Measurements of snowpack properties using structure-from-motion photogrammetry

Prag

2015

Poster

MonographieSonderdruck

192298

Spatially continuous information on snowpack properties such as snow depth or snow grain size is the basis of investigations in numerous snow-related research fields. Such parameters have traditionally been measured at discrete point locations by automated weather stations or observers in the field, with the drawback of insufficiently capturing the high spatial variability of snow in alpine terrain. Remote sensing techniques allow gathering spatially continuous information over large areas (up to more than hundred km2), without the necessity of having personnel in potential avalanche terrain. Technical advances and a rising number of data and system providers have eased remote sensing data collection and availability throughout the past decade. Today they are widely available and affordable for research and practical applications. In this investigation, we present the potential of airborne optical sensors to map snow depth and different snow types such as windblown snow at the surface with high spatial resolution (5 200 cm) in the mountainous region of Davos, Switzerland. We discuss advantages and disadvantages of different available measurement and data processing approaches. A key for mapping snowpack properties with optical sensors is the application of near infrared bands (wavelength 780 1400 nm). In these wavelength snow absorbs more solar radiation than in the visible wavelength (380 780 nm) and is sensitive to snow grain size. This is an advantage for the generation of digital surface models (DSM) using structure-from-motion photogrammetry. We validate the remotely sensed snow depth maps using independent snow depth measurements generated by hand probing and terrestrial laser scanning.