According to the London 2050 report’s forecasts, the demand for the Underground will rise by 60% in the next thirty five years. That’s a challenging target to address with additional capacity, given the pressures the network is already experiencing. Growth will not be limited to peak hours – nevertheless how do you get 60% more, even as a basic target, with varying additional capacity on different lines?

High pressure tubes

The capacity differentials between lines will become critical at interchanges, with extra passengers off a busy but not jammed-full Tube being the straw on the camel’s back for a busier line that people then seek to board. Station and interchange capacity will become at least as important as train capacity and frequency.

A partial solution set out in London 2050, and already part of TfL’s thinking, is a heavily automated tube and sub-surface system to maximise train frequency. A ‘New Tube [train] for London’ (NTfL) is another, such as the ‘EVO’ train with an up-to-11% capacity gain from the design, on a like-for-like basis with existing tubes – mainly because it has continuous internal space, not offering capacity on a car-by-car basis.

London Reconnections has covered elsewhere the detailed issues that the design objectives create in planning and equipping lines and stations, the requirement for new trains, and the extent of future driver involvement in train operation. See the New Tube for London, Automatic for the People, the New Sub Surface Timetable and the Sub-surface Railway Big Plan, as examples for Underground operations, and the discussion on Bank Station as a token of the complexity of rebuilding busy tube interchanges.

Getting geographic

Higher housing densities, geared to Public Transport Accessibility Level (PTAL) indicators or town centres, or to inner London (itself a combination of both those), will cause large-scale capacity pressures on London’s rail. This applies for inner London with all of London 2050’s Central Activity Zone options (we’ve covered these previously here) in the absence of new surface transport intermediate capacity solutions in inner London. Maximum ‘train slot’ capacity and route efficiency must be sought, whether on tubes, main lines, DLR, Overground or Tramlink in Croydon.

It is considered sustainable and efficient to create this maximum capacity rail infrastructure, to support high workplace densities in some very accessible locations, with less car-related travel to work. This is despite the knock-on effects on surface transport capacity (car/bus/two wheels/on foot), which must then be managed better with these higher densities, to minimise additional congestion on the surface access routes and to share available road space more rationally.

Tube Capacity

It is not intended in this article to discuss each scheme in detail, as many are already the subject of individual commentaries in London Reconnections and other websites. Instead an overview is offered about the outcomes and issues arising with the combination of tube capacity schemes.

For example the Victoria Line is targeted towards 36 tph by 2016, which is a tremendous task to deliver consistently well, with seconds counting at each station stop and at the termini. It is already at 34tph (up from 33tph a couple of years ago), so there is less capacity gain (only 6% at 34tph) but every extra train is well worth having, representing another 1,100+ passengers at any one location.

Given the 2050 end date, it’s useful to corral projects in terms of schemes already under way or subject to significant planning, and to highlight the new PPCs (‘prospective project corridors’) on the blocks in this General Election year, and their general rationale.

Since the ‘shopping’ list’ of Underground improvements was set out in July 2014 (see the appendices to London 2050’s Transport Supporting Paper (TSP), there have been changes to the priority and delivery date of some projects. Also the London 2050 list of Underground schemes discussed tube projects and capacities, but not the sub-surface network.

The principal sequences are set out below, as they were in July 2014. The changes will then be summarised.

Central London schemes already funded or partially planned

(MPPA = million passengers per annum, Yellow = longer term aspirations)

