Thursday 19 January 2017

Metropolitan Line Southern Extension?

The Transport Secretary, Chris Grayling, recently killed off the plan to allow the London Mayor to take over the Network Rail lines in South East London. This was despite a high benefit:cost ratio and broad agreement between all groups concerned earlier in 2016.

Given this, and the difficulty more generally of providing enough capacity in south London, this blog considers an alternative plan to provide additional capacity for South East London.

Metropolitan Line Southern Extension

South East London has a reasonable network of rail lines, and they reach two good central London terminals at Cannon Street and Charing Cross. But the time is fast approaching when that network is simply beyond capacity. Rather than look at Crossrail 3, or other super-expensive solutions requiring a new tunnel across Central London, it is time to consider extending the Metropolitan Line. Such an approach would complement the proposed Bakerloo Line extension to Lewisham.

Having looked at the options, an extension to Abbey Wood seems to make the most sense. In this scenario, the Elizabeth line (Crossrail 1) would extend to Dartford (or beyond) using the existing tracks:

The full route would be as follows:

  • Aldgate - rebuilt, interchange with District line
  • City Hall - new underground station
  • Bermondsey - underground interchange with Jubilee line
  • Surrey Quays - underground interchange with Overground
  • Deptford
  • Greenwich - interchange with DLR
  • Maze Hill
  • Westcombe Park
  • Charlton - interchange with service to Lewisham & Victoria
  • Woolwich Dockyard
  • Woolwich Arsenal - interchange with DLR
  • Plumstead
  • Abbey Wood - interchange with Elizabeth line (Crossrail 1)

For comparison, here is the existing map:

The first question to ask is whether it is possible to extend the Metropolitan Line at Aldgate, the current terminus.

Given this is a site at the edge of the City, it is certainly tricky, but it seems that it should be feasible. Just south of the existing Aldgate station is a large bus station (behind the buffer stops in the second picture). The extension would use the bus station site to develop the new interchange station using standard top-down construction. To provide space, the Metropolitan line and Hammersmith & City line trains would have to terminate at Liverpool Street during the works, with the Circle line ceasing to run. The goal of the construction would be to build a four platform station - two platforms for the District line above two platforms for the Metropolitan line.

The Metropolitan line trains would need to descend from the current level to be beneath the District line. This would be achieved using the site of the existing Aldgate station (the first picture). Once complete, the tracks would be covered over, and the station turned into a bus station. The site of the new station (the existing bus station) would be developed.

Once complete and open, the service patterns of the existing lines would change radically. There would be no Circle line and no Hammersmith & City line. District line trains would all run from Earls Court through Victoria and Tower Hill to Whitechapel. Similarly, Metropolitan line trains from Hammersmith and Harrow would all run through Liverpool Street and on to Greenwich and Abbey Wood. In the west, a shuttle service would run from Edgeware Road to Gloucester Road in place of the existing Circle line. This service pattern eliminates most of the flat junctions from the District and Metropolitan lines, making services much more reliable and able to run at a higher frequency. Note that the loss of services from Liverpool Street to Whitechapel (the current Hammersmith & City line) is mitigated by the Elizabeth line, which runs on exactly that route.

Constructing the rest of the proposed extension is relatively easy by comparison with Aldgate station. It would involve two tunnel boring machines and three underground stations. The station at City Hall would be entirely new. The station at Bermondsey would be an underground interchange, designed for ease of use, while the station at Surrey Quays would require a rebuild of the Overground station as well.

The Bermondsey interchange is key to the success of the plan, because it provides passengers from the extended Metropolitan line a simple change to reach the West End. Journeys such as Greenwich to Bond Street become a pleasure, with one simple well-designed interchange. This is vital, as it greatly increases the time benefits to passengers, boosting the business case.

The final piece of the puzzle will be a tunnel portal location to access Deptford. One possibility would be to use the land of the New Cross branch of the Overground. Such an approach would allow more Overground services to run to Clapham Junction, Crystal Palace and/or West Croydon, but would also require the Metropolitan extension to have a short branch to New Cross.

Costs and Benefits

The benefits of this scheme are not limited to the line through Greenwich, because the scheme would free up paths into Cannon Street. (No services would run from Greenwich to Cannon Street - the line through Greenwich would be transferred from Network Rail to TfL.) Currently, there are 7 trains per hour from Greenwich to Cannon Street between 7am and 9am. These paths would be reallocated to other services, benefiting passengers on other lines, including the routes to Sevenoaks, Hayes, Bexleyheath and Sidcup.

