Archive for Rail

A Cunning Plan

How UK railways improve – by carefully developing what is already there, rather than building miles of new lines.

The West Coast Main Line has dedicated fast lines (shown in red on the map below) for most of the way from London to Crewe, but there is a gap in the Stafford area. Fifteen years ago, the plan was to build a Stafford bypass, taking fast trains past three flat junctions on tight curves, at Colwich, Stafford and Norton Bridge. This consisted of a flyover at Colwich junction, so all fast trains would head towards Stoke, then a new line north of Stafford for trains to switch back to the line to Crewe (shown in yellow). But the economics didn’t stack up.


Instead, what is currently being built is much simpler – a flyover at just one of the three junctions, Norton Bridge (shown in green). Doesn’t this just leave the other two problem junctions unresolved? No – and this is the cunning part.

While ideally the fast trains would have a pair of lines to themselves, this isn’t actually essential unless the railway is close to maximum capacity. You can get a decent capacity for high speed trains just by eliminating conflicting moves. The odd two-track bottleneck is quite tolerable, as long as there are no flat junctions.

This has already been used to good effect between Rugby and Nuneaton. The fast and slow lines are combined in the northbound direction for a few miles, because a motorway bridge was built too narrow. This also, very cleverly, avoided the need for a new flyover at Rugby.

The cunning part of the Norton Bridge project is that the four-track railway between Crewe and Stafford will be paired by route, with services for the Trent Valley using the fast lines, and services for Birmingham using the slow lines. The speed of the slow lines has been increased to 100mph, and there are no intermediate stations, so journey times will be unaffected. This creates a conflict-free path for northbound fast services, increasing capacity. It also means there can be fewer sets of points south of Stafford station, allowing the track to be canted (tilted) for higher speeds.

In terms of existing passenger services, the southbound Trent Valley stopper will use the main line from Norton Bridge to Colwich, stopping at Stafford. Northbound fast services to Stoke will divert via Stafford and Norton Bridge, allowing Colwich to be reconfigured for higher speeds.

Freight is a bit more complicated, but the main requirement for freight is capacity, not speed. To remove the conflict, freight can be diverted away from the Trent Valley, via Penkridge, Willenhall, Walsall, Sutton Park and Water Eaton to Nuneaton.

This is how the conventional railway increases its capacity and speed, by carefully organising slow traffic out of the way, and removing conflicts, leaving a clear route for fast traffic. In effect, the two-track section is turning red.

The next cunning plan will probably be to reconfigure Crewe. The slower services can be routed to use the freight underpasses, though they will need new platforms and signalling. This will allow the constrained junction north of the station to be moved to the south, where there is much more room. This doesn’t need to be a flying junction, because capacity isn’t as tight when most of the trains are stopping, but the speed for Manchester services can be increased to a lot more than the current 20mph.

Map data © OpenStreetMap contributors. Base map is here.

Integrated Transport

The railways should be given the role of coordinating Integrated Transport. Information about buses (and cycling and walking) needs to be considerably improved, so people can readily understand and begin to trust the quality of connecting services.

When I’m going to visit a new city, I look to find out how to get from the station to my destination – be that somebody’s offices, where I’m staying, a conference venue or the city centre. In a small city, I’ll probably just walk. If there’s obvious public transport – a tram or a metro, I’ll use that. It’s usually next to impossible to decipher the buses. If I’m travelling on my own and feeling adventurous, I might look up potential cycle routes. But failing that, I’ll probably resort to a taxi.

Basel All ServicesWhy is it that integrated transport is so good in some countries, but so hopeless in Britain? A large part of it is down to our lack of trams. Tram and metro systems are generally simpler to understand, because the cost of infrastructure limits their complexity, and most routes have a good frequency. There is a strong tradition of diagrammatic mapping that people are used to, and which works well. Because it’s a diagrammatic map, you do then have to go to the effort of finding your station/stop on a street map, but the diagrammatic map inspires confidence, and makes that worthwhile. Here’s an example from Basel (there’s an even simpler version which just has the coloured tram lines, and leaves out the buses).

