Monday 31 October 2011

The Heathrow tail wagging the HS2 dog

Update 2011-11-02: This blog was written just before Labour suggested routing via Heathrow and Lord Foster release the Thames Hub proposal. Read on to understand why it is essential that the UK decides on Heathrow vs Thames Hub *before* deciding on HS2 (if Heathrow declines or is closed, the sensible HS2 route would be very different)

For HS2, a number of groups have come together in the Right Lines Charter to encourage Government to think a little more deeply. The first principle of the charter is that High Speed Rail should only be introduced within the scope of a national framework. This is especially important when considering access to Heathrow.

As I've noted before, Heathrow is a major factor in the current HS2 route proposal. The original terms of reference framed the route planning with a "via somewhere near Heathrow" approach, partly to replace domestic air routes by High Speed Rail. Those opposed to Heathrow's third runway were in agreement as removing domestic air "obviously" helps Heathrow.

Thus, HS2 route evaluation was constrained to include only those routes running to or near Heathrow. (Old Oak Common was the choice made).

This constraint had some big effects:

Firtsly, the Heathrow constraint forced the route to Birmingham and the North from London to start in West London, rather than in Central London. A simple look at the map shows how the proposed line from Euston spends a long time heading west and south(!) to reach Old Oak Common, before it starts any travel to the north. Clearly this is only justified if Heathrow is a big enough target.

Secondly, the Heathrow constraint lengthened the route from the shortest possible (by adding the dog-leg via Heathrow). As such, it became necessary to make the trains go faster to meet the journey time reductions that Britain's broken transport economics require. Hence the 400kph design speed, which in turn lead to the straight line route through the Chliterns.

Thus, the net effect of the terms of reference and focus on Heathrow was to compromise sensible evaluation of other options. For example, the M1 route was rejected partly because the evaluation still required the line to go via Old Oak Common, which is clearly a crazy direction if you want to go from Euston to the M1 (just look at a map!!!)

To be clear, the HS2 route is a result of the constraints applied - of "somewhere near Heathrow" and "serving Birmingham in phase 1" The constraints became locations - Old Oak Common and Birmingham Interchange - and between these two constraints, the HS2 route chosen is one of very few options available. It is the broken constraints I object to, not the Chiltern routing per se. (ie. if I agreed with the constraints, and thus the constraining locations at each end, I would have relatively few qualms about the chosen route.)

But is Heathrow important enough to justify all this?

Well, I started this article with a link to the Right Lines Charter and its desire for a national framework for the investment. An absolutely key part of that framework is a national policy for airports.

Put bluntly, the UK and London airport situation is a national disgrace, with Heathrow merely one element. Its the classic example of how the UK fails to properly provide national long term plans that make sense and aren't overtly political.

Most MPs, and perhaps many people in the UK, still see Heathrow as something important. A vital hub airport in the world's network.

But the truth is that Heathrow is no longer in the first tier of world hub airports.

A two runway airport simply doesn't cut it anymore as a hub. Paris, Amsterdam and Frankfurt now clearly lead Heathrow in Europe (with four or more runways), with British Airways now looking at Madrid (four runways) as the solution to the lack of space at Heathrow. This matters in that there are far fewer direct flights to destinations in China, India and other upcoming destinations - from London you have to change at another hub. Over time, this acts as a major disincentive for multi-national inward investment based growth in the UK.

So, the solution is to remove all domestic flights from Heathrow, leaving more space for international flights?

Well no. Only around 10% of all take offs and landings at Heathrow are domestic, and some of those are flights across the water to Belfast. So, while HS2 could remove some domestic flights from Heathrow, it certainly wouldn't remove enough to convert it into a viable first tier hub.

Thus, a key national question for politicians is whether they want the UK to have a globally significant first tier hub airport. Heathrow isn't one (and a third runway probably wouldn't have been enough, as four runways is the current "standard" for a hub).

