Flood Modelling for Cities using Cloud Computing

Lead Research Organisation: Newcastle University
Department Name: Civil Engineering and Geosciences

Abstract

Cities are concentrations of flood risk because they are densely developed and tend to be situated in low-lying areas. Floods in UK cities of 2005 and 2007 have been a major driver for planning and research. However, risk from flooding from intense rainfall is particularly difficult to assess because it is sensitive to the spatial-temporal characteristics of rainfall, local runoff and surface flow processes, the performance of urban drainage systems and the exact location of buildings. Detailed models of water flow over complex surfaces are needed so that the effects of buildings, roads, vegetation and other features can be captured. Such models take a very long time to run on conventional computers and therefore need special facilities which have hitherto not been readily available. Fortunately, Cloud computing is now available. Clouds are systems where very many large computing jobs can be submitted to flexible, large facilities at relatively cheap cost without the user having to buy and manage the facility themselves: they simply pay a fee for time and storage of data.The outputs of flood models for cities are of great interest to a range of end-users in local government, the Environment Agency and water utilities. Until now they have been unable to generate and use such models because they could simply not be run for the large areas needed or for the number of cases of rain storms required to cover all eventualities. This is now a real possibility with the Cloud, so we plan to demonstrate its use not only for flood mapping for a large city, but for a wide range of rain storms covering possible climates for now and the future to allow planning and adaptation to take place.

Planned Impact

There are beneficiaries in three main areas of industry and government. 1. Climate change impact assessment and adaptation. The advent of the UKCP09 probabilistic climate change scenarios has brought a step-change in the level of computational resource required for carrying out future impact assessment and planning for adaptation. Some sectors are more receptive and geared up for this but there are some such as the water industry and engineering sectors where the paradigm shift is not being achieved. A recent workshop run by the Newcastle CEG water group brought together Ofwat, the Environment Agency (EA) and the water industry to discuss barriers to implementing such approaches under the NERC Climate Change Network programme. They found the key issues to be lack of computational resources for implementing large ensemble assessments using computationally heavy impacts models (such as hydraulic flood models) as well as the lack of readily available expertise. A further case in point is the currently ongoing national Climate Change Risk Assessment (CCRA) which is tasked with carrying out meaningful assessments over a wide range of sectors and faces an essentially impossible task of carrying these out with any statistical rigour due to lack of resources. The water group in CEG is already engaged with various key hubs in the climate change impact and adaptation sector. Kilsby (co-I) is a member of the UKCP09 Project Management Group and will arrange for the results to be included on the UKCIP portfolio of exemplar case studies for probabilistic risk assessment. Colleague Prof Jim Hall is a member of the Adaptation Sub-Committee of the government Climate Change Committee. The work will be promulgated through these and other offices to show that very large and statistically rigorous assessments may be carried out at affordable cost. Additionally, the CEG water group runs influential research led MSc programmes (Hydroinformatics, Hydrology and Climate Change) providing training for the water industry where leading edge practice is introduced. The pilot results would provide a compelling case study. 2. Flood risk assessment and urban development. Local government and regulators are charged with producing Surface Water Management Plans under new legislation. Such plans required detailed mapping of city-wide flood risk and these are currently of great concern as engineering consultants are unable to provide these at suitable levels of detail and at affordable cost. The outputs of the project will therefore be of great interest to these users. The CEG water group also have a formal research collaboration with Northumbrian Water Limited (utility covering NE England). The CEG water group runs a flexible part-time MSc programme for the water industry Flood Risk Management which again provides a key platform accessing influential practitioners. 3. High performance computing industry The pilot serves as an excellent case for opening up a major computing resource to a historically reluctant set of customers and users.

Publications

10 25 50
 
Description Flood modelling for cities has become a key technology as increasing flood risk has become apparent in the UK and worldwide.
Conventional models operated using coarse resolution data (50m elevation models) and inaccurate approximations for the representation of flow and buildings because insufficient computational power was available.
This grant has shown that new high resolution data can be used (e.g. 2m Lidar elevation maps) with accurate hydrodynamic models which properly represent the complex surface flow fields in cities and the effects of buildings and sewer networks.
Exploitation Route The work has been taken forward by several research projects (EPSRC, NERC and EU funded), results have been made available to city authorities worldwide (Melbourne, Antwerp, and 500+ other EU cities). These results continue to change the way flood riosk information is used.

Further possibilities using building information are a future development area.
Sectors Digital/Communication/Information Technologies (including Software),Energy,Environment,Transport

URL http://www.ncl.ac.uk/ceser/research/software/citycat
 
Description The pioneering work using the Cloud for flood risk computing has led to its use on a range of further projects funded by EPSRC and EU. The recognition that both large memory and fast CPU are available has led to further developments capitalising on the Cloud capability. So, the CityCAT flood model developed at Newcastle has been set up as a multi-thread system allowing large domains to be modelled by splitting across multiple machines, as well as multiple simulations being performed, e.g. for different size rainstorms and for multiple cities. Simulation results performed on the Cloud have been used not only in academic work but also in support of planning and adaptation work in Newcastle associated with sustainable drainage programmes in collaboration with Newcastle City Council, the University's Estates team and Northumbrian Water.
First Year Of Impact 2014
Sector Construction,Environment
Impact Types Policy & public services

 
Description Cloud computing award
Amount £50,000 (GBP)
Organisation Microsoft Research 
Department Microsoft Research Cambridge
Sector Private
Country United Kingdom
Start 06/2014