Ephemeral snow is a critical water source in the North American Southwest, and globally seasonal snowpack is converting to ephemeral snow under climate change. Better understanding and timely detection of ephemeral snow dynamics are important for forest management and climate adaptation. Unfortunately, studying ephemeral snow presents significant challenges because of high heterogeneity over small spatial and temporal extents. Leveraging novel CubeSat imaging technology, we created a dense time series of PlanetScope imagery to monitor ephemeral snow dynamics across three winters (2022-2024) in thinned and nonthinned ponderosa pine forests in northern Arizona to evaluate forest structure effects on snow cover and persistent snow patches. We paired these observations with a large network of soil water potential sensors, facilitating unique insights into the ecologically relevant connections between ephemeral snowpack dynamics, forest structure, and soil water availability. A random forest classifier applied to PlanetScope imagery and validated against imagery from an uncrewed aerial vehicle was effective at classifying snow-covered area, with a balanced accuracy of 0.92 and F-score of 0.90. Thinned forest had greater snow cover and more persistent snow patches than nonthinned forest. Importantly, the dense PlanetScope time series revealed that the more-prevalent snow patches in thinned forest were associated with higher root-zone soil moisture, likely reducing water stress in subsequent dry seasons. The excellent performance of PlanetScope in this application highlights its novel potential for monitoring important ecohydrological processes under climate change and informs directions for future research.