compared to observations.
Mesoscale Convective Systems tracking Method (MCSMIP) provides a framework for a more robust evaluation of model performance to guide future model development to improve predictions of storms and their attendant impacts.