Child pages
  • Chapter 3: How to setup a flood impact model
Skip to end of metadata
Go to start of metadata

This chapter describes the model configuration, object maps, impact functions, maximum object values, hazard maps, default data and settings needed to set-up a flood impact model.

This chapter describes step-wise the requirements to set-up the flood impact model.


Figure 3.1 stepwise approach to set-up a flood impact model

3.1       Model configuration


A model configuration file is a spreadsheet and includes any references and settings necessary to make a successful flood impact model run. The non-bold cells are configurable.

Figure 3.2 flood impact model configuration file

3.1.1       Header


The top 5 rows defines a header and describes general i/o file and folder locations:

Cell B2 – name of the model run

Cell B4 – location to input waterdepth file or list of waterdepth files (see for the latter section 3.6)

Cell J3 – location to flood impact functions folder

Cell J4 – location to object maps folder

Cell J5 – location of an empty output folder


All above are required for each flood impact model run.


In Column D and E several optional settings can be used (leave blank if not):

Cell D1 – location to incremental file (Delft-Sobek or Delft-FLS incremental files describing flood simulation patterns)

Cell D2 and D3 – location to rise rate file and location to arrival time file (instead of incremental file)

Cell D4 – location to flow speed file

Cell D5 – location to zonal statistics input shapefile

Cell E1 – the offset to t0 (time of breaching), to be used when using the incremental file

3.1.2       Base file


Rows below row 6, describe impact categories and their specific characteristics:

Cell below A6 – specify whether the impact category should be used in the model (1) or not (0);

Cell below B6 – specify the name of the impact category;

Cell below C6 and E6 (default = blank) – specify the maximum damage, as used by the damage functions;

Cell below D6 (default = 0) - specify whether both (1) or only the first (0) damage function is to be used;[1]

Cell below D6 and F6 (default = blank) – specify the name of a damage function file (without file extension and folder name)

Cell below H6 (default = 0) – specify whether an individual impact category grid is to be output

Cell below I6 (default = 1) – specify an overall weight factor per category

Cell below J6 – specify category grid file name (with file extension)

Cell below K6 – specify the units of the category

3.2       Object maps


The first step (A) of your flood impact model is to acquire spatial information on the location of objects, vulnerable to floods when exposed. Delft-FIAT requires one or more object raster files of the format Tiff or Ascii.


Figure 3.3 Requirements to object raster file dataset


The object raster files need to have a identical coordinate reference system (also identical to the hazard files), an identical extent, an identical cell size, an identical origin, a clear unit value (e.g. number of m2) and should be stored in one folder location.


On purpose, the modeller can choose its own comfortable tool to generate these object maps. One can choose for ArcGIS or QGIS packages or use the free available pre-processing python scripts based on the GDAL library coming with Delft-FIAT (see wiki and annexes).


Several scripts exist to efficiently convert (large) vector files (point, lines, polygons) into required object maps. Further, several GDAL commands are provided to easily transform and efficiently store raster files to the required Delft-FIAT settings.


Figure 3.4 Pre-processing python scripts available for straightforward vector-raster conversions

3.3       Impact functions


The second step (B) in the set-up of your flood impact model is the description of impact functions. The functions are easily described as comma separated value files using a transparent and flexible template. Delft-FIAT comes with an extensive library of functions used in different damage models in several countries and researches. The library records are described in the tab sheet ‘functions’ of the configuration file.


A function basically includes two columns: one describing the waterdepth and one describing the accompanying damage factor (between 0 and 1). Sofar, it starts with 0.01 m, it has a stepsize of 0.01 m, saved in one location and multi-dimensional functions are yet also possible[2]. In this way, new functions can easily be configured by the user.


Figure 3.5 Describing flood impact functions

3.4       Object values


Related to the flood impact functions and object maps are the maximum values of one unit within the object maps (step C). They are often obtained from research in the field or macro-economic figures. They are set in column C of the configuration file below the header. Be sure the units are correctly scaled (e.g. cell size 50x50 m, unit per ha).


Figure 3.6 Maximum object values

3.5       Hazard maps


The last required input for a flood impact model to run are the hazard maps (Step D). They describe flood characteristics as waterdepth, rise rate, time of arrival, velocity, duration etc. Delft-FIAT is able to run default output from SOBEK, Delft-FLS but also Mike 11 or GIS calculations.[3] Only a correct waterdepth file which extent is partially overlapping or completely within the object maps extent will run. Be sure the other requirements (coordinate system and cell size) are identical.


Figure 3.7 Hazard maps

3.6       Advanced settings


When optional cells of the header of the config file in columns D and E are used, specific casualty information (related to function 2_0001) can be calculated. The incremental file (or rise rate and time of arrival file) provides temporal information for calculating casualties in different time zones for that specific simulation.


The zonal shapefile referenced in cell D5 of the config file is used to define zonal summary statistics for each impact category in the config file and for each individual record in the shapefile. As such, the total damage or casualties are summed for each region you are interested in.


Delft-FIAT can also perform risk calculations. In the config file, the user can fill in in Cell B4, the location of a .csv file describing between which lower and upper bound the integration of impact maps should be done, as well describing the location (and return period) of the waterdepth files considered. Risk calculations are done for total damage figures and maps, as well as casualty risk (using the special function 2_0001).


Figure 3.8 csv file for risk calculations


On the first line the boundaries of the risk graph are defined. The other lines give a location to the hazard grids and a return period (1/n), which is used to calculate the risk.

3.7       Default models and data


Delft-FIAT comes with a default flood impact model and data from HIS-SSM2000. Further, Delft-FIAT comes with an extensive library of functions used in different damage models in several countries and researches and which is updated with each release.

[1] This setting is optional and only needed in the dutch setting (default = 0)

[2] An example is the mortality function (e.g. 2_0001.csv) with two variables: waterdepth and rise rate. This csv file consists of three columns: one describing waterdepth, one describing the damage factor related to a low rise rate value (< 0.5 m/s) and one describing the damage factor related to a high rise rate value (> 4 m/s). Delft-FIAT interpolates between those two as the rise rate is between the lower and upper value. This function is very specific and should be used with case, as the damage factors are given in percentages instead.

[3] A default handler of 3DI output (NetCDF) is being developed

  • No labels