LuxembourgBusMapBus maps are generally more complicated. The norm in the UK is to use a different colour for each route in the suburbs, in an attempt to mimic metro maps, but to give up in the city centre. Buses often run in complicated loops in the centre, and if you have one colour per bus route, you would end up with multiple parallel lines, and an unintelligible tangle. Mostly, this type of map never sees the light of day, but here’s an example from Luxembourg.

Oxford Smart Zone MapThe best you can usually hope for in the UK is a central street map showing the stops, and a guide which lists the stops you can use for each destination. There might be another list that tells you how frequent each service is, but not how frequent they are in combination. You can work it out if you need to, but most people will have given up by then. Sometimes the city centre is just a mystery, like in Oxford.

paddingtonIf I’m visiting a city, I want to know whether there’s a frequent service from near at hand to roughly where I want to go. British bus maps are mostly useless for this purpose. The best in the UK are TfL’s spider maps, which are produced individually for each small area, and rely on the fact that most services are high-frequency and planned as a system. Even these maps resort to destination lists and stop codes, so take a bit of deciphering. Here’s an example for the area around Paddington.

An alternative approach is to treat buses like low-frequency transport, and provide a journey planner. This will give you a list of departures, but it doesn’t give you a feel for how good the service is. It might tell you there’s a journey every few minutes, but they might take different routes, or involve changing in different places. The great thing about a map is that you can see this straight away, not have to decipher it from a list.

BusMap_150_CentrePlusI’ve been developing a solution for my Oxford bus map, which is to use a small range of colours to show the routes that buses take in the centre. There are usually only a few distinct routes through the centre, which then spread out in the suburbs. I’ve limited the map to the routes taken by higher-frequency services, to keep it reasonably simple. Lower frequency branches in the suburbs are shown as dashed lines, to give a clear visual indication of where the quality drops. In this way, the central map blends seamlessly into the whole-city map, so it can be zoomable, and overlaid on a street map.

Why can’t we get good information already? Well mainly because the private bus operators are focused on selling services from the suburbs to the centre, not on providing a service that works for visitors. Some operators provide excellent information, others little more than a printout of a spreadsheet. The councils have responsibility for public transport in general, but are very patchy in the information they offer. Mostly they have subscribed to providing journey planners to discharge their basic responsibility, and left the private operators to provide further information.

This is why I think the responsibility for Integrated Transport needs to be formally given to the railway. The railway already takes on part of the responsibility, but it’s still rather patchy. The railway companies are the key beneficiaries of good integrated transport – they are the ones that get the income for the majority of the journey. So the railways are best placed to make a system that works well for visitors.

Integrated transport is mostly about information. In an ideal world, the bus companies would be improving their connecting services, and the local authorities would be improving conditions for cycling and walking. But for the moment, it is mostly a matter of providing the information, good and bad, and letting travellers choose accordingly. When we can easily see the good examples, there will be more incentive for improvements to be made.

In recent years, there’s been a lot of enthusiasm for providing bike parking and bike hire. In countries like the Netherlands, there are huge cycle parking facilities at stations, and the two modes definitely complement one-another: 40% of train trips in the Netherlands are combined with cycling. The railways certainly have an interest in providing cycle parking and cycle hire, because there is never going to be enough space for commuters to take their bikes on the train. So this is a significant part of the market, and it does need to be provided for. But unless the conditions for cycling in our cities improve dramatically, it’s hard to see cycling becoming a major choice for visitors.

I think the emphasis needs to shift to promoting walking and local public transport. The NationalRail website needs to be much clearer about providing onward travel information, starting with walking. There should be a link to a zoomable street map for each station, perhaps as an icon next to the arrival and departure station, as on the Swiss Railways site.

Reading Station PosterAt the moment there is a link about three levels down which takes you to a pdf of the station information poster. While it’s good that these posters have been produced, they are unwieldy and a long way from best practice. Outside London and the major cities, the posters are very basic (in London and the major cities, they use TfL or PTE maps, which are slightly better, but still limited). The maps don’t give you any feel for what the city is like – the cartography is minimalist, and the cycle and walking routes seem to be almost random. They don’t really tell you whether this is the sort of station where you walk straight out into the city centre, or whether cycling is fairly safe, let alone where buses run.

