Research

Society faces the urgent challenge of developing sustainable, equitable, and scalable solutions to climate change, driven by the critical need for evidence-based management of geo-ecological life support systems like carbon, water, and heat. To address this need, we will develop scale-aware benchmarks for land carbon sequestration by leveraging atmospheric CO2 flux ground truth data alongside a wide range of in-situ and remotely sensed earth observations. We will provide an independent, impartial and actionable benchmark with uncertainty estimates to link activity-based bottom-up inventories and top-down atmospheric concentration inversions. This benchmark is essential for accurate global-to-ownership-level carbon accounting, and unlocking the development and improvement of effective nature-based climate solutions. Our project will initiate a paradigm shift from the current practice of scale-agnostic data joins to a scale-aware methodology from local to global levels. This new machine learning benchmark will maximize the use of scarce flux ground truth observations, consistently integrate multi-source earth observations, increase statistical power, and reconcile flux estimates across scales.

Vegetation mapping is an essential component in the domain of nature and environmental protection. Traditional approaches, aligned with current guidelines, necessitate high research specifications typically fulfilled through terrain mapping efforts. Despite its efficacy, this method encounters limitations in terms of seasonality and the timely processing of extensive areas. In contrast, remote sensing-based models offer noteworthy advantages under their season-independence and rapid large-scale processing capabilities. The present initiative seeks to leverage new technologies, such as cloud computing, to augment conventional supervised remote sensing classifications by merging ecological expertise with the sophisticated capabilities of cloud computing technology. For a few years now, the “Google Earth Engine” (GEE) platform has made it possible to carry out geospatial processing data and analyze them based on a huge time satellite imagery series at large study area in combination with multivariate statistical methods. It also enables the integration of location and laser scan data as well as geospatial information systems data. Challenges and Research Requirements in the project: Integrating the processing chain into a cloud platform poses considerable challenges and necessitates extensive research for a coherent, smooth, and consistent adaptation. From the initial model selection to the subsequent post-processing phase, passing through exhaustive feature selection analysis and model evaluation, which are crucial phases of adaptation, require exhaustive research to ensure the accomplishment of expected outcomes in the project. Recognizing the complexity of this task, the Egger Natural Space Planning Company requires the expertise of a remote sensing scientist specializing in cloud computing platforms. The scientist, their expertise in the field, and their scientific input are vital for providing a perspective and conducting the needed research to identify optimal approaches that align with the project's expected outputs.

Groundwater recharge in forest areas: in parts of Austria - specifically Weilhartsforst - it has been shown that annual precipitation totals are decreasing (e.g. Braunau area). The aim of this study is to analyse the impact of this on groundwater recharge, especially when there is no significant surface water supply. For the future water management utilisation of such groundwater bodies, the complex interaction between percolating precipitation and underground groundwater inflows - taking into account different forms of land use such as forest, meadows and arable land - is to be investigated using a specific area (Weilhartsfort in the Braunau area).