Transfer design

Or an attempt to explore how well feeder bus services actually work, in promoting accessibility to the MRT network. It’s tangentially related to the Newton case, so you might want to take a look at that first. There is also more than enough literature out there on rail-rail transfers, so that won’t be the topic of this post.

The recent bus service changes in Bukit Panjang have given me a pretty good alternative — instead of entering the bus interchange to take 973 or 976, I can just take 171 directly from the MRT exit. This reminded me of another reddit argument I got into a while back, where someone wanted to know why I felt that a feeder bus was preferable (from a policy perspective) over detouring an MRT route to provide a missing link.

It may surprise you that this is actually an urban planning issue and not one directly related to transit service, even if it does have implications on rail ridership and thus rail sustainability.

Not exactly “integrated” transport hub

The opening of Woodlands Regional Interchange in 1996 was heralded as Singapore’s first integrated transport facility, with the bus platforms directly beneath the MRT in the same building — though at a stretch you might argue that it is Yio Chu Kang that deserves this honor with the bus facilities wrapped around the MRT station concourse from day one in 1987. Six years after Woodlands, Toa Payoh bus interchange was also upgraded with two exits leading from the MRT underpass directly into the bus interchanges; the same also being done at Hougang Central to the new NEL station below. Sengkang Bus Interchange is also directly accessible from the NEL ground-level fareline.

But why not any more? I partially blame the private sector. Integrated bus interchanges are now to be built by the private sector as a condition of government land sales. This means that the private-sector development would logically have an incentive to force passengers to detour through their commercial developments, in order to drive foot traffic. The weakness of this model is apparent given current events and access control policies implemented as part of public health measures.

There are exceptions to such private sector involvement, of course. Toa Payoh’s upgrade came with the HDB’s new offices. Bidadari interchange is being built by the HDB as part of its Woodleigh Village project. Sometimes, it can also be poor land availability — such as the upcoming Beauty World ITH, located at the land lot opposite the hawker centre, and easily a 100–150m underpass walk from the MRT fareline. This is because the area around the station is occupied by shophouses and Beauty World Centre, making it difficult to site a decently large bus terminal. Or Yishun, where there wasn’t a good place to site the bus interchange; remember having to walk through Northpoint for the temporary interchange?

Of course, it doesn’t help that MRT lines are historically routed along roads, which means that access from the underground train station to the at-grade bus terminal, even if it had direct pedestrian access from the pavement without having to walk through a mall, can still require quite a few level changes. Even an aboveground line is not spared — from JRL Boon Lay to the bus interchange looks like it’s going to be hell on earth, though its access to the EWL is far better, which presumably can drive walk > JRL > EWL type trips instead of taking a bus to the interchange then to EWL.

Down and up and down again

So what are the practical implications of such designs? This is what I personally like to call the “access penalty” — the time taken to descend up/down to the train platform. The larger the station and/or the more escalators are needed, the larger the penalty. This is important because it adds to perceived travel time.

Let’s look at this in the context of TEL Stage 1. It would theoretically be possible for bus planners to reduce service to the areas in the vicinity of Woodlands South station such that they can take the TEL to connect to the NSL instead. Furthermore, bus 902 should have been killed a long time ago, with Republic Poly students told to either take the train to Woodlands North, or walk, or take 169. You can take a stopwatch and repeat this experiment if you don’t believe me. The walking time from the NSL to the TEL should be about the same as that to 902’s boarding berth at the temporary interchange.

This indicates to me that the access penalty from changing to TEL to get to Woodlands North, and the walk from Woodlands North to RP proper, is enough to dissuade potential TEL riders from breaking their old habits and using the TEL for this stretch of journey. It could have other implications such as at the relatively deep Tampines East and West stations, where the poor transfer at Tampines — in other words, high access penalty to EWL from DTL — means that you might as well stick with the DTL all the way to town. And if you want EWL, you might as well take a bus to the interchange.

At Bukit Panjang, though, things are quite different, yet the underlying principles are still the same. The local sections of 922 and 972 are almost practically the same, yet the 922 + DTL combination is not that popular and people take 972 instead. Modifying 922 to serve the Senja Road BTOs could work, but why do that when 972 takes you direct to Orchard? My personal suspicion is that the access penalty from the LRT and the bus interchanges, requiring crossing the canal at the big empty plaza before going down into the DTL station, is simply too high and takes too much time for a lot of people — with 972 the time taken from this area to Stevens and Orchard is actually lesser than with the ostensibly faster MRT, because of the poor connections.

Solutions?

This is where route planners can consider taking a more holistic solution towards the problem. Route design can take into consideration not only the presence, but also the quality of transfers provided at a location, and especially at rail stations and for “rail feeder” services consider the use of stops more conveniently located to the rail service instead of the interchange being a be-all and end-all. URA can also prevent this from happening in future by developing dense, walkable communities around MRT stations first and foremost and reducing the need for feeder buses to developments further out in the first place.

One might even want to be more ambitious by not having feeder services turn into the interchange if it’s a significant detour from the MRT station. Bus drivers can simply be relieved at a chosen bus stop, that is either near an interchange (allowing them to use the interchange for breaks), or a roadside terminal provided for them (a practice called “jumpbus”). With the current way schedules are set up, I feel that there is a lesser risk of continuous looping resulting in delays building up along the route as the day goes on. Continuous looping might even make some trip pairs possible that weren’t possible before, or at least no longer a major pain in the ass. But it may also bring unintended consequences such as needing more drivers to maintain the service level, which we can ill afford.

The other alternative, of course, is brute force — by removing or reducing connecting bus service within the 400m radius circle of the rail service, force people to walk to a rail line. This can be acceptable in certain situations, such as densely populated inner towns and transit-oriented development where feeder service to meet the high expected demand would be impractical, or even the low-medium density sprawl between Bedok and Kallang where bus services may not be able to run too frequently and stops more spaced out (outside the 400m radius of a station) due to the lower demand. But of course, this always runs the risk of angry residents resulting in politicians losing elections, and thus political pressure being applied to reverse unpopular initiatives.

Is there any other way?

Here’s a question I think needs to be gotten out of the way as well — why not just get URA to intensify the land use around stations, have more people within the 400m walkshed, so we don’t need to provide feeder service at all? That’s one thing that the URA should consider, no doubt, but you can’t really have 40–50 floor high massive towers clumped around train stations while still maintaining some kind of liveability. At the end of the day, we also need an air force which means airbases all over and pesky height limits. And if you still talk about bus service, it may not really solve the problem at hand. There are such developments already planned, like the grade separated pedestrian networks in the CBD and Jurong are, encouraging walkability instead of last-mile bus transfers.

There’s also the possibility of encouraging last mile active mobility. Biking around the neighbourhood, things like that. The good news is that a bicycle is more versatile than a bus, and with extra space in train stations some space can be made for bicycle parking much like at Keppel and Marine Parade, if not the bike racks already available at existing train stations.

The spokes, after all, are only held together by a strong hub. The best service standards won’t save you from poor connections dissuading passengers from using the service. It might even make things worse and result in losing money on the routes in question, as we’re seeing now.

And no, my answer remains the same as Ong Teng Cheong’s back in 1982 — there was, and still remains, a very good reason why we did not just retain the major bus trunk routes as is when developing the MRT Initial System. It thus remains important that the feeder system be strengthened especially around new MRT line openings, if it has to be done at the expense of long, unreliable trunk routes. Interestingly, this is something that even KL leads us by (they’ve even gone to cashless flat fares!).