Managing River levels with the help of Data Science

Critically low river levels in Europe's major rivers are putting vital shipping routes in serious trouble. Can data science help?

As summer 2022 continues to grip the European continent with exceptionally hot and dry weather, its major rivers are experiencing unprecedented low flows. From an ecological perspective it can be distressing to see nearly dry riverbeds, but major rivers such as the Rhine are also important commercial shipping routes that can be severely disrupted when river levels drop to historic lows.

As the final downstream recipients of the river Rhine, the Netherlands - home to some of the world's busiest waterway networks - is quite familiar with the challenge of both too little river water at times of drought, as well as too much during times of flooding.

In 2016 HAL24K responded to a data challenge set by The Netherlands water agency, Rijkswaterstaat, to predict river levels across the run of the river as it crosses the Netherlands. The Challenge - to predict the effects and timing of remedial actions they might take to remove or hold back water to manage the river during extremes without needing to rely on information from the nations upstream.

The competition was won by HAL24K (of course!) and was an early stage demonstration of the real potential for considering a data science approach to environmental management where normally a traditional hydrological modelling approach might have been the only option. The data science approach did not need to model the complex riverbed profile nor make finite element calculations of the river sections that might be more common in traditional techniques. Instead a data model was produced that was able to accurately predict levels and consequences of interventions by statistical analysis of combined large historic data sets. The model was also able to consider river water quality predictions, a correlation that fits comfortably together when approaching from a data science perspective.

Some six years later and the use of data science and data modelling to predict our environment continues to grow exponentially as managers like Rijkwaterstaat seek to unlock the hidden potential in the years of data they already have.

At HAL24K Water we love to work on predictive data models like this and are proud of doing our bit to support organisations find the best ways of working out how to live with the effects of climate change.

Why not get in touch to run through your own challenge, maybe there's a data science-led answer out there that can help you.

#watermanagement #climatechange #datascience