Modelling takes on the urban flooding menace
What role can technology play in reducing the risks of flooding in the UK? Is it possible to be more precise in forecasting where the problems may lie? To find the answer, Mike Woolgar discusses the role of mapping and modelling in combating flood risk.
In 2007, floods caused insured losses of close to £3B, according to the EA, flooding more than 55,000 homes. Floods are beginning to supersede all other disasters in terms of frequency, impact on people, property and land degradation, prompting government to increase investment in flood and coastal erosion risk management from £600M to £800M in 2010.
While floods themselves are natural occurrences, it is human settlement in flood-prone areas that turn them into disasters. And with experts warning that the impact of climate change will only serve to heighten flood risk, flood management and protection strategies will be essential to mitigate the consequences of inundation of urbanised floodplains.
Flood-risk assessment forms an essential part of any such strategy, and computer-based modelling is increasingly being used to investigate different scenarios - from how a river flood propagates, to the flood return period, and flood depths and velocities.
Conventional, one-dimensional models are now less practical for simulating propagation over complex terrains, such as urban areas. This has led to the development of two-dimensional depth-averaged models, and even a step change in run time of whole-system simulation. Real-time modelling enables risks to be analysed, the performance of management options to be explored, and efficient solutions selected.
As the accuracy of models increases, so does their business value. The power of computer technology means that water distribution systems are being modelled with more accuracy.
Models that may have been skeletonised in the past can now be made to represent more complete systems, such as all pipes 100mm below or above the flood line.
In fact, water modelling has grown so dramatically over the past ten years, moving from singular-network modelling for hydraulic analysis, into multi-user and multifunctional company-wide systems they can now work hand-in-hand with more established water industry systems, such as GIS.
The term integrated models is perhaps the ultimate expression in modelling trends. This is because, at present, many models only cover discrete parts of the water system - such as watershed boundaries or company limits. One such emerging trend is an all-in-one model that encompasses all of a company's network assets.
Taking this a step further, there are models that combine key data across several companies, such as urban drainage with wastewater, flood impact or river models, across multiple industries.
One such technique that the Atkins Water & Environment team has been pioneering, is the first coupling of Infoworks and CS models using open MI, which not only allows greater discrimination and variability of data, but is also helping to advance future integrated river modelling.
Atkins was commissioned by the EA to carry out a flood mapping study in Havant in Hampshire, a highly populated town in the middle of a dense stream network. The study was to predict the extent of flooding in the town under a range of fluvial conditions.
The project required consideration of two opposing hydrological processes. The majority of the 80,000ha catchment is composed of highly permeable chalk formations, while impermeable surfaces in the urban areas drain into the highly culverted stream network, affecting stream levels during floods.
Flooding in Havant has tended to happen in isolated pockets - rather than as widespread floodplain inundation - when short and intense rainfall events drain into the stream network from the urban drainage systems. These already have high water levels from increased baseflow of the upper catchment area.
Atkins built a detailed representation of the permeable upper catchment, simulating historic and design flood events, using CEH Wallingford's Probability Distribution Model. This allowed greater flexibility in simulating a complex hydrological response, as opposed to the traditional Flood Estimation Handbook approach, which only considers the catchment area and streams as a whole.
A suite of software packages was then used for further analysis, including Infoworks for hydraulic modelling, InfoWorks River Systems - to construct a one-dimensional hydrodynamic model of the stream network - and a river channel survey, combined with LiDAR for topographic data.
Meanwhile, a hydrodynamic model of the closed-pipe drainage system was constructed, using Info Works Collection Systems, combined with detailed pipe survey data from Southern Water. The Infoworks models were configured so that that flow from urban outfalls corresponded to inflows in the river model.
To reduce data processing, the Infoworks models were combined using the Open MI modelling standard for data exchange. The outfall-inflow relationships were configured as links in Open MI, removing the need for iterative model runs.
Following calibration of the hydraulic and hydrological models, a range of extreme fluvial design scenarios was simulated. Flood extent outlines were produced, using InfoWorks built-in automatic ground-model capabilities, marking another departure from traditional data-intensive, semi-automated techniques.
While coupling Infoworks and CS models using open MI more effectively represents the feedback between the river and surface-water systems, the challenge now is to use the same technology to investigate the urban-flooding system. If drainage systems can be better linked to the river, then in future it will be possible to identify ponding more accurately in paved-surface areas.
Despite the growth and accuracy of modelling for urban-flood propagation, there are still a number of technical challenges, such as representing the complex urban terrain in a form that can be used in computation.
Urban terrain is not in homogenous formats, and there are also other factors to consider - such as canals that snake through cities and serve as both drainage and transport systems. Again, these are poorly understood in terms of how they affect fluvial flood propagation.
