Flood Forecasting and Risk Assessment
Flood ForecastingFlood forecasting relies on the fact that a computer can calculate the response of a river basin to extreme rainfall or snow melt - using an appropriate model - in a shorter time than the response takes in real life. Rainfall, other climatological data and river levels must be transmitted to the computer system in "real-time". Forecasts can be updated and the time horizon extended continually as new information becomes available.
The hydrograph shows a sample forecast for the Niger at Niamey, a rather unusual case where the response time of the river is extremely slow - around three months.
Flow forecasting programs should be sufficiently robust to continue functioning when a failure in the transmitting network reduces the amount of data received. They should also adapt to variation in the values of the model parameters as indicated by continual self-checking of forecast performance.
The expertise of Water Resources Associates covers both the design of the computer programs and the specification of the hardware for measuring and transmitting data.
Specific WRA ExperienceSouthern Region, Environment AgencyThe Southern Region of the Agency wished to improve the accuracy and lead time of forecasts with the use of PC based models linked to their real-time telemetry system. Reliable software, delivered on time and to budget was required handle the interaction with the user, telemetry data files, execution of an arbitrary number of hydrological models and presentation of results. Time and budget were both critical for this project and WRA provided the required software within the imposed constraints.BangladeshDesign and implementation of a networked PC Unix-based system for forecasting operations at the Flood Forecasting and Warning Centre in Dhaka. A microTIDEDA hydrological database and MIKE 11 hydraulic model were configured to enable the seasonal forecasting of floods on the Ganges and Brahmaputra rivers. A training programme was provided for counterpart staff.NigerA flow forecasting system for the Niger river basin - contributing basin 1.2 million km2 - was funded by WMO. The network comprised 65 stations measuring river level and rainfall, linked by satellite to an international forecasting centre in Niamey and to national forecasting centres in each of the eight participating countries. Each of the stations also transmitted diagnostic data to facilitate preventative maintenance. The network was able to forecast flows up to 3 months in advance, making it a valuable tool for water management and agricultural production.IndonesiaA pilot forecasting system was established in the Cimanuk basin in Java to improve the country’s capability in flood forecasting. The system comprised 30 stations; all measured rainfall and nine of them measured flow. Three of the stations also measured turbidity on an experimental basis. The data were transmitted by UHF radio links. As the two receiving stations were outside of the basin two repeater stations were needed. The system was designed to provide forecasts to towns and to the operators of major hydraulic structures diverting water for rice irrigation.UKThe river Dee in north Wales was a test bed in the 1970s and 1980s for telemetry systems and forecasting models. It provided the first UK demonstration of real-time linkage between radar-measured rainfall linked to a rainfall-runoff model. Three of our principals had key roles throughout the project, which was carried out by IH and the Welsh National Water Development Authority.
Flood Risk AssessmentWater Resource Associates [WRA] continues to carry out flood risk assessments for a wide range of clients, developers and owners of property in southern England. Since the January 2003 floods, the provision of flood risk reports has been a growing area of our business, and the company is well-placed to provide a rapid-response to householders and developers in the process of buying and selling houses, or developing new infrastructure.
The service provided by WRA can be particularly helpful to owners facing difficulties in obtaining insurance of their house and belongings, as a result of flood risk, as well as owners and developers wishing to progress building plans.
In cases where there is doubt over the level of risk or whether a particular house is liable to flood, insurance companies have required preparation of a flood risk report by professional hydrologists, to clearly identify the risks of flooding at the site in question.
WRA can carry out a flood risk assessment for clients, which comprises topographic survey of the site and building levels [supported by Sitech Survey Services, when more extensive areas of survey are required]. For borderline cases of flood risk, the detailed studies often show that the property under investigation lies above the maximum flood level, yet lies within the floodplain envelope defined by the Environment Agency Flood Map. The WRA report provides a fully-documented basis for the insurance company to then provide cover for the property in question.
WRA is a small specialist consultancy comprising experienced and nationally-acknowledged scientists at the forefront of work involving flood hydrology and flood forecasting. Directors and associates have been involved in the development of flood prediction methods and software since the 1970s, which now form the basis for flood estimation in the United Kingdom.
