Water accumulating on microstructural transitions inside a snowpack is often considered a prerequisite for wet snow avalanches. Recent advances in numerical snowpack modeling allow for an explicit simulation of this process. We analyze detailed snowpack simulations driven by meteorological stations in three different climate regimes (Alps, Central Andes, and Pyrenees), with accompanying wet snow avalanche activity observations. Predicting wet snow avalanche activity based on whether modeled water accumulations inside the snowpack locally exceed 5–6% volumetric liquid water content is providing a higher prediction skill than using thresholds for daily mean air temperature, or the daily sum of the positive snow energy balance. Additionally, the depth of the maximum water accumulation in the simulations showed a significant correlation with observed avalanche size. Direct output from detailed snow cover models thereby is able to provide a better regional assessment of dangerous slope aspects and potential avalanche size than traditional methods.