There are various possible solutions to the hub issue, including an island to the east of London, greatly expanding either Stanstead or Gatwick to 3 or 4 runways, or expanding both Stanstead and Gatwick to 2 runways each - sending each airline alliance to a different airport. Another solution is to say that the UK doesn't need a first tier hub. (HS2 implies use of Birmingham to relieve Heathrow, which is fine if everyone travels to airports by train, but we should recognise that simply isn't the case, thus Birmingham Airport is a bit of a red herring here.)

The key point I'm making is that Heathrow is a bit of a dead duck. Its not going to grow any more. In fact its going to reduce in importance over the next 50 years, not increase.

So, the truth is that Heathrow simply doesn't matter enough to the UK over the 50 year investment lifetime of HS2 to force the entire High Speed route strategy to serve it. The vast majority of passengers from the northern cities want to go to London (or somewhere around London) not Heathrow. Airport traffic as a proportion of the total is actually very small (HS2 Ltd estimate less than 2000 passengers per day).

And its this uncomfortable truth that the Right Lines Charter's "national framework" drives at. Without an airports policy that underlines the lack of importance of Heathrow, its difficult for politicians and civil servants to realise that the HS2 focus on Heathrow is plane stupid!

(As it happens, Old Oak Common serves a second purpose - linking to Crossrail, but that has generally been seen as a secondary item compared to serving Heathrow. In reality from a transport perspective, the Crossrail distribution effect is far more important than Heathrow, but getting that message across to non-rail audiences seems to be hard. However, if you remove Heathrow, it also turns out that the distribution effect could be fulfilled in other ways, such as linking to Thameslink if routing via Luton/West Hampstead.)


Accepting that Heathrow is on the decline is sometimes hard, especially when it seems like a permanent fixture in UK geography. But it is simple logic to see that Heathrow will over the next 50 years come to matter to UK air travel less and less.

But if UK air policy decides to shift to an estuary airport, or a hub at Stanstead/Gatwick, then it completely changes the rationale for a High Speed Rail route directly to "somewhere near Heathrow". Heathrow is no longer important enough to the UK to justify that.

And as I've described before with HC-Midland, removing that broken constraint can result in a much more sensible routing for the new lines the UK rail network desperately needs.

Friday 28 October 2011

Birmingham Interchange is not an interchange

In my ongoing evaluation of HS2, its time to write about Birmingham Interchange.

Birmingham Interchange, is the HS2 station located on the edge of Birmingham. The site is next to the M42 1 mile east of Birmingham International station on the existing main line (the airport station). The intention is to add a cheap "people mover" to transfer passengers to the NEC, airport or existing station. There is also vague talk of a tram line (which doesn't materially affect the following conclusions).

The executive summary is that the plan for Birmingham Interchange is dire. Read on for the detail.

The key issue is that the station is completely disconnected from any form of public transport, meaning that the vast majority of passengers will be expected to arrive by car, for which 7000 parking spaces are planned. But the M42 in that location is one of the busiest stretches of Motorway in the entire country and well beyond its design capacity. This combination is a recipe for disaster. HS2 Ltd comment as follows (pdf, chapter 8):

Our estimate is that in 2033, the interchange station would attract a net addition of 1500-2000 cars trips to the area during the morning peak hour.

The chapter is worth reading to understand the analysis done. It makes no reference whatsoever to access by public transport. Only access by car is considered.

The station is intended to be used by passengers living in various parts of the West Midlands conurbation and in Coventry (nobody lives close enough to walk to the station). Its only fair to look at what these passengers do today, and what they will do in the future.

Many in the West Midlands conurbation currently use local rail services to access London trains at Wolverhampton, Birmingham New Street, Birmingham International or Coventry. With HS2 there will be no direct access to the new line by local rail at any of these locations.

