The Himalaya provides hydrological and cultural resources to a growing downstream population of billions of people. In acknowledgement of this fact, the region has become a hotbed for research into the changing land surface systems of the mountain range. Such studies often rely upon high-resolution forcing datasets which can capture the spatial heterogeneity of mass and energy inputs into the land surface. Yet downscaling efforts over the Himalaya face serious challenges, including the scarcity of in-situ measurements for statistical downscaling, and the complexity of the topography for dynamical downscaling, a technique which has been further constrained by computational demand. Here we present a novel approach to dynamic downscaling over the Himalaya using the High-resolution Intermediate Complexity Atmospheric Research (HICAR) model. HICAR’s approach to atmospheric modeling reduces the need for terrain smoothing to retain numeric stability, allowing for a better representation of the steep orography of the range. Importantly, the HICAR approach also results in computational efficiency orders of magnitude greater than traditional atmospheric models, allowing for dynamic downscaling at higher horizontal resolutions and longer time scales. Prior studies have shown that the HICAR model improves estimates of seasonal snow relative to snow model simulations driven by coarser resolution forcing data from a traditional atmospheric model. We show preliminary results from an effort to run HICAR over the Himalaya at sub-kilometer resolutions. This effort is undertaken with the goal of improving long-range snow forecasts over the mountain range, a focus of study where simulations at scales of tens of kilometers using simplistic snow models currently offer the best predictions. A comparison of predicted snowcover from HICAR relative to current methods is shown, detailing which processes resolved by the high-resolution model contribute to differences in accumulated snow. HICAR relies greatly on input data from coarser resolution NWP models, and limitations of this modeling approach will also be discussed. This work also seeks to provide improved forcing data for the land surface modeling community and is intended to spur discussion about how sub-kilometer forcing data could best support their modeling efforts.