Operationally the Northern Line is probably the most interesting tube line. As the line’s signal upgrade programme comes to an official end, we look at the recent Northern Line timetable improvements and how frequency seems to be slowly but steadily improving.
We recently looked at the Subsurface Railway (SSR) timetable commencing in December 2014 which contained some significant improvements off-peak. These improvements may not have been publicised but if one knew there were to be an off-peak SSR timetable upgrade one could guess at what they might be. In a sense then they did not come as a surprise.
In contrast the December 2014 Northern Line improvements were quite widely known about, expected and they applied to both peak and off-peak periods. The peak period improvements were not what we had been led to believe they would be. Some off-peak improvement was expected but probably not to the extent that has actually happened. As we shall see, the Northern Line has truly joined the league of deep tube lines with an intensive frequent service, albeit one which is a bit patchy.
Asymmetric in the Peaks
The Northern Line is almost certainly the trickiest of the deep tube lines to timetable. It is not just the complexity of having two separate routes through central London, nor the famous complex junction at Camden Town. What also distinguishes it is the asymmetric nature of the peak period service which is mainly driven by a need to provide as intensive a service as possible from Morden to Bank in the morning and vice-versa in the evening. Whilst the recent introduction of ATO seems to have eliminated the need for some of these complexities there are still residual features remaining.
The peak period demand for trains between Morden and Bank leads to more trains running via Bank in the direction of peak flow than running via Charing Cross. It also leads to a slightly more intensive service in the peak flow direction between Morden and Kennington than in the contrapeak direction.
To help cater for the desire to run extra trains from Morden via Bank in the morning peak period, there used to be a strategy that sent all northbound trains via Bank in the morning peak period to High Barnet and all the ones via Charing Cross to Edgware. One could get away with this because traffic would be relatively sparse in the contrapeak direction in the morning. This increased capacity through Camden Town junction in a northbound direction in the morning peak period. What this meant was that, during this period in this direction, the junction was effectively no longer a junction. To cater for the lesser capacity in a southbound direction trains departed from High Barnet at a slightly slower rate than they arrived – presumably a couple got put away in the sidings there.
In the evening peak passengers are travelling in large numbers in all directions. This made providing as good a service in the evening peak as in the morning quite impossible as trains had to serve both northern branches in both directions.
In the latest working timetable we are told that:
The northbound service segregation in the morning peak has been reversed with Bank branch trains generally now operating to the Edgware branch and Charing Cross branch trains operating to High Barnet/Mill Hill East.
Quite why the segregation was reversed is hard to understand. A “heat map” shows passenger overcrowding is higher on the High Barnet branch than on the Edgware branch.
What is more interesting and curious though is an earlier statement in the working timetable introduction that:
The “tidal” peak service on the Morden branch (northbound in the morning peak; southbound in the evening peak) has been increased to provide 30 trains per hour for approximately 90 minutes.
Now interestingly enough in the evening peak period there is an identical frequency of service as in the morning of 30tph between Bank and Morden in the direction of flow. This suggests that nowadays, with the latest signalling, Camden Town Junction is not currently a limiting factor in providing the maximum service one can. Presumably we are now at a point where the limiting factor on the Northern Line is the number of trains available. If this is correct, then one must ask why it is worth bothering to send all northbound trains via Bank in the morning peak to Edgware. Why not alternate (with one or two trains off-pattern) between the branches?
High Barnet and Mill Hill East branches
A further complication for the Northern Line timetable relates to the High Barnet and Mill Hill East branches. Normally where branches exist the general strategy is to give them both the same level of service. Clearly this is not desirable on the one station Mill Hill East branch which is very lightly used. In any case the branch is currently single track so it is not possible to run more than about 5 trains per hour (tph) on the branch. This leads to a slight anomaly on the High Barnet Branch from East Finchley to High Barnet. The stations involved (High Barnet, Totteridge & Whetstone, Woodside Park and West Finchley) actually have a better (both more frequent and more regular) service in the off-peak, when the Mill Hill East branch is reduced to a shuttle, than they do in the peak hours. If more trains were available it would be possible to run all trains in the peak to High Barnet and have the Mill Hill East branch run as a shuttle in the peaks too.