Scheme Timescale MPPA Output Benefits Business Case # £m Capex
Victoria Line capacity upgrade 2016 200 33 to 36 tph, +9% capacity Better jny times, reduced crowding, wider econ. gains 1 – London growth, best use of resignalling 30
Jubilee Line capacity upgrade 2019 214 30 to 34-36 tph, +13-20% capacity Better jny times, reduced crowding, wider econ. gains 1 – London growth, best use of resignalling 200
Northern Line capacity upgrade 2 2022 252 24 to 30-36 tph, +25-50% capacity Better jny times, reduced crowding, wider econ. gains 1 – London growth, best use of resignalling 560
Piccadilly Line deep tube project (DTP) 2026 a2 2029 a4 210 24 to 33-36 tph, ca +60% capacity Better jny times, reduced crowding, wider econ. gains 1 – London growth, end of life assets, full automation 7,700 at 2012, 12,600 outturn
Central Line deep tube project 2029 a2 2032 a4 261 30 to 33-36 tph, ca +15-25% capacity Better jny times, reduced crowding, wider econ. gains 1 – London growth, end of life assets, full automation 7,700 at 2012, 12,600 outturn
Bakerloo Line deep tube project 2035 a2 NOT full automation 111 22 to 27 tph, ca +30% capacity Better jny times, reduced crowding, wider econ. gains 1 – London growth, end of life assets 7,700 at 2012, 12,600 outturn
Bakerloo Line upgrade 2 2045 a? [supports SE extn] 200 27 to 33-36 tph, ca +22-33% capacity Better jny times, reduced crowding, wider econ. gains 1 – London growth, best use of extra capacity 200-400
Bakerloo Southern [SE] extension 2040 200 Tube Old Kent Rd, Lewisham, Catford 27 tph, Hayes 15 tph, Beckenham 6 Max. capacity on main line network, reduced crowding, wider econ. gains 2 – Free NR slots, intensive developments along route 2,000 to 2,600 Time benefits 400 pa
Northern Line extn phase 2 2045 (after Crossrail 2) 50 Tube extension 30-36 tph Battersea to Clapham Jn Better jny times, reduced crowding, better connections 1 – case context depends on Crossrail 2 300
Tube platform edge doors Programme 450 Safer platforms, facilitates tube automation Network-wide not VfM. Vic & Jubilee Ctl Lon benefits 3 – needed with full automation. Others specific 400 if £3-4m per platform
Long term service levels offpeak on TfL railways Programme 735 Day offpeak 27-33 tph by mid-20s. Night networks towards 6-10 tph Tube user demand requires this (and long term on SSL, Overground, DLR) 3 – demand case but with operation/maint /quality risks 35 pa opex, benefits 50 pa, revenue 15 pa

(MPPA = million passengers per annum, Yellow = longer term aspirations)

Scheme Timescale MPPA Output Benefits Business Case # £m Capex
Tube future stations programme phase 1 2017-2024 300 Tott Ct Road 2017
Vauxhall 2017
Bond Street 2018
Victoria 2019
Elephant & C 2020
Bank 2022
Holborn 2023
Camden Tn 2024
1 – demand case for extra passenger throughput Not stated
Tube future stations programme phase 2 2021-2024 300 Old Street 2021
Paddington Bakerloo 2024
Associated with development opportunities No priority yet, demand growth projected Not stated
Tube future stations programme phase 3 2025-2050 300 By 2035:
Baker Street
Piccadilly Circus
Liverpool Street
High Street Ken
By 2050:
Waterloo, Earl’s Ct
White City
Green Park
Warren Street
No priority yet, demand growth projected and London 2050 economic growth impacts 5,000-10,000 some scope for 3rd party funding contributions
Tube future stations programme (accessibility) Programme 300 12-15 discrete accessibility schemes to make significant addition to step-free jnys (assumed) London 2050 projections for more elderly travelling more often (TSP p.48) 3 – subject to prioritisation, some schemes will secure 3rd party funding Not stated

(MPPA = million passengers per annum, Yellow = longer term aspirations)

Scheme Timescale MPPA Output Benefits Business Case # £m Capex
Crossrail 1 HEX merger & serve LHR T5 2030 5 Up to 6 tph semi-fast to T5, + 50% Better connections for Heathrow and Crossrail network 1 20 for relief lines
65 for 5 trains
Crossrail 1 WCML Tring via Watford 2026 18 Through Crossrail-WCML commuter trains Congestion relief Euston, better use Crossrail capacity, HS2 benefits 2 – demand and population and econ. growth 150
Crossrail 1
30 tph
CP7 / 2029 243 +25% capacity in central area Congestion relief, extra destinations eg WCML 1 – demand and population and econ. growth 100 (new trains)
Crossrail 1
Dartford, Ebbsfleet, poss. airport
CP8 / 2030 9
(excl airport)
Through Crossrail-North Kent trains [serves Ebbsfleet Garden City] Better connections for Crossrail network, reduced crowding Lon.Bdge 2 – demand and population and econ. growth 100, plus trains
Crossrail 2 Early 2030s 269 24 tph + better main lines = 100k extra peak trips LU reduced congestion, extra SW, WA capacities 1 12,000-20,000

