Length matters
Or why simply lengthening trains may not work.
A common complaint about the DTL and perhaps TEL is that they may be underbuilt to handle the expected passenger loads. This is a trope happily plugged by vested interests but I’m afraid it may not be so simple.
How true is such a claim? Sure, having only three or four cars may be a pittance compared to other radial lines, and sure, frequency may not be able to fully make up for the load. But before we can even expand on the form such a service expansion might take, we may need to look at whether the need even exists in the first place.
As things stand
I would, in fact, go as far as to argue that ridership projections have so far been quite spot on. In 2011 the DTL was claimed to be able to serve around half a million passengers a day. In 2019, 477k passengers in a typical day were reported by SBST. After accounting for differences in traffic between weekdays and weekends, we could be there or even well past the 500k mark.
But in 2019 SBST also claimed at their Annual General Meeting that their break even point was 650k passengers a day. I don’t see why they would lie to their investors. This figure may raise some questions, but they may be easily answered. One potential explanation is that the rail reliability drive and resulting insistence on staffing every train has resulted in increased operating costs, resulting in a greater hurdle when thinking about cost per passenger.
A post-Khaw reassessment on the true cost of sky-high reliability should help, apart from the contracting structure being changed to being paid based on mileage operated instead of passengers being carried. Also consider the DTL’s role in bringing passengers to the New Downtown — if new work policies create a situation where it’s not so important for many to travel down to the New Downtown every day, it may not be possible to once again reach that 500k value.
And lastly, that 500k number may have been reached merely with a two and a half minute train frequency as well, considering that that’s the train frequency operated today and we’re supposedly operating 2019’s peak hour schedules. Adding that few more trips per hour at the peak of peak may also increase the possible carrying capacity by around a hundred thousand daily, assuming there’s riders to actually justify the increased service level. Sure, it may not be so necessary to hit 650k with the LTA taking over revenue risk, but it’s still important to keep passenger counts up to minimize the subsidies needed.
The Future
There are sources of additional passengers. Firstly and most immediately, the removal of highway bus services along the route from Bukit Panjang, Tampines, and Bedok, or “promotion” of such services to the express fare band, may also help in pushing just several thousand more to the rail service as a relatively more affordable alternative. If they wish to pay more for the added comfort of a seat, let them.
Next we have the potential for developments along the line, but these may be limited, though. What I can think of are the Tampines West BTOs, some redevelopment in the Bukit Panjang area, and then the southern edge of the Turf City redevelopment that touches areas near Sixth Avenue station. By interchanging at King Albert Park from the CRL, it may also be possible that the DTL forms the main trunk for residents here — but how useful public transport may be for them remains to be a mixed bag.
Still, I have a hard time believing this can generate 25% more riders to justify 25% more capacity. As a radial, many DTL passengers may be on the line for more than eight stations — or roughly 10km, which means that 500k passengers will generate more passenger-km than on a comparable line. Trains currently start out empty nearer to the ends and then fill up as they approach the new downtown, with long cross-island trips being diverted to the CCL and perhaps in future the CRL as well.
This may change, too. The DTL3 Extension allows Tampines and maybe Bedok residents to use spare reverse-peak capacity to get to the TEL at Sungei Bedok station and avoid the not-interchanges at the central area, should that option be more convenient for their final destination. Likewise, whilst the DTL2 Extension may capture a larger portion of Yew Tee residents than what 979 can do now with more favourable station placing, there could also be some reverse-peak travel to the Sungei Kadut industrial area, from either Bukit Panjang or King Albert Park. This is because the CRL doesn’t have decent connections towards the northwest portion of the NSL — the best being to change to the EWL at Clementi and then go 1 stop to Jurong East to get the NSL from there.
The stretch
A train every 2.5 minutes is 24 trains per hour. 25% more capacity means 6 more trains per hour, which means only 30 trains operate in an hour — that is, a train every two minutes, similar to the pre-2017 NSL status quo. Considering that design capacity is 90 seconds, and we can get much closer to that, two minutes isn’t really a lot.
As we’ve seen, a considerable amount of the potential for growth comes from the reverse-peak traveller instead of those journeying to the New Downtown. Adding reverse-peak travellers may help to increase absolute passenger counts and make better use of empty trains going the other way, which may mean less service increase is actually needed just to see more passengers on the line. The same argument can be used to justify the Punggol Digital District, where passengers going to PDD would be using otherwise-empty northbound trains returning to Punggol to pick up more passengers anyway.
Consequently, to me, it looks more likely that running 36 trains per hour — or more — may only really happen should 50% more capacity in a single given direction be justified. With long-term redevelopment, this could happen, but it may be more on the side of unlikely. And even if it does happen, it may be sufficiently far away enough in the future that technological advances permit running even more trains on the same infrastructure, to the extent seen in Taipei or even on the VAL systems in France.
Plus, there still remains a possible route along Holland Road and to Tengah that can be used to build a high-capacity MRT line which could take some load off too. This means current passengers who need a bus ride to the DTL could instead take a bus ride to this other line instead — mainly passengers in Bukit Batok and Choa Chu Kang.
