Human Transit, a very technical and theory focused transit blog by Australia based Jarrett Walker, this weekend posted and article about the unique benefits that Seattle enjoys when it comes to transit. He uses Seattle as an example of how chokepoints, in Seattle’s case geographically created chokepoints, creates very valuable opportunities for transit to becomes more competitive with cars, compared to cities that have few chokepoints. Jarrett writes:

If you want a real focus for sustainable transport improvements, however, look for chokepoints.

A chokepoint is anywhere in the transport network where many different trip paths have to go through the same point to get past a geographic barrier. Bridges and tunnels are chokepoints. So are mountain passes. Wherever a steep hill is right next to a body of water, the little ledge in between them is a chokepoint, as it often only has room for one road, or a road plus a single track of rail.

No North American city has more chokepoints than Seattle. The city itself consists of three peninsulas with narrow water barriers between them. Further barriers are created by steep hills in most parts of the city. Nowhere in Seattle can you travel in a straight line for more than a few miles without going into the water or over a cliff.

Seattle’s geographical isolation from its suburbs, of course, means it is also surrounded by chokepoints. There are only two bridges across Lake Washington to the east, and to take your car across Puget Sound on the west you have to use a car ferry, which means your trip will be no faster or more frequent than that of a transit passenger.

Transit planning is frustrating in such a place, but road planning is even more so. Ultimately, Seattle’s chokepoints have the effect of reducing much of the complex problem of mode share to a critical decision about a strategic spot. If you give transit an advantage through a chokepoint, you’ve given it a big advantage over a large area.

He continues:

The lesson of Seattle is that successful transit infrastructure responds to demand, and what drives transit demand is high overall travel demand plus serious barriers to driving. In Seattle, a lot of the barriers to driving take the form of hassle and delay, due to limited capacity through chokepoints.

In other words chokepoints act as an equalizers between transit and cars. Give priority to transit at the chokepoint and you improve its relative attractiveness over cars to a large area. The westbound 3+ HOV lane on SR-520 is a great example.

The relative attractiveness between cars and transit is most commonly model using a logit model. As the attractiveness of transit and cars begin to equalize something interesting happens. Each time transit becomes more attractive, through a reduced travel time of 2 minutes for example, the marginal mode share increases even more than it did for the last 2 minutes of travel time savings. Simply put a travel time savings of 2 minutes is more important when the travel time of car and transit are already close or getting closer. If a car takes 30 minutes but transit takes 50 minutes a 2 minute improvement doesn’t mean much. But if a car takes 30 minutes and transit takes 34 minutes a 2 minute improvement does much more to gain new riders.

This is why Chokepoints like SR-520 eastbound, where transit can travel as fast or faster than cars, especially during peak periods, are places where transit is so competitive with cars. And as Jarrett points out chokepoints are where this confluence of events naturally occur. Read his full post here. There are some good things in it!

UPDATE 1:15 – Jarrett has posted a follow-up piece here.

98 Replies to “Human Transit: Chokepoints and Seattle”

  1. The link to his full post is not working.

    Feel free to delete this comment once it is fixed.

  2. Another “chokepoint” that Seattle has addressed well, unlike a certain large urban area in the Southern West Coast north of San Diego, is the issue of parking pricing and the ability of employees to “cash-out”.

    Also having the trip-reduction act really helps.

    Mention that in “the Southland” and people think you are from Mars. Or the Soviet Union circa 1956. Donald Shoup tries (on the parking issue), but his proven methods are viewed as heresy.

    And then there is the Growth Management act which ensures that 40-story office towers don’t pop up in, say, Sunset Hill, ala Century City (which, I must admit, was supposed to be hooked onto the real/originally-planned “Santa Monica Freeway”, which was never built)

    1. The Soviet Union didn’t have large quantities of cars in 1956. I really doubt that someone from there at that time would have been concerned much with traffic management….

      1. Anyone discussing the topic of land use controls in much of the USA was declared a “Communist” as late as 1961. That was my point.

  3. Of course if transit takes 50 minutes then saving 2 minutes isn’t as much as if it was 34 minutes and you save 2 minutes. In the former case, 2 minutes represents 4% of the total travel time. In the latter, it represents 5.9%. The real question should be posed given these premises:

    Scenario 1: A transit trip takes 50 minutes, while a car trip takes 30 minutes.

    Scenario 2: A transit trip takes 50 minutes, while a car trip takes 45 minutes.

    If you can shave two minutes off a transit trip, which scenario will gain the most riders from that times savings?

    1. Scenario 2.

      Potential transit users are much more likely to switch when the difference is three minutes than when the difference is 18 minutes.

      The big reason here is that the car trip’s time variability comes into play – the car trip could take an extra three minutes regularly, but it’s unlikely to take an extra eighteen minutes regularly.

    2. It’s not always about shorter commutes but also quality of commute time. Time spent on a bus or light rail car can be spent reading, zoning out, or doing work on your laptop (especially if transit is WIFI equipped). Many people prefer a slightly longer commute because they don’t have to worry about driving.

      This is also true of cycling, especially during good weather. Not all of the cyclists on I-90 on a warm sunny day are cycling across that bridge to save the planet. Most are doing it because it’s more pleasureful than driving to work in a steel box. They are also multi-tasking by getting exercise on their commute.