So what has changed? Let’s start with the sub-surface railways (SSR) – the Circle, District, Hammersmith & City and Metropolitan Lines – which are not listed above. Essentially the signalling deal with Bombardier has been cancelled because of technical inadequacies. A new deal is in final stages of discussion with Thales. This is described in detail in the LR article Mixed Signals. The consequences are considerable. The Circle route will be the first to benefit, now in 2021, not 2018. The rest of the SSR network will follow in 2022, assuming that no other unforeseen complexities are identified. The capacity implications are considered below.

Because the Piccadilly Line is closely dovetailed with the District Line in West London, including track sharing, this project delivery date in turn delays the re-equipment of the Piccadilly Line. This is still intended to be the first line to receive the ‘New Tube for London’. Pedantic commented in the Mixed Signals article that:

Once these four lines have been completed, LU will then move on to buying new trains and control systems for the Piccadilly, Central, Bakerloo, and Waterloo & City lines.

This replaces the previous plan, largely as set out in the London 2050 material (see the No-do-nothing-option article), that:

TfL have confirmed that the current schedule for deployment begins with the Piccadilly Line in 2022, to be finished by 2025. This is to be followed by the Bakerloo Line, completing in 2027 and the Central Line/Waterloo & City to be completed by 2032

The consequences were described by Pedantic in the Mixed Signals article:

The true ramifications of this should be plain to see. If true, this means that there will be a substantial further delay to New Tube for London (NTfL) programme with all the consequences that will bring. Currently, according to the latest published Fit for the Future plan, installation of the new Piccadilly signalling will commence in mid 2019. If TfL are only going to start buying new control systems once the SSR upgrade is complete then that looks like a three year delay, although time could be made up by the resignalling and delivery of new trains going hand in hand.

So all the tube line upgrades could be delayed in project delivery by up to three years, on current estimates. This has already had a further knock-on effect, with reprioritisation of the tube train upgrade programme. The Central and Waterloo & City Line trains will be re-furbished, and re-equipped with new motors and control gear, for further years’ work.

The Bakerloo Line will now see new stock sooner than the Central and W&C, but even so requires major repair work (now authorised) to keep the 1972-vintage fleet in safe working order until the late 2020s or 2030. The Central and Waterloo & City Lines would then follow with new trains by 2035. This assumes that a three-year delay remains in place – a not-unreasonable assumption as some of the rephasing appears to be for budgetary reasons rather than technical issues.

We can now portray the changes in availability of Central London Underground and Crossrail capacity, comparing the plans last year and now. The figures below are based on London Underground estimates of seating and floor space. These can vary depending on source and context, so we have endeavoured to maintain a consistent basis for the comparisons.

The calculations for last year’s planned outcome include the following assumptions:

  • Seating includes tip-up but not perches.
  • It is intended to demonstrate capacities at full stretch, so standing passengers are at 6 per sq.metre, compared to 5 at planning standard and 7 at crush loading.
  • Crossrail train capacities are taken at a nominal 150 passengers per car as if they were 10-car trains. The variation with the intended 9-car trains having a similar overall length is not substantial. Later Crossrail capacity upgrades are shown in this article as: 2030 (same year as CR2 opening), CR1 to 30 tph; 2040: CR1 to 11-car, CR2 to 30 tph; 2050: CR2 to 11-car.
  • Implementation dates for some tube capacity improvements are taken (where not stated in London 2050) as three years after the previous initiative. So the Jubilee Line is shown as reaching 36 tph in 2022, and the two Northern Lines in 2025 (assuming those lines must wait for Bank and Camden Town rebuilding).
  • The numerical evidence on line capacities with a NTfL train, comparing nominal capacities before and after, is that adopting 6 passengers standing per sq.metre plus a 5% overall gain in capacity from the new train, achieves close similarity with the numbers stated in London 2050.
  • The Waterloo & City Line capacity improvement (not stated in London 2050) is assumed to increase frequency from 22 to 33 tph, after noting possible frequency limits with slow entry/overrun issues at Bank terminus. This might be implemented in 2032 if phased at the same time as Central Line upgrading.