A frequent Northern Line service
Until mid-2014 the level of service on the Northern Line was really not what you would expect from a tube line in central London. Even the Bakerloo and Waterloo & City could do better. In rough terms each central Northern Line section and northern branch could only manage around 20tph in the peak – a figure to bear in mind for later.
Things were improved dramatically by the full introduction of ATO (automatic train operation) on the line by June 2014. This on its own allowed the frequency in the central area to go up to 22tph without needing any extra trains.
Then in December 2014, with no apparent further changes, a better timetable still was introduced. In fact though there has been a lot going on since June. With the ATO signalling installed it was time to increase speeds on the Northern Line by means of replacing, reprofiling (rail grinding) and realigning track. What has been impressive is that all this work, a lot of which was expected to take place during weekend closures, has in fact been carried out in “engineering hours” – overnight when the trains aren’t running.
What was expected for December 2014 was a timetable where there was roughly 24tph on the two northern branches, 24tph on each of the central London routes and 32tph between Kennington and Morden.
What has actually happened in the peak period on the Northern Line is that the Charing Cross branch now runs at roughly 24tph in the peak and the Bank branch runs 26tph in one direction. 26tph on the Bank branch is up 4tph from just six months ago and may even be the most frequent service it has ever seen. To put it in context, this is better than the Piccadilly Line can manage and only 4tph short of what runs on the Jubilee Line. All the talk might be of the New Tube for London but it is the oldest tube for London that is currently delivering service improvements.
30tph on the Kennington – Morden section
Running ‘only’ 30tph between Morden and Kennington in the morning peak in the current timetable is a slight disappointment as 32tph was expected but it still is 3tph better than before. It would appear that the reason for this is a combination of lack of trains and a decision to prioritise improving the peak service on the Bank branch. This has a secondary benefit as far as the timetable writers are concerned. Every extra train on the Bank branch removes the need for a train to depart from Morden destined for the Charing Cross branch and every train removed from the Morden – Kennington section also removes a train from Morden via Charing Cross. Consequently the current timetable requires only 4tph to travel from Morden via Charing Cross rather than the previously envisaged 8tph. This reduces the need for critical pathing at Kennington Junction. The ultimate desire is to eliminate trains to and from Morden running via Charing Cross – this already happens off-peak but cannot currently be avoided in peak hours.
More trains needed
The timetable already requires 96 trains out of a total of 106 trains in service in peak hours. To get another train in service would probably require negotiation with Alstom who still maintain the trains under a pre-PPP contract. This is by no means impossible. One more train in service would still only require an availability of 91.5%. An extra two trains in service compared to today would require an availability of almost 92.5% – tough but not unrealistic in this day and age. The running time from Kennington to Morden is around 22 minutes and 21 minutes in the reverse direction so if 2 extra trains could be made available then two, or possibly three, extra trains via Charing Cross could be extended from Kennington to Morden should that be a desired objective.
As an alternative to finding extra trains one can run in service, one could push for a bit more speed. If London Underground could further reduce the travelling time between Morden and Kennington by just one minute in each direction then that would be a equivalent to an extra train. A thread on the District Dave’s website suggests that there could be further improvements in this area in future. Here, seconds and fractions of mph matter would matter, and luckily one advantage of ATO is that speed limits are not rounded down to the nearest 5km/h so once a section of track is reassessed it is quite possible that the speed limit will be slightly raised.
