In the spring of 1995 there was a large flood in the rivers flowing into Lake Mjøsa in Norway. The flood caused extensive damage by submerging houses and infrastructure by the lake. There was also concern about the water quality in the lake during the flood, as the lake water was used in the water supply for many of the villages around the lake.
Lake Mjøsa is the largest lake in Norway. The modelled part of the lake is approximately 60 km long and has a maximum depth of 400 meters. The bathymetry of the lake is shown in the figure below.
The colours show the depth, where blue is the largest depth, and red is the smallest depth.
The SSIIM model was used for the CFD calculations. The lake was modelled with an unstructured grid with hexahedral cells. The grid is seen from above in the figure below:
A detail of the grid is shown below:
The water flow in the lake was calculated based on recorded inflow and outflow. The flood started on the 27th of May 1998, when the water discharge increased from 300 qm/s to 600 qm/s. It increased steadily to 2885 qm/s on the 3rd of May, before decreasing slowly. It was assumed that the water quality deterioration started simultaneously with the flood increase.
The Norwegian Institute for Water Research measured the water quality in the lake during the flood. Several parameters were measured. Colour maps based on the measured turbidity was made. These are presented below. The turbidity was also calculated by solving the transient convection-diffusion equation. The calculated results are presented together with the measurements below:
Calculated (left) and measured (right) turbidity at the water surface at 6th of June 1995.
Calculated (left) and measured (right) turbidity at the water surface at 9th of June 1995.
Calculated (left) and measured (right) turbidity at the water surface at 12th of June 1995.
Calculated (left) and measured (right) turbidity at the water surface at 15th of June 1995.
The animation above shows the bed and the surface of the computational grid, with a vertical profile at the upstream part of the lake. The colors show the turbidity concentration at 15. June. The final profile in the animation shows that the contaminated flood water is confined to the layer close to the water surface. This is due to the temperature stratification of the lake. The water intakes to the main towns around the lake, for example Lillehammer and Hamar, are located at in the deeper parts of the lake. The low quality of the flood water therefore had less effect on the water supply for the local population. Measurements of the vertical profiles also showed this effect.
This project is carried out in cooperation with Gøsta Kjellberg at The Norwegian Institute for Water Research
The Norwegian Research Council is partly sponsoring this project.
Back to CFD page
Nils Reidar B. Olsen