Taking the plans intended last year, the one hour maximum line capacities in 2014 are estimated as 560,600, combining both directions of travel, with 425,800 on the tube lines and 134,800 on the sub-surface railways. A further 28,500 hourly capacity is available on the west-side Circle and District lines through High Street Kensington.

The tables below show the hourly capacities by line and date of upgrade completion. Automatic train driving level ‘a2’ still requires train staff to manage door opening and closing; ‘a4’ would include platform doors and automatic door opening/closing. The red years below are an estimate that the later phase would be three years beyond the initial technical update.


With the upgrades foreseen in 2014, the hourly capacities are raised to 737,600 overall (+32%), with 581,100 on the tube lines (+36%) and 156,500 (+16%) on the sub-surface railways. In absolute numbers, the gains offer capacity for a further 177,000 journeys in one hour, 155,300 on the tubes, 21,700 on sub-surface.

The capacity gains with Crossrail can best be appreciated in this context. Crossrail 1 on its own would be worth 72,000 passengers hourly two-way, at 24 tph. This grows to 90,000 passengers if services were either increased to 30 tph (allowed for in the design specification) or to 11-car trains, while the second element grows the hourly two-way capacity to 108,000. Adding in Crossrail 2 doubles these numbers, and would bring the total passenger distribution capacity in Central London to almost a million (952,000) per peak hour!

The two Crossrail lines plus the tube and sub-surface plans will increase rail capacities within Central London by 393,000 hourly passengers, a 70% increase over 2014. These figures exclude Thameslink, which will offer an additional 40,000 hourly two-way capacity on top of 2014 capacity (this assumes a future 130 passengers per car, and 16 x 12-car, 8 x 8-car).

The gross increase of over 430,000 passengers in one hour sounds very useful, and if multiplied up to a three-hour peak would be worth 900-950,000 passengers in that period. That should be matched against the London 2050 expectations of 1.3 to 1.4 million jobs, with the bulk of those forecast to be within the Central Activity Zone (CAZ) plus any satellite areas such as Canary Wharf, Old Oak Common and Stratford.

So already there could be a mismatch in extra capacity versus foreseen jobs growth. Expanding National Rail commuter capacities by the desired 80% stated in London 2050, might help for jobs located near to main line termini, but it still doesn’t sound that the tube at maximum stretch plus two Crossrails plus one Thameslink will be enough to do the distribution job properly. There will also be a requirement for some capacity margin to absorb unforeseen service interruptions, which otherwise could lead to a ripple of temporary station closures as passenger handling issues got out of hand with the volumes encountered.

Worse, there are three other issues with the Central London capacity statistics:

  • They assume that one extra Central London job will require additional capacity on only one line. In practice, passengers interchange between lines, so that one extra job could require, say, 1.2 x line capacity as passengers switch between lines. The scale of this is discussed below.
  • It is not yet clear how passenger usage of the two Crossrail lines will interact with available railways in Central London, and to what extent Crossrail will permanently relieve flows on specific corridors or add to Central London passenger distribution burdens, with each Central London Crossrail station acting in effect as a mini-National Rail terminus with the connecting tubes having to handle the local distribution loads within the CAZ. Yes there has been modelling, but how will passengers actually behave?
  • What is the practical effect of the reshuffle of project timings for Underground line capacity updates?

Impacts of Underground project reshuffles

It is easiest to tackle the last point first, about the Underground project reshuffles. With the new project schedules caused by the SSR changes and the knock-on effect on the tube network, the changes in capacity are set out below diagrammatically.

