Abstract. Increasingly advanced and affordable close-range sensing techniques are employed by an ever-broadening range of users, with varying competence and experience. In this context a method was tested that uses photogrammetry and classification by machine learning to divide a point cloud into different surface type classes. The study site is a peat scarp 20 metres long in the actively eroding river bank of the Rotmoos valley near Obergurgl, Austria. Imagery from near-infra red (NIR) and conventional (RGB) sensors, georeferenced with coordinates of targets surveyed with a total station, was used to create a point cloud using structure from motion and dense image matching. NIR and RGB information were merged into a single point cloud and 18 geometric features were extracted using three different radii (0.02 m, 0.05 m and 0.1 m) totalling 58 variables on which to apply the machine learning classification. Segments representing six classes, dry grass, green grass, peat, rock, snow and target, were extracted from the point cloud and split into a training set and a testing set. A Random Forest machine learning model was trained using machine learning packages in the R-CRAN environment. The overall classification accuracy and Kappa Index were 98% and 97% respectively. Rock, snow and target classes had the highest producer and user accuracies. Dry and green grass had the highest omission (1.9% and 5.6% respectively) and commission errors (3.3% and 3.4% respectively). Analysis of feature importance revealed that the spectral descriptors (NIR, R, G, B) were by far the most important determinants followed by verticality at 0.1 m radius.
<p>As the climate warms, the earths&#8217; cryosphere melts. Among the regions with the highest sensitivity to recent climate change are the high altitudes of the European Alps. This can be seen most clearly in the melting of glacier ice. Most glaciers show a strong receding trend since the last maximum extent during the little ice age (LIA) around AD 1850. When glaciers retreat, they leave behind a characteristic paraglacial landscape in a transient state from glacial to non-glacial conditions. Dominated by large amounts of unconsolidated glacial sediments they show an extremely high geomorphic activity.</p><p>However, these proglacial areas can still hold ice even decades after the glacier has left. In a simplified manner, this can be conceptually described by two main mechanisms: i) When glaciers retreat parts of the glacier front are often decoupled from the main glacier. These so-called dead ice bodies can remain for years, especially when they are buried by a thick debris cover and thus protected from atmospheric conditions. ii) Particularly in high-elevated glacier forefields, the thermal regime can be suitable for the direct transition from a glacial to a periglacial environment, compassing the aggradation of permafrost ice in areas that have been released from the glacier.</p><p>Climate warming speeds up in recent times, related with an enhanced receding of glaciers and growing alpine proglacial areas. Ground and dead ice are among the most important drivers of geomorphic activity in these regions, though in the long-term it is most likely, that it will melt out as well. How fast this will happen and in what stage it may play a role in stabilizing these environments is not yet fully clarified. Therefore, a better knowledge on ice distribution and dynamics in alpine proglacial regions is needed. Additionally, the quantification of ice and water contents is crucial in terms of potential hazardous processes, regarding the supply of (drinking) water and hydropower.</p><p>Here we present a new (PhD-) project in close cooperation with the DFG-funded research unit SEHAG, which is at the beginning of its implementation. Focussing on ground and dead ice we aim i) to assess the current distribution, ii) to reconstruct dynamics since the LIA, iii) to reveal recent and future trends (aggradation, degradation and persistence), and iv) to quantify effects on sediment dynamics in three Central Alpine proglacial areas. We combine different geophysical techniques with a focus on electrical resistivity tomography, water isotope analysis and ground (surface) temperature measurements with high-resolution geomorphic change modelling.</p>
During the last glacial maximum (LGM), large parts of the European Alps were occupied by an extensive interconnected system of valley glaciers. At lower elevations on the eastern fringe of the Alps, only a patchy pattern of isolated glaciers developed. Erosion by many of these smaller glaciers was however sufficient to excavate subglacial basins ('overdeepening') large enough to act as sedimentary archives of deglaciation processes. To constrain the deglaciation history and the post-LGM evolution of an isolated glacial system, we analyzed the geomorphology and the sedimentary record of a local overdeepening at the northern fringe of the Northern Calcareous Alps. We focus on the heavily silted lake system Taferlklaussee (TKS) (Höllengebirge massif, Austria) confined by a succession of terminal moraines and a steep headwall. Field techniques (mapping, Direct Current Resistivity, and core drilling), lab techniques (lithostratigraphic analyses and radiocarbon dating) and Geographic Information System (GIS) analyses are employed to document dynamics of erosion and sedimentation after deglaciation. We discuss drivers and controls of paraglacial landscape evolution. The almost complete postglacial (lake) record indicates abundant sediment input during the lateglacial and early Holocene, but strongly reduced dynamics since the Mid-Holocene. Sediment storage volumes range between 3.83 Mm³ and 5.75 Mm³ and lithology-specific mechanical denudation rates range between 0.8 mm/a and 1.7 mm/a depending on the scenario used. The study provides a well-constrained reference for the depositional dynamics of isolated LGM glacier systems of the Eastern Alps. It highlights the potential of using related sedimentary records to constrain the local variability in postglacial landscape evolution.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.