Reading OSMA proper zoomable street map would at least give you a feel for what the area around the station is like: are there major roads to cross, do you have to walk along a main road or cross at a major junction to get into the centre? Is there a pedestrianised area and shops close by, or is it industrial? Open Street Map already gives a good impression of this, though it could be improved (it doesn’t show pedestrian crossings, for instance).

In the same way, you could have a map focused on cycling – one that shows where the routes are (not just odd bits of cycle path), whether there’s any provision on main roads, and whether you can cycle in the pedestrianised area. There can also be links giving more information on cycle hire and how it works.

The biggest challenge is to map the central bus services, but that’s not impossible, as I have shown. Some of the TfL and PTE maps are pretty good, though they are often focused on taking you off to the suburbs, rather than getting about the centre. And of course, every map is different. There needs to be a consistent zoomable map, showing clearly where services run in the centre.

In the UK, I think the starting point is to make the railways responsible for providing good consistent readily-accessible information. We have the technology; all we need is an informed customer.

Railway Boundaries

Whether trains are run by public or private companies, most services will be specified by government and organised by geographical area. But where should the boundaries be?

There’s often a tendency to run the long-distance services separately, because they are less dependent on public subsidy. Arguably, long-distance and local services need a different style of management, and the conventional wisdom was that it would be hard for one company to do both. But having separate companies means additional interfaces, unresolved conflicts, bigger meetings, and meetings instead of phone calls. Alternatively, it might be better if companies were small and focused: they certainly seem to perform better, especially if they are self-contained.

Since privatisation, the boundaries between franchises have changed. Some franchises have been combined to make for a single large operator in a region – creating Greater Western and Anglia. A franchise for Wales was created, to match the political geography, and Central Trains was split up because it was sprawling, unfocused, and performing badly. The local services in the north east were combined with those in the north west, but Transpennine was separated out.

Organising services by region seems to be effective. There doesn’t seem to any real problem with combining local and long-distance services. Nobody seems to place any great value on franchises competing directly with each other. So for a lot of the country, the question is pretty much settled, from Anglia clockwise around London to Chiltern, plus Wales, Merseyrail and Scotland.

But the organisation of services in the rest of the country is unresolved. This comprises the services currently run by East Coast, Northern, Transpennine, East Midlands, West Coast, CrossCountry, and London Midland. The perennial problem is the local services, which require substantial subsidy. It makes sense to coordinate them with the mainline services, to maximise connecting revenue, but it also makes sense to focus on local journeys into the major cities, and follow a regional political agenda. Maybe one day the cities will restrict car traffic, reducing the subsidy required.

One key problem is Birmingham, where CrossCountry and London Midland are both significant operators. I would suggest transferring the West Midlands local services to CrossCountry, so that there is one main operator. The remaining London Midland services would then transfer to West Coast, so there’s one main operator for all the services into Euston.

In the North, despite a certain amount of political cooperation, the two sides of the Pennines are really quite separate. I think it makes more sense to combine the local services around Manchester with West Coast, and the local services in Yorkshire with East Coast.

You’d then end up with four franchises: West Coast based in Manchester, East Coast based in York, East Midlands based in Derby and CrossCountry based in Birmingham.

There are a number of options with Transpennine services. In some ways it makes sense running services for long distances. Passengers make all sorts of overlapping journeys, and through-running reduces the need for interchange. But it adds to operational complexity – trains have to be fuelled and cleaned away from their home depot, and there have to be small driver and conductor depots. And the further you run, the more chance of a delay. This isn’t a sensible way to operate low-margin low-frequency services.

It’s better to split operations at natural thresholds, either places where there is little traffic, or where almost everybody changes trains anyway. Manchester provides a natural hub, so I think it is better to split services there, with East Coast providing the main services across the Pennines, and down to Nottingham. East Midlands would then concentrate on covering the east-west flows further south.

What people want from the railways is a well-managed public service. A proliferation of operators doesn’t help, and we should move to a situation where each region has one main operator, with Network Rail organised in the same way.

The map shows a possible split.

Map data © OpenStreetMap contributors. Base map is OpenStreetMap Transport layer.