  • The SSR output is shown as for 2022, because interim changes to the Circle line in 2021 might have adverse consequences for the interim capacity of the other SSR routes, if Circle frequencies were temporarily to be increased.
  • NTfL trains are assumed as follows: Piccadilly Line 2028, Bakerloo Line 2030, Central and Waterloo & City Lines 2035. No change is assumed to the introduction of automatic train driving level ‘a4’ on the Bakerloo in 2045, or sooner if a SE extension is opened earlier. The Central Line is assumed to go straight to 36 tph in 2035, with experience gained in 2028-30 with raising the Piccadilly Line frequency from 33 to 36 tph.

The diagram below has been estimated, to show the yearly change in projected capacities after allowing for the new project delivery dates. It also assumes that Crossrail 2 opens as Phase 1 in 2030 – to time with the current plans for HS2 Phase 2. The green boundary line and related numbers show the comparative capacity offered in the 2014 plans, and the capacity gaps which are opened up.


There appears to be a marginal overall effect, amounting in some years to a loss of projected total capacity of about 20-22,000 passengers hourly across the Central London Underground and Crossrail networks. The direction of travel which loses out might be more significant.

In the SSR case, Crossrail 1 is fortuitously also East-West, so can alleviate capacity shortcomings in the years until 2022 – providing Crossrail 1 doesn’t fill up instantly. The delay in achieving the 21,700 hourly SSR capacity gain can be offset partly by the 72,000 Crossrail capacity on offer from 2018-19. However, by location Crossrail 1 is more oriented towards relief of the Met/H&C/Circle side of the SSR, than the District/south-side Circle (albeit Ealing users may divert to Crossrail).

If the Circle Line were tackled first, with higher frequencies on that line – which is understood to be the case – then there is a risk that temporarily reduced District services on the south-side Circle is exactly where one wouldn’t want to put the suburban SSR capacity reduction! How many District trains will run to Edgware Road in that period, to interchange with Crossrail 1 at Paddington? Will the Circle run also to Wimbledon via the south-side and Earl’s Court in that interim period? At this stage, these are questions rather than answers.

In the mid-to-late 2020s, the capacity shortfall is a consequence of the NTfL programme starting late and continuing late, beginning with the Piccadilly Line and then the rephasing of the Bakerloo, Central and Waterloo & City Lines. This will increase capacity pressures on the North-South tube lines, post HS2 Phase 1 opening (currently due in December 2026), and ahead of Crossrail 2 opening. Any delay to Crossrail 2 could see major problems arising at Euston (HS2 Phase 2), and on the existing North-South lines.

A further consequence of rescheduling the Waterloo & City upgrade could be a shortfall in capacity to the City, with that scheme deferred until about 2035, and the only relief being offered via Crossrail 2 and the Crossrail 1 interchange at Tottenham Court Road, from the inner SW suburbs.

Another point demonstrated on the diagram is the rate of planned hourly capacity gain (the green diagonal) which averages about 11,000 a year. If actual demand occurred at a faster rate, such as the orange diagonal which shows the outcome if the 2050 output were required by 2043 (a Network Rail planning date), then the tube capacity would be particularly stressed for at least half of the 2020s, before Crossrail 2 opens.

Getting Crossrail 2 authorised, funded and built as soon as possible looks like a desirable large-scale objective in the face of difficulties with the present Underground upgrades, in order to build in adequate capacity headroom. A Crossrail 3 and/or Thameslink 2, etcetera, for later headroom, might be pointed to for the 2030s/40s. This will also depend on the other two topics – interchange impacts on required line capacities, and Crossrail relief and distributional effects.

Effects of interchange volume on line capacities

It is possible to get an understanding of the scale of through travel across Central London, and particularly the extent to which Underground lines have to offer capacity for passengers interchanging with other Underground lines. This is achieved from an analysis of the maximum hourly capacity offered in 2014, against the actual numbers of passengers EXITING at each Central London station and assigned to individual lines for the final leg of their journey.