Passengers currently using services from Greenwich would still have direct trains to the City, but would have a choice of Aldgate, Liverpool Street and Moorgate instead of Cannon Street. Passengers for London Bridge would get off at City Hall, while passengers for the West End would change at Bermondsey. As such, existing passengers would not see major changes to their journeys causing disbenefits.

Costs are always hard to estimate, but a rough guess can be based on 5km of tunnelling, four underground stations and line conversion works on takeover from Network Rail. Say £500m for the tunnels, £1bn for Aldgate, £2bn shared between the other three underground stations, and £1.5bn of other work, suggests a possible total of £5bn. This compares with £3bn for the Bakerloo line extension, so the cost estimate seems sound enough.

The cost of extending the Elizabeth line to Dartford would need separate examination. I'd note that initially, the Metropolitan line extension could run to Charlton, rather than Abbey Wood to side-step that problem.


This is a proposal to extend the Metropolitan line from Aldgate to South East London, taking over the line from Deptford to Abbey Wood via Greenwich and Charlton. It provides a step change in service to that line, a radically simpler and more reliable service on the District line, and an additional 7 peak-hour paths into Cannon Street for the rest of South East London and Kent. And all for around £5bn.

Given that there won't be a TfL run South Eastern Overground any time soon, a plan like this may well be the best way to improve rail services in this part of London. Thoughts welcome!


  1. I love the idea in principle, and I hope it is examined. But the sums you've used should add up to £6bn, I believe, not 5. That's double the cost you've cited for the Bakerloo Line conversion. If that's close I reckon it'll be a tough sell, politically, with most regeneration along the route already well down the line.

    On the plus side, the tunneling would be about 2 miles rather than the 4 required for the BL extension (I've just done that using an "as the crow flies website"). Not sure if that alters the figures. And I suppose it'd be 3 new stations/platform compared to 4 on the BL line. Are you sure those figures are close?

    1. Oops, it was meant to be £1.5bn of other work, which I've fixed now. (Tunnelling is cheap, so I'm not worried about the exact distance).

      The cost to compare with is the cost of providing increased capacity across the entire South Eastern network. How much would it cost to get seven more paths into Cannon Street without this scheme?

      The problem with premature regeneration is a more general one - the solution has to be to tax the value increase in housing due to public investment in transport.

  2. Wouldn't it be both better and cheaper to dig a deep level tunnel (actually two bores) for fast and outer suburban services from Dartford to Lewisham (-ish), leaving the existing lines for all-stopping services run by TfL?

    Since there is a capacity shortage on c2c it seems better to dig a tunnel from Aldgate eastwards via Stratford to somewhere around Forest Gate, Wanstead Park and Woodgrange Park to take over the southeastern part of Goblin and with some rework at Barking be extended both to Barking Riverside and Dagenham Dock (or maybe even switch services with some existing line, perhaps switch with the District at Barking making the Upminster slow line a semi-fast line and Barking Riverside (which is closer to central London) the slow line.

    The problem with extending Metropolitan or any other SSR line is that the stations is far shorter than most rail stations, giving shorter trains and thus not as good value for the money. IMHO it should be investigated what it would cost to extend the Metropolitan station platforms.

    1. The fast line from the Dartford area already exists - HS1 - and while it is far from perfect, it seems unlikely that another from that area would happen politically. Plus a fast line tunnel alone doesn't solve the central London problem, where there are no free platforms at the termini. What this plan does is use an existing tunnel through central London to dramatically reduce the cost of a new line in South East London. Compare Crossrail 2 at £30bn to this at £5bn. The Bexleyheath and Sidcup lines would probably get at least one more peak-hour train to Cannon Street with this plan.

      This is why the plan finds a way to allow Crossrail 1 to be extended to Dartford without needing two new tracks all the way. While I understand there are needs on c2c too, the routing you propose would not be competitive for passengers when compared with Crossrail 1.

      The fact that Metropolitan line trains are shorter is not a big deal if they run frequently enough. The current 7 trains per hour of 12 car trains is equivalent to 11 trains per hour of 8 car trains, and this plan actually results in around 18 trains per hour through Greenwich. Thus this plan would actually be a massive capacity boost for the area despite the shorter trains. No more looking at the timetable either - just turn up and go.