Birmingham New Street

Birmingham New Street is the hub of the national rail network. It has long been a dark and confusing place, buried under a shopping centre, with dim lighting and lots of blank walls that all look the same. But a glance up at the departure board always gave me that “where shall we go today” feeling, and suddenly the world seemed a brighter place.

It was hard to recommend New Street as a place to change between trains, so the railway made a virtue of running through services. Which is a pity, because sitting in a train in the bowels of the station is hardly the best way to see Birmingham. In fact, unlike many cities, the station is in the heart of the city centre – in two minutes you could be on the main shopping street, or sunning yourself in Victoria Square.

This comprehensive pattern of through workings comes at a cost. It means that services are a bit irregular, don’t depart from consistent platforms, and are prone to delay. In fact, small delays are endemic. With some trains running head on, some reversing, and some terminating, they inevitably get in each other’s way. The reversals in the high-numbered platforms, in particular, have never really worked.

New Street is currently being rebuilt, with a huge atrium to let in the light. It will be a much more pleasant place to change. At platform level it will be brighter, though still pretty disorientating. But the main improvement is that there will be a large central concourse, big enough to let you get your bearings, rather than just follow the signs.

We should take this opportunity to simplify the pattern of services. Instead of alternating between head-on and reversing, the main CrossCountry services would all run head-on. This will allow departure platforms to be standardised, and substantially reduce delays. It will inevitably force a few people to change who did not have to before, but the overall effect should be substantially positive.

All services from Reading would run to Manchester, and all services from Bristol would run to Sheffield. This is mostly just a replatforming exercise, but it would mean the service to Reading leaving slightly later (as it often does anyway).

The great virtue of simplifying the service pattern is that it will demystify New Street. The staff will soon get used to the idea, and be able to tell you which platform you need without a second thought. But it will also make it a lot easier for everybody to understand. Rail travel shouldn’t be a great mystery. It should be as easy as ABC.

The platforms would be roughly as follows:

High Speed or not High Speed

High speed rail has its benefits, but we don’t have to build it all straight away: use high speed to enhance the existing rail system, not replace it.

Normally, it’s a good idea to separate fast and slow railway services. There are multiple intermediate stops on almost all corridors, so you need four tracks to run fast services. Approaching London, you ideally need six tracks, so the urban stopping service is separate. That’s the situation on the West Coast Main Line – six tracks to Watford, four tracks (mostly) to Crewe.

Around London, many of the main lines were four-tracked decades ago, in the days when there were far more freight trains. This typically extends beyond the limit of outer commuter services, allowing the fast trains a clear run. This is why there hasn’t been the pressure to build high-speed lines – we already have an extensive network of fast lines.

With the exception of Kent (where there had never been much four-tracking), the sensible option has been to increase linespeeds, build flyovers, and infill gaps in the four-track network.

And so it might continue if it weren’t for Milton Keynes.

The two largest rail markets in Britain are London to Manchester and London to Birmingham. The number of people wanting to travel would justify half-hourly non-stop services, but the infrastructure does not allow this. At the London end the fast lines are getting close to capacity. We are getting to the point where either everything has to stop at Milton Keynes, or nothing.

Approaching Birmingham and Manchester, there are sections of two-track railway, and the fast services have to run between stoppers. To make this fit, the fast services have to stop a couple of times. These extra stops do mean that there are more services to Birmingham/Manchester, and the high frequency partly compensates for the loss of speed, but there isn’t much value running more than three trains an hour.

With Milton Keynes planned to expand, the simplest option is to stop everything at Milton Keynes, in much the same way as everything out of Paddington stops at Reading. This slows all the fast trains by about five minutes. So the high speed line isn’t really about solving a capacity problem, it’s more that the capacity problem reduces speed, making a high speed line more worthwhile. Stopping everything at Milton Keynes would probably deliver enough capacity for a few years, but it wouldn’t take much more growth before something further would be required.

The engineers have identified a route for a new line, and have ended up proposing a tunnel almost all the way from central London to Amersham – about 40km. To justify the cost of this, the route to Birmingham and Manchester needs to be ultra-fast throughout, and the line needs to be full from the outset, with branches to anywhere big enough to fill a train. This intensity of service also requires a major expansion of Euston station, demolishing several blocks of social housing in the process.

