How to create a 1D River - FAST linked simulation

2D Solver (FAST) is an innovative flood inundation modelling tool designed to allow quick assessment of flooding in coastal, estuary and floodplain environments using simplified hydraulics. It provides results in seconds or minutes as opposed to hours or days, which is up to 1,000 times faster than traditional two-dimensional models.
FAST can be used to run linked 1D-FAST simulations. This is similar to running linked 1D -2D models and can be achieved by specifying an *.ief file, which contains the information on the 1D model required by Flood Modeller. The geographic locations where the 1D and 2D solvers can exchange water (river banks, defences, etc.) are specified in a shapefile, which is prepared using the tools in the Flood Modeller Interface or third-party GIS software.

Why link these two pieces of software?

To represent the exchange of water between channel and floodplain.

Which software is linked?

In this situation, Flood Modeller 1D Solver and 2D Solver (FAST) software are linked.

Types of links

There are three ways in which 1D and FAST models can be linked:

• Two way dynamic link : water can flow both ways, i.e. water can flow from the 1D channel to the 2D domain at each time step and vice versa. You might need to troubleshoot model instabilities for two way dynamic linked models.
• Quasi two way link : water cannot flow back to the 1D channel. This will improve model stability but may overestimate the flood volume on the 2D domain. FAST will use water level from the 1D channel as a weir type boundary and subtract water from the 1D channel although water will never flow back to the 1D channel.
• One way only link : FAST will use water level from the 1D channel as a weir type boundary but does not subtract water from the 1D channel. As such, water never flows back to the 1D channel. This might greatly overestimate the flow from the 1D channel to the 2D domain but it also greatly improves the linked model’s stability. This is suitable for unattended batch simulations where the linked model will never crash as long as the 1D model is running OK. This effectively means that the two models run in parallel (uninterrupted).

Flows between 1D and FAST domain are calculated using the weir equation.

Water level linking (type H): Water levels from 1D nodes are taken by the model, and imposed as a boundary condition on the 2D model. The 2D part of the computational calculates the flow using the drowned weir equation at these linked boundaries, and passes this information back to the 1D part, where it is added/removed from the channel, to ensure mass conservation.

Polylines where this method of linking is used are denoted by 'H' in the 'type' field.

Two 1D nodes are also specified in the shapefile (in the 'node1' and 'node2' fields). Along the polyline, the water level imposed on the 2D model boundary is interpolated linearly between the water levels at these 1D nodes, allowing the sloping water surface in the model to be represented.

Where the interpolated water level is below the FAST model ground level, no flow can occur from the channel onto the floodplain, representing the case where channel water levels are below bank or defence level. Water can flow from the floodplain into the channel when the water level on the floodplain is above ground level at the link, allowing the floodplain to drain back into the channel. Water levels and flows in the 1D model are updated at each 1D time step, however, water levels and volumes in the FAST domain are updated only once every FAST time step, i.e. 'Refresh Frequency').

Flow calculated by the FAST model is added/extracted from the 1D model from the reach downstream of 'node1' (i.e. between 'node1' and 'node2').

The ground level for linked FAST model cells can be modified in the link boundary shapefile using the 'height1' and 'height2' fields, to represent defences not present in the model DTM (e.g. because the DTM cell size is too large). Defence levels will be interpolated between these two height values. Where the height field values contain no data, represented by a -9999 value, the defence level is taken to be the same as the ground level.

Flow linking (type Q): Flow from 1D nodes is taken from the 1D model and imposed as a flow boundary condition on the FAST model. Maximum water level along the link line is returned as level to the 1D. This can be used to represent a 1D reach flowing into a FAST model, or flow over a defence represented by a lateral spill unit. An HTBDY node is used to link to the FAST model, and flows/levels are transferred via this node.

Polylines where this method of linking is used are denoted by 'Q' in the 'type' field.

A single 1D node is also specified in the shapefile (in the 'node1' field, the 'node2' field is ignored). Along the polyline, the flow imposed on the FAST model boundary is accumulated in a virtual tank and applied to the FAST domain at the end of each FAST time step, i.e. 'Refresh Frequency'. The water level calculated by the FAST model is passed back to the 1D model (at each 1D time step), where it is used to provide, for example, the downstream water level needed for calculation of flow over the spill.

'Height1' and 'height 2' information is not used in Q type linking as it is assumed that the levels in the spill unit (specified on the 1D model side) is the data that will take precedence.

How can this be done?

Linking a FAST model can be done by enabling the Linking option in the 2D simulation window. To make the 2D model compatible with FAST you should first open the 2D simulation window. This can be done from the Simulation tab of the main toolbar using either (Load Simulations (for existing 2D simulation files) or New 2D Simulation (to create a new 2D simulation). In the 2D interface go to the Domains > Options tab and change the Scheme setting (in the Solver box, lower left corner) to FAST. Then choose a suitable value for the 'Refresh Frequency' in the ‘Outputs’ tab.

What you have to be careful with when creating a 1D-FAST Link

When creating a 1D-FAST Link, instabilities may appear in your simulation. This can have a serious impact on model build times and quality of outputs. You should think carefully before opting to establish a 1D-FAST Link model. An unlinked model could represent a considerable saving in time and resources.