Correcting UAV derived winter snow depth on glaciers by modelling the evolution of the No-Snow glacier surface 

Abstract

<p>Spatially distributed winter snow accumulation over glaciers is an important information for a lot of purposes. Typically, snow depth on glaciers is measured by manual snow probing or ground penetrating radar. The point measurements of snow depth and snow density are then used to calculate the winter mass balance of the glacier.</p><p>In the last decade remote sensing techniques such as LIDAR and structure from motion (sfm) photogrammetry in combination with unmanned aerial vehicles (UAVs) have become more frequent to reconstruct snow surfaces providing a better spatial coverage and spatial resolution. Snow depth is calculated by DEM differencing of a No-Snow surface (summer surface) and the snow surface (winter surface).</p><p>However, using DEM differencing to extract snow depth over glaciers introduces the problem, that the No-Snow surface is not constant, as (1) the glacier is moving between the survey dates and (2) the surface possibly undergoes surface lowering due to melt after the summer survey.</p><p>In this study we present measurements on two small mass balance glaciers in the Austrian Alps (Goldbergkees and Kleinfleißkees). We account for the evolution of the No-Snow surface by (1) applying a simple model of the vertical ice movement and by (2) calculating the surface lowering due to melt using a distributed mass balance model.  The effect of both corrections is then validated using a dense network of manual snow depth measurements across the glacier.</p>