WRA’s clients include the Environment Agency, water companies, research agencies, and the tourism industry. After the 2003 floods, WRA provided expert support to The Upstream Group [TUG] during the deliberations of the Flood Risk Action Group [FRAG] for the Thames valley from Hurley to Wraysbury.
Here is a selection of sites studied since 2003: Ascott-under-Wychwood, Northmoor, Woolstone Mill & Sutton Courtenay in Oxfordshire, Bisham Brook in Berkshire, Thameside properties in Bourne End, Marlow & Temple [Bucks], Chertsey [Surrey], Downham Market [Norfolk], Gt Finborough & Gt Gabbard in Suffolk, Longwick Mill in Buckinghamshire, Chertsey & Nutfield in Surrey, and Bocking & Wix in Essex.
WRA prepared guidelines to help TUG members in acquiring property insurance, and this document was adopted in 2005 by the Environment Agency, as a roadmap which would guide property owners through the insurance maze.
Download the flow diagram, which shows the stages at which WRA expertise may be sought.
Groundwater FloodingEynsford Flood Mitigation Measures, Kent, 2014-2015
The property at Eynsford is located on a 4.5 m thick alluvium consisting of sandy clays overlying sand and gravel, before striking Seaford Chalk. Following an appraisal of local geology and available data, two hand-augered holes were drilled and equipped with piezometers, and two 200 mm production boreholes were drilled and tested, taking soil and water samples for laboratory analysis. A data logger was installed in one borehole to monitor water levels. Appropriate measures were identified to lower groundwater levels in the alluvial deposits underlying a property at Eynsford, to prevent recurrence of flooding which affected the house ground-floor in Jan-Feb 2014.
In 2014, water welled up through the joint between the kitchen and the adjacent room foundation, and a sump pump was used to evacuate groundwater collecting on the kitchen floor.
The aquifer causing groundwater ingress in the cottage is the shallow sand and gravel, fed from high groundwater levels in the Chalk. The overlying clay acts as a semi-confining layer, with rest groundwater level at 0.25 m below ground level near the kitchen door. Pumping tests showed that the aquifer had a hydraulic gradient of 0.0205 and transmissivity across the eastern edge of the property of 200 m3/d.
Evidence showed that groundwater levels rose by at least 0.75 m in extreme winters such as Jan-Feb 2014. Putting the flooding at Eynsford into context, winter 2014 was the wettest in records dating from 1776. Over the past 40 years, winter rainfall led to high groundwater levels in the North Downs on seven occasions: 1975, 1988, 1994, 1995, 2001, 2003, and 2014. In practical terms, groundwater flooding events may occur in this chalk block with an average frequency of between 5 and 10 years, and that events appear to becoming more severe, in a changing climate. Duration of groundwater flooding varies significantly, but was as long as six months in 2001.
The work provided detailed design, costing and construction supervision of a three-borehole scheme removing 3 l/s which would maintain water level below the kitchen floor during periods of high groundwater, and a monitoring system was left in place to trigger pumps.
Hambledon groundwater flood mitigation, 2014
Groundwater flooding of a basement was investigated on a large estate farm in Hampshire to assess the best engineering solution for mitigating the problem. Groundwater ingress was related to high groundwater in the Hampshire Chalk recharging water circulating in the sand and gravel alluvial deposits.
A solution was recommended involving the installation of well-points which would discharge to a nearby stream culvert.
Investigation of groundwater ingress at a property in Malvern, 2014
WRA investigated the source of water discharging in October 2012, through weep-holes in a retaining wall bounding a mews-style residential housing development on the east side of Malvern Hills.
Most of the Malvern urban area is located on clay, silt, sand and gravel superficial head deposits, classified as stony gravelly clay, which overlie sedimentary bedrock of the Sidmouth Mudstone Formation. In contrast, the Malvern Hills are the result of an igneous intrusion primarily of diorite and tonalite with some granite, with low porosity and permeability where groundwater responds quickly to rainfall and storage is confined to fissures that are the result of extensive fracturing of the rock. The fissure system discharges on meeting the less permeable sedimentary rock where over 130 springs have been recorded around the margin of the hills. Flow can cease in times of drought.