The expensive parts of the high speed line are the new terminal stations and the tunnelling into the cities. The French typically don’t do either of these – they concentrate on building new lines between cities. Now that we’ve identified just how expensive the full scheme would be, I think it’s time to consider intermediate options that leaves the most expensive sections to later.

Looking at the problem incrementally, the starting point is finding a way to serve Milton Keynes using the slow lines. There is plenty of spare capacity on the slow lines, it’s just that the intermediate stops get in the way of running fast services. The incremental solution is therefore to provide a way for fast services to overtake slow ones. The sensible place to do this is on the stretch between Kings Langley and Tring, where there are five stops fairly close together, and a major road to build alongside.

Six-tracking between Kings Langley and Tring would allow fast services to run on the slow lines to serve Milton Keynes and Northampton. Unfortunately, the slow lines through Northampton aren’t fast enough to provide a good alternative for destinations further north (eg Euston-Rugby, or Milton Keynes-Birmingham). The loop through Northampton is two miles longer, generally has a lower speed (75mph), and a severe restriction through Northampton (30mph). It might be possible to upgrade the existing line, but it wouldn’t be cheap. To provide for fast services to destinations further north, it would probably be necessary to build a new line bypassing Northampton.

The approach to Birmingham is two-track from Rugby, and completely full with a mix of fast and stopping services between Coventry and Birmingham. Running more services would require four-tracking or building a parallel line for at least some of the route.

So, catering for growth will need additional infrastructure at several points along the route. Joining these up and building a completely new line is clearly a sensible alternative, especially if it means that fast trains can then go even faster. This pretty much makes the case for a new line from the outskirts of London to the outskirts of Birmingham (eg Kings Langley to Berkswell).

The HS2 scheme for the Birmingham approach is proving to be quite complicated and expensive, having to slot between canals, railways and motorways. Speeds cannot be particularly high, since the trains are about to stop, and there is little advantage compared to the existing (100mph) line. It would be more effective to four-track the existing line, and provide capacity for other fast services as well. There is also, broadly, enough capacity at Birmingham New Street, assuming fast trains don’t get any longer than at the moment (260m), and slow services run through, rather than terminating.

The HS2 scheme for rejoining the West Coast Main Line is to head north from Berkswell, rejoining beyond Lichfield. This is duplicating the existing 125mph four-track route through the Trent Valley. It would be substantially shorter to diverge further south, and rejoin near Rugby.

The current works at Norton Bridge are essentially completing the separation of the fast lines as far as Crewe. When Crewe is resignalled in the next few years, it’s likely that linespeeds through the station will be increased for trains that don’t stop, particularly for those going to Manchester.

While it’s then two-track to Stockport, it’s fast, and not too many other services to get in the way. Through Stockport it’s slow, but there are four tracks, and potential to dedicate two of them to fast services. One possibility would be starting the four-tracking at Cheadle Hulme, eliminating the junction, and then providing a flyover just south of Stockport to take the slow services over to the west side without conflict. This would allow linespeeds to be increased, and provide enough capacity for some fast services to run non-stop.

At Manchester Piccadilly, the new platforms on the through lines will provide capacity for more terminating services in the main station. Further platforms can be constructed on the east side of the station, when they become necessary.

So, there is a coherent package of improvements that can provide substantial new capacity, and faster journey times, without the expense of tunnelling into London, Birmingham and Manchester, and without wholesale redevelopment of stations. The full HS2 package would provide even more speed and capacity, but it’s not at all clear that this is required immediately. Better to do the partial scheme first.

In a similar way, services to Leeds and Sheffield can be improved incrementally. The key bottleneck is south of Hitchin, where there are too many outer-suburban services on the main line. One possibility is to bypass the bottleneck, by using the fast lines out of St Pancras to Bedford, and then a new line alongside the A421 to St Neots. This would provide enough capacity to run additional trains non-stop to Leeds, and to run services via the East Coast to Sheffield. Unfortunately, there is a shortage of platforms at St Pancras, but it should be possible to create a link to the St Pancras lines from Kings Cross, at the north end of the railway lands.

