Centre for Earth Systems Engineering Research

CityCAT: Urban flood model

CityCAT: Urban flood model

CityCAT (City Catchment Analysis Tool) is unique software tool for modelling, analysis and visualisation of surface water flooding.


CityCAT enables rapid assessment of combined pluvial and fluvial flood risk. It allows assessment of the effects of different flood alleviation measures. 

It uses a unique combination of the following factors:

  • efficient software architecture throughout the tool and especially in the numerical part
  • use of standard, readily available data sets and sophisticated and efficient algorithms for grid generation
  • robust and accurate solutions of flow equations and computational algorithms for other processes

Input Data

CityCAT uses standard datasets. It uses a Digital Terrain Model (DTM) for the topography. It uses the OS MasterMap data to delineate the urban features such as: 

  • buildings
  • roads
  • permeable surfaces

Simulations of different flood events can be driven by rainfall, flow and/or water depth time series.


The computational grid is generated automatically using the DTM. 

The OS- MasterMap is used to exclude the buildings’ footprint from the grid the buildings layer. This improves the ability of the model to realistically capture the flow paths in urban areas. 

It reduces the simulation time due to the reduction in the number of computational cells. The removed cells form the 'buildings' layer used in the roof drainage algorithms. 

Computational Algorithms

Simulation of free surface flow is based on the full 2D shallow water equations. The solution is obtained using high-resolution finite volume methods with shock-capturing schemes.

These are able to accurately capture propagation of flood waves as well as wetting and drying of the domain.

The Green-Ampt method is used to estimate infiltration in pervious areas as a function of soil hydraulic conductivity, porosity and suction head.  

The time dependent solution is obtained using an iterative method. 

Two types of roof storage are applied to the buildings layer of the grid. These are 'blue roof' which is based on the available volume of storage and 'green roof' which uses the Green-Ampt algorithm. 


The model provides two types of visual outputs:

  1. Time series of water depths and flow velocities at selected locations.
  2. Snapshot maps of water depths and velocities at different times during the simulation. 

These maps are then combined to produce an animation of the flood propagation. 

Software architecture

The architecture of CityCAT is based on the object-oriented approach. This offers great flexibility in  development and allows rapid extension of functionality. 

Additionally, due to this architecture, there are no software limitations on the size of the domain. 

The computational efficiency is improved considerably by removing the decisions ('If Then Else' statements) during run time. 

This is achieved by making all the decisions during the initial set up which is a main feature of object-oriented design. 

Cloud computing

CityCAT was originally developed as a PC application. Now, two  versions without the GUI are also available for deployment on servers or Cloud computing. These are a 32bit/64bit Windows and a 32bit/64bit Linux. 

High resolution large scale modelling of flooding at city scale was achieved by adopting a high throughput model of computation on the Cloud. A a Condor cluster of nodes were deployed as a set of virtual machines instances on the Amazon Cloud. 

Here's an example of the Cloud computing runs for 3 different domains.

 No. cellsCell sizeRequired memoryPC instancesNo. jobs submitted per instance
1 1,000,000 2m 3 GB Standard instances - 7.5GB 2
2 4,000,000 1m 11 GB High-memory instances - 68.4GB 5
3 16,000,000 0.5m 40 GB High-memory instances - 68.4GB 1