Abstract
The EnvironTwin project seeks to enrich the Environmental Data Platform (EDP) [1] by implementing a Digital Twin (DT) service to represent, model, and forecast key alpine environmental scenarios. EDP was previously developed for the FAIR management of environmental data resources at Eurac Research, to provide stakeholders with actionable insights into ecosystem dynamics, risks, and management strategies. EnvironTwin leverages technologies such as in-situ sensors, proximal sensing, satellite imagery, cloud computing, and advanced data modeling. The project addresses critical challenges posed by human activities and their impacts on agriculture, forestry, and environmental conservation as well as some limitations of technology. Objectives: The project aims to: (ι) Commission advanced instrumentation and sensor technology for environmental monitoring and digital shadowing. (ιι) Establish and integrate computing infrastructure within the existing Environmental Data Platform. (ιιι) Implement a Continuous Integration and Continuous Development (CI/CD) environment to streamline digital twin creation. (ιν) Combine heterogeneous data sources into a unified framework for robust digital twin modeling. Challenges: Mountain and Alpine environment monitoring systems face technological, organizational, and operational limitations. The sudden change in orography, the weather variability, different climate zones, and a wide variety of human activities make it difficult to use one single monitoring strategy. EnvironTwin integrates different natures of instrumented systems, from ground to remote sensing approaches. However, one of the primary challenges is acquiring high-quality, heterogeneous data and building an operative infrastructure to effectively integrate these data into simulation models to suggest different possible scenarios in four key alpine use cases: Grasslands, Forestry, Agrovoltaic and Natural Hazards.. Above all, every use case is followed by a scientific supervisor, making environTwin a unique multi-disciplinary approach. Use Cases: (ι) Grassland management detection: Grasslands in South Tyrol are managed by small farming businesses. Grazing, mowing and harvesting, and fertilization are some of them. However, every farmer follows a different management strategy, making it challenging to generalize monitoring systems. EnvironTwin evaluates the effectiveness of high-resolution Planet satellite data in overcoming spatial and temporal monitoring events in South Tyrol. The optimization will be achieved by incorporating Sentinel-2 imagery and webcam-derived reference data, thereby refining the spatial and temporal resolution for agricultural applications. (ιι) Forest structural diversity and modeling: Ground and proximal sensing-based data on forest structural diversity foreseen the monitoring of individual trees distributed in a 1500-meter elevation profile. EnvironTwin integrates heterogeneous data sources into forest dynamics models, enabling predictions of forest adaptation and growth under changing climatic conditions. This data supports the creation of a digital twin of forests for current and future scenario adaptation. (ιιι) Agri-Voltaic Systems: By modeling the interactions between photovoltaic (PV) systems and agricultural practices, digital twins optimize dual land use for energy generation and crop cultivation. Climatic and weather variability, including droughts and extreme temperatures, are considered to ensure resource efficiency and sustainability. The main objective is to forecast energy and fruit production based on IoT data. (ιν) Rock glacier deformation: Traditional methods to monitor rock glacier deformation rely on the monitoring of medium-sized boulders using GPS. Nevertheless, the integration of proximal sensing (LiDAR, Thermal, and RGB) allows the possibility to identify hot spots, as well as monitor rotational, gravitational, material deposition, and geological structures at a very small scale. In this use case, these technologies are applied in the Lazaun Senales Valley, Italy, to extract environmental variables such as temperature and time. The collected data is incorporated into a digital twin (DT) model, which provides valuable insights for developing strategies to mitigate and manage natural hazard risks. Pilot Study: Alongside the four use cases planned for the project, a pilot case has been developed to test the infrastructure from which to start and then adapt to other application fields. The pilot study was set in the Laimburg field, South Tyrol in Italy, and it focuses on predicting soil moisture and precipitation using in situ sensors, satellite data (temperature, water, and soil moisture indices), and weather station observations. Algorithms such as LightGBM, CNNs, LSTMs, and regression models are applied to analyze data, predict irrigation needs, and enhance water management. Impact: EnvironTwin fosters collaboration among researchers, public authorities, and businesses by creating synergies with predictive modeling experts and showcasing the potential of digital twin technology through the research and development ecosystem. The project’s outcomes include scalable, data-driven tools for managing environmental challenges, validated through interdisciplinary case studies. By demonstrating its applicability across diverse fields, EnvironTwin establishes a foundation for sustainable, resource-efficient environmental management. The research leading to these results has received funding from the European Regional Development Fund, Operational Programme Investment for jobs and growth ERDF 2021-2027 under Project number ERDF1045 Service for the development of digital twins to predict environmental management scenarios through dynamic reconstruction of the Alpine environment, EnvironTwin. [1] FAIRsharing.org: EDP; Environmental Data Platform, DOI: 10.25504/FAIRsharing.e4268b, Last Edited: Monday, October 9th 2023, 18:54, Last Accessed: Wednesday, November 29th 2023, 10:19