Forest structure and topography can influence the ecohydrologic function and resiliency to drought and changing climate. It is, therefore, important to understand how forest restoration treatments alter snowpack distribution and design the treatments accordingly. We use a combination of aerial lidar, multi-temporal terrestrial mobile lidar, and UAV photogrammetry to estimate rapidly changing snow depth and cover in northern Arizona, USA. We then examine the impact of forest structure and topography on snow depth and snow cover persistence to inform forest restoration treatments. Our results show that mobile lidar data can be used to estimate snow depth with standard errors of 8 cm when differenced with snow-off airborne lidar data. UAV-based Structure-from-Motion data can be used to estimate snow cover persistence with 92–97% overall accuracies in forested ecosystems. Random forest models indicate spatially varying importance of forest structural and topographic variables in predicting snow depth and cover persistence, when summarized at different spatial scales (from 5 m to 250 m) and with variable directional location offsets. Forest snow depth was best explained (R² ≈ 0.46) by canopy height metrics at summary scales of >75 m, while canopy cover was most important at summary scales of <40 m (R² ≈ 0.3). Snow cover persistence was best explained at very local scales by canopy cover (R² ≈ 0.38) and less so at larger scales (>75 m) by topographic and forest patch characteristics (R² ≈ 0.34). Our results demonstrate that 3-dimensional datasets are critical in rapidly characterizing changing snowpack to better understand the impacts of forest structure and topography to inform forest restoration treatment designs. The relationships observed in our study can inform currently ongoing regional-scale forest restoration in the southwest to improve forest health and resiliency.