After reading Martin’s post about the SW Seattle service changes here, here and here I wanted to add a few comments about transfers. Basically, not all transfers are equal. I know this is pretty basic but this point gets overlooked a lot, especially in places like Seattle, where few routes have less than 10 minute headways. Ever ask how to get somewhere by bus and be told to “take route # to ___ and transfer to blah blah blah blah blah”? Yeah you kind of glaze over and zone out when they say transfer. In these circumstances lots of people decided to drive, take a taxi or don’t even make the trip.
In technical circles this is call a “transfer penalty” and is a number that is factored into a transportation demand modeling disutility function (for more info go to one of my old post on mode choice modeling over at oraphanroad). This transfer penalty is relatively high and is why many small and medium sized transit systems try to maximize the number of trips that can be made without transferring. This is also one reason, among many, why many transit planners don’t believe in circulator routes. They force riders to transfer and riders hate that. So its pretty easy to see why both riders and transit planners trend to be skeptical about changing one seat based systems to transfer based systems.
However, like most things, its not that simple. First, the mode has an impact on transfer rates. Riders are more likely to prefer a rail-to-rail or bus-to-rail transfer over a rail-to-bus transfer even if rail and bus take them to the same place. Riders are simply more comfortable on trains. You might become acutely aware of this while you are traveling. Tourist hate buses and love trains. While in NYC a few weeks ago, I took the subway from Brooklyn into Manhattan, and then back out to Brooklyn. I know there were buses which probably were faster, but it was so much easier and I knew I wouldn’t get lost.
Much more significantly though, the headway length of both transit lines has a huge impact on the willingness (or disutility) of riders to transfer.
When transferring to a line with short headways (ie it comes often, generally less than 6-7 minutes) from a line with short or long headway (Case A and B) the transfer will always be quick and easy. If there is no transfer coordination the average transfer time will be 1/2 of the destination line’s headway. When transferring from a line that has short headways to a line with long headways (Case C) the transfer becomes more problematic. In these circumstances rider information is critical because the total trip duration is determined by the transfer time which is a function of the destination line headway. This would be akin to a trip from Downtown to Seward Park in the afternoon. The last possibility, Case D, can only be address through timed or coordinated transfers. Transfer coordination can be very extensive requiring significant planning or it can simply favor a travel direction.
I point all this out because smart transfer planning and rider information can go a long way in solving some of the concerns that riders have, especially when making significant changes. The major transfer “pain point” for most riders will be the afternoon transfer from LINK to buses (Case C). By providing coordinated transfers between LINK and local buses (with priority in peak direction) and pre-trip information about which LINK train riders should take to minimize transfer time, the transfer experience can be significantly improved without improving bus service headways. At the very least Sound Transit and Metro should work together to provide real-time information at LINK stations for riders not only transferring to LINK but also to buses. Onebusaway already provides this information and hopefully Metro will finally wake up and appreciate how important a service Onebusaway provides. Metro should keep in mind that it is selling a product and tools like Onebusaway, especially in this context, add significant and tangible value for riders. I know it does for me.