LUL’s Railway Origin and Destination Survey (RODS) has been used for this analysis, based on the 2012 station exit numbers in the period 08:15 to 09:15. This also avoids the dominance of National Rail termini entry volumes, and allows a clear measure of end destinations (and, unavoidably, interchanges with main line stations for flows heading outwards from Central London). None of this will be unfamiliar to TfL. They will have much more sophisticated data and planning tools to hand.

There are detailed tables which there’s no space to show here. However line and station exit summaries by direction of travel are shown below, based on three categories:

  • West End plus Victoria Street stations
  • City stations
  • Other Zone 1 stations.

Which station is in which group is shown in the final map at the end of this article. Temple overlaps the City and West End zones, so has its exits allocated 50% to each.

A sensitivity test has been run, using the planning standard of 5 passengers standing per square metre, in comparison with the more crowded 6 per sq.m used above. If you take a nominal 20% of capacity as seems reasonable in most circumstances for interchange and through passengers in both sets of outputs, then this quickly shows some lines which are already under capacity pressure at 5 per sq.m and the scale of margin at 6. As we would expect, these point to the pressure points being the Central Line and Northern Line via City in both directions, and the southbound/ eastbound flows on the District, Jubilee, Met/Circle North, Piccadilly and Victoria Lines, and the northbound Waterloo & City. The Jubilee data includes Canary Wharf, as a City-type destination.

Also telling is that there are tubes such as the Victoria Line, already very busy in practice, which on a nominal 20% usage for through or interchange flows appear to have useful residual capacity at present. Clearly the actual usage for through or interchange flows varies significantly between lines, and the Victoria Line is one which is popular for such journeys – possibly as high as 30% through or interchange – because it provides a convenient diagonal cutting across a more simplistic east-west and north-south ‘grid-iron’ network. Remember what the tube map used to look like before it existed.

There are also lines under less pressure, such as the Bakerloo and Northern via Charing Cross, and the opposite directional flows on other lines. A Bakerloo extension to Lewisham and maybe Hayes is being considered (see Death Taxes and Lewisham: Extending The Bakerloo). The Northern via Charing Cross is being extended to Battersea Power Station (Northern Line Extension Public Consultation).

peak 4 revised
peak 5 revised

Several observations can be made on these figures.

  • It is slightly comforting that the received wisdom about some lines’ limited margins for more passengers is indeed borne out.
  • Moving the planning standard from 5 to 6 standing passengers per square metre can give some extra margin throughout the Central London network. In practice passengers will do this themselves if that were their least worst travelling option.
  • Some lines such as the Victoria Line may be popular because of their through travel and interchange capability rather than the inherent density of exit demand at their Central London stations.
  • The sensitivity of the Waterloo-City corridor to new capacity initiatives is worth noting. Even at 6 passengers standing per sq.m, there isn’t much capacity margin left with the present service, once a nominal level of interchange and through travel is taken into account. Crossrail 2 helps indirectly, routeing some NE and SW passengers onto City services at Tottenham Court Road (probably of more use from the SW, though North and NE residents working near say St Pauls or Moorgate could find TCR useful in future).
  • The future, beyond the Underground upgrades set out in TfL plans, will have to be with new lines offering new capacity. ‘Peak Tube’ means just that, for the existing tubes and SSR.
  • With the London 2050 central forecasts of 1.3-1.4 m jobs mostly in Central London, these imply crudely that line capacity may be required for 20-30% extra, because of the through/interchange factor. This is 1.5-1.8m capacity in the worst case, over three hours. It is at least one extra travelling space for every extra job even if a proportion of those jobs were outside the CAZ, which points to two further new lines being required.*
  • Any desire to improve peak time travelling quality would raise the game above that level of additional capacity.
  • Careful definition of the routes for additional lines could yield more utility from the existing network, by removing some of the current through or interchange flows which currently help to crowd out some parts of lines – providing that the new lines didn’t cause net detriment by in turn unloading their passengers into the existing network to an undue extent.
  • A study of interchange flows – taking the Crossrail plans into account (and those in turn should also be assessed for their interchange and distribution impacts) – is therefore critically important in order to define the greatest line reliefs that may be feasible with further new cross-Central London railways.
  • Existing stations with high levels of interchange flows, and lines with multiple heavy duty interchange volumes, should be high up in that study.

