Unlocking Herne Hill and the Kent route to the City

The extensive RUS process has already yielded a long shopping list of potential projects, including some big ticket items such as Crossrail 2. We are more likely, however, to see a package of more modest investments in CP5 (Control Period 5, 2014-2019) which will pave the way for more ambitious projects in CP6.

Following the government’s recent High Level Output Statement (HLOS), the “rail industry” will need to prioritise the key problem areas around the London network that need to be resolved in order to release capacity, and to identify where infrastructure can be better used.

This analysis will be challenging, because resolving one problem tends to push the stress down the line to the next pinch-point. An integrated assessment needs to look at the network as a whole rather than looking at each area in isolation, and analyse alternate portfolios of investments in order to define which package has greatest merit. Here we explore one key bottleneck: at Herne Hill, where the Wimbledon-Blackfriars route crosses the busy Kent Mainline on the flat.

The 2008 South London RUS identifies Herne Hill as one of a small number of “key localised capacity constraints which restrict the overall capacity of the whole wider network”, but no solution is presented in the 2011 London & South East RUS. Resolving Herne Hill will unlock capacity on a number of routes, but it is clearly a thorny problem.

We explore some options below, and find that major investment here needs to be driven by a clear vision of what we want the south London network to look like. It should also be determined by what TfL expects of its “strategic interchanges” such as Herne Hill, and what this entails in terms of service provision. These will dictate the extent of fundamental restructuring we are prepared to tackle, and the level of investment required. There are four broad options:

• Do nothing or, at most, modest investment to relieve station overcrowding
• Reconfigure the track layout, potentially alongside rebuilding to expand station capacity
• Reconfigure the track layout and the service pattern, with significant knock-on effects across the network and probably also requiring grade-separation at Tulse Hill
• Invest in Herne Hill alongside new infrastructure at Lewisham, Brixton and elsewhere as an integrated program to relieve other bottlenecks and maximise the potential of the underutilised City route to Blackfriars.

Which of these is tenable, and what about grade-separation at Herne Hill itself? Read on…

Nowhere to run

In our post on safeguarding to expand the Blackfriars terminal bays, we noted how the ex-London Chatham & Dover Railway (LCDR) route from Herne Hill to Blackfriars is probably the most under-utilised stretch of mainline in central London. In 2018 the four-track viaduct from Loughborough Junction to Blackfriars will carry just 10tph each way off-peak and 14tph in the peaks. Arguably this is an inexcusable waste of infrastructure. More intensive use of the LCDR mainline through Elephant & Castle, however, is limited by bottlenecks north, south, east and west.

Heading north, only so many trains can go through the Thameslink core and the 2018 service pattern favours the London Bridge route rather than Elephant & Castle. For the rest there are but two terminal bays, and no prospect of any more unless land is safeguarded as part of the Ludgate House redevelopment. Failure to achieve this will relegate the route to a has-been ― all for the sake of a small triangle of land on the South Bank.

Looking east from Loughborough Jn, there may be potential to run a metro service on the Catford Loop through Peckham Rye, say to Bellingham or Bromley South. This is hard-fought territory, as attested by the efforts of London Travelwatch and TfL to develop alternatives to the doomed South London Line service between London Bridge and Victoria. The options studied by TfL highlighted the lack of capacity on the Kent Mainline, and the under-utilised section from Nunhead to Lewisham is limited to a meagre 3tph due to the complex flat junctions at Lewisham.

Looking west from Loughborough Jn, the chord around to Canterbury Road Jn and Brixton is not used by scheduled services. The complex Kent Mainline service pattern through Brixton effectively shuts out any other service options, and it will need a major rebuilding of Brixton Jn to open up the route to Blackfriars. However, rebuilding at Brixton should be explored further as it could offer multiple benefits, not least platforms on the Denmark Hill lines. The area has development potential and this may help offset the costs.

So east and west are limited largely by the bottlenecks at Lewisham and Brixton. But what of the route south through Herne Hill?

A crossing point

Back in the early LCDR days most trains had separate City and West End portions, which joined/divided at Herne Hill. As well as being an important junction for passenger trains, Herne Hill was also a sorting point for freight and extensive sidings were provided either side of the line to Loughborough Jn. A link was also added to the London Brighton & South Coast Railway’s London Bridge route at Tulse Hill, which now carries the Blackfriars-Wimbledon Loop service.

The station building is attractive and Grade II listed, shown in the old photos below with the original Up platform alongside the upper level of the building. The station was rebuilt in 1925 to provide the layout we have today, with two island platforms providing cross-platform interchange. Herne Hill has a very detailed Wikipedia page, which we have gratefully plundered, including a useful animation which shows the evolution of the station layout.