  3. A couple of questions...

    How many trains per hour would you envisage running to Abbey Wood in your proposal... and secondly... would this line not represent a good opportunity to connect Thamesmead to the network, by extending the line from Abbey Wood [I know it increases cost, and its only crayonista thinking but...] it would offer a better [more tph] option to the likely never to be extended Overground from Barking riverside ...

    1. I'd imagine there would be 30tph through central London in the peak, with perhaps 12tph to New Cross and 18tph to Abbey Wood. I suspect an extension to Thamesmead would be very expensive for the benefits. Thamesmead is currently looking likely to get the DLR, and I suspect if that happens, that is all it will ever get.

  4. Whether the down ramp, to take the Met deeper so as to safely cross under the Thames, can be fitted into the footprint of the existing Aldgate station, strikes me as a bit doubtful. I don't know what is underneath the tracks, but I suppose the descent could start earlier, just east of Liverpool Street station.

    Another snag is the loss of through trains from Euston (square) and Kings Cross to Mile End and everywhere east of that on the District line. Not just no through trains, but during the works, quite a long diversion (via Embankment or Victoria or with two changes).

    Snags like this (which may not be fatal) might show up better if your maps showed a bit more of central London. And you really need two maps, one during the works, and one when it's finished. Still, maps don't cost much!

    1. To build the new Aldgate station (under the current bus station) requires a length of 133m for the platforms. It also requires the Met to drop by at least 4m, which requires a further 160m (at 1 in 40). The total 293m is of course very tight, but assuming some commercial demolition, ought to be feasible.

      I'm not too worried about the loss of Kings Cross and Euston journeys. The Elizabeth line will capture many of those displaced, as it serves all four stops from Liverpool Street to Farringdon.

      PS. when drawing maps by hand, each extra bit is a lot of effort!

  5. Henning Makholm2 March 2017 at 17:57

    It's not clear to me why interrupting the Circle Line would entail a shuttle between Edgware and Gloucester roads. Instead, the disappearance of H&C from Barking would be filled with more District line trains into Tower Hill and continuing westwards in the same paths that the Circle Line uses today. As a baseline those would end up at Edgware Road. If a reform of service patterns in the Kensington area is desirable, this could be undertaken independently of any construction at Aldgate.

    1. You are correct that the two elements are not directly connected. However, once the Circle ceases to exist, it makes more sense to me to run extra services to Wimbledon and Richmond together with a shuttle, than running services from Barking to Edgeware Road.

  6. You say that the Bermondsey interchange is 'key to the success of the plan', but isn't the Jubilee already completely full westbound in the AM peak by Canada Water?

    1. Pretty much every tube in London is full at peak time! Some people would get off the Jubilee and change to the Metropolitan, while others from Greenwich would fill the space. Without detailed modelling it is hard to predict the exact flows (presumably these passengers today would change to the Jubilee at London Bridge, once it reopens properly). If it wasn't possible to board the Jubilee there, passengers for the West End could change to Crossrail 1 at Liverpool Street.

  7. This is a good proposal however I think the tunnelled section should carry on longer. Maybe end between Greenwich and Maze Hill? I think Woolwich Arsenal should be in its own tunnelled section as well to allow cross platform interchange between DLR or Crossrail rather than National rail services. It is interesting to think about how far the tunnelled sections of tube lines carry on for outside Central London.

    1. A key goal here is to remove Greenwich services from Cannon Street, to provide more services from other destinations. Tunnelling further would simply waste the existing Greenwich line, which would be extra expense for no real benefit. (While extra tunnelling is cheap, extra underground stations are expensive). In fact, the common feature of many of my proposals is to use existing infrastructure more optimally, to reduce the cost.