The major Central London termini will all feature among the busiest interchanges, although the way most of those journeys are measured in RODS is entries/exits, so they do not feature in the interchange volume analysis below. This may point to some transposition of National Rail terminating lines into through cross-London railways – where there is then a moot issue about whether inner suburban services should be favoured for through running over outer suburban. There might be simple or complex reasons for one or other or a mix.

The harder test may be which interchange stations internal to Central London should be prioritised for relief with new lines. There will be plenty of choice. Crossrail 1 should relieve some, but may create new pressures elsewhere, for example at Bond Street, Tottenham Court Road, Farringdon, Moorgate and Whitechapel, on routes which might then offer a feeder and distributor function. The norm with Crossrail 1 would be for East-West capacity to be relieved, for example on the Central Line, but for the potential of greater demand on feeder and North-South lines.

2019 / 2030 modelling should show those consequences for Crossrails 1 and 2. On the face of it, the Central London element of Crossrail 2 looks useful, by providing an ‘Albert’ for the Victoria on the NE-SW diagonal. Clearly CR2 in its safeguarded form should relieve the Piccadilly and Victoria Lines. What might be the other consequences, however, with extra interchange loads at a ‘Euston Cross’ and Victoria?

Meanwhile, in 2014, the most pressured LUL interchanges are set out below, down to just under 4,000 interchanges during the 3 hour AM peak period – so something under 2,000 interchanging passengers in the busiest single peak hour. Main National Rail termini interchange volumes are excluded, but other National Rail and DLR interchange volumes are included. Bus interchange volumes are also not shown in the 2012 RODS data, although this will be very important at some suburban stations.


From the point of view of new line planning beyond ‘Peak Tube’, it is the ‘Top 22’ which may be critical. These are highlighted above in yellow. They include 4 National Rail termini (Kings Cross through to Euston – the latter also influenced by HS2 planning), 9 major interchanges in Zone 1, and, for suburban railway planning, 9 in the inner suburbs, including Canada Water as well as Canning Town, because of capacity issues on the East London Line.

Let’s acknowledge here the shortcomings in the available data, especially with National Rail termini information. For example, Waterloo looks trivial, but the figure shown above is just the LUL internal interchange flows. In practice Waterloo is the busiest interchange around when you count in the bulk of transfers between National Rail and the Tube.

The Waterloo numbers gross about 222 million yearly when you count in entries, exits and interchanges at Waterloo, Waterloo East, Southwark and all the related Tube volumes. Much of that of that number is double counting, as a National Rail passenger changing twice a day with the Tube will be counted four times not twice, as 2 x entries, 2 x exits. For Waterloo, Crossrail 2 is supposed to relieve the complex, via Victoria and Tottenham Court Road. Its efficacy has yet to be proved. The sheer volume of rail and tube passenger throughput in the Waterloo area may make a case for a CR3 or CR4 heading that way.

Similar calculations have been undertaken for other main London stations and their related Underground and DLR flows. To conform with the limited ORR public data, these are shown on an annualised basis. No attempt is made to combine the Tube and main line numbers, except where LUL data includes the National Rail volumes. Where the LUL RODS data does also count the NR flows, this is indicated.


Even if ORR data understates matters somewhat, the point that only four National Rail stations anywhere in the rest of Britain would qualify for inclusion in the table above makes a profound point about the sheer volume of rail-based travel and related investment requirements for the London area.

The simpler peaks-only interchange table above may remain useful in considering new lines for London beyond Crossrail 2. After allowing for CR1 and CR2 effects, there could still be challenges to address at:

Zone 1:

  • Kings Cross/St Pancras/Euston, with CR2 adding interchange flows as well as relieving
  • Similarly at Victoria
  • Bond Street with CR1 distribution
  • Green Park, because the Piccadilly axis appears to be ignored by CR2.


