SEHAG 2 - Historic image processing Phase 2

SEHAG is a multi-partner project with the aim to understand the relation between processes in the geosystem of Alpine catchments and climate. While the first SEHAG project phase offered insights into the past, collecting and transforming historic data to draw conclusions on the former state of the geosystem, the second phase is dedicated to the future. Besides inferring information from photos and other sensors (e.g. LiDAR, GPR, SONAR), research will focus on the collaborative exploration and visualization of predicted changes in 2D- and 3D-models of our environment. In order to broaden the basis of available datasets, the collection and georeferencing of newly acquired and historic geodata will be continued. For historical terrestrial images, research on improved methods for the automatic orientation will increase automation, reliability, and accuracy. Besides airborne laser scanning (ALS), data acquired with additional sensors will be used to derive surface models in specific areas: (i) ground penetrating radar (GPR) to derive subglacial topography (ii) echo sounding for seafloor bathymetry. With the focus on the prediction of changes of our environment, the role of images is reversed in comparison to the first project phase: Instead of solely extracting information from images, information generated from models will be visualized. The generated visualizations are not merely a means of dissemination, but more importantly trigger scientific discussion on the interaction of processes. Focus will be laid on the evaluation and development of methods for the combined visualization of the predicted changes from various geoscience disciplines (glaciology, geomorphology including the cryospheric component, hydrology, and vegetation science). Furthermore, 3D georeferenced historical images will be used, to visually represent the connection between the past and future by projecting predicted features into the images. The Corona19-pandemic had drastic impacts on the way we communicated and interacted. In the new collab- oration forms induced by restrictions, we see a chance for sustainably transforming (parts of) our work. There- fore, new possibilities to collaboratively analyze, interact and interpret data, used within the SEHAG project, will be investigated. Special attention will be laid on the question how to remotely interact with each other and the data. This will be investigated by means of so called virtual field meetings where the basic elements of real world field meetings will be made available virtually. With the developed tool researchers from various research areas will be able to collaboratively analyze and interpret environmental past, current and future data and states remotely. While the need for such possibilities became apparent through the Corona-19 pandemic, they will be relevant in the future as well, supplementing and extending current fieldwork.

Within SEHAG, the influence of climate change onto alpine landscapes was investigated in three selected study areas (Kaunertal, Horlachtal - both Austria; Martelltall - Italy). Within this interdisciplinary research project, together with various research partners across Germany, we developed a new approach to derive accurate (3D) datasets describing both the current and historic conditions of the selected study areas. To represent the current conditions of landscapes in 3D, Airborne Laser Scanning was used. This required new processing pipelines to derive accurate 3D point clouds and digital models from the raw measurements acquired under challenging conditions (e.g. weather, topography). To describe the past conditions in 3D, historical aerial images, dating back as far as 1950, were used. Within SEHAG, optimized workflows for the processing of historical aerial images and the deviation of 3D models were developed that allowed us to reconstruct several snapshots for the selected study areas (i.e. Kaunertal - 1953|1969|1982|2003, Horlachtal - 1953|1973, Martelltal - 1996|1982). Together with the current ALS datasets, these snapshots formed the basis to quantify landscape changes over the last 75 years, which helped to address to overall research questions of quantifying the change in the geo-system. In addition to historical aerial photographs, terrestrial images taken by locals and mountaineers are another important source that provide us with additional insights prior to the first aerial images. However, in contrast to aerial images, no information (e.g. camera model, location, focal length) is known for collections of such images, which makes their processing very challenging. Accordingly, within SEHAG new methods were invented to automatically estimate the position and orientation of such images, facilitate derivation of the historic state by projecting the image content onto the terrain surface (monoplotting) and address the uncertainty of such results. All of these methods were integrated into our free software (monique - https://github.com/smfloery/moniQue), which was developed as QGIS plugin to make it available to any interested researcher. Beyond scientific exploitation, historical (terrestrial) images offer intuitive ways to communicate landscape changes with the public. To enhance this potential, we further explored and developed various ways to visualize historical terrestrial images together with current renderings of the landscape (represented through digital terrain models and orthophotos) to visually emphasize landscape changes. The potential of these visualizations for engaging the public was further investigated and used within a citizen science extension to this project funded by the FWF through its Top Citizen Science program. Through all of these activities, we substantially contributed to SEHAG through the optimized processing of ALS data, (automatic) processing of historical aerial and terrestrial images, explored intuitive ways to communicate landscape changes with the public and facilitated working with historical terrestrial images through the developed QGIS plugin.