A further change to the Northern Line timetable is the re-introduction of stepping back at Morden. Stepping back is when the driver does not leave on the same train they brought in but “steps back” and leaves it for another driver. They then take out a subsequent train. Strangely Morden is a three platform terminus yet the working timetable tells us that operators stepping back every four trains has been implemented. One would have expected this to be every third train or possibly every sixth train but it seems that it is more operationally convenient to actually use just two of the three platforms for terminating trains.
It would appear that stepping back is generally disliked on the Northern Line as it creates performance risk on such a complicated line. It was implemented a couple of timetables ago at Morden but then removed from the last one. Nevertheless it would appear that the Northern Line has reached the point where this is unavoidable in the peak hours. It may well be that one of the principal future motivations to completely separate the Northern Line into two distinct tube lines will be the ability to operate stepping back without fear of total chaos when things go wrong.
As good as it can get for the moment?
No doubt the ATO system will continue to be refined but the dramatic time reductions already achieved are unlikely to be improved on much more. Unless more available trains or speed can be coaxed of the existing fleet it is hard to see how the peak timetable can be improved until new trains arrive. Unfortunately these have not even been ordered yet. They are thus not expected to be delivered for a few years.
Northern Line off-peak
Where the Northern Line gets really impressive is the off-peak service. At the beginning of 2014 this was 15-16tph so roughly a train every four minutes on the Edgware and High Barnet branches. For each of those branches alternate trains would go to Morden via Bank or Kennington via Charing Cross.
Now the off-peak service is 20tph, seven days a week. This is roughly a 33% percent improvement in trains leading to a train every three minutes instead of every four minutes – a 25% reduction in average waiting time. What makes this even more impressive is this is roughly what a year ago was being run in the peak period. It is true that the Victoria Line and to some extent the Central Line run at a higher off-peak frequency in the inner zones, but to run 20tph out to Morden (zone 4) and Edgware and High Barnet (both zone 5) on a Sunday is quite remarkable.
From a passenger’s perspective the off-peak service can be viewed as two different distinct services that might as well be different tube lines. From an operational point of view this is not the case as trains arrive at High Barnet having been routed on one line through central London and return on the other one. It is just how the timetable has ended up. No doubt this situation at High Barnet could be avoided if there was any good reason to do so.
TfL and London Underground have never really spelt out exactly what the long term timetable is for improvements to the Northern Line when it comes to new trains and rebuilt stations. However they have now said enough and quoted dates where necessary to make it possible to get a pretty good idea of how this will pan out for the future.
Between 2015 and 2018 we cannot expect to see any dramatic change to the Northern Line. Any limited improvement in service will be dependent on better rolling stock availability or reduced journey times. Priority of new rolling stock will be given to the Jubilee Line. These trains will superficially be clones of the existing trains with similar, if not the same, performance characteristics but will be more advanced below floor level. The new stock on both lines will be visually very similar as far as passengers are concerned though it will have 7 cars on the Jubilee Line and 6 on the Northern Line. The cabs will be appropriately configured to match existing cabs on the line they are going to.
Battersea Branch Frequencies
It may seem strange to talk about frequencies on the forthcoming Battersea Branch when construction work has not yet started but the project is due for completion by the end of 2019 so we expect trains to be running within five years from now. The projected initial service seems to be largely based on extending some or all Charing Cross branch trains currently terminating at Kennington. It is of course presumed that by then there will be sufficient new trains to run the service required.
The original expected number of Kennington terminators in the peaks was 16tph. This figure was arrived at because there would be 24tph on the Charing Cross branch but 8 of these would continue to Morden to provide an additional 8tph south of Kennington. In fact, as we have seen, the timetable currently implemented is more complex than the one anticipated and it would appear that only 4tph start from or continue to Morden. Based on the timetable currently running it could well be the case that a slightly better initial peak hour service of 20tph may be run on this new branch.
One can also speculate on the initial off-peak frequency to Battersea. Whatever may have been suggested in the past, it would not be surprising if London Underground kept things simple and just extended all off-peak Kennington terminators to Battersea. This would mean a frequency of 20tph – the same as in the peak.
