Understanding the accumulation and melt of snow is essential to capture the hydrology of mountainous areas. Forests impact significantly the amount of snow on the ground through diverse processes, including the interception of snowfall. The Multi-Energy-Balance (MEB) module within the SURFEX/ISBA surface model has been designed to capture these impacts for climate applications. However, the original version of MEB coupled with the physically-based snow model Crocus (MEB/Crocus) fails to reproduce the subcanopy snow dynamics observed at Col de Porte, a mid-altitude alpine snow observatory. In this study, based on new observations at Col de Porte and recent literature, we propose an enhanced version of MEB/Crocus by revisiting solid interception processes: maximum interception, interception rate, unloading and melt. This new version reduces the positive bias of subcanopy simulations by at least 40 % for snow water equivalent and snow depth, through an enhanced interception, longer interception duration and associated sublimation and melt. The wide range of observations at Col de Porte furthermore confirms that MEB/Crocus has a satisfactory representation of the different processes involving the forest cover, including vegetation temperatures, ground temperatures and radiation under forest. With this new version of MEB/Crocus, the model’s net performance remains unchanged -or is even slightly improved- at boreal sites. We also take the opportunity to use new observations to explore the sensitivity of the model to a top-of-canopy forcing in wind and air temperature. While forcing the model with air temperature from above the canopy is found to have no significant impact, having a top-of-canopy wind is key to the representation of the evolution of snow under forest as it strongly affects the sensible and latent heat fluxes. The improvements and enhanced understanding of the model sensitivity pave the ground to extending the applications of MEB, from climate to the hydrology of mid-latitude mountainous areas.