Residents in the west of the conurbation around Wolverhampton can currently travel to Wolverhampton station for their long distance service. HS2 does not serve Wolverhampton, so those passengers will either use the existing service (assuming it continues) or drive to Birmingham Interchange. (The official document assumes Wolverhampton will use the city centre station, but they can't as New Street and Fazely Street aren't integrated)

Residents in the centre of the conurbation can currently use local rail and buses to access New Street for their long distance service. Since the HS2 station in Birmingham is disconnected from the local rail network, there will be far fewer uses of the local network to access the long distance services. Thus again, passengers using HS2 are most likely to use a car to access Interchange. (The official document assumes these travellers will use the city centre station)

Residents of Coventry currently have an excellent three trains per hour service to London. This will be scaled back after HS2. Passengers in Coventry will be encouraged to travel to the Birmingham Interchange station instead. Which again means a car. Certainly no sane person would take a local train to Birmingham International, then a "people mover", then an HS2 train. (The official document assumes these travellers will use cars to access Interchange station)

Taken together, HS2's Birmingham stations are a recipe for a massive increase in car journeys.

However, there is a catch. 1 in 3 households in the West Midlands don't own a car. So, by forcing car access, HS2 manages to reduce the overall mobility of the population!

Moreover, what has long been understood by those in the rail industry is that once people start their journey by car they are more likely to just keep on going by car. If you're driving from Wolverhampton to eastern Birmingham, why not just keep on going and drive all the way to London, or if you're in Coventry why not carry on driving to Manchester or Leeds?

It turns out that the rail network is absolutely dependent on good interchanges and junctions, where local journeys can be fed into national ones. HS2 ignores over 100 years of experience of what makes railways work by not buliding proper interchanges.

Interestingly, the London stations on HS2 are planned to be integrated with the local networks, at Euston and Old Oak Common (although the Old Oak plan is far from certain at this point). It is clear to me that Birmingham has drawn the short straw here. The effect of HS2 is to boost car journeys within the West Midlands area and reduce demand on the local rail network.

Given the lack of integration, especially relative to London, it is patent nonsense that Centro, the publicly funded body for public transport in the West Midlands, is supporting the current HS2 scheme.


Overall I think HS2 is a bad project and there is a better and more cost-effective approach - HC-Midland. However, my position is far more nuanced than that of the "anti" brigade. For starters, while the route through the Chilterns isn't my preferred choice and is far from ideal, I don't think that aspect of the scheme would be disasterous if it were to go ahead.

However, I am utterly clear that both HS2 Birmingham stations are unfit for purpose on basic transport grounds.

It is simply unacceptable and irresponsible in today's world to build a station whose only sensible access is by car and then have the cheek to call it "interchange". That mistake is compounded by the lack of integration in the City Centre.

In summary, there has been a lot of arguing about rail capacity, the business case and the importance of the environment. In my opinion, not enough focus has been placed on whether the line's stations meet basic transport criteria. Sadly, HS2's Birmingham stations fail spectacularly on those basic transport criteria.

Monday 17 October 2011

Swanlink at Waterloo - photos

This blog expands a little on the Swanlink scheme around Waterloo.

A key part of Swanlink's relatively low cost is using surface access as far as Waterloo. Here I show some pictures that aim to explore that area in more detail:

At the south-west end of Waterloo is Westminster Bridge Road. The road is directly under the station throat. Swanlink proposes to build a new bridge over the road, at least two tracks wide. This requires the demolition of some non-listed commercial properties and relocation of railway equipment.

Moving north-west along the south side of the station the Swanlink project would remove the taxi road completely, lowering this to ground level or slightly below. The Spur Road ramp from Baylis Road would be redeveloped to replace the facilities lost underneath the current taxi road (such as deliveries and Waterloo & City line access). The taxi road (which is about 5m above ground level) is shown here:

East of the station a row of non-listed commercial buildings would be taken:

This provides access to the Waterloo (Cornwall Road) bus garage which would be the tunnel portal site:

Finding a route from there to Blackfriars Road through Southwark station will not be easy, especially given some buildings which are likely to have piled foundations. A tight curve may be needed, but speeds can be relatively low just outside Waterloo. The proposed Bankside station would use this land on Blackfriars Road:

Overall, the route on the map has been chosen to avoid tall buildings, but without detailed sub-surface maps it is impossible to confirm the route. Finding a route for tunnelling in London is hard!

