Machine Learning for Hydrological Impacts Mapping
This course introduces the application of Machine Learning (ML) techniques for mapping and predicting hydrological and environmental impacts such as soil erosion, flood susceptibility, and groundwater variations. Participants will learn how to collect, preprocess, and integrate geospatial and remote sensing data into predictive models. Through practical exercises, the course will demonstrate how ML algorithms, particularly Random Forest and other ensemble methods, can be used to generate high-resolution impact maps and support sustainable water and land management. The training emphasizes the methodological workflow: data preparation, feature selection, model training, validation, and map generation. By the end of the course, participants will be able to design and implement ML-based mapping workflows adapted to hydrological and environmental challenges in data-scarce regions.

Integrating LiDAR Data for Monitoring High-Resolution Elevation Changes
This course covers the principles, tools, and advanced techniques of terrain digitization using airborne LiDAR and photogrammetry. It emphasizes the acquisition, processing, and 3D modeling of geospatial data to depict and analyze natural and built environments. Participants will learn how to handle point clouds, generate Digital Terrain Models (DTMs), and create detailed visualizations for various environmental and engineering uses. The training also explores open-source workflows rooted in Structure from Motion (SfM) photogrammetry, multi-source data fusion combining LiDAR, aerial imagery, and satellite data, and the application of Machine Learning for extracting and classifying environmental objects from LiDAR. Through practical exercises, participants will develop skills in planning missions and creating end-to-end workflows for 3D data modeling and visualization.

Hydrological forecasting and the Global Flood Awareness System (GloFAS) from the Copernicus Emergency Management Service
This course introduces the use of ensemble hydrological forecasting for flood risk prediction, based on the Global Flood Awareness System (GloFAS) from the Copernicus Emergency Management Service of the European Commission. The training is designed to be interactive and hands-on, alternating between presentations and practical exercises using the GloFAS Interactive System and Jupyter notebooks. Participants will learn how GloFAS global hydrological forecast are generated, how the forcast products portfolio can be used for example to prepare flood forecast bulletins for disaster risk reduction, what type of data are freely available for operational and research use and how they can be accessed. Throughout the session, participants will practice mapping gauge stations to the GloFAS river network, accessing, downloading, and analysing GloFAS data from the Copernicus Early Warning Data Store (EWDS), and assessing the forecast performance. By the end of the course, the participants will be able to design and implement their own workflow to produce hydrological forecast products tailored to their needs from the GloFAS predictions

Use of Radar Imagery for Monitoring of Ungauged Rivers
This course provides the specialized skills to measure water levels in remote, ungauged rivers using satellite radar altimetry. You will learn to process and analyze satellite data to create hydrologic insights where ground-based monitoring is impossible. The curriculum covers the core principles of altimetry, the technique for establishing and validating Virtual Stations (VS) at river crossings, and the methods to transform raw data into reliable water level time series. Empower yourself to perform discharge estimation and robust hydrologic analysis in any basin, anywhere.

Deep Learning for Hydrological Time Series Prediction
This course provides a practical introduction to the use of Deep Learning, particularly Long Short-Term Memory (LSTM) networks, for the prediction of hydrological time series. It is designed for professionals, PhD candidates, and students with basic knowledge of AI and Python. At first, participants will explore the key challenges and fundamental concepts of time series modeling, as well as the reasons for applying neural networks in hydrology. Then, they will learn the complete theoretical approach, including data preprocessing, LSTM model construction, and performance evaluation using indicators such as R², RMSE, and NSE. Finally, a guided hands-on session will allow participants to apply these concepts to a real case study using a Python notebook (Google Colab), with result visualization, interpretation, and manual optimization. By the end of the training, participants will be able to design, train, and validate a simple LSTM model while understanding its limitations and potential improvements.