Model helped to play it safe
The proposal: developing a residential site on a flood plain. The problem: the development is to be a retirement home with sleeping quarters below the floodline. The solution: 2D modelling, then analysis, followed by building flood defences.A BREACH analysis using Wallingford Software's InfoWorks CS 2D has been undertaken for Truro in Cornwall. This is because it is proposed to develop a residential site on the flood plain of a large tidal river - with sleeping accommodation below the predicted one-in-200-year tidal flood level.
According to Wallingford, any proposals had to demonstrate the building and its occupants would be safe should a breach of the river flood defences occur simultaneously with such an event.
The analysis showed that the site would be subject to gradual low-level flooding during the event. Maximum flood depths were predicted to vary across the site between 100mm and 400mm as ground levels vary.
The study concluded that measures could be taken to ensure the building and its occupants are protected from flooding during a breach.
The purpose of the Truro high-level study was to investigate whether a proposed 0.4ha development site would be at risk of flooding from a small dock about 0.3 miles away. This, says Wallingford, involved an unusual use of InfoWorks CS 2D to model the impact of such a breach.
The area is scheduled for development as a retirement village. A flood-risk assessment was ordered to ensure the building and its occupants would be safe should a breach occur, and also to satisfy the requirements of PPS25, the government's Planning and Policy Statement on development and flood risk.
This policy ensures flood risk is taken into account during planning, and that development is directed away from high-risk areas.
The proposed development site is within the Environment Agency's (EA) Flood Map Zone 3, which indicates a high risk of flooding with an annual probability from tidal sources - confirmed as the main source of flood risk by the regulator - of 0.5%, or a one-in-200-year return period. But the EA's flood maps do not take into account the presence of flood defences.
Historically, there has only been one recorded flooding incident in the area - in January 1928 - but this was a fairly serious event affecting the park immediately to the east of the development site. The flooding was the direct result of high rainfall run-off and melting snow coinciding with a storm surge from the sea.
The current defences at that location were considered to be reasonably robust, providing protection against a one-in- 1,000-year fluvial flooding event and a one-in-100-years tidal flooding event. The EA regularly inspects the defences to ensure they are fit for purpose - the current condition score is two, valued as good on a one to five scale, with one being the best.
InfoWorks CS v9.0 2D was used to undertake the analysis of flood risk at the proposed site. A simple model was created to model the breach.
This consisted of two nodes and one conduit. One node was an outfall located on the river side of the breach, and the other was a dummy node, with the flood type set to 2D to allow overland flooding. The shaft plan area was set to 31.8m2 to represent a 20m-long weir to mimic a 20m-wide flood defence breach.
The 2D software uses the circumference of the shaft plan area as the weir.
A global Manning's roughness of 0.015 was applied across the catchment, a fairly arbitrary value for paved and tarmac surfaces. An EA ground model of 0.5m resolution LIDAR data was obtained, in both DEM (digital elevation model) and DTM (digital terrain model) formats.
The DTM was used as a basis for the 2D above-ground flood model. This model has the buildings filtered out, so it was therefore necessary to represent these as voids to ensure that any overland flooding routed around them. OS master map digital mapping was used to import the building polygons.
A number of inconsistencies in the dataset were identified where predicted depths were unrealistically large. But their impact on the overall result was assessed to be negligible.
A 2D mesh was created around the entire potential flooding area, and a smaller detailed mesh added to accurately model the flood route from the breach location to the development site. With InfoWorks CS 2D it is possible import the DEM model to display the results even if a DTM was used as a basis for the modelling. One of the advantages of this software is that it is possible to import ground models, making it possible to obtain a good visual representation of the real situation in a catchment.
Three key model assumptions were made for this study. The first was that the lock gates at the dock entrance offered no flood protection, and the second was that buildings in the study area would be modelled as fully-impermeable voids. But in reality a proportion of the floodwaters would have entered some of the buildings.
Thirdly, no allowance was made for any below-ground storage offered by existing drainage systems and/or underground subways or railways.
The assumptions can be considered to be conservative - therefore the predicted flood depths would be slightly greater than the actual situation.
A one-in-200-year tide level hydrograph was used on the outfall. The predicted peak depths were found to vary between 50mm and 300mm above current ground levels around the development site, and the floodwaters from the breach were predicted to reach the development site from the breach within 15 minutes.
As the plan was for a home for the elderly, and there would be some sleeping accommodation in the basements, there was some concern about how the residents would exit the lower floors of the building.
The main proposed mitigation for the predicted flooding was the construction of a flood wall around the property with an additional 150mm freeboard to take into account any potential uncertainties in the model. It is proposed to locate a car park and service area at basement level with an entrance lower than the highest predicted flood level. This was protected by a floodgate.
In addition to external flood defences, ground floors will incorporate flood resilience design measures as described in the EA's publication Preparing for Floods such as the fitting of flap valves on drainage pipes and sanitary connections.
The 2D software proved relatively easy to use and some answers were obtained within hours of beginning work. LiDAR data is becoming cheaper and more widely available, and data for this study was bought from the EA for a reasonable price.
It is recommended the 2D software is used more widely, particularly within the UK water industry, when overland routes need modelling.
There have already been a number of case studies involving the use of InfoWorks CS 2D, and understanding flood routes and flood mechanisms is an important issue for the industry. It can be difficult to understand what is happening in a catchment without the use of such software.
Pell Frischmann, with its work with South West Water, has discovered that understanding overland flood routes using InfoWorks 2D is an important factor to ensuring sufficiently robust schemes are evaluated, designed and implemented.