Machine Learning for Environmental research, Extreme Events and beyond

Hydrological Events

Accurate forecasts of precipitation is the foundation of countless use cases revolving around water management, from civil and disaster management to  commercial applications such as agriculture. The AML group has expertise in modeling hydrological data, across various forecast horizons. Our state-of-the-art ML models leverage the potential of multimodal data from a wide range of sources to gain novel insight.

GeoAI/Remote Sensing

GeoAI is an interdisciplinary subject of AI and geoscience, which is using AI models to solve problems in geoscience. The general AI models are required to be re-designed to meet the geo-tasks (e.g., applicable for various types of training data). Remote sensing images are common geo-data that record the multiple bands of observed information in the study area. The AML group has developed AI models to classify scenes based on remote sensing images.

Air Quality & Heat Islands

Air quality

AI-based air pollution modeling for the purpose of real-time display, prediction, and simulation is one of the AML Group's areas of expertise. Within the mFund-funded SAUBER project, a spatially high-resolution real-time system for visualizing air pollution levels was developed using the example of a region in North Rhine-Westphalia and the city of Stuttgart. This system enables 48-hour forecasts and the simulation of the effects of structural changes on pollution levels.

Heat islands

Another area of expertise is calculating a spatially high-resolution temperature distribution (10mx10m) from a bundle of data, thus determining and predicting heat islands in real-time and simulating them for urban planning purposes. The sensor network was installed in the two cities of Dresden and Langenfeld. Combined AI-based methods for time series analysis (RNNs) and structure-based spatial interpolation (attention mechanism) are used. 


  1. Petry, Lisanne und Herold, Hendrik und Meinel, Gotthard und Meiers, Thomas und Müller, Inken und Kalusche, Elenav und Erbertseder, Thilo und Taubenböck, Hannes und Zaunseder, Elaine und Srinivasan, Vignesh und Osman, Ahmed Mohamed Magdi Mohamed und Weber, Beatrix und Jäger, Stefan und Mayer, Christian und Gengenbach, Christian (2020). Air Quality Monitoring and Data Management in Germany - Status Quo and Suggestions for Improvement. In: The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, XLIV-4, Seiten 37-43. 5th International Conference on Smart Data and Smart Cities, 30. Sept. - 2. Oct. 2020, Nice, France. DOI: 10.5194/isprs-archives-XLIV-4-W2-2020-37-2020 ISSN 1682-1750
  2. Petry, L., Meiers, T., Reuschenberg, D., Mirzavand Borujeni, S., Arndt, J., Odenthal, L., Erbertseder, T., Taubenböck, H., Müller, I., Kalusche, E., Weber, B., Kaeflein, J., Mayer, C., Meinel, G., Gengenbach, C., & Herold, H. (2021). Design and Results of an AI-Based Forecasting of Air Pollutants for Smart Cities. ISPRS Annals of the Photogrammetry, Remote Sensing and Spatial Information Sciences, VIII-4/W1-2021, 89–96.
  3. Ximeng Cheng, Marc Vischer, Zachary Schellin, Leila Arras, Monique Kuglitsch, Wojciech Samek, Jackie Ma (2023). Explainability in GeoAI, In: Handbook of Geospatial Artificial Intelligence, CRC Press, Boca Raton, USA, pp. 177-201, December 2023, doi: 10.1201/9781003308423-9, Book Chapter