2020 – a critical year for the Northern Line
2020 is going to be an important year for the Northern Line. In the spring we will see the full then partial closure of the Bank branch for 117 days whilst a new southbound platform is brought into use at Bank station. This closure will finish in August 2020. It is presumed that London Underground would not want to improve the service on the line, beyond what is run today, prior to then as this is a major closure and one would not want to build up traffic only to subsequently be unable to handle it.
Once trains are running once more after the full reopening of Bank station we now know that TfL plan to run at least 30tph (presumably on both central sections) because this press release states that:
By 2020, LU is planning to increase train frequencies on the line even further to at least 30 trains per hour by introducing new trains.
As usual the off peak frequency isn’t mentioned but if the peak frequency is 30tph then one could reasonably expect to see an off-peak frequency of 24tph by 2020, which does seem to be the current minimum standard for central London on lines where it is feasible to implement it.
To consider what will happen next we need to look at Camden Town.
Camden Town Junction
The Northern Line has two critical junctions – Kennington and Camden Town. The long term plan is for all trains from Morden to continue via Bank and all trains from Battersea (or using the Kennington Loop) to continue via Charing Cross. So effectively the Northern Line has only one critical junction to worry about – that at Camden Town.
Unfortunately that junction is really complicated as a famous diagram originally published in the Eagle Magazine shows. ATO on its own probably isn’t really enough to optimise this junction and the 30tph quoted and planned for 2020 probably represents the limit of throughput at this complex junction.
To get beyond 30tph through Camden Town probably requires some kind of centralised supervising software (generally known as Automatic Train Control – ATC) to optimise train speed on approach to junctions to ensure maximum throughput. The current signalling contract for the Northern Line did not include ATC which has yet to exist on this manufacturer’s product. The SSR resignalling will have to include it to cater for the complex multiple junctions such as Praed Street.
It should be noted, as regular readers are aware, that the LU preferred solution appears to be to completely separate the Northern Line into two independent tube lines. These lines could then potentially be run at 36tph without resorting to ATC. Because separating the branches will result in a lot of people changing trains at Camden Town station, this cannot take place before Camden Town station is rebuilt in 2024.
The fallback position as an alternative of separating the lines is to implement ATC (probably after it has been implemented successfully on the SSR) to try to get significantly more then 30tph through Camden Town Junction. In this scenario, as in splitting the line, it should be possible to maximise capacity on the Northern Line well before HS2 arrives at Euston in 2026 which would no doubt put additional demands on the line.
The Northern Line – our best hope?
The London Underground line upgrade is starting to separate into three very distinct parts. There is the Subsurface Railway which will have some peak hour benefits when the resignalling is completed but off-peak already offers a similar level of service to that planned. There are the lines that will be eventually served by New Tube for London but little improvement can be expected until we are well into the 2020s. That leaves the “traditional lines” of the Northern, Jubilee and Victoria Lines. A extra 2tph will be squeezed out of the Victoria Line in the peak period and an already frequent off-peak will improve slightly (27tph from 24tph). The Jubilee Line will, in the next few years, improve the peak service slightly from 30tph to 36tph and, like the Victoria Line, slightly improve the off-peak service to 27tph. It is however the Northern Line which gives cause for hope. Not only is there a plan to improve peak period frequency by around 50% for much of the line but because it actually runs in two sections through central London, the benefits are effectively doubled.
The future Northern Line upgrades may just be unexciting implementations of existing technology but, in railway planning terms, they are coming soon and they are significant. Whether they do any more than keep pace with homebuilding along the route is another matter. The Northern Line in the past probably has not been seen as major tube artery in the way that the Central, Piccadilly, Jubilee and Victoria Lines are but perhaps it should be. With passenger journeys already substantially more than 900,000 per day it won’t be long before a million daily passenger journeys are made on the Northern Line each weekday and that number is only going to rise.