On the east side, there’s the TEL Stage 4 and 5 aka ERL. One might argue that 4 cars are overkill for what the ERL proper passes through, but connecting the DTL 3 Extension to Sungei Bedok also allows for routing around the DTL to use destinations along the TSL — or one can also take a bus to the ERL instead.
Getting dirty
Only after such “easier” options have been exhausted is one then able to justify the extensive works that will be necessary to permit longer trains to operate on the line. This may be a thing only in the 2040s or the 2050s, especially with the possibility to build a new line still around.
It is probably going to be difficult to avoid any platform extension works. Even a short-lived flirtation with 9-car trains on the London Tube, which only lasted around two years, required extensive modifications to signalling systems to run the nine-car trains. The need for platform extensions was avoided with a complicated mix of stopping either the first two or last two cars in tunnels, with passengers told to use those cars depending on how far they were going.
It might say something that this was quickly abandoned a few years later and that London, like most other European cities with small subway trains, has mostly chosen to keep frequency higher instead of revisiting the prospect of longer trains again. It only did so with the introduction of the S7 stock, where walk-through trains mitigated the effect by allowing passengers to proceed to an adjoining car to exit the train. In this case, the perks of having a common fleet far outweighing the cost of whatever modifications were needed for short stations. And even so, where possible, they attempted to extend platforms.
In Singapore the works will likely be similar and yet considerably different. The most immediately obvious to anyone is the expansion of platforms into what are now platform-end ancillary areas, which mostly already exist here and lead to service rooms; instead of having to dig them out once again as London and New York had to. The extent to which this can be done, being limited by space provision for emergency escape staircases and even the station sizes themselves, will become a control on train size and may even force us to have to considerably redesign trains to meet the space constraints.
Even if you wanted to learn from London and use the weird stopping patterns tried in the 1930s to eke out savings on station modifications, extensive upgrading will still be needed to the depots and sidings in order to even store and maintain longer trains in the first place — storage tracks and workshops will need to be expanded, and the like. If necessary, it might be possible to tweak track geometries like using narrower crossovers in order to squeeze in the additional space needed to park the extra cars; but doing that may require full line closures or even single tracking for weeks (if the Marina Bay works are any indication).
The use of selective door operation, or keeping certain doors closed, should not be an excuse for not lengthening platforms where possible either. Any door kept closed impacts passenger exchange times and thus frequency, and may also pose difficulties in crowded trains where passengers have to push their way through the train towards a door that will open. This is similar to what you see on a bus where people congregate around the exit door in the middle despite repeated exhortations to move to the rear portion.
Having to keep a longer train less crowded to allow passenger flow within the train may reduce the additional passenger carrying effect of longer trains. At the very least, even if the ends of the longer train can be thought of as similar to the upper deck of a double decker bus, it will still be quite necessary to extend platforms at major downtown stations and interchanges to allow these parts of the train to be accessed.
Elsewhere
What about other lines? For the TEL, a lot of the arguments are similar. Prospects for future growth on the TEL look better. Interestingly, the eventual daily ridership of the TEL is actually claimed to be a million in the long run — can the one additional car contribute to double the amount of trips? It’s quite possible. The TEL runs through generally quieter areas, with medium term growth mostly around Lentor, Mount Pleasant, Bayshore, and perhaps the Havelock/Kim Seng area. However, there’s the 900-pound gorilla that is the RTS; at 10k passengers per hour per direction, if fully used the RTS may be able to contribute over 100k daily trips to the TEL.
But how would this work with four-car trains? I suspect this may be a combination of proper use of the 5th door keeping dwells down and thus frequency up, as well as planners thinking that the average TEL trip generally is shorter, meaning more people use a given space on a train.
The second argument also applies for the Circle Line, which carried a similar amount of passengers in 2019 (176 million) as the DTL over 20% less absolute track length. As an orbital, journeys are expected to be shorter, and the current lower demand for the city and CCL5 areas may result in actual heavier loading on the north side of the Circle. This is the justifying case for the CRL, especially the CCNR tunnel, which can funnel away passengers who would otherwise use the stretch from Serangoon to Buona Vista to change lines.
But of course, implementation is always the issue. The CCNR tunnel is at least 10 years away, and the 9th MRT line through Seletar and Sengkang West is probably around 15 years away too. Operating more trains on the NEL could be a short term measure (and in fact is already happening), and there may be more room to improve that with the completion of refurbishment works and the entry into service of the C851E trains. But that said, I can’t help but wonder whether a 7th car would be viable there as well.
As previously explored, apart from full 2-car operation, the Sengkang-Punggol LRT could also benefit from longer trains. This could be either fixed 3-car sets, which would require modifications to existing facilities, AND for the new depot to be built to cater to 3-car trains from the beginning to avoid a second round of invasive modifications — which it doesn’t seem to be. Or like in Taipei on the Brown line, a pair of 2-car trains could be linked together to operate as a 4-car train — while possibly needing more platform modifications, a 4-car train could be split into its constituent pairs within the depot for easier maintenance and avoid the depot modifications.
Finding a way to lengthen the LRT trains first, including the preparation of necessary infrastructure such as shifting of plant rooms, could give us experience in what we would need to do, should it ever be necessary at any point in the future to lengthen trains on MRT lines.
