Case Studies

The advantage of our flood warning systems is that they can determine water levels up to a week ahead of time. This gives communities enough time to prepare themselves for floods by securing the area and livestock. Our systems consider both regulated and unregulated catchments and consider the snowmelt, an issue in Sauda and other catchments.

Tovdal Flood Warning System

Client:

Kristiansand Kommune

Contact Person:

Sigurd Paulsen (Beredskapsleder)

Tovdal

This pilot system offers predictions of the maximum daily water level 4 days ahead between two bridges (Drangholtsmyra and Boenfoss) located in Tovdal River (Agder, Norway).

This modelling strategy consists of a hybrid model composed of several AI/Machine Learning models coupled with a statistical regressor emulator of 2D hydrodynamic simulations.

The predictions are updated 4 times a day in accordance with weather forecast updates. The corresponding warning levels are then displayed in a webapp and can also be obtained through an API.

Sauda Flood Warning System

Client:
Digi Rogaland (Sauda kommune)

Contact Person:
Hugo Kind (Berdskapsleder)

Sauda

This pilot system offers predictions of the maximum daily water level 4 days ahead in four cross-sections (Sauda center) located in Nordelva and Storelva rivers (Rogaland, Norway).

Forsyst has developed a hybrid model composed of:

  • A statistical hydrological model for predicting water flow in Nordelva where there are no level measurements.
  • Several AI models for predicting water flow in Storeva (considering snowmelt), which is a strongly regulated watercourse with hydropower reservoirs
  • A hydrodynamic model that uses the predicted upstream water flows and the predicted sea level downstream to predict water level in four locations.

The predictions are updated 4 times a day in accordance with the weather forecast updates.