In 2023, the project Algorithmic determination of water levels with Remote Sensing received seed money from DigiShape. It was the start of a search for a new way to visualize water levels using lidar data. The Hollands Noorderkwartier Water Board (HHNK) took this up, together with TU Delft and other partners. A toolbox has now been developed and further developed, but the step to structural application is still pending.
From bycatch to toolbox
The reason was a concrete bottleneck. There are several water level areas in North Holland, some of which have a permit to lower the water level. These sub-drainages are often small ditches on farmyards, where there are no sensors. There is no supervision, while the risks can be high if the water is structurally too low.
Lars Tesselaar, geo-developer at the Hollands Noorderkwartier Water Board, was asked if remote sensing could help. Coincidentally, in his study, he came across lidar, a technique that collects elevation data with laser pulses. “Actually, lidar is not intended for water, because the points bounce off. But with drone measurements, especially at banks, we saw that there were still useful signals. Enough to be able to say something about the water level, with ifs and buts.”
The idea was further developed with seed money from DigiShape. Together with TU Delft and partners such as Rijkswaterstaat and Wetterskip Fryslân, a toolbox was created that makes it possible to estimate water levels.
Further development at HHNK
After that first phase, HHNK picked up the thread itself. All water level areas have been calculated with AHN data, the Current Altitude File Netherlands. That provides useful material. “For a lot of places, you don’t know the water level. With interpolation you can still make a covering picture. The result can eventually be used by water consultants who want to know what a water level would be somewhere.”
However, the application is still mainly in the experimentation phase. The toolbox is not used as standard in the work processes. “We have something beautiful on the shelf, but it often takes a while before it really lands in the organization.”
Opportunities for collaboration
Lars is positive about the future. “The toolbox is on GitHub and is therefore available to other water boards. We just share that, I think that’s logical.” The experiment continues at HHNK and he sees more opportunities if other organizations join.
“We now have experience with setting up the tool and running it on a fairly large scale. But it only becomes really interesting when the tool is actually embedded in the processes. If another water board can already do that or can do it faster, it can accelerate the entire development. And that would also be very valuable to us.”
Lars sketches a picture of the future that makes the application tangible: “The idea is that you test digitally and get a kind of traffic light system. Green means: this is fine, red means: this is where we have to look. Then an inspector no longer has to visit an entire area, but can set off in a targeted manner. That saves time and makes the work more effective.”
A joint approach can also reduce costs. Aerial photos with lidar are expensive, but if several parties such as Rijkswaterstaat, the province and water boards purchase data together, the costs can be shared.
Potential, but not yet fully developed
Where is the project now? Lars is down-to-earth: “We are working on it. It can fail, it can be a success. We don’t know yet. What we do know is that the potential is great. Once it’s widely adopted, you can really take steps with improvements that users find necessary. Sometimes it only takes something small to happen and then the project gains momentum again.”
In this way, the project clearly shows what DigiShape can mean: driving innovation with seed money and collaboration, even if the outcome is not yet certain.
How do you measure water levels remotely?
The project combines several techniques:
- Lidar (Light Detection and Ranging): laser pulses from aircraft or drone that are normally used for altitude maps. By chance, it turned out that it also contained useful signals from the water surface.
- AHN (Current Altitude File Netherlands): national dataset with lidar measurements, collected every three years from an aircraft.
- Drone images: suitable for smaller areas, provides very detailed data but is impractical on a large scale.
- TMX sensors: fixed measuring points in watercourses that register current water levels.
- Interpolation: statistical methods that combine different data sources into a smooth result.
By cleverly combining these sources, an increasingly reliable picture of water levels is created, even in areas where there are no sensors.
