Abstract We present a comprehensive evaluation of 13 global storm‐resolving models participating in the DYnamics of the Atmospheric general circulation Modeled On Non‐hydrostatic Domains (DYAMOND) Winter intercomparison project, focusing on their ability to simulate key atmospheric fields, such as precipitation rate, outgoing longwave radiation, and precipitable water, during January–February 2020. All models permit explicit convection and have grid spacings of 5‐km or finer. Comparisons against observational and reanalysis data sets (CERES, IMERG, ERA5) reveal substantial inter‐model spread and systematic biases. While large‐scale features such as the Intertropical Convergence Zone (ITCZ) and South Pacific Convergence Zone are broadly captured, many models exhibit persistent problems, including the double ITCZ structure, overestimated OLR, and inconsistent diurnal cycles of convection. Precipitation and PW over land show particularly strong disagreement with observations. Comparison against coarser resolution simulations of select CMIP6 experiments for short‐term period show improved performance of DYAMOND winter models on most atmospheric fields. We also assess deep convective system sizes using brightness temperature proxies and clustering analysis, finding wide variability in convective extent, sensitivity to threshold choice, and possible biases in vertical structure. These findings highlight both progress and persistent limitations in kilometer‐scale global modeling, emphasizing the need for targeted sensitivity studies, improved observational constraints, and coordinated diagnostic frameworks to advance understanding and simulation of organized tropical convection.
In et al. (Sun,) studied this question.