Pre-Processing

Background

The Flood Modeller 2D solver (FAST) - quick flood inundation method routes water over the floodplain through a series of depressions. These depressions can fill with water, either from sources (for example rainfall) or by spilling in from neighbouring depressions. A depression is defined in terms of its lowest point, and all water with a source within that depression will drain into that point.

To run FAST solver, a description of the floodplain is required in terms of these depressions. It is generated by a pre-processing stage, generating the following information:

• A map giving depression ID for each DTM cell
• For each depression:
  • A stage versus area relationship
  • A stage versus volume relationship
  • A list of neighbouring depressions
  • The lowest elevation at which water can flow between the depression and its neighbour

The following section describes the methods used in FAST solver to generate the required model input datasets.

Method

DTM preparation

The DTM code works on the assumption that the natural surface has variation in elevation from cell to cell. Thus, the code will have difficulties working on flat surfaces. This is avoided by modifying the DTM using this formula:

The averaging window distorts the flat surfaces slightly so that points are not on the same elevation. The parameter defining the scale of distortion is provided as the filter parameter.

In addition to preparing input data, the pre-processor checks that the allowable number of active computation cells is not exceeded. If this limit is exceeded, the simulation will be aborted and a message explaining which limit is exceeded will be written to the FAST solver log file. Unless you are using a Professional edition licence with the unlimited cell 2D solver enabled, the number of computational cells in Flood Modeller are limited as follows:

• Standard – 400,000 computational cells
• Professional – 1,000,000 computational cells

Identify depression low points

This step marks the start of depression delineation by identifying points that have neighbouring points all at a higher elevation.

In addition to preparing input data, the pre-processor checks that the allowable number of depressions is not exceeded. If this limit is exceeded, the simulation will be aborted and a message explaining which limit is exceeded will be written to the FAST solver log file. Unless you are using a Professional edition licence with the unlimited cell 2D solver enabled, the number of depressions in Flood Modeller are limited as follows:

• Standard – 4,000 depressions
• Professional – 10,000 depressions

Find drainage directions

For each cell of the DTM, a drainage direction is then determined. This can be any of eight possible directions into any of the surrounding cells (including diagonal directions), namely: north, south, east, west, north-east, north-west, south-east or south-west. The selected direction is based on the highest gradient with neighbouring cells.

Determine depression ID map

Ultimately, water will follow these drainage directions towards a low point and this is used to define the initial 'depression area' of each low point.

Merge neighbouring depressions

Many of the depressions generated above will be separated by a barrier with a small elevation difference. In reality, it might not block the exchange of water and it would be easier (computationally) to merge these depressions. This minimum elevation difference (shown below) is specified as the 'merge parameter':

Compile stage-area-volume tables

This penultimate step uses the catchment boundaries, set in the above step, to find each depression's stage-area-volume relationship along and the number of neighbours. The discretisation parameter sets the minimum distance between recorded stages for the relationship.

Write results to files

The final step in the pre-processing stage is to write the results of this stage to the files. The depression ID grid is written as an ASCII raster file. The depression information is written as separate text files for the depression neighbours, stages, areas, volumes and connection levels.

Computational engine

Following the pre-processing stage, the computational engine is used to introduce water into the depressions linked to the boundary conditions specified in the model.

FAST solver follows these rules to determine flows over a floodplain:

Rule 1: Water instantly appears at the lowest point in a catchment.

Rule 2: The water level in any given depression can only rise to the lowest connection level with its neighbouring depression. A further rise will cause a spill.

Rule 3: Rule 2 is exempt if the water level in the neighbouring depression is also above the connection level. Under this condition, water levels in both depressions must be equal.

Rule 4: If a depression of the ground grid has a boundary condition of the type 'water elevation' imposed on it, then the water level in the depression is fixed to that level. For neighbouring depressions, two cases are possible:

• If the water level in the neighbouring depression is above the minimum connection level and the water in the depression with the boundary condition of the type 'water elevation' is below the minimum connection level, then the water level in the neighbour is set to the height of the minimum connection level.
• If the water level in the depression with the boundary condition of the type 'water elevation' is above the minimum connection level, then the level in the neighbouring depression is made the same and the boundary condition of the type 'water elevation' is also imposed on the neighbour.

This implementation does not take into account any dynamic elements of hydraulics like momentum or velocity. Hence, you can only visualise the progress of flood propagation by using a series of steady-state snap shots.

Duration of 2D FAST models

A 2D FAST model has no time element. It outputs a single flood map solution to represent the final distribution of water across all defined storage areas within the input ground grid. However, FAST models still require a non-zero time element to be set in the simulation interface (on the 'general' tab). If no time is set (it is left at the default zero hours) then your simulation will not run.

In some cases the time value set is used by a FAST simulation. The options for simulation times are as follows:

• If you are using a boundary condition that is linked to a hydrograph, then the simulation time you set will be used to calculate the total volume of water entering via the boundary, the area under the (flow vs time) hydrograph.
• If you are using a boundary condition with no time component (for example a global water depth) then you can set the FAST simulation time to a nominal 1 hour.

Note, these definitions relate to a 2D FAST and not a 2D FAST dynamic simulation. FAST dynamic simulations do have a time element and produce time series of flood data.