Objectives of the service
Water levels in lakes and reservoirs drive hydropower planning, water supply and flood preparedness. Many lakes, especially small or remote ones, have no gauge. In Norway, installing a gauge station costs about 75,000 to 80,000 NOK, plus around 10,000 NOK a year for control visits. In Greenland, logistics is the main cost of running stations. Gauges that exist can also fail or drift without anyone noticing.
WAVES gives users a regular water level series for each lake, measured from space, with no field work. The service:
-
Measures levels on lakes from about 0.2 km² to over 300 km².
-
Merges all usable satellite passes into one series per lake, with an uncertainty for each value.
-
Gives levels in the national height system, so they can be compared directly with gauges.
-
Flags when a gauge and the satellite series disagree.
The Kick-start tests this with real users and real gauges in Norway and Greenland: how accurate the levels are, how often they update, and what users need before they pay for such a service
Users and their needs
The target users are hydropower producers and public water and survey authorities, first in Norway and Greenland, then in the other Nordic countries and the rest of Europe. Energy analysis houses are a further group.
Three organisations take part in the project:
-
NVE, the Norwegian Water Resources and Energy Directorate, provides gauge data and validates the results.
-
Lyse, a Norwegian hydropower producer, supplied seven reservoirs as test sites.
-
ASIAQ Greenland Survey supplied two lakes in Greenland.
What the users need:
-
Levels for lakes and reservoirs without a gauge.
-
An independent check when a gauge is faulty, drifting or being replaced.
-
Weekly updates or better.
-
A few centimetres of accuracy for some uses, while decimetres are enough for hydrological models and ungauged lakes.
-
One easy-to-use series per lake, with uncertainties, in their own height system and through a web service.
The main challenges are reservoirs with large drawdowns, where accuracy drops when the reservoir is low, and the millimetre accuracy needed for core hydrology, which satellites cannot reach.
Service/ system concept
For each lake, users get a time series of water level with an uncertainty and a quality flag for every value, in the national height system and on the global reference ellipsoid. Where the user has a gauge, the service also reports the difference between gauge and satellite.
How it works:
-
Satellite data for each lake is downloaded automatically.
-
Each satellite pass is reduced to one water level for the lake.
-
All values are converted to the same height reference.
-
Quality checks remove bad passes. A large jump is kept when another satellite confirms it, so real floods stay in the series.
-
All sources are merged into one series per lake, with each satellite track calibrated separately.
-
The series is compared with gauges and delivered as files and reports.
New lakes are added through configuration, without code changes. A delivery API is part of the next phase.
Space Added Value
WAVES uses Earth observation satellites:
-
SWOT (NASA and CNES) measures water height across a wide swath, so it sees the whole lake, including small and narrow ones. It is the main source and reached about 6 to 15 cm on the five NVE test lakes.
-
Sentinel-3 (Copernicus) adds independent measurements over lakes on its target list.
-
Sentinel-1 radar images were tested for detecting the shoreline. This approach is kept for later study.
Compared with gauges, satellites need no installation, no field visits and no local power or communication, and they cover remote lakes in the same way as accessible ones. Compared with optical satellites, radar altimetry works through clouds and in the polar night. Free public databases of lake levels cover pre-selected lakes. WAVES serves the lakes the user chooses, checks each one against the user's gauges and combines several satellites into one series.
Current Status


Status at the Mid Term Review, November 2026:
-
Met NVE, Lyse and ASIAQ to define needs. They supplied 14 test lakes in Norway and Greenland.
-
Built a processing chain for SWOT and Sentinel-3 that runs on all 14 lakes.
-
Validated against gauges on 12 lakes for 2024 and 2025: 6 to 15 cm on the NVE and Greenland lakes and 7 to 24 cm on three Lyse reservoirs, with a new value every 3 to 8.5 days on the NVE lakes.
-
Found that SWOT pixel data can place parts of a lake tens of metres off, and built a fix.
Next: better results for reservoirs with large drawdowns, tests on lakes the method has not seen, a delivery interface and price feedback from users.