Leigh flood barrier weathers the storm

16 March 2001

The UK is well known for its annual rainfall and the wet winter months of 2000 put the country’s flood defences under extreme pressure. Suzanne Pritchard met with the manager of the UK’s largest flood relief scheme, and discovered that risk assessment played a vital role in the fight to tame the river Medway

Tonbridge in Kent is a small shopping town in the southeast of England. Only 30 miles from London it houses a population of over 30,000; boasts the historic remains of a thirteenth century castle; and has been home to me for most of my life. But why should this interest you?

Tonbridge is situated on one of the longest rivers in the southeast of England; the 110km long river Medway divides the town and has played a significant role in its development. Two road bridges had to be constructed to allow passage across the river, while the dramatic events of September 1968 prompted construction of a flood relief scheme: the largest of its kind in the UK.

A history of flooding

Historically, Tonbridge and its surrounding industrial and agricultural areas have been flooded at regular intervals by the river Medway. Records dating back to 1897 show that cataclysmic floods have inundated the town. Significant flooding occurred in 1958, 1960 and 1963. In July 1968 consultants Mott Macdonald were appointed to study the problem and recommend solutions. Ironically, shortly after this, a major flood event occurred.

‘In statistical probability terms,’ says Richard Francis, team leader of flood defence operations at the Environment Agency in northwest Kent, ‘a 1 in 100-year flood event was experienced in September 1968. The flow in the Medway just upstream of Tonbridge was 226m3/sec – enough to create deep flooding in the town.’

Over 2150ha were flooded, causing millions of pounds of damage (at 1968 prices) to domestic property, industrial areas and agricultural land. Certain points throughout the high street were inundated by over 2m of water.

Mott MacDonald’s reports of 1969 and 1974 looked at various channel improvements and schemes to protect Tonbridge, and neighbouring villages, from further devastation. The construction of an on-line flood storage reservoir immediately downstream of the confluence of the rivers Eden and Medway was considered the most effective option. Impounding flood water behind an embankment, which would incorporate a reinforced concrete flow control structure housing radial gates, would increase the duration and depth of flooding upstream. While flooding downstream, upon the return period of the event, would be eliminated or reduced significantly.

The best position to site an on-line flood storage area was 2km upstream of Tonbridge in Leigh. Geographically, this was a convenient place for a barrier as the river valley narrows and an embankment of just over 1km long could be constructed. Land ownership issues also played an important role in the decision. Most of the land to be inundated by the flood storage area around the village of Penshurst was owned by one person, Viscount de L’Isle. This would make negotiations to purchase the right to flood the land much simpler, while the area was already an established river corridor floodplain.

The decision to build the Leigh flood barrier was underpinned by the River Medway Flood Relief Act of 1976. The Act stipulates that certain flows must be maintained in the Medway but, if they exceed 35m3/sec, impoundment can take place. ‘This is the cornerstone of the scheme,’ Francis says. ‘It does not mean that we always have to impound at this flow.’

Site investigations began in the autumn of 1976 and the barrier was completed in 1981. The 1.3km long embankment has a maximum height of 5.7m. It consists of a clay core supported by gravel shoulders, and is founded on existing ground for most of its length. The flood storage area covers 278ha and the catchment of the barrier is 500km2. At maximum storage the scheme can store 5.77M m3 – making it the largest of its kind in the UK.

The barrier is classified as a Category A facility under the Institution of Civil Engineers’ definition in Floods and Reservoir Safety; meaning that a breach will endanger lives in the community. It was also designed for a 1 in 100-year event, as had been experienced in September 1968. ‘I think it was mainly economic justifications that led to this design specification and standard of protection,’ says Francis. And at the time no one knew that, after just 20 years of operation, Leigh flood barrier would be put to the ultimate test: facing another 1 in 100-year flood event.

A terrible winter

Throughout 2000 the UK experienced some extreme weather events. ‘In Kent we had a particularly wet spring,’ says Francis (picture above), who is responsible for operating the Leigh flood barrier. ‘Flooding occurred in May and, although June was dry, it was on the cards that if we had a lot of rain in the autumn, we were set for a terrible winter. And it came.’

Crowborough in East Sussex is on the border of the catchment of the rivers Rother, Uck and Medway. Four days prior to 12 October over 236mm of rainfall was recorded in Crowborough. ‘Consequently,’ says Francis, ‘the river Uck flooded the town of Uckfield, the Rother flashed down and flooded Robertsbridge and the Medway came thundering down towards us. We were the only ones to have control on any of these three rivers, and we had to set about exercising that authority.’

