How do changes in the water cycle affect biodiversity? What information do we need to identify threatened ecosystems at an early stage? And how can artificial intelligence help us translate scientific findings into concrete conservation measures?

BioWaWi-AI takes an interdisciplinary research approach that combines an understanding of natural processes, model development, and practical application. The focus is on five research questions.

Water Cycle and Biodiversity — What Processes Determine the Condition of Our Ecosystems?

How do meteorological conditions, soil properties, groundwater, and land use influence water availability and, consequently, the biodiversity of different habitats?

Water availability is a critical habitat factor for many plant and animal species. However, it depends not only on precipitation but also on the water-holding capacity of the soil, topography, vegetation, and the interactions between soil and groundwater.
BioWaWi-KI investigates these relationships using an extensive environmental monitoring network and detailed studies at selected intensive monitoring sites. In particular, the project examines the spatial and temporal dynamics of soil moisture—ranging from small-scale differences in the root zone to the water balance of entire watersheds. The measurements are integrated with pedological analyses as well as floristic and faunistic surveys.
Particular emphasis is placed on moisture-dependent biotopes and indicator species, whose occurrence and development allow conclusions to be drawn about changes in the water balance. The target is to identify the key environmental factors and interactions that determine the stability of sensitive habitats.

Detecting Changes Early — Can AI Warn Us About Drought Stress and Biodiversity Loss?

Can critical changes in the water balance and vegetation be detected before they lead to a permanent loss of species and habitats?

Drought stress often develops gradually. While short-term changes in soil moisture are already measurable, effects on vegetation or species composition sometimes do not become apparent until much later. BioWaWi-KI is therefore investigating which measurable environmental changes can serve as early indicators of threats to biotopes and species.
To this end, continuous soil moisture and climate measurements are combined with vegetation surveys, automated imagery, and satellite-based vegetation indices such as the Normalized Difference Vegetation Index (NDVI). With the help of AI, the project aims, for example, to investigate whether changes in vegetation structure, species composition, or seasonal vegetation development can be linked to water availability.
Based on this, the project aims to identify critical thresholds and characteristic patterns that indicate increasing drought stress. The long-term target is an AI-supported early-warning system that identifies at-risk habitats and species as early as possible, thereby facilitating timely conservation measures.

Intelligently Linking Environmental Data — How Can Spatial and Temporal Data Gaps Be Filled?

How can AI methods use spot measurements, time series, photographs, and satellite data to generate the most comprehensive picture possible of the state of the environment?

Research on ecosystems poses unique challenges for data analysis. While weather stations and soil moisture sensors provide continuous measurements at individual locations, satellites cover larger areas but with different temporal and spatial resolutions. Added to this are species surveys, habitat maps, and photographs, which in turn contain very different types of information.
BioWaWi-AI investigates how these heterogeneous and, in some cases, incomplete datasets can be linked together using modern AI methods. A key focus is on the spatial and temporal interpolation of environmental parameters, particularly soil moisture. The research explores the extent to which AI can derive reliable conclusions about areas not directly studied based on available measurements and additional site information.
The automated analysis of image data also plays an important role. The research examines whether changes in vegetation and habitat condition can be quantified using recurring photographs and remote sensing data.
Another research question concerns the optimization of future monitoring strategies: Which environmental parameters must be recorded at what spatial and temporal resolution so that AI models can deliver the most reliable results possible?

Prozesse modellieren und Zukunftsszenarien entwickeln – was passiert, wenn sich Umweltbedingungen verändern?

Wie lassen sich die Auswirkungen von Klimawandel, Landnutzungsänderungen und wasserwirtschaftlichen Maßnahmen auf den Wasserhaushalt und die Biodiversität vorhersagen?

Aufbauend auf dem gewonnenen Prozessverständnis werden bestehende physikalisch basierte Wasserhaushalts- und Grundwassermodelle weiterentwickelt und mit KI-Verfahren verknüpft. Die Prozessmodelle beschreiben unter anderem die Grundwasserströmung, die Grundwasserneubildung und den Wassertransport im Boden. KI soll helfen, komplexe Zusammenhänge in den verfügbaren Umweltdaten zu erkennen und die Berechnung unterschiedlicher Szenarien zu beschleunigen.
Ein zentraler Forschungsansatz sind sogenannte Was-wäre-wenn-Szenarien (What-if-Szenarien). Dabei werden gezielt einzelne Einflussfaktoren verändert, um ihre möglichen Auswirkungen auf den Wasserhaushalt und die Lebensbedingungen empfindlicher Arten und Biotope zu untersuchen.
Beispielsweise soll abgeschätzt werden, wie sich längere Trockenperioden, veränderte Grundwasserstände, forstwirtschaftliche Eingriffe oder gezielte Bewässerungsmaßnahmen auf die Wasserverfügbarkeit auswirken könnten. Ein besonderes Potenzial liegt darin, aufwendige numerische Simulationen durch KI-basierte Ersatzmodelle zu ergänzen, um unterschiedliche Entwicklungen und Handlungsoptionen schneller vergleichen zu können.

Ziel ist es, wissenschaftlich fundierte Prognosen zu entwickeln und mögliche Folgen von Umweltveränderungen bereits im Vorfeld bewerten zu können.

Translating Research into Practice — How Does Knowledge Lead to Concrete Biodiversity Conservation?

How can AI-based forecasts and scientific findings be translated into effective measures for the management of water protection areas?


The methods developed are intended not only to provide new scientific insights but also to be applicable in water management practice. In collaboration with Stadtwerke Bühl, we are investigating how monitoring data, model calculations, and AI-based forecasts can be integrated into an early-warning and decision-support system.
A key component is the establishment of a digital control center that provides current environmental measurements and, in the future, model and forecast results for the assessment of water protection areas. Based on this, recommendations for action will be developed—for example, to protect biotopes that are particularly vulnerable to drought or to evaluate potential adaptation measures.
Regional stakeholders, government agencies, schools, and the public will also be involved in the project. Their experiences, requirements, and perspectives will help to present scientific findings in an accessible manner and develop practical measures.
Another target is to integrate suitable results into Stadtwerke Bühl’s environmental management system in accordance with ISO 14001. At the same time, the project will examine under what conditions the developed methods and concepts can be applied to other water protection areas.