Monday 10 October 2011

Swanlink - Crossrail for SW London

This is the third blog in a series of three on Crossrail, the major new rail tunnel running under London. The first and second entries were written to set the scene for this entry, which focuses on solving the major problem of Crossrail - two eastern branches - with the Swanlink proposal.

Low frequency branches

The big problem with Crossrail is not that there are two eastern branches, but that the dividing point of those branches is so close to central London. This means that of the 24tph serving the central section from Paddington to Whitechapel, only half can continue to Stratford and half to Canary Wharf. However, there is clearly demand at both Stratford and Canary Wharf for all 24tph to go there.

The official response to this appears to be threefold. Firstly, that the capacity of each train is large, secondly that the trains can be lengthened by 2 carriages, and thirdly to design the signalling for future running of 30tph. The first and second points are fine, but miss the fact that trains of similar size already run via Stratford and are already full when running at a higher frequency. The third point would still only provide 15tph down each branch, but more importantly removes resilience from the network meaning worse recovery from incidents.

So, I've looked at what is needed to fix the problem, making the Crossrail investment dramatically more effective to the whole of London.

Proposal: Swanlink

This proposal is designed to solve multiple problems with London's rail network in a single hit. As such, it should have a good rate of return.

Clearly, the only way to solve the eastern branch problem is to run a full 24tph service on each branch, to at least Stratford and Canary Wharf. Simple logic therefore requires some new tunnelling from the Stepney Green junction to provide the necessary tracks. But the question is where should they go?

At this point, I switch to the recent London & South East RUS, which identifies capacity shortfalls in the railways around London:

With respect to longer distance services the RUS therefore notes that a significant peak capacity gap may arise, with a forecast shortfall in capacity for some 7,000 passengers in the busiest peak hour; this figure includes capacity required on today’s already overcrowded trains, along with the 3,500 resulting from future growth.....

An alternative way to increase capacity on the route would be to increase the number of tracks from the Surbiton area to central London from four to six, but this is only realistically achievable by means of tunnelling over a long distance. Such a tunnel would need to fit into a cross-industry strategy for future underground lines in the capital in general. The RUS has therefore worked closely with Transport for London to identify a variant of the currently safeguarded Crossrail line 2 route, and this forms Option F7 in this RUS.

Clearly, the SWML (South West Main Line) has serious capacity issues and Crossrail 2 is being mentioned as a solution. But looking at the problems with Crossrail 1 as I have done above, its clear that the SWML can be relieved at a lower cost by reusing part of Crossrail 1 being built now.

The Swanlink proposal connects the SWML to the Stratford branch of Crossrail.

The principle is to acknowledge that Crossrail 1 is actually "one and a half crossrails". By building the aditional "half a crossrail" linking to the SWML, it is converted to 2 independent Crossrail lines each with 24tph - one from west of Paddington to Canary Wharf and Abbey Wood ("Wharflink"), and one from the SWML to Stratford and Shenfield ("Swanlink" - South West ANglia).

Swanlink detail

The junction at Stepney Green is to the east of Whitechapel so thats where the proposal must start. The first point to note is that there will need to be an interchange between Wharflink and Swanlink. The second point to note is that many commuters will want to travel from Stratford to Canary Wharf, changing at Whitechapel (it will be faster than changing to the Jubilee line at Stratford). As such, an interchange at Whitechapel makes sense. This should be designed so that passengers can easily travel from Stratford to Canary Wharf and vice versa. This requires a same-level interchange between passengers travelling west on Swanlink and east on Wharflink, and vice versa. The best design for this station is to place the two new platforms directly beneath the platforms being built now, reusing all of the surface access facilities.

The next point to note is that most travellers from the Shenfield/Stratford route currently travel to Liverpool Street. As such, removing their ability to travel to the Liverpool Street area is unacceptable. Thus the Swanlink route needs to run to Liverpool Street. Two more underground platforms will be needed, again directly underneath those being built now reusing the surface facilities. This time, the same-level interchange must be between passengers travelling east on Wharflink to east on Swanlink, and vice versa.

