A comparison of snow depth scaling patterns from TLS, UAV and Pleiades observations

Understanding the evolution of snowpack in heterogeneous mountain areas is a highly demanding task and requires the application of suitable observation techniques to retrieve snow properties at distinct spatial scales. In turn, once the reliability of these techniques is established, the comprehension of snowpack scaling properties helps to determine which processes are more relevant on the control of snow distribution and its temporal evolution. Previous studies have reported detailed observational datasets and insights on the main drivers of snowpack distribution through variogram analysis up to 500-800 m, identifying scale break lengths and their anisotropies. Here, we examine scale breaks derived from variogram analysis applied to snow depth observations at the Izas Experimental Catchment (located in Central Spanish Pyrenees) and the surrounding area for the period 2019-2023. To this end, we use data retrieved with three observation techniques: Terrestrial Laser Scanning (TLS-LiDAR, 12 acquisitions), Unmanned Aerial Vehicles (UAV-SfM, 20 acquisitions), and satellite stereo images (4 Pléiades acquisitions), covering different domains around the experimental site. First, we analyze the consistency among the observational techniques, and then we explore possible drivers explaining detected scale breaks through variogram analysis up to 4000 m. Overall, similar results were obtained with the three observational techniques, with a very high temporal consistency for the first detected scale break length and little variations with direction. We also found good agreement between the search distance used to compute the topographic position index (TPI), the first scale break length, and the mean distance between peak snow accumulations, which vary between 15 and 25 m, not only for the entire study domain, but also in manually delineated Hydrological Response Units.