A lack of correlation between rainfall and the discharge from the retaining wall weep holes indicated that spring flow was not a source of the water discharging from them. Abruptly, discharge stopped in July 2013 on the same day that a leaking private mains supply pipe was repaired about 28 m up-slope of the weep holes. There was therefore strong circumstantial evidence that the origin of the water discharging from the weep holes was a leaking private mains supply pipe.
The repair was effected and the flow stopped before any direct measurements were made that would have provided definitive evidence of a link between the leaking pipe and the flow from the weep holes
Groundwater Control Measures at Caudle Green, Cotswolds, 2017
The project involved the investigation, design and construction of groundwater control measures at a rural location in the Cotswolds, near Stroud in Gloucestershire.
Hydrogeological appraisal of problems associated with springs and dampness at the rear of a property at Caudle Green identified three main points where remedial measures would be required to bring water under control at the property.
- Boundary stream diversion or culverting
- Excavation of a 6 metre limestone cliff-wall
- Construction of a horizontal aquifer drain to capture and manage the spring-flow.
The final phase was supervised construction and commissioning.
The work on the Boundary stream reduced the groundwater recharge of springs appearing behind the house, and a phasing of the activities simplified the excavation work so that work could be carried out in drier conditions.
The investigation work included construction of two shallow boreholes to establish the piezometric head in the buildings rear corridor and to determine the different lithological layers which cause water to emerge in this area.
Flooding - a Guide for Insurers, Lawyers, Developers and BuildersAt WRA we like to keep our clients "in the loop" rather than presenting ourselves as superior beings with esoteric knowledge grudgingly (and expensively) revealed. This guide is aimed at professionals, who are concerned with legal, insurance and development aspect of floods, but who have no specialist knowledge of hydrology. It aims to explain how floods are caused, the extent to which they can be considered as "acts of God" and how the activities of man can make floods more or less severe. Please view our flooding guide below.
Flooding Guide (Printable)
a) The drainage basin
Water flows downhill. If you were to stand on the top of a hill and drop some water just to one side of the highest point, the water would run in one direction; if you were to drop the water to the other side it would flow in the opposite direction. By walking over all the hills around a town subject to flooding you could work out, very laboriously, where the water that flooded the town came from. Fortunately using maps, contour or digital maps, you can get the same information. The total area of land from which water can contribute to the flood is called the drainage basin. The size and shape of the drainage basin are two factors which help to determine the severity of flooding.
b) The flood plain
In its natural state a river will over-flow onto land next to the river in most years. This area of flat land is called the flood plain.
c) Flood statistics
The pattern of flooding at a particular town or village is normally described statistically. As an analogy consider the National Lottery in the United Kingdom. Fifty percent of the money staked goes as prizes: one can therefore say the average punter wins 50p for for each pound staked. But this does not make sense as there is no 50p prize. What there is, is a series of prizes of different values and different chances of winning. Most often the £1 stake does not get a prize at all, there is a one in 50 chance of winning a small prize of 10 pounds and a one in several million chance of winning the big prize, the jackpot.
Flooding is bit similar. Most of the time rivers are not in flood, fairly often there is a small flood not large enough to cause significant damage, rarely there is a big flood which cause a lot of damage. And just as the average lottery prize is not very meaningful so it is with rivers. What is meaningful is the pattern of frequent and rare floods. That is why hydrologists will talk about a 1-in10 year, or 1-in-100 year flood; it is a way of quantifying the pattern of flooding.
The analogy should not be taken too far. The chance of winning the lottery is completely random whereas flooding can be, and often is, influenced by man. Another difference is that a draw on Wednesday is not influenced by what happened the previous Saturday - they are independent events. With rivers what has happened over the drainage basin in the days preceding a rain storm can affect the size of any subsequent rain storm. A final difference is that the lottery follows a well defined statistical pattern, that is why the lottery company was able to define the distribution of prizes before a single draw had been made, particularly since they knew they would be dealing with millions of participants so that any random differences would soon even themselves out. With rivers the statistics are built up after several years of observations; the greater the number of years of observations the more accurate the statistics are likely to be. Whilst flood statisticians would like to be able to achieve a similar degree of precision to the lottery this is not really possible - major floods are often caused by an unlikely combination of rare events and over time the pattern of flooding can be altered by the activities of man.
