Key points are not available for this paper at this time.
Field teaching and research in the High Arctic is costly, both economically and in terms of environmental impact. In the Norwegian archipelago of Svalbard (74-81N, 15-35E), the field season is defined by the extreme annual daylight cycle, with a 4 month long polar night and midnight sun from late April to late August. The University Centre in Svalbard (UNIS) in Longyearbyen serves as Norways field university. Arctic Geology courses at UNIS make use of the excellent vegetation-free outcrops exhibiting various lithologies and tectono-magmatic structures. At UNIS, geological fieldwork is mostly conducted in the snow free summer with focus on coastal localities accessible by hiking and boats. The snow-covered spring season in March and April allows access to inland locations on snowmobiles, but only steep cliffs are snow-free and accessible for investigations. Irrespective of the time of the year, the harsh Arctic weather conditions, presence of wildlife (i.e. polar bears) and safe access to outcrops (e.g., strenuous hiking in steep terrain) routinely requires adjusting the field plans. In addition, the short field season(s) often only allow relatively short single field site visits. To make most efficient use of the field excursions, we have since 2016 built and developed Svalbox an effort to systematically digitize Svalbards outcrops through digital outcrop models (DOMs) and photospheres integrated with other geoscientific data sets (maps, geophysical data, terrain models, borehole data etc.). Starting as an educational tool at UNIS to facilitate year-round (digital) fieldwork and quantitative analyses on outcrops, Svalbox has in recent years become an important resource for the wider geoscientific community. Svalbox comprises three main elements; 1) A database of freely accessible DOMs. 2) An open portal for visualising drone-based virtual field trips (VFTs) and 3) Thematic data packages integrating various data sets for UNIS courses or specific research projects. In this contribution, we present all these three Svalbox elements. The database itself, visualised via www.svalbox.no/map, offers a growing number of DOMs and photospheres (i.e. 360 images) from all over Svalbard. Both incorporate georeferenced photographs acquired with consumer cameras and drones (DJI Mavic 2 and 3 in particular). The DOMs are processed with structure-from-motion photogrammetry using the Metashape software. DOMs and photospheres are fully downloadable, including the processing reports and all associated input data , facilitating reprocessing if needed. Curated data packages are integrated as VFTs and are accessible through a dedicated web portal, www.vrsvalbard.com/map with thematic groupings for specific UNIS courses and larger-scale research projects. Thematic data packages are generated from UNIS-internal databases (onshore and offshore seismic data, boreholes, digitized maps, cross sections, and sedimentary logs in publications etc.) and openly available datasets (digital terrain models, bathymetry, geophysical grids, maps etc.) where data are spatially connected in single software projects (e.g., Petrel, GPlates) for specific courses or projects. At this stage, we strive for expanded usage of Svalbox beyond UNIS. In particular, we invite the geoscience and data analytics community to use the exponentially growing number of DOMs to test and train machine learning algorithms for (semi-)automatic interpretation of DOMs.
Senger et al. (Fri,) studied this question.