One aspect of fluvial urban flooding that is least understood is the relationship between the flood wavelength, and spatial and temporal characteristics of the inundation - such as water level, flow velocity, inundation duration, and flood extent. Lack of such understanding could weaken the effectiveness of flood mitigation strategies.
Changing weather patterns are also influencing the frequency and variability of extreme flood events so much, that climate change is now one of the key considerations for flood mitigation. You only have to look back to last summer to see the impact of flash flooding.
This affected more than 55,000 homes and 7,000 businesses and, according to the Pitt Report, was the wettest May- July period since 1766, with 414mm of rainfall - two-and-a- half times more than the 186mm average 1971-2000.
Part of the challenge is that current measures to mitigate flood impacts, particularly in the urban environment, are based on previously planned flood-risk intervals and, in certain instances, they might no longer provide sufficient protection. Being prepared for unexpected changes and extreme flood events asks for a paradigm shift in current strategies to avoid and manage flood disasters, especially with large populations living near coasts and rivers. In order to stem the increasing impact of urban floods, a major rethink of current planning and flood management policies and practice is required.
To be effective, flooding clearly needs to be managed in a more integrated way to enable risk-based assessment of urban flooding to come together. At the moment, responsibility for urban flooding is too fragmented, divided between the water companies, drainage authorities, the Highways Authority, local authorities, private landowners and the EA.
More integration would bring together flooding from fluvial, coastal and pluvial sources, and all possible management responses, including modification to surface and subsurface infrastructure.
Indeed, many influential figures within the water and civil-engineering industries have called for the establishment of a single flood authority - whether as part of the EA, or as a standalone body. Such an organisation would be empowered to establish a coherent and integrated approach to the complexities of flood prevention and management. This is, not least, in resolving the apparent conflicts of interest that exist between those currently responsible for the issue. The creation of such a unitary organisation would also help in addressing the obligation that the EU's Water Framework Directive (WFD) places on member states to take measures for flood management and land-drainage schemes to ensure compatibility with the new WFD environmental standards.
This thinking has been reinforced through a number of recent consultations, such as the government's strategy for flood and coastal-risk management, called Making Space For Water. The first report, of March 2005, said the EA should adopt a strategic overview of all flood-risk management issues. Now, as a result of PPS25, they play a key role in the control of new developments in flood plains.
In January 2007, 15 integrated urban-drainage pilot studies were launched, looking at ways to better prepare urban areas to cope with flood risk from overwhelmed drains and sewers.
Atkins currently supports Denbighshire County Council with its pilot scheme in Prestatyn in Wales. And, more recently, an extensive report on groundwater has just also been completed. The programme board is considering the results with further studies, to investigate how to map the level of risk from all sources of flooding.
This is a step in the right direction, but the key innovation lies in the development of a risk-based method of approach, capable of exploring the performance of multiple flood-management strategies within a single framework.
Achieving this remains a significant challenge. For example, systematic techniques to enable options to be tested and appraised - within the context of large-scale and complex system of sources - still needs to be developed and proven. Understanding how the system and management options behave, over time, will also need to be considered for developing strategic management choices.
Improved urban and flood management will have significant environmental and social benefits. The techniques, for example, that Atkins is helping to pioneer, enables the performance of urban-flood systems to be better evaluated within a virtual world and volumes of water more precisely modelled.
The total UK exposure to flood risks from all sources is difficult to calculate, but is known to be substantial. This lack of base information can hinder effective policy making, and prioritisation of expenditure and can lead to financial inefficiency.
Three aspects control the probability of a flood occurring: the characteristics of the storm conditions, the performance of the surface and sub-surface infrastructure under that load, and the local topography. Traditional mapping and modelling methods fail to capture this interaction, as they represent one state for analysis and provide decision makers with limited information.
What we are now seeing is advances in modelling, where information can be run more frequently, and system failures can be quickly identified by comparing observed and calculated pressures. The normal process would mean companies only becoming aware of the problem when the water rises to the surface.
It is precisely innovations like this that will help to support water companies to make more informed investment decisions, as well as government and its agencies in working together to deliver efficient and effective flood-risk reduction measures.
It will also support the delivery of both national and international policy, and places UK practice at the cutting edge of flood management technology.
Some of the challenges laid down to develop detailed and reliable flood-modelling tools have largely been met, but future research will need to continue to concentrate on this area. While modelling does not hold all the answers, it can help to facilitate:
- Procedure, methods and tools for exploring the effectiveness of future flood-management policies
- Better targeted and integrated private and public sector funding
- Insight into uncertainties regarding the behaviour of the urban-flooding system, through more sophisticated techniques
- The delivery of an integrated approach to urban-flood management