WOBO acknowledges the information and links provided by Fathom to support professional learning and development.

Artificial intelligence, whichever way you feel about it and whether you are aware of it or not, is an integral part of our daily lives. It can, apparently, do everything humans can do, only better and faster, from diagnosing illness and driving trucks to updating your CV when you’ve just lost your job to a chat bot.

With flood risk growing thanks to climate change and increasing urbanization, the race is on to apply AI and its subset, machine learning (ML), to the problem of predicting and managing flood.

Because of its ability to process massive datasets without the high computational cost, ML has the potential to transform flood modeling, filling in data gaps and helping reduce uncertainties.

Fathom’s new research into a ~30m digital elevation model, FathomDEM, uses a cutting-edge hybrid computer-vision approach to correct inaccuracies in data and arrive at a more realistic representation of the Earth’s terrain. With firm guardrails and principles to guide it, ML is a useful addition to our flood modeling toolbox.

However, it can never be a silver bullet.

One of the limitations of the most popular forms of ML is that it functions like a ‘black box’. You input data and the model “learns” from it to arrive at an output. But it can be difficult to decipher how it arrived at that output and, therefore, whether the results can be trusted. Ultimately, fully understanding its limitations and appropriate techniques to solve specific, well-understood problems is the key to harnessing the enormous power of ML.

To find out more about machine learning and how we apply it to Fathom’s products, join our upcoming webinar, An Introduction to Machine Learning on April 29, 2025.

In partnership with CIWEM, Fathom is pleased to invite you to our new CPD-accredited* three-part series exploring the transformative role of machine learning in flood modeling.

We’ll highlight how machine learning improves uncertainty management, enhances modeling accuracy and supports early decision-making for adaptation and resilience.

Part one – An introduction to machine learning

Kickstarting the series, we’ll explore machine learning –what it is, how it works and why it’s gaining traction.

We’ll cover its fundamentals, key principles, strengths and limitations, along with its applications in flood and climate modeling, as well as emerging trends and future uses.  Register for part one

Part two – The future of terrain data

Join us as we dive into how machine learning is reshaping the way we model our planet.

We’ll explore the intersection of machine learning and terrain data, revealing how cutting-edge techniques are reshaping the way we analyze and utilize geographical information. Register for part two

Part three – Unlocking the power of global flood maps

Our final webinar in the series will share how machine learning is rapidly advancing flood and climate modeling.We’ll examine the integration of machine learning and hydrology, focusing on global-scale flood and climate modeling.  Register for part three

If you have any questions about the content presented or would like a more detailed discussion about our Global Flood Map with a member of the team, please get in touch here.