Many of us are experiencing it right now: Europe is suffering from a long lasting large scale drought (Fig. 1), with record heat and low-flow levels that probably nobody of us has seen before; and current forecasts do not announce any relieve for the coming weeks. The pictures of record low levels in the river Rhine are scaring – knowing how strongly Europe depends on this fluvial transport ...[Read More]
What machine learning can and cannot tell us about floods
Floods are shaped by complex and nonlinear interactions between weather patterns, rainfall, soil properties, topography, land cover, and how wet a catchment already is. This complexity is one reason machine learning (ML) is increasingly used in hydrology. ML models can learn patterns from large and complex datasets with many variables. Early in my PhD, while exploring machine learning models for ...[Read More]
Communicating and Managing Residual Risk with Perfect Storms and Other Counterfactual Stories
The aim of risk management is to prepare society in order to limit loss and damage when an extreme event occurs and to restore the functioning of society afterwards . While current risk management practices are helpful in many regards, they fall short when it comes to unprecedented events. An analysis of event pairs and recent disasters show that societies often fail to cope with events that are l ...[Read More]
Can Machine Learning Help Us Monitor Streams?
Picture this: you’re hiking through a dry landscape when suddenly you hear it—the serene sound of a babbling brook. You round a corner and discover a small waterfall cascading into crystal-clear pools, surrounded by lush green ferns and wildflowers attracting buzzing bees. It feels like stumbling upon a secret oasis. These magical streams that appear and disappear with the seasons are called ...[Read More]