Filling the Holes in the Space-Time Cube of Snowpack Evolution with Lasers, Cameras, Computers, and Snow Shovels

TitleFilling the Holes in the Space-Time Cube of Snowpack Evolution with Lasers, Cameras, Computers, and Snow Shovels
Publication TypeConference Proceedings
Year of Conference2018
AuthorsRaleigh, Mark S., and Deems Jeffrey S.
Conference Name86th Annual Western Snow Conference
Conference LocationAlbuquerque, New Mexico

An information revolution in high-resolution snow observation is producing unprecedented applications in snow research and management of snow-dominated watersheds. Lidar technology has transformed our view and understanding of the spatial nature of snow depth. Between and after lidar flights, sparse in situ measurements are the only direct snowpack data. The central problem is that large gaps in the evolution of snowpack properties through space and time remain even when high-resolution repeat lidar are combined with time-continuous spare point observations. What is needed are additional tools and sources of information to “fill in” the snow data cube. In this project, we assess how time-lapse photogrammetry can help guide modeling decisions and assess model realism. The study focuses on the well-instrumented Senator Beck Basin, a small headwaters basin near Red Mountain Pass, Colorado. Daily maps of binary snow cover and a new dust-on-snow index are derived from a time-lapse camera at a 3 m resolution through the 2013 snowmelt season. A high-resolution (20 meter) snow model is applied using five methods for distributing precipitation spatially, including terrain-based approaches (e.g., lapse rates) and data-driven approaches (e.g., scaled by snow depth patterns from airborne lidar data). The different modeling approaches show a range of structure in spatial snow patterns, with data-driven approaches yielding higher spatial complexity, which are also seen in the camera data. Relative to camera observations, the models tend to estimate higher snow-covered area in the snowmelt season. As suggested by the camera-derived dust indices, this result may be explained by the lack of dust-enhancement in the modeled snowmelt. This study demonstrates the value of time-lapse cameras for identifying strengths and shortcomings of snow models, thereby guiding snow estimation approaches when high resolution snow depth data are unavailable. (KEYWORDS: snow distribution, time-lapse, modeling, dust-on-snow)