Nowadays, Herne Hill is a thorn in the side of railway planners. Capacity is significantly limited by the established service pattern of two routes crossing, hence the perennial interest in grade-separating the two routes, options for which we explore below.

The station is hemmed in by flat junctions at either end, and by a road underbridge at the southern end. Being on a viaduct from the station southwards, it will be very hard for it to be anything other than four-platforms with four running lines.

The flat junctions limit the platforms to 8-car trains ― a trait shared with Tulse Hill a mile away. The platforms could be extended at the north end, but this would require substantial track layout changes.

The 2008 South London RUS recommended that land was safeguarded for grade-separation (Option 17.3 p145), and word on the grapevine is that land has indeed been set aside. But what land is that, if any? The fact that it is not public knowledge suggests that Network Rail may have safeguarded on its own land, which lies to the north and is shown on the map below, but is this is even technically possible?

Hemmed in at the north end

South Junction is built on a viaduct and is very close to the station. An underpass is unworkable and a flyover would require a high-level platform on the east side, which will be obtrusive in a Conservation Area, technically challenging, and expensive. It is a highly unlikely prospect.

So, to the north end we look. The railway lands here are fairly generous, a vestige of the extensive Herne Hill Sorting Sidings. These lay either side of the main line to Loughborough Junction, extending westwards to Shakespeare Road. Most is now lost to housing, but a rump lies in the cleft of the Herne Hill North Jn where there is a Network Rail depot and a waste transfer station.
Grade-separation might be possible here, but alas it is hard to see how. The photo below is looking north from the Up Kent platform, with the signal box in the distance on the left and beyond that the rail depot marking the divergence of the two routes. It is clear from this and the map above that grade-separation would be an engineering challenge, if it is even possible.

Two example alignments for the Up Elephant & Castle line are shown on the map. It is likely that the Kent Mainline tracks would have to be realigned eastwards, and even then there would probably be a need to demolish the signal box and compulsorily purchase property in Mayall Road. If the Kent Mainline was raised then this might reduce the distance required for the Elephant & Castle tracks to descend, but the track would still require a steep gradient north from the platforms to clear the Kent lines and then a sharp curve north-east. Indeed, the curvature required may preclude a flyover and entail burrowing beneath the depot and then the new housing along Shakespeare Rd, to arise somewhere south of Loughborough Jn.

These works would preclude the potential for platform lengthening on the Kent lines. Platform extensions would be out of the question on the Elephant & Castle route platforms, although this is less of a concern as platform lengths at Tulse Hill and Elephant & Castle are also limited to 8-car.

This, of course, only provides grade-separation on the Up line. Providing grade-separation for Up and Down tracks will require more land, and the platforms tracks would then need to be reconfigured so that Tulse Hill trains used the westernmost island platform. The easternmost island platform would become Kent Mainline, and you’d lose cross-platform interchange.

Grade-separation at the south end

Back in 1989 the Central London Rail Study proposed an enhanced ‘Thameslink Metro’ via Elephant & Castle on the newly-opened cross-city route. However the service would require grade-separation at Herne Hill, not least because the Kent Mainline was also hosting Eurostar services into Waterloo. British Rail considered four options, and the preferred option in engineering and cost terms was to provide a new viaduct south-east of the station with a flyover to Knights Hill Tunnel on the Tulse Hill-London Bridge line.

It would be interesting to know what the other three options were. Perhaps TfL should review these given the capacity issues across the network ― a very different situation now compared to 1989.

The proposed viaduct and flyover at Herne Hill would have been built through a residential area, requiring the demolition of four houses and a number of commercial premises including a petrol filling station, and significant land-grab from the gardens of around thirty houses. The environmental impacts were considered worse than at Borough Market and this was a key driver to focus the Thameslink core route on London Bridge. A reappraisal during the preparation of the Thameslink 2004 Environmental Statement found that the works required for grade-separation essentially remain unchanged, but that some newer buildings would now require demolition and therefore the environmental and community impacts would be worse than identified previously.

There are also specific local issues arising from land ownership and planning. Aside from being in a Conservation Area, where buildings and trees are protected, the Dulwich Estate retains significant planning powers and is probably the freeholder of the rail lines alongside the estate. Network Rail will be cognisant of the potential level of opposition from local residents in this affluent part of south London, as well as the likely difficulties in negotiating with the Estate, including any legal wrangling which may necessitate revisiting the 1870 Act of Parliament curtailing the Estate’s powers. This probably explains why the default position is to keep this project firmly in the pending tray ― something needs to be done, but not right now.

Are there other options for grade-separation at the south end? Given that a flyover will be an unacceptable eyesore that enrages residents for generations, how about an underpass or tunnelling?