The combination of these two interchanges allows a passenger to change to the other line in any direction without needing to use any stairs, escalators or lifts. This saves time for commuters, building the business case.

After Moorgate, the proposed route takes a sharp turn under Guildhall to a station underneath Queen Street north of Southwark Bridge ("Queen Street" station). This would serve the western side of the city including Cheapside, Queen Victoria Street and Mansion House. Options exist to link this station to Mansion House, Bank and Cannon Street tube stations, but modelling would be needed to determine which if any makes sense. As this section is mostly under streets it should avoid foundations.

After Queen Street, the line would turn under the Thames to reach a new station under, or just to the east of, Blackfriars Road ("Bankside" station). This station would be linked with Blackfraiars Thameslink station (at the new Bankside entrance) thus provding another key link as at Farringdon. Land is currently available south of Southwark Street to assist building work here. This station may be able to link to Southwark tube station, but this may be of limited value. From there, the line would turn towards Waterloo.

The key to the affordability of the proposal is at Waterloo. South of the station there is a taxi road at the same level as the station itself. I propose replacing this with a new through platform. My analysis suggests that the limitation of 1 in 30 for gradient means that the line would need to go under Waterloo Road, not over it, so the new platforms would be at ground level. The tunnel portal would be on the site of the Cornwall Road bus garage requiring limited demolition of 2 or 3 non-listed commercial buildings. From there it routes to the "Bankside" station described above. West of Waterloo, a new bridge over the A23 would be needed to clear the Waterloo station throat and link to the existing SWML Wimbledon slow lines, again taking 3 to 5 non-listed buildings.

The advantage of this route is that the line requires an absolute bare minimum of tunnelling, probably with just one TBM drive for each running tunnel. The platforms at Waterloo would be above ground aiding access and construction. And the route serves two stations providing good access to Bankside (where large developments are proposed) and the western City area. The problems lie with the steepness of the grdients and navigating through the maze of underground tunnels, notably at Southwark tube station.

If the City route is not possible, then there is an alternate alignment, with a single station lying between St.Pauls and Blackfriars (see map). If it is not possible to build the Waterloo surface station, then a longer tunnel would be needed with a portal potentially as far out as Clapham Junction.


Swanlink dramatically takes the benefits of Crossrail to a whole new level. Pasengers from the entire South West corner of London would be linked in with big journey time savings.

Services would run from the existing Wimbledon slow lines (including Guildford, Epsom, Hampton Court, Chessington & Shepperton) through to Shenfield. It is intended that some Woking semi-fast services would move to Swanlink, freeing up the SWML fast lines as needed by the RUS mentioned above. (The Woking semi-fasts would stop more often than now, but they would gain time not having to wait for a platform at Waterloo, and in journey times beyond Waterloo.)

By itself, linking Wimbledon to Stratford is of limited value to commuters (although it would boost travel to Westfield Stratford City). The bigger benefit is for SWML commuters to reach the City and Canary Wharf directly. City commuters currently change at Waterloo to the Waterloo & City line, the 521 bus or a bike. They would now be able to travel directly to the City without changing, perhaps saving 20 minutes or more in each direction at peak time. Commuters to Canary Wharf from the SWML currently take the Jubilee line, whereas with Swanlink they would simply change at Liverpool Street (on the same-level) to access the 24tph Wharflink service, again with a big saving of time and hassle. Its important to note that the journey from Waterloo to the north and east sides of the City are very difficult today, so there is a big benefit there too. Finally, Stratford/Shenfield line passengers get much better access to the South Bank.

Estimated journey times
Journey*Now (off-peak, TfL website)With SwanlinkSaving
Wimbledon to Liverpool Street392217
Wimbledon to Stratford503020
Wimbledon to Canary Wharf39318
Wimbledon to Hackney573522
Wimbledon to Hoxton513219
Wimbledon to Woolwich513813
* Note that Wimbledon is used as the base location, but journey time savings apply over the entire suburban route network via Wimbledon.

Clearly these new links would greatly relieve the Waterloo & City line and the Jubilee line. Given the queue for the Waterloo & City (and crush conditions), this is a very good thing.