2. Factors naturally affecting the size of a flood
a) The rain storm
It is obvious that this is the main determinant. If there is no rain there is no flood; if there is heavy rain there might be a flood. It is important to appreciate that the type of rainfall also has a bearing. On a largely urban area with a small drainage basin bad flooding may be due to an intense thunder-storm, on a river like the Severn with a large drainage basin it might be due to several days of less intense rain. The weather pattern in the days before a storm can also influence the size of the flood. If rain falls on wet ground it is more likely to produce a flood than if it falls on dry ground.
b) Snow melt
Melting snow can also add to the size of a flood. In some countries, though not the UK, it can be the single major cause of flooding.
c) Drainage basin
i) Size, shape and steepness
As we said above there is a link between the size of a drainage basin and the type of storm which causes flooding. Every drainage basin has a characteristic called the "Time of Concentration". This can be considered as the time it takes for rain from the most distant part of the basin to reach the place where the flood occurs. Generally speaking, a storm which lasts for the same length of time as the time of concentration will produce the worst flood.
The size of the drainage basin is not the only factor. A drainage basin with steep slopes will have a shorter time of concentration than one with shallow slopes. A long thin basin will also have a longer time of concentration than a more compact one.
The time of concentration can vary from less than an hour, for a compact urban basin (a "flash flood"), up to a few days, for a large basin such as the Severn, Trent or Thames.
A basin with sandy permeable soils will tend to have smaller floods than one with impermeable clay soils.
For places near to the coast at risk from flooding tides can affect flood levels in two ways: Firstly, if flow in a river is an important component of the flood then high tide levels, whether increased by other factors or not, can increase the level in the river and hence the severity of a flood. Secondly, in many places the risk of flooding will come entirely from the sea. Where this is the case the risk would not come simply from a high tide but from one whose levels had been increased by other factors, e.g. wind strength and direction, which can lead to a "funnelling" effect, and low atmospheric pressure. At some sites the worst flooding may occur when high river flows, a high tide, low pressure and strong wind from a particular direction happen to occur at the same time. Some of these factors are independent of each other, a particular combination of the position on the sun and moon which might give rise to a high "natural" would not effect the strength and direction of the wind. On the other hand some factors can be linked, for example low pressure which might cause tide levels to rise can also be associate with heavy prolonged rain. This gives an indication of how complex calculating the statistical likelihood of flooding can be.
3. Factors artificially affecting the size of a flood
a) Land use
As a very general rule the less vegetation there is in the drainage basin, the more severe the flood will be. A basin with heavy tree cover will trap a lot of rain in the leaves of the trees, even grassland will slow down the runoff whereas bare soil will not. Industrial or commercial development with extensive paved areas of roads and vehicle parks, will lead to all the rain becoming runoff and contributing to the flood. Most housing development, with impermeable roofs and roads but with areas of garden and other open spaces, is somewhere between these two extremes.
b) Flood embankments
The effect of flood embankments, or other developments, on the flood plain is one of the most important factors in the pattern of flooding and also one of the least obvious. It may at first seem that building an embankment around a town or village liable to frequent flooding is simply a way of reducing flooding. Locally this is true but downstream the opposite is the case. If development takes place which blocks off all or part of the flood plain the downstream flood peak will be more like the higher upstream peak with, as a consequence, more severe flooding. As an analogy consider a wash basin. If the taps are partly open the drainage pipe can carry all the water easily - just like a river most of the time. If the taps are turned on full and the pipe can't get rid of all the water the wash basin will start to fill up - just like the flood plain. When the taps are turned off the excess drains away again. In this case it is easy to appreciate that the larger the basin, the longer the taps can be full on without the basin overflowing. Similarly with rivers, the more storage there is in the flood plain the less danger of the capacity of the river channel downstream being exceeded.
4. Things which don't affect a flood
This section may seem a bit out of place but it is important to recognise that some things which might be suspected of affecting flood levels generally do not. The first of these is pumping. If a river is used for water resources purposes water may be pumped into it. Not only is it extremely unlikely that pumping of this type would take place during a flood, the rate of pumping would be several orders of magnitude less than that of a major flood.
The operation of sluice gates might have a local effect for a short distance upstream (a few hundred metres at the most) but would have little effect further up or downstream of the gates.