The Kent Mainline rises at a gradient of 1 in 120 south of Herne Hill, which would reduce the distance needed for lines to rise up from tunnels. So far so good, but the problem lies in getting the lines down into tunnel from the Tulse Hill end. Emerging north from Knights Hill tunnel, the cutting quickly becomes embankment and then crosses two roads, so there is not sufficient distance to descend.

Tunnelling would therefore have to start back at Tulse Hill, which is feasible in engineering terms but the business case is unlikely to be favourable. This scale of works may only make sense if part of a new strategic route to, say, relieve the Brighton Mainline. In this case the outcome may be a tunnelled route all the way through underground platforms at Herne Hill.

So, surreptitiously relinquishing grade-separation, as indeed Network Rail appears to have done, let’s look at other options.

Reconfiguring the tracks, or a new service pattern?

The more you look at Herne Hill the more intractable the problem appears to become. We’ve few options left. One is to rethink the pattern of services through Herne Hill, and another is to reconfigure the track layout.

Changing the service pattern will of course mean that some passengers would lose their direct trains, and therefore this option is utterly untenable, nay heresy. So we’ll tentatively park this for now and look at the track layout.
The layout at Herne Hill is designed to allow two separate routes to cross, yet still allow traffic movement between any of the branches. It is very flexible and able to accommodate the diversity of services, plus diversions or empty stock movements that don’t fit the normal pattern.

As always we are grateful to the LR commentariat for boldly exploring the options, entirely unprovoked – in this case buried in the fray of comments on a recent post on Finsbury Park no less… such is our penchant for perambulation down the side-track. One suggestion was to reconfigure the running lines so that the Kent Mainline services run through the outside platforms and the Tulse Hill route take the centre platforms. The advantage with this is that Tulse Hill route trains don’t need to cross both the Up and Down Kent tracks in one move, but can ‘hop’ the first track and await a path away while at the platforms.

This is an intriguing proposition. As the map above shows, there is space at the north end to accommodate a realigned Down line into the westernmost platform. At South Junc, although the current layout narrows to three tracks, the viaduct is built for four tracks, so there is space to expand.

In the photo below the relocated Down Kent line could slot in where the speed limit signs are on the left side. The photo also shows how this arrangement would straighten the Up Kent line by returning to the original LCDR alignment onto the track on the right.

The diagrams below show the current arrangement alongside the potential new layout. This new configuration still allows Kent Mainline tracks to use all four tracks, which provides an opportunity to regulate Slow services and allow Fast services to pass. A further benefit of this new layout also becomes clear: it will allow platform extensions. Given that there are significant problems in extending the platforms at Tulse Hill and Elephant & Castle, this may not be such a priority on the current Blackfriars route, but it will be beneficial on the Kent Mainline.

The siding on the east side used to be connected at the south end but was relaid in 2008 as a Thameslink reversing siding connected at the north end. There have been suggestions to bring this siding into use as a platform line, though it’s not clear what service would use it. A new platform here would be expensive to provide: a new access would need to be dug from the subway, the platform would need to cantilever out over Milkwood Road for over half its length, and it would not provide the benefit of cross-platform interchange.
As noted above, the alignment towards Loughborough Junc is generous and a new reversing siding could be provided here, sitting between Up and Down tracks to avoid crossing movements. This would release precious space on the Herne Hill viaduct to widen the easternmost platform ― an important factor as the interchange becomes busier.

Perhaps this alternative layout should be explored by Network Rail or TfL, to see whether it provides more capacity and/or operational resilience. It may only yield a small number of additional train paths, if any, and the extensive works required will need to demonstrate significant benefits for it to pass muster. We include it here, though, to illustrate how constrained the Herne Hill site is and how few options there are, hence the need to consider more radical alternatives such as changing the service pattern.

Uncrossing

Perhaps we are expecting too much of this diminutive station caught in the cross-traffic. Given that the fundamental problem is that of two conflicting service patterns, is there a case to reconfigure the pattern to minimise crossing movements?

The myriad route options across the south London network are almost unfathomable. This author is still recovering from the mind-numbing task of teasing out the Herne Hill train movements from the various diagrams in the 2008 South London RUS and other sources.

The diagram below presents a snapshot of this small nub of the rail network, showing indicative services in 2015. Immediately it is clear that the service pattern is about as far removed from the typical London Underground pattern of lines and interchanges as it is possible to be. No wonder casual passengers prefer to look for the nearest ‘tube’.

The throughput of the station is 18tph in the morning peak on two running lines, a relatively modest utilisation because of the crossing pattern. If trains did not cross, what could the throughput of the station be?

Our attempt to untangle the service pattern is shown below. This option would require a simpler track layout, shown in the diagram above. Tulse Hill services would use the eastern island platform and go on towards Brixton and Victoria. Kent mainline services would use both platforms depending on whether they are destined for Victoria or Blackfriars. You’d lose cross-platform interchange, but the service pattern would no longer cross, which in theory should allow a more intensive service, particularly on the Blackfriars route.

