Ph.D. in Geography and Cartography, Meteorology, Geographical and meteorological support at University of Defence, Czech Republic (2021)
M.Sc. in Geography and Cartography, Meteorology, Geographical and meteorological support at University of Defence, Czech Republic (2017)
Machine Learning in Meteorology, Data Quality and Forecast Verification, Automated Weather Systems, Geospatial & Climatic Analysis
Duration
4 months, January 2026 - May 2026
Host at the University of Passau
Prof. Dr. Matthias Kranz
(Faculty of Computer Science and Mathematics)
The increasing integration of drones necessitates highly accurate and tailored weather information for safe and efficient operations. The forecast must be reliable and issued immediately, especially during critical scenarios (e.g. natural disasters). Dr. Sládek's research addresses the critical challenge of error propagation in automated weather forecasting systems designed for drone applications and beyond. We decompose the forecasting process into key stages:
1. Input Data Sources (observations, numerical and AI models, human analyses),
2. Automated Processing Pipelines (AI-based/ traditional forecasting methodologies),
3. Nowcasting and Assimilation (incorporation of near real-time observations to update and refine forecasts post-issuance),
4. Information Delivery (visualization and data interfaces for optimal decision-making).
The primary objective is to quantify the inherent uncertainty and error characteristics of individual data sources and to model their cumulative impact throughout the automated forecasting workflow. To achieve this, these will be employed:
1. a detailed statistical analysis of historical meteorological observations at selected European location,
2. the development and comparative evaluation of diverse modelling approaches leveraging observations, numerical models, and supplementary datasets,
3. the application of Explainable AI methodologies to identify and interpret the data sources that contribute most significantly to forecast errors.
System development will require robust database management, spatiotemporal meteorological data, efficient data storage optimized for real-time processing, intuitive visualization interfaces, and cybersecurity protocols to ensure data integrity and system resilience. This research aims to establish a framework for a reliable and objective weather forecasting system capable of providing high-quality, drone-specific meteorological support with potential for widespread application across European drone-based industries.