      It takes me 50 minutes to bike one-way, 40 minutes to bike/bus, and about 20-25 minutes to drive (off-peak). Depending on weather, my commute preferences are usually an inverse of commute time.

      1. I too bike to work, and it’s 65 minutes bicycle, 50 minutes car/bus, 45 minutes car but if I drive I have to take my wife directly to her work (car pool), so her commute goes to 45 and mine becomes 50 again!

        The other factor not mentioned is parking fees. A $20/day parking fee is enough to get me to look for alternative ways to get to work.

        The third is at work showers. Nothing like an hour ride in the rain and no hot shower and a clean dry towel.

        At one time I looked at the “save the planet” issues and tried to add in all the non planet saving things that bicycles do vs my car, and the one final one is “not burn gasoline”. The rest is a wash, as I shower more (twice instead of once), I replace tires more frequently as bicycle tires don’t last near as long as auto tires. Those rechargeable batteries for my light system aren’t recycled etc etc. The overriding thing that keeps me riding is the exercise and the view. This is a beautiful city and you can really appreciate that when you are out in it and not inside a bus or car.

      2. I assure you, no matter what goes into your bicycle commute it will *NEVER* compare to your automobile commute in it’s impact. Yes, there are impacts to cycling as well. Tires and brake pads, materials used in the manufacture of the bike, water and energy for showers, extra food that needs to be grown to fuel the cyclist, etc. But don’t forget the indirect impacts of an automobile such as concrete and/or asphalt to park it on, materials and energy to manufacture it, oil to lubricate it with, or the long list of replacement parts that need to be manufactured. Heck, it takes 40,000 gallons of water to manufacture the average car. That’s a lot of showers.

        It’s also easier to mitigate the impacts of a cyclist. Wastewater from showers can be (and is) recycled. Heating water for a shower is far less energy intensive than propelling a 2 ton vehicle around for 45 minutes. Food can be grown organically.

      3. Bicycle tires may need to be replaced more often than car tires, but consider the difference in mass. My car tires weigh something like 40 lbs each, there are four of them, and they wear out in around 50,000 miles. 0.0032 lbs of tire per mile.

        My bicycle tires are 590 grams each, so a pair is about 2.6 pounds, and they generally last around 5,000 miles of commuting. 0.00052 lbs of tire per mile.

        Almost an order of magnitude less tire rubber per mile for the bike.

        Of course, you have to consider the long-term environmental impact of cycling. Bicycle commuters have an average lifespan several years longer than motorized commuters, according to peer-reviewed studies in Europe. All those years of food, clothing, hot showers….

    3. the point that should be made is that as Transit travel time gets closer to car travel time, the relative popularity of transit travel increases at a higher rate. Using his example, if a trip takes 50 minutes via bus and 30 minutes via car, an improvement of 2 minutes means a 10% decrease in the time penalty the bus rider pays. On the other hand, if a trip takes 34 minutes via bus and 30 minutes via car, an improvement of 2 minutes cuts the time penalty by half.

      Also, if the bus trip is only 4 minutes slower than the car, it is possible that all the low hanging fruit has been already gobbled up, and the bus route is already quite efficient. So, like an object approaching the speed of light, increasing speed (cutting the difference between car and bus) is going to become more and more expensive. So, while return (increased ridership) vs. time saved will go up, return vs. cost or effort expended might go down for those last 2 minutes.

      1. Yeah exactly and that is why it is so important to try to target markets where cars are slow and transit can be faster. I think Ballard to UW is a great example. Driving is slow and buses are even slower, but if some form of transit could do the NE 45th/46th journey faster than cars, by having exclusive ROW then you’ll pick up a lot of ridership.

      2. The problem is exclusive ROW between Ballard and the UW means either an elevated line or a tunnel. An elevated line down 45th would face huge opposition, especially in Wallingford. A tunnel is going to be hugely expensive even if you only put it in the most congested part of the corridor. Even an elevated line may be more expensive than the potential ridership can really justify.

        If such a line is ever seriously on the table I’d say a possible alignment via Fremont should be considered. While it is a bit of an out of direction jog there is a fair concentration of jobs, residents, and activities that make it an attractive destination for transit.

  4. I find it curious that this post by Adam Parast does not even mention the main point of the article referred to, which is this:

    “Yet bus-dependent Seattle has a higher transit mode share to downtown (transit trips as a percentage of all trips) than rail-oriented Portland does.”

    This article makes a good argument for bus raped transit in Seattle, instead of light rail. And Jarrett Walker, who wrote the article on choke points in Seattle, doesn’t even consider that Sound Transit’s plan for light rail over I-90 is a large DEcrease in capacity compared to just the current bus service over that bridge.

    1. Norman there were a lot of things that I didn’t talk about. I chose to focus just on the chokepoint idea because I think that is the most interesting point in a purely technical way.

      I find that when you talk about more than one topic at a time you lose any level of clarity and detail in a conversation about it, and at that point the discussion becomes predictable and pointless.

    2. Norman, his point is that chokepoints create the opportunity for transit usage, and Seattle has more chokepoints than Portland. Seattle’s transit usage might be much higher