Comment: How water can be a destructive force on the railway

Philip Haigh examines recent incidents in London and in the West Midlands, and discovers links with the results of water ingress that led to fatalities.

Water, water, everywhere.

And for the railway, often in the wrong place. City Thameslink station in central London closed in early February after a sprinkler system sprang a leak and flooded the electrics of two escalators at the Holborn end of the station.

The station normally closes on Sundays, and the leak happened sometime between the last train on Saturday February 1 and its discovery the following evening. By that time, it had filled the equipment spaces at the bottom of two escalators - which now need to be repaired, if not replaced.

Thameslink managed to have the Ludgate end of the station open by Thursday evening (February 6), but it could only allow trains to call after 0900 the following day.

This was because before 0900 the station is normally so busy that having just one exit was not enough to clear the platforms before the next train disgorged its load of City commuters.

Meanwhile, the Severn Valley Railway is celebrating its 60th anniversary this year. But instead of putting all its efforts into this great event, it’s working out how to cope with a crippling landslide found by its permanent way team on January 29, between Hampton Loade and Bridgnorth.

The slide has cut the line just south of the site of Eardington station, and has left its steam motive power depot isolated from the rest of the 15-mile route that runs into Kidderminster, where the SVR has its carriage shed and diesel depot.

The slip sits where the railway crosses a tributary of the River Severn on a single-arch bridge, with high approach embankments on either side.

In the days that followed, a group of Network Rail staff visited the site near Sterns as part of the partnership between NR and SVR.

In a BBC report of the incident, NR Central Route Infrastructure Director Adam Checkley noted: “It’s the river and the sheer power of the river that’s probably taken the wing wall of this particular structure away, and the earthwork has followed it through.”

Wing walls serve to retain the fill behind them, but also to protect that fill from the erosive effects of flowing water. In this case, it seems that recent heavy rain has contributed to a faster flow of water, and that has likely scoured beneath the foundations of this wing wall, leading to its collapse.

Scour from fast-flowing water is an old and well-known problem. And pictures from the scene show sections of the waterway’s banks that have partially collapsed, which all suggests a rapid flow of water.

I’m hugely relieved that the railway’s permanent way team found the problem, rather than a service train.

Sadly, that wasn’t the case on October 19 1987, when British Rail’s 0527 Swansea-Shrewsbury ran onto a collapsed bridge over the River Towy (Afon Tywi) at Glanrhyd, between Llandeilo and Llangadog.

Three passengers and the train’s driver drowned when the front car of the two-car Class 108 diesel multiple unit (DMU) was washed away, even as they were being helped into the rear car that remained on the bridge.

Flowing water had scoured a hole under one of the bridge piers, leading it to collapse sometime between the passage of a locomotive at 2120 the night before and the DMU’s arrival in the morning.

BR was using the train to examine the line, having received reports from that locomotive’s driver of floods along the line. The driver had even stopped his locomotive on the bridge, but (aside from seeing the river running very high) saw nothing amiss.

In those days, the Railway Inspectorate investigated accidents as part of the Department of Transport. Its reports are an odd mix of formal (we learn that the locomotive’s driver was Relief Driver E J Rossiter, for example) and informal.

So, Deputy Chief Inspecting Officer of Railways Alan Cooksey records the evidence of passenger Mrs A B Angus, who had boarded at Llandybie: “When it came to that part of the line where the river flows close to the right hand side of the track she saw the river was a ‘raging torrent’. She told me that there was water on both sides and Mr and Mrs Evans and herself had risen from their seats and that ‘we were terrified’. She said that they spoke about their fears, because they had never seen flooding like it before with the water running so very fast.”

Cooksey continues: “In order to distract herself from her apprehension she began to read a book.”

She survived, but Sarah Evans, William Evans, Simon Penny and Driver John Churchill did not. A coroner’s jury later ruled them unlawfully killed.

Cooksey comments that BR staff were concentrating on areas where floods had been reported this previous evening, not appreciating the severity of the flooding of the river. Inspections after the accident revealed more problems when the area civil engineer ordered checks of other structures.

“Later that day one of the patrolmen carrying out the task had the misfortune to actually fall through the track and down behind the abutment of a cattle creep which had been washed away.

“The cattle creep was a small bridge of about eight foot span some 70 yards from the River Towy and 100 yards from the River Dulais, and through which normally there was no water course. The effect of the flood had been to scour deep holes under the bridge and collapse both abutments inwards leaving the track and its ballast suspended above.”

Cooksey dismisses as unfounded suggestions made during his inquiry and at the coroner’s inquest that the train was driven at a speed inappropriate for the conditions.

Reading that reminds me of suggestions following August 2020’s fatal accident at Carmont that the HST involved should have been cautioned to run at reduced speed rather than line speed, as it returned towards Aberdeen after floods blocked its proper journey south to Glasgow.

Heavy overnight rain had blocked the line, but there was no specific reason to think that water gushing from shoddily built drains had washed debris onto the track.

It’s easy to say after the event that someone should have realised there were likely to be more problems than had been reported. But those railway staff on duty were too busy dealing with their known problems.

Cooksey says of the Glanryhd bridge and the cattle creep: “It was not known that either structure was suspect and the question of a railway structure being in doubt was not part of the decision-making process in this accident.”

Nor was the shoddiness of Carmont’s drain known at the time of that accident. But Cooksey concludes that BR’s engineers did not understand sufficiently the complex behaviour of rivers when in flood - just as Network Rail did not properly supervise the building of those drains at Carmont.

One of the twists in Carmont’s story is that NR had contractors on site at Bridge 325 where the derailed HST plunged down the embankment. They were installing scour protection to the bridge’s abutments.

And elsewhere, NR has flood markers on many bridges, with a yellow mark showing when restrictions should be imposed and a red marker above which NR will close the bridge to traffic.

It has also improved its weather forecasting, generating warnings when required.

Back in 1987, British Rail was not even on the distribution list for flood warnings from the Welsh Water Authority, so did not receive the most serious ‘Red 2’ warning issued on the evening before the accident.

Water also derailed Northern’s 0518 Preston to Barrow-in-Furness as it approached Grange-over-Sands at around 0605 on March 22 2024 (RAIL 1028).

In this case, a tamper damaged a pipe carrying pumped water from a golf course beneath the track to the sea. Tamper staff reported the incident immediately on March 20, but NR did nothing to stop the subsequent leak.

This leak washed away material from under the track, opening a void into which the track dropped as Northern’s train passed over it.

The pipe was temporary, although it had been in place since 2016, and it wasn’t marked on NR’s engineering plans, so the tamper crew did not know it was there. After the tamper pierced the pipe, there was what the Rail Accident Investigation Branch (RAIB) describes as “ineffective communications”.

NR’s senior asset engineer (SAE) thought the pump supplying the pipe was no longer in use, having arranged for it to be collected after the golf club told him it no longer needed the pump.

Reports that water coming from the pipe didn’t reach the SAE, but RAIB suggests that had they done so, then the SAE would have realised the risk.

In addition, when the incident controller first tried to contact the SAE as the first line on-call manager, they could not get through. The IC didn’t try the second line on-call manager, believing that they would not want to be disturbed just because the first line manager was not answering.

When the SAE called back, it wasn’t the incident controller but the incident support controller that answered - and they didn’t know the pipe was pressurised because it wasn’t mentioned in the control room log.

Hence RAIB’s three learning points from the accident: track quality supervisors should mark obstructions to tampers; on-call managers should be immediately contactable when on call; and incident controllers should properly escalate problems when they can’t reach first line on-call staff.

Water and railways really don’t mix.