Recent advancements in remote sensing of snow (e.g., lidar) have allowed for the characterization of mountain snowpacks at higher spatial resolutions (< 10-m), with higher vertical accuracy (< 20-cm), and covering entire catchments repeatedly throughout the snow season. Here, we use distributed snow water equivalent (SWE) maps over the Tuolumne River Basin (1180-km2, 1150-4000 m elevation range) in California, USA, from the Airborne Snow Observatory (ASO) program for the period 2013-2017 (48 flight dates, 50-m resolution) to characterize the spatial and temporal variations in SWE distribution in a headwater catchment. Peak basin snow storage ranged between 142 M m3 in 2015 and 1467 M m3 in 2017 covering one of the widest ranges in recorded history. Basin empirical distributions vary between mono- and bi-modal distributions earlier in the season to exponential decaying distributions later in the ablation season. Snow storage peaks at elevations between 2750-3250 m, a consequence of increases in SWE with elevation and basin hypsometry. The date of peak SWE varies several weeks across the watershed and between years, according to accumulation and melt patterns partially explained by elevation and aspect. These variations in peak timing lead to an underestimation of spatial peak SWE when a single date of peak SWE is assumed across the watershed, with an underestimation of hundreds of mm possible depending on each seasons weather patterns. Our results illustrate how understanding these spatial and temporal dynamics in snow accumulation and melt can improve reservoir operations though understanding where the snow is and when it melts. The dynamics in peak accumulation and timing should also be considered for survey design and planning, and which is particularly relevant for future global snow mapping efforts.