  • The figures give a sense that, post-CR2, there might still be significant shortfall east of Central London at main interchanges, as the sub-region continues to grow
  • The frequent visibility where a Jubilee Line axis (NW-E/SE) pops up as under pressure at main interchanges is interesting. Will CR1 be adequate to relieve that, or do we have an incubus remaining to be addressed?

These are not profound points, but just observations. The comments also exclude the potential for CR1 and CR2 to fill up much sooner than expected, with consequential impacts on how piqued a ‘Peak Tube’ could be, before further relief arrived.

Crossrail interaction with existing tubes

This is the third area of interest. Fundamentally the question is to what extent passengers choose in the future to transfer back to a tube for a stop or three beyond the Crossrail railhead, with possible overload on an existing tube which becomes a local distributor railway within Central London, or to what extent people are attracted to walk to the final destination from a Crossrail station – where the quality of surface walking and convenience of multiple station entrances/exits will be very important.

The risk is that, with Crossrail 1, there will be undue additional passenger distribution load on North-South lines at main interchanges, quite apart from strategic changes in London-wide passenger flows because it then becomes more efficient to use a North-South line to connect to/from Crossrail 1 for journeys further afield. Crossrail 2 could be similar in style, though for different directions of travel.

The Crossrails will also bring current and new main line rail users directly to different interfaces underground with the tube/SSR networks, than those which the Underground has been used to accommodating up till now. The advent of radically new passenger flows implanted within Central London is not something that the Underground has had to address in recent decades, with the possible exception of the Jubilee Line extension.

The previous occasion was when the Victoria Line opened and infused the Underground with an influx of NE Londoners, whose gateway was previously Liverpool Street and suddenly became Kings Cross, Euston, Oxford Circus, Green Park and Victoria. The transformation of such accessibility to the currently ‘distant’ suburbs of London West-of-Ealing, London South-of-Wimbledon, and London South-East-of-Canary, should be modelled with trepidation. Even in 1969, the public take-up was massive, as shown in post-Victoria Line-opening demand assessments, when a new London travel orthodoxy was established.

Add Crossrails 1 and 2, ‘Thameslink 1’, and maybe more in the future. The shock and awe may be less these days, because Oyster exists now. However there is nothing like reality emerging and intruding, to challenge preconceptions and preferences. Any Crossrail or Thameslink line is inherently likely to attract a greater volume of rail-based journeys from their suburban and commuter catchments, to and via Central London.

Back to the core challenge. Would you prefer to stay underground via a Central London interchange, for a stop or three, or emerge into daylight/night/rain/sun etc and walk? This could be important for the vitality of Peak Tube downstairs, potentially to be burdened with short-distance interchangees. Alternatively, perhaps there will be demand for a new distribution line for Central London, linking selectively with Crossrail stations and avoiding overload of existing capacity-sensitive Underground interchanges – a ‘Heineken Line’, if you like, getting to the parts other tubes can’t reach.

Let’s end this article by just posing the surface option, and accompany it with a map showing 800 metres walking distance from planned or nominal Crossrail 1, 2 and Thameslink stations. 800m in a circle is roughly equal to 960m street walking round corners, so matches London’s rail PTAL limits. Maybe part of optimising Peak Tube capacity could be the quality of surface access within an effective Crossrail station catchment, and how that might be enlarged through subsidiary entrances / exits.

Peak 8 revised

Nominal statistical zones for stations in West End + Victoria Street, and the City, are shown on this map.

Update: Since initial publication of this article, TfL has advised that design options for a southern access for Tottenham Court Road CR2 station are being considered. The above map has been amended to show an approximate location on Shaftesbury Avenue, in pink. This would extend the CR2 catchment towards Piccadilly Circus, St James and Trafalgar Square. It is hoped to consult on station design options this autumn.

* 1.3m capacity value including through/interchange impact, less 0.9m planned new capacity over 3 hours = 0.4m, revised to maximum intensity single hour = less than 0.2m, at a Crossrail-style overall capacity of less than 0.1m per line two-way per hour.

Written by Jonathan Roberts