The 2011 London and SE RUS is actually projecting a 16% drop in peak demand between Victoria and Kent by 2031, in the expectation that there will be a substantial shift to HS1 services into St Pancras (Table 6.2, p94). It also expects an increase of 15% on the Blackfriars route through Elephant & Castle. The pattern below therefore sends more Kent services to Blackfriars, which reduces crossing movements at Herne Hill. Thameslink services continue to run via Peckham Rye, onto the easternmost pair of tracks on the long viaduct through Elephant & Castle.

The service frequency on the Kent Mainline will still be limited by the two-track sections sharing Fast and Slow services. Although this simplified service pattern and track layout will help resolve the Herne Hill bottleneck, it pushes the pressure down the line, in particular Brixton Junc and the two-track section from Shortlands.

There may be potential to provide crossovers at Herne Hill North Jn to allow Fast trains to pass Slow, but in practice the crossing movements may actually reduce capacity, and the junction will also limit platform extensions. Plus, it is not apparent whether the crossovers should be arranged to provide layover for the Victoria route or the Elephant & Castle route. It is likely that careful timetabling is the most efficient option.

The problem with uncrossing the pattern at Herne Hill is that the knock-on effects through other routes may be widespread and unwelcome. Beleaguered Wimbledon Loop passengers now find themselves heading to Victoria on a circuitous route, although the trade-off is an increase in service frequency and improved interchange.

There is a better option: trains also cross at Tulse Hill and here there is potential for grade-separation at the SW end which would allow the service pattern to cross. This would allow Wimbledon Loop services to run to London Bridge, and a Croydon-West Norwood-Herne Hill service to Victoria, which could help provide relief to the Brighton Mainline.

There will be other permutations. The aim with this flight of fancy is not to find a magical perfect solution, but merely to illustrate some of the options – and we stress that it is merely a sketch, rather than a detailed look at service paths and capacity.

An improved TfL “strategic interchange”?

The 2011 Stations RUS identifies eight stations where:

there is some uncertainty whether current improvement plans will fully address congestion issues in the future, and it is therefore recommended that the situation at these locations be kept under review.

Seven out of these eight are in London and five are on the Kent/Southeast Mainline, including Herne Hill.

The Mayor’s 2011 submission on the HLOS notes that Herne Hill is one of the key stations with significant congestion problems projected in 2014-19, and recommends specific improvement works: new entrance doors, removal of an interior wall, wider stairs to the platforms and a second station entrance. These improvements may not be that straightforward given the cramped nature of the station.

There may be benefits in a more fundamental rebuilding, which could be combined with track reconfiguration and associated works. This could include new subways, and relinquishing the Thameslink siding in order to widen the easternmost island platform.

The options outlined above aim to provide at least 6 tph on all routes, which could be regarded as a minimum requirement for a TfL “strategic interchange” with turn-up-and-go services. But for Herne Hill to realise its value as an interchange, should Fast services stop here? We have included this in our indicative service pattern above, but it is a thorny question which also applies to interchanges such as Finsbury Park, Wimbledon and Lewisham, and to potential new interchanges such as Old Oak Common.

Creating an additional stop at Herne Hill will need to demonstrate significant benefits to passengers in the form of useful route options with good frequency. Herne Hill does not currently have that, not least because it is not served by the Underground.

But what of the putative extension of the Victoria Line? This would extend the 1200’ overrun tunnels at Brixton into a long loop with Herne Hill at the south-eastern end as a one-way station. The key benefit is to provide faster turnaround at the southern end of the line. The distance is modest and building works could use land at the Network Rail depot at North Jn. It would benefit passengers on local journeys, and become useful as more Kent Mainline trains are routed to Blackfriars.
The downside is that a Victoria Line extension may exacerbate overcrowding on the line, and it also duplicates the surface route to Victoria. In 2005 TfL confirmed that the extension was not a priority as it had a weaker business case than other infrastructure projects.

This may, however, be worth revisiting in the light of new demand forecasts, improved service frequencies and the opportunities presented by the Blackfriars route. Not least, TfL needs to ask whether Herne Hill is indeed a strategic interchange, and the base level of investment required for this to be tenable.

So to conclude, it is clear that Herne Hill will always a problematic node in the south London network. Grade-separation appears to be illusory, unless tunnelled all the way from Tulse Hill to Loughborough Jn as part of a grander scheme. But there are other options.

The task now is to weigh those options alongside other potential investments nearby, such as Brixton and Lewisham, and assess alternative portfolios of investments for CP5 and beyond – and, of course, to safeguard an alignment for further bay platforms at Blackfriars, so the additional trains have somewhere to go.