If there is a reservoir upstream of the point of flooding, releases from the reservoir would also be unlikely to affect the size of a flood. Indeed releasing water when a flood was predicted to provide extra storage during the peak of the flood can be an effective way of reducing flooding.
5. Ways of reducing flood damage
a) Flood embankments
These are the most obvious way of protecting property however, as indicated above, they can have a negative impact further down. There may also be other complications, such as getting rid of runoff behind the embankments. It should also be recognised that flood embankments cannot provide protection against any eventuality. A normal standard is against a flood estimated to occur once every hundred years. As is probably clear from the above discussion on statistics it is not always easy to know to a high degree of accuracy what the 1-in-100 year flood is. This however is the solution which has been most often adopted in the past. Were it be adopted in the future great care will have be taken to avoid exacerbating downstream flooding.
b) Channel improvements and diversion channels
Whilst the option of enlarging a river channel, or providing a flood diversion channel, to carry more water seems to be an obvious solution it has a major drawback. The reason for this is due to something called the "regime" of a river. The size of a river is not an accident - it is a function of the pattern of flows in the river and the amount of sediment it carries. Over time the dimensions of a river will reach a state of equilibrium; no erosion will take place in a flood and no sediment will be deposited during periods of flow flow. Whilst natural rivers might have some growth of water weeds in the summer generally they will not become overgrown. Experience has shown that if a river is enlarged beyond its natural size it will soon start to silt up and vegetation may also increase, both of which reduce flow capacity. Similar problems occur with flood diversion channels, these can also be blocked with vegetation (and supermarket trolleys, old mattresses etc!). An additional problem with diversion channels is that suitable land has to be found to construct them and a mechanism devised to allow flow to enter during a flood but not at other times. If either of these solutions are adopted it means that regular, and potentially expensive, maintenance is needed.
c) Upstream storage
It is possible for a reservoir to be built upstream of a town or village liable to be flooded which can be operated in such a way as to absorb much of the flood and only release water at a rate which does not cause flooding downstream. It is, in effect, an artificial way of recreating the benefits of a natural floodplain. This solution is not one that is often adopted. However many reservoirs built primarily for water resources purposes are operated in such a way that they can provide some benefit in relation to downstream flooding.
d) Flood warning
Although this option does nothing to reduce the size of a flood it can be very effective in reducing the cost of damage. If house owners have time to take consumer electronic equipment, carpets and some furniture to an upstairs room it can easily save hundreds or thousands of pounds. For many people their car is their most valuable possession after their house so having time to drive the car to high ground can also save a lot of money.
It has to be recognised that the length of advance warning varies from place to place. We mentioned earlier the "time of concentration" of a river basin. It is not possible to give an accurate forecast further ahead than the time of concentration. This is because flood forecasting is based in part in estimating how soon water in the upper parts of a basin will reach the point where there is flood risk and in part on estimating how much the falling rain will add to the flood. As a result, whilst forecasts of a day or more might be possible at the downstream end of large river basins, the maximum period of forecasts may be less than an hour in small steep basins.
In addition to saving money we must not ignore the saving of lives. Even a short period of warning can enable people to leave areas where flooding is about to take place. Experience, incidentally, has shown that very few deaths during a flood are a direct and unavoidable result of the flood. Some of them are caused by heart attacks due to the shock of flooding or the exertion of trying to move too much in too short a time and others are caused by people taking unnecessary risks.
6. Where to go for more advice
If you are planning a development, the first place, obviously, must be the local planning authority. They will be able to tell you if a particular site is at risk from flooding and what measures might be required both to prevent flooding at a particular site and to prevent flooding upstream or downstream of the site.
The next source of advice is the Environment Agency in England and Wales (EA) or the Scottish Environmental Protection Agency (SEPA). The EA and SEPA have a "Floodline" on 0845 988 1188 which can advise on the risk of flooding at particular site (based on post code). Both agencies have websites with a lot of very useful information on floods including maps of areas at risk of flooding, advance preparations to take if you live a flood prone area and what to do during a flood. The addresses are:
For England and Wales: http://www.environment-agency.gov.uk/homeandleisure/floods/default.aspx
For Scotland: http://www.sepa.org.uk/flooding.aspxS