Finally, the plan permits 24tph between the City and both Stratford and Canary Wharf, a key design goal itself, reducing the inevitable overcrowding that will occur when Crossrail opens. Beyond those two stations there is no need for 24tph. Therefore, investment could be made to add new branches - to Barking, Upminster, Brimsdown or Chingford beyond Stratford, and to Dagenham/Grays or Thamesmead/Grays beyond Canary Wharf. These would be independent investments and are not included in the Swanlink case presented here.


The cost of Swanlink can be roughly estimated by comparison to Crossrail. That project has a budget of about £15bn. Swanlink is probably between a quarter and a third the size of Crossrail. It would have less initial setup costs (as it would follow on from Crossrail's up front investment), and also re-uses facilities at Liverpool Street and Whitechapel. The relatively short tunnel and surface station at Waterloo help keep costs down. An initial estimate suggests a figure around £4bn.

The quantifiable benefits arise mainly from journey time reduction, which would be considerable given current change to the crowded Waterloo & City line, 521 bus or Jubilee line. Further benefits accrue from the Wider Economic Benefits regime (Moving to more productive jobs, agglomeration effects, etc). Serving the Nine Elms regeneration area is clearly a huge boost.

Direct revenue can also be raised. The proposal also serves some London boroughs currently paying the lower levels of the Mayor's Crossrail Levy - increasing this to the higher level would also directly raise cash.

Swanlink also reduces the pressure on the SWML outlined in the London & South East RUS which advocated building Crossrail 2. Since Crossrail 2 (Clapham - Hackney) would be a £15bn+ project, and Swanlink is a £4bn project, so there is clearly a saving to be priced into the business case. Potentially, that £11bn difference is enough to justify Swanlink by itself. In addition, 5 terminus platforms at Waterloo are freed up, which can be used by other services, reducing the need to invest in bringing the Eurostar platforms back into use for SWML services, another saving that can be factored in.

Overall, I'm confident that the CBR for Swanlink would be good.


Swanlink is a logical extension to the current Crossrail scheme that suffers from having two eastern branches which divide too close to the City. By linking Stratford eastern branch and the RUS identified gap on the SWML a coherent transport solution is obtained. By reusing the Stratford tunnel and portal large cost savings are obtained, ideal in the age of austerity. The benefits to South West London are huge, with the Stratford/Shenfield and Canary Wharf/Abbey Wood lines also benefiting from more frequent services.

I hope that readers use their imagination and see how this investment makes sense and could really make Crossrail shine! There really is no cheaper way to get a second Crossrail line in London.

If you back the proposal, or have any other opinions, why not leave a comment!

Sunday 2 October 2011

Crossrail, east of Whitechapel

This is the second blog in a series of three on Crossrail, the major new rail tunnel running under London. The last entry focussed on the route West of Paddington, while here I'll focus on the two Eastern branches. The final entry will examine the central section and put forward a new proposal to improve the return on Crossrail dramatically.

The two eastern branches divide at Stepney Green, east of Whitechapel, where a shaft is being dug to allow access. The north-east branch goes to Shenfield and the south-east branch to Abbey Wood.

Shenfield branch

The Shenfield branch is designed to improve capacity and quality for the line that currently runs from Liverpool Street to Shenfield. Trains exit the central tunnel at Pudding Mill west of Stratford and then run on the slow line, leaving the fast line for services from Southend, Chelmsford and Colchester.

The slow line consists of existing stations spaced in a very similar way to tube lines. As such, it is clear that the plan for this branch is to operate it very like the District line to Upminster, or the Central to Epping. Although not 100% clear, it appears that the plan is to operate all services as "all stations" to Shenfield.

The Shenfield branch is due to operate at 12tph in the peak and 8tph off-peak. This is one train every 5 minutes in the peak, and one train every 7.5 minutes off-peak. Unfortunately, the route from Shenfield to London already has 15tph in the busiest peak hour. The current plan is for the additional "residual" 3 or 4tph to continue to run terminating in the main Liverpool Street station as they do now. This is undesirable as it means some platforms at congested Liverpool Street still have to be reserved for the local services. Off-peak, the current service is 6tph, so 8tph would be a welcome enhancement.