Rather than build an all-new high speed line, I think a greater improvement can be made by selective new build and four-tracking more of what we already have. That way the benefits will be felt more widely, and the new capacity will be put to good use, rather than duplicating what we already have. In a small country, it’s the ability to run non-stop that makes the critical difference to speed.

The map shows the existing four-track railways (in yellow), the suggested high-speed lines (in red), the suggested four-tracking on the approach to Birmingham and Manchester (in green), and the suggested link between the Midland and East Coast main lines (in blue).


Map data © OpenStreetMap contributors. Base map is OpenStreetMap Transport layer.

Lower Fares / South London Overground

To really reduce fares, we need to make rail into an all-day transport service. South London is a key opportunity for transformation.

It’s become an annual tradition: when the inflation index is published for July, the campaigners start putting out press releases about what that means for rail fares in January. Eventually the government agrees to reduce it by a little. This makes the campaigners seem important, and the government seem generous, so everyone’s a winner. But it detracts from the underlying issues.

To get fares down properly, we need lots more people using the railway off-peak. The cost of running the service is mostly fixed, so the more people travel when there’s spare capacity, the lower the cost per journey. In due course, this will feed through to lower fares.

Longer-distance services into London roughly pay for themselves. This is where rail is strongest – running fast into the centre of London. It’s also where improvements to rail have the greatest effect, because getting to and from the station has (proportionately) least effect. Inevitably, railway management tends to focus on these services, rather ignoring the shorter-distance ones.

This is a problem, because it is the short-distance services that require the subsidy. But it’s also an opportunity, because short-distance services are more likely to work for leisure journeys. All you need is a high-frequency service, easy access to and from the station, and something worth going to.

Putting this together, there’s a distinct virtue in short-distance services being managed separately from long-distance services, and integrated with connecting services in the centre. If you can combine that with improving the attractiveness of the city centre, all the better.

For a number of reasons, the short-distance services in south London have remained part of the national rail network, rather than being incorporated into the Underground. The network is more interconnected, so it couldn’t readily be linked into the Underground one line at a time, as was done in north London. Instead, the whole inner network needs to transferred in one go, with just a few fast lines left for the outer services.

That wasn’t feasible until a few years ago, but the advent of HS1, and Thameslink begin to make it possible. There will need to be some further alterations, but on nothing like the same scale. The outer services need sufficient capacity so they can largely avoid interacting with the inners. I’ve identified the key improvements that will be required:

New flyovers north and south of East Croydon. These allow fast services to run from Victoria and London Bridge through East Croydon without conflict, but with cross-platform interchange. This increases the capacity for outer services, allowing the inner services to terminate. The first flyover is at Windmill Bridge (just north of East Croydon), and will take the southbound Victoria fast line (blue) over the London Bridge fast lines (orange). This reduces the inner services to a single line for a short way, but this is sufficient.

The second flyover is at South Croydon, and will take the southbound Victoria fast line (blue) back over the London Bridge fast lines (orange).

A new bridge, and a rearrangement of lines at Chislehurst. This diverts the Kent mainline outer services (orange) away from the stopping services (green), so that they can take over the lines to Chatham and Maidstone. A new curve between the fast and slow lines allows the stopping services from Bromley South to run to Orpington without conflicting with the fast services.

Six-tracking of the section from St Johns to New Cross. This keeps the outer services (orange) separate from the Charing Cross and Cannon Street locals (green). The outer services can then run into Thameslink using the new flyover at Bermondsey. This requires the widening of the cutting under a local street, and rebuilding a number of bridges. There are a couple of tight spots, but it looks feasible.

With a few fairly simple changes, all the local lines out of Victoria, Charing Cross, Blackfriars, Cannon Street and London Bridge can be separated from the outer services, and integrated with the Underground. Off-peak service frequencies can be increased, making rail into a true all-day transport service. If this is combined with making central London into an attractive leisure destination, then more people will travel off-peak, spreading the costs, and (ultimately) reducing fares.

Original imagery courtesy of Google Maps.