Urban water system simulation using a dynamic coupled Triwaco-DuFlow model, Nijmegen (Netherlands)      
 

Framework

The city of Nijmegen is located near the German border in the east of the Netherlands and lies at the river Waal (one of the big rivers of the Netherlands). The municipality of Nijmegen (Netherlands) initiated a study for the optimisation of the urban water system of the city district called the 'Waalsprong'. This study is part of the ongoing Environmental Impact Assessment of the city district.

The urban water system should at least:
  • be able to keep sewage water and rainwater separated as much as possible
  • contain small lakes outside the city distric for seasonal surface water storage
  • contain surface water storage in ponds and canals within the city
  • use of water of secondary quality with respect to drinking water to be used for flushing the lavatory, cleaning and use in the garden

The municipality have to decide on several scenarios to achieve an optimal urban water system. Royal Haskoning was asked to set up a model which can be used to weigh these scenarios. The different scenarios include displacement of the Waal-dike, surface area of the lakes (surface water storage) as well as the dimensions of the canals and ponds. In addition possible consequences for changes in water quality caused by inlet of river water are included in the modelling.

At a first glance the use of a standalone surface water model seems to be sufficient. However the constantly changing water levels in the river Waal is of great influence the on the urban water system. During periods with a low river level the urban water system will lose substantial amounts of water caused by downward seepage to the river. During periods with a high river level the urban water system has to cope with substantial amounts of water received from seepage of river water. Therefor with regard to the urban water system groundwater is as important as surface water.

To simulate the urban water system a dynamic coupling was realised of the surface water model DUFLOW and the groundwater model TRIWACO-FLAIRS.

How it was done

Starting point for the coupling in this study was the surface water model DUFLOW. The watercourses, lakes and ponds as schematised in DUFLOW are directly imported into triwaco and incorperated in the finite element grid. triwaco has the advantage that watercources are schematised as line elements each having a unique ID which corresponds to the ID given in DUFLOW. So a dynamic coupling between the two models is relatively easy established. The actual coupling and exchange of data between the two models is provided for by the sofware-application TRIDUFLO developed by Royal Haskoning. TRIDUFLO uses the standard DUFLOW and TRIWACO-FLAIRS simulation packages so it is universally adoptable and also will work for a coupling with MODFLOW.

During the simulation TRIDUFLO is in control of both the surface water and groundwater model. The fluxes for the entire simulation period from TRIWACO-FLAIRS watercourses are transformed by TRIDUFLOW to boundary conditions for the DUFLOW model. This flux can be negative or positive depending on the nature of the watercourse which may be draining or infiltrating. Next the waterlevels calculated in DUFLOW are used as a boundary condition for the TRIWACO-FLAIRS simulation. These steps are repeated several times for the entire simulation period until it reaches the criterion for convergence.

Results

The coupling between DUFLOW and triwaco has shown that the interaction between groundwater and surface water plays an important role in the urban water system. The dynamics of the modelled water levels in a canal shown in the graph emphasise the crucial importance for coupled groundwater-surface water simulation in this study.

The resulting model has shown to be accurate in predicting changes in the urban water system. It is set up in a flexible manner so that it can be used as a decision tool by the municipality to test other scenarios to be formulated in the near future.

The coupled model was used to gain insight into the dynamics of the urban water system. The subsequent scenario simulations were used to determine dimensions of canals, ponds and small lakes with respect to different climatic influences. In addtion the model was used to predict the future surface water quality.

 
  Location of the area modelled.

  TRIWACO-FLAIRS model and DUFLOW model showing the correspondig watercourses.
  Difference between a coupled and stand-alone calculated waterlevel in a watercourse. The graphs show that with a dynamic coupled model the changes in the waterlevel are simulated much better.