The whole situation was worsened by the fact that the land was incredibly wet already. ‘As a result of the rainfall we had experienced the soil moisture deficit was zero,’ Francis explained. ‘So there was immediate runoff. There was no ability of the catchment to retain any water. It was super saturated and all water was in the river – heading towards us.’


River level data for operating the Leigh barrier is obtained continuously by telemetry links from gauging stations, two of which are upstream on the rivers Medway and Eden. Automatic 1mm tipping bucket rain gauges also trigger alarms at 10mm and 18mm. However, the gauging channel at Colliers Land gauging station, 2km upstream from the barrier, became flooded. It is now apparent that only 30% of the flow was being gauged. Consequently on 12 October, Francis and his team did not have an accurate idea of how much flow was heading towards the barrier. ‘We only had visual reports of what was happening upstream,’ he says, ‘and we knew there was a colossal amount of water there.’

For the 72 hours leading up to noon on 12 October typical flows into Leigh reservoir were 50m3/sec. ‘Under previous management regimes they would have started impounding,’ says Francis, ‘as this would have been reservoir operation according to the Act of 1976. But I chose not to. I said that if a really big storm hit us, and the reservoir was already half-full, we would not have sufficient capacity to take the storm. It was purely my judgement.’

So three days prior to the big rainfall event Francis started his risk assessment. ‘Flood management is always about looking ahead. You don’t react to it, you have to be pro-active and anticipate the possibility of a big rainfall event following lots of smaller events, which is what happened.’

Francis allowed the river to flow naturally through the control structure. ‘If you looked at the river generally it was just full to the brim,’ he said. ‘At this point there was some flooding downstream in the recreation ground in Tonbridge, but this was just an inconvenience to the public. It hadn’t damaged property or endangered lives.

‘I decided to continue like this. We kept getting multiple plugs of rain: 10mm here and 15mm there just kept coming at us. Then suddenly on the night of 11 October 136mm fell. But I had an empty reservoir – it would have been a disaster if it had not been empty.’

Reflecting on his operation of the barrier, Francis is realistic. ‘To be honest,’ he says, ‘I don’t know whether it was luck or judgement. Amusingly the Met Office (the government weather station) informed us 12 hours before the big event that we were going to get 20mm of rainfall – when 136mm actually fell.

‘Even with this knowledge I knew I didn’t want to fill my reservoir then. I know that risk assessment comes into play at a much earlier stage than just the day of the crisis. So I had started to lay the foundations for successful flood management several days earlier.’

Prior to 12 October 2000, maximum flows in the river Medway, as experienced at the Leigh flood barrier, were up to 140m3/sec. ‘These were sizeable flows,’ says Francis, ‘and sufficient to generate a fair bit of impounding.’ On the morning of 12 October the flows started to rise steeply. It is now known that flows into the reservoir were in excess of 250m3/sec, exceeding the 1968 flows of 226m3/sec.

‘I gave instructions to follow the discharge as closely as possible to the inflow, up to an approximate 80m3/sec,’ Francis said. ‘But things became critical and I was aware that if we started discharging more than this, without preparing the public, emergency services and the local authorities, there would be widespread mayhem downstream.’

So the emergency planning department of the local authority was informed and plans were set in motion to evacuate vulnerable parts of Tonbridge. The services needed four to six hours to get people out and into safe areas, and Francis knew that this meant taking the risk that too much water would fill the reservoir without sufficient discharge.

‘You must remember,’ he said, ‘we didn’t have gauging information of any reliability. We didn’t know how much water there was until it hit the reservoir.’ Effectively working blind, Francis and his team decided to monitor inflows into the reservoir at 30-minute intervals. By measuring changes in the reservoir water level, changes in volume, and assessing discharge, they had an idea of how much water was coming in. ‘We got a good feel of the inflow,’ Francis said. ‘By repeating this every 30 minutes we could compensate for any mis-measurements. But we only knew the inflow in retrospect at half-hourly intervals.’

Evacuation was well under way in Tonbridge by now. Francis was advising the emergency services about which areas could be at risk from deep flooding, and was assuming there would be every possibility of extensive flooding in and around the town. ‘I was doing risk assessment all the time,’ he said. ‘Although I attempted to give it my best shot to manage every conceivable risk, I could not afford to expose the public without warning.’

‘We were now in a land that no man had trod before in terms of operating this flood barrier,’ Francis says. ‘Previous maximum inflows were 110m3/sec less than we were now experiencing. This was just a different league.’