  8. Have you presented this proposal to TfL? This will be good for London.

  9. Looking at the tube map today I thought that there was scope for extending the Metropolitan line further past Aldgate, I am slightly miffed to see I didn't have an original thought! However, my idea was to extend the Met a little further east, towards the Isle of Dogs. I live in Island Gardens and my struggles with the DLR (speed, links to west end etc.) had me thinking that it could be a good idea to see some new lines through the Docklands. It takes me over 15 minutes to even reach an underground station. I thought extending the Met through possibly Tower Hill then possibly to Wapping/St. Katherine's, Shadwell/Limehouse/Westferry, possibly Blackwall then to Canary Wharf, Island Gardens (forgive me for being so selfish), Greenwich, Deptford Park and then Lewisham. This would serve the south east and Isle of Dogs much better, having a direct line all the way through London, rather than the rather slow DLR and the often long walks to a different station. Connecting through the Docklands may be a much harder task thought, as underground stations would probably have to be built at nearly all stops (and is it even viable) as well as possibly new tracks. Still, one can dream. Loved this post!

  10. I like this idea because it is an idea I have roughly envisaged myself. It could alleviate the stress on the other lines and makes the Met Line a proper train service from Northwest to Southeast. However, I would be interested to know how this would affect the services to Chesham, Amersham, Uxbridge and Watford.
    The station at Uxbridge is nearing capacity.
    Watford only has 2 platforms so there is a slight problem with expansion, especially seeing as it is outside central Watford and a fair distance from the other stations (Overground and Main Line).
    Amersham has 1 southbound platform used by both the Met and the Chiltern to Marylebone, while there are 2 northbound platforms for the chiltern to aylesbury: 1 platform for regular use and another for an alternative platform when the other is occupied.
    Chesham is a very small station with only 1 platform, small ticket room, not enough ticket barriers and almost no staff. Possible expansion of Chesham station would require full redevelopment, meaning more than 1 platform, more ticket machines, more barriers, more parking in a different layout, and by now, there should be a consideration for extension beyond Chesham to build potential new stations at Ley Hill and Bovingdon (both underground, possibly underground at Chesham as well with overground terminating services as they are now) and an interchange with the main lines at Hemel Hempstead, Tring and/or Berkhamsted. This would result in more links to London from rural areas, while also eliminating the need for any bus travel between stations. It would also complete the extension from Chesham that was originally intended since many years before the Chesham branch opened in 1889.
    This is an idea TfL should really consider and would make the Metropolitan line the true connector service it deserves to be.

  11. I’d been having similar thoughts except I thought more of a trajectory from Aldgate to Shadwell then parallel with DLR to Limehouse then curving southeast to interchange with the Jubilee line at Canary Wharf.then continuing via new stations at Kitts Town and Greenwich Central to terminate at Charlton.

  12. I was thinking smaller, wanting to spend less. I'd divert the Metropolitan Line, letting it follow the Thameslink alignment from Farringdon, letting it go underground south of Farringdon and come up again between Blackfriars and London Bridge, taking over some suitable services there.

    For a bigger plan, I'd let it go deeper north of Baker Street but follow the same route, in which case I'd use the freed capacity to extend the Wimbleware to Aldgate.

    1. Right now we've no idea where the demand will be in 10 or 20 years time, but its more likely to be in the eastern City fringes than the western ones IMO. As far as I can tell there is nowhere to bring a tunnel out between Blackfriars and London Bridge - if there were it would be better used to build a tunnel via Charing Cross to Euston replacing the surface tracks IMO.

    2. You are making a good point about demand.

      As for a tunnel to Euston, it would bring up the old problem of getting yet more lines with mixed overhead and third-rail electrification. The lines into Fenchurch Street have 25kV AC, just like those into Euston, so it would probably be better to link those, if (a big if) a good place could be found to take the Fenchurch Street lines underground.

    3. What service would take the place of the far western leg of the Hammersmith and City line?

    4. A (22/1/2021) The western leg of the Hammersmith and City would have the same frequency as it does now (the Circle and H&C would cease to exist, but the same frequency would run to Hammersmith branded as the Met instead. Edgeware Road to High Street Ken and south would run as part of the District line, again with the same frequency as today)

  13. Nice idea. I note you have a new Aldgate Station with interchange to District. Would this be a new station for both lines? If positioned further south it could have an interchange with Fenchurch St Station (would need a new access and gateline at Eastern end) and also with Tower Gateway DLR. The existing Tower Hill could be closed.

    1. Yes the new Aldgate would be a District/Metropolitan interchange. That is what makes the proposal feasible in terms of not losing journey opportunities. The proposed location was a bus station when this page was written, so could be developed, whereas a site further south would need more complex land purchase (an interchange with Fenchurch Street is of course desirable, but not at any price).