Workshop on Mining Imaging Data
for Hydrological and Environmental Modelling

  Keynote Address

"From Fields to Fires to Fens: Integrating Deep Learning and Multi-Sensor Earth Observation for Global Environmental Resilience"

Prof. Kevin Tansey · University of Leicester, UK

Abstract

The convergence of climate change and population growth necessitates advanced Earth Observation (EO) frameworks to monitor terrestrial ecosystems and ensure food security. This talk explores the evolution of remote sensing methodologies — transitioning from traditional statistical models to sophisticated deep learning architectures — to address three critical environmental pillars: agricultural productivity, wildfire dynamics, and peatland preservation.

In the agricultural domain, the research takes two pathways. The first is a multi-scale deep learning framework integrating time-series satellite data, meteorological variables, and soil characteristics. Using models such as LSTM and neural networks, high-precision county-level yield estimates of crops such as winter wheat are achieved. The second pathway, developed with industry partners in the UK and Nigeria, predicts crop status, harvest windows, and crop type for the determination of crop waste availability for biofuel through a pyrolysis process.

The increasing frequency of extreme weather events has triggered an increase in wildfire activity outside of normal fire seasons. Work on automated processing pipelines using high-resolution PlanetScope and Sentinel imagery to characterise burned area will be discussed. Using adaptive machine learning classifications, 3-metre resolution burned-area maps are generated, enabling rapid response and long-term ecological impact assessment.

An overview of findings from the ESA WorldPeatland project — an international initiative dedicated to EO-based mapping and monitoring of peatland hydrology — will be presented. Peatlands are vital carbon sinks, yet highly vulnerable to drainage and fire. The catalogue of products created by the project team and insights into peatland extent, hydrology, and surface motion will be discussed.

Finally, Prof. Tansey will outline the changing direction of remote sensing and Earth observation research as seen through his role as Editor-in-Chief of the International Journal of Remote Sensing, and the challenges faced by the community regarding the use and misuse of AI in scientific publishing.

About the Speaker

Kevin Tansey is a Geographer who has worked as a remote sensing scientist since 1995. He has published 112 papers in international peer-reviewed journals across application fields including agriculture, wildfire and hazards assessment, forestry, and renewable energies. He works with all forms of EO data — optical, thermal, radar, and LiDAR — from the ground to space.

He is Editor-in-Chief of the International Journal of Remote Sensing. He has previously held positions on the United Nations World Meteorological Organization (WMO) Terrestrial Observations Panel for Climate (TOPC) and has been co-chair of the Committee of Earth Observing Satellites (CEOS) Land Product Validation sub-group on fire disturbance. He has supervised more than 25 PhD students and has over 20 years of teaching experience in remote sensing, GIS, and statistics for data science.

  Workshop Organizer

Mourad Oussalah

Full Professor · University of Oulu, Finland

Leads the Social Mining Research Group at the University of Oulu. Expert in computer vision, machine learning, and remote sensing for environmental monitoring, with over 300 publications. Involved in the Finnish Research Council Digital Water Flagship and Hybrid Intelligence Flagship.

  Workshop Organizer

Olof Mogren

Research Director · RISE Research Institute of Sweden

Works on deep learning and foundation models for climate adaptation and environmental monitoring. Published at ICCV and CVPR. Leads Climate AI Nordic, a Nordic initiative promoting AI research in climate change and environmental process monitoring.

  Workshop Organizer

Jukka Heikkonen

Full Professor · University of Turku, Finland

Leads the Algorithms and Computational Intelligence Research Group at the University of Turku. Expert at the intersection of computer vision and GIS data analysis for geographical applications and autonomous systems, with over 200 publications. Involved in the Finnish Research Council Digital Water Flagship.

  Workshop Organizer

Getnet Demil

Doctoral Researcher · University of Oulu, Finland

PhD researcher at the University of Oulu, specializing in GeoAI, remote sensing, and AI-driven hydrological modeling. His research involves creating deep learning frameworks to estimate snow hydrology parameters using Earth observation data from multiple sensors. He is dedicated to merging multi-sensor imagery to enhance the accuracy of snow cover fraction and snow water equivalent estimates, which are crucial for hydrological modeling and analyzing climate impacts.

Research Interests

• AI for hydrological modeling
• Multi-temporal and multi-sensor data fusion
• Snow water equivalent and snow cover estimation
• Foundation models for Earth observation
• Climate-focused computer vision

  Workshop Organizer

Farhan Humayun

Doctoral Researcher · University of Turku, Finland

Works on computer vision and SAR-based remote sensing, with focus on object detection in multi-resolution SAR satellite imagery and UAV-based orthomosaic generation. Broader interests include deep learning for geospatial analysis and autonomous navigation.