Gradually Francis was increasing discharges from the barrier from 80-100m3/sec. ‘Things were becoming rather pressing,’ he explains. ‘And I started to ask emergency services for a timescale of when I could discharge what I wanted. I was concerned that if I didn’t have some flexibility then I would risk overtopping the dam. Strictly speaking it’s not designed for this and so there would be the attached risk of actually losing the whole dam. The worst case scenario – so things were hotting up for me.’

Reservoir levels continued to rise steeply. In one minute, a 30mm increase was recorded over the 278ha area. ‘I had to see this to believe it,’ Francis said. So he left the control room and walked down to the embankment. ‘Sure enough the grass was quickly disappearing further and further up the embankment.’

Francis was now pressurising the emergency services. He knew that there was tremendous organisation to be done downstream but if he lost the dam, Tonbridge would be lost anyway. As the pressure mounted a supervising engineer was invited to the dam, according to regulations of the Reservoir Act of 1975. When he saw the situation at 5pm on 12 October, he was also concerned about overtopping the structure.

By 7pm, Francis was allowed to discharge what was needed. By this time there was surface water in areas of Tonbridge high street due to surcharging of the drain system, but the water was still in the river channel. It was recommended that 200m3/sec should be discharged to save the dam. However, Francis decided to incrementally increase output through the gates at a steeper rate than previously, all the time monitoring the situation downstream.

Overtopping the dam was not Francis’ only concern. He was worried about the stability of river retaining walls around Tonbridge, and these were beginning to leak. ‘From my risk assessment I’d calculated that if I kept the river level 40mm from the top of these walls they shouldn’t topple over. However, what I couldn’t calculate was if the pressure of the water could actually push the wall sideways, and then the river would burst. These were, after all, just 34cm wide brick walls resting on a mortar foundation. They were built in the 1960s purely for ornamental purposes, not flood defence.’

At this point Francis decided to take the river as close to the top of the wall as possible, and watch to see if the inflow and outflow equated at the reservoir. At 11pm on 12 October the retaining walls were still standing, and the peak reservoir level for this flood event was recorded.

Discharge (150m3/sec) from the reservoir continued into the night until 4am on 13 October. Then the reservoir proceeded to empty during the next two days.

‘This was truly a 1 in 100-year flood event,’ says Francis. ‘I measured peak outflow at 150m3/sec but, due to insufficient gauging data at the time, we now believe that it was probably closer to 170m3/sec. We also reached the peak reservoir level within 14 hours, previous events have taken over 24 hours. And this made me realise how fast events were moving.’

Tonbridge and its surrounding areas were not saved entirely from flooding. A lot of damage occurred but no where near the scale as experienced in 1968. The benefits of the barrier are demonstrated by the difference in river levels between the flooding events of 1968 and 2000. ‘In October 2000 the river level was 90cm lower than it had been in September 1968,’ says Francis. ‘When we talk to members of the public that were flooded out of their homes or businesses, and we show them how much water we actually saved them from, they really begin to appreciate what could have happened without the barrier.

I can assure you,’ Leigh’s flood defence manager says, ‘it was dramatic.’

Colossal storms

Following the flood event on 12 October 2000, Francis and his team were unable to rest. More storms followed. On 31 October a big depression and very high winds brought in a ‘gentle giant’ of a storm. Inflows of 200m3/sec were recorded but the reservoir did not fill for over 24 hours.

Up to the end of December 2000, Leigh flood barrier witnessed a combination of a 1 in 400-year event. Four big storms over a two-month period proved to be the barrier’s biggest test of its 20-year existence. And so far this winter, as of the beginning of March 2001, the barrier has impounded 11 times. The barrier usually only impounds twice during an average winter.

‘We now have a lot more experience of what this barrier can and cannot do,’ Francis said. ‘It is 20 years old and showing signs of ageing, and there are things we need to do to strengthen its operation. For example we have to consider the retaining walls around Tonbridge – if they seriously limit what we can discharge then they are putting the dam at risk. We also know that, due to insufficient gauging instrumentation, with flows over 100m3/sec we are blind upstream. And it’s not ideal to rely on the calculations we were using. It has all been a steep learning curve.’

‘The important thing,’ Francis adds, ‘is that the events over the past few months have really demonstrated just how critical the barrier is. We can only store 25% of total water in the catchment during such storms, but this is enough to save Tonbridge from utter devastation.’

Privacy Policy
We have updated our privacy policy. In the latest update it explains what cookies are and how we use them on our site. To learn more about cookies and their benefits, please view our privacy policy. Please be aware that parts of this site will not function correctly if you disable cookies. By continuing to use this site, you consent to our use of cookies in accordance with our privacy policy unless you have disabled them.