So, clearly the 12tph service is insufficient for the demands of the Shenfield route, however the situation is worse than that. Crossrail will be the fastest way to travel from Stratford to the West End, thus demand from Stratford in is likely to be higher than 12tph can handle. Many commuters from further out (Chelmsford, Colchester, Wickford, Epping, etc) are likely to want to change onto Crossrail at Stratford in order to avoid the change onto the Central line at Liverpool Street. This will place a lot of pressure on the service. Retention of the remaining services terminating at Liverpool Street shows that demand on this route is already high, so it seems unlikely that 12tph will be enough.

There is also difficulty with the timetable detail. The current peak timetable mixes services starting from Ilford, Gidea Park and Shenfield, with some of those from further out skipping some of the stations close to London. These commuters will probably want to keep this, but with no passing loops, it will be tricky. It may be that the service becomes tube-like "all stations" once Crossrail starts.

Looking at extensions, with the limitation of 12tph in the peak, I don't believe that it is feasible to servce any more destinations. Were the service receiving all 24tph from the central tunnel there might be more possibilities, but as it stands, Shenfield is all that is possible, and even that is troublesome.

In summary, the Shenfield branch is simple, self-contained and about the right length. The main problem is that there needs to be 24tph between Stratford and the City, and probably 18tph beyond that, yet only 12tph will be available. The "residual" Liverpool Street service is also a bit of a mess.

Abbey Wood branch

The prime reason for the Abbey Wood branch is to serve Canary Wharf. It is likely that the scheme only got political/business/financial support by routing via Canary Wharf. That said, it is certainly the case that Canary Wharf could do with another transport link. The Jubilee Line upgrade has shown how fragile transport to the area is and how the DLR cannot cope on its own. Once Canary Wharf was served, the line needed to terminate somewhere, and Abbey Wood was the option chosen.

Abbey Wood itself is effectively a convenient end-point however, rather than a true destination. It is basically a housing area, including access to Thamesmead area which has no rail access at all. Early planning did not terminate the line at Abbey Wood, instead it was intended to continue via Dartford to Ebbsfleet, see this 2002 pdf map. An extension from Abbey Wood to Dartford and Gravesend is still officially considered possible in the future.

The choice of Abbey Wood does have benefits to the services into Charing Cross and Cannon Street too. Some passengers will now travel via Crossrail rather than the existing route. This will provide some welcome relief on those lines. It should also have a effect on the DLR, as passengers switch from DLR to Crossrail.

The Canary Wharf station is being built in the dock just to the north of the main towers. This approach is perhaps the only possible option given the depth of foundations of the skyscrapers.

However, as with the Shenfield branch, the Abbey Wood branch will only be receiving 12tph - one train every five minutes. Given the crush on the Jubilee Line and DLR today, and massive growth in new buildings in the area it seems likely that 12tph simply won't be enough between Canary Wharf and the City. Beyond Canary Wharf to Abbey Wood the 12tph service will probably be enough.

Looking at extensions, the line to Dartford does make sense to be converted to Crossrail. In my opinion, other services should then terminate at Abbey Wood, such that only Crossrail serves the line between Abbey Wood and Dartford. If regeneration at Ebbsfleet takes off, then I think serving there rather than Gravesend would be the best choice.

In summary, the Abbey Wood branch exists to servce Canary Wharf and provide a useful link to the south east corner of London. The destination of Abbey Wood is clearly a cost-saving exercise as opposed to the more suitable destinations of Dartford and Ebbsfleet/Gravesend. As with the Shenfield branch, the main problem is that there needs to be 24tph between Canary Wharf and the City, but only 12tph are being provided.


The eastern side of Crossrail is a fudge. The two branches are fine in isolation, but trying to serve both will result in a lower quality railway and overcrowding. In the next blog I'll look at the choices for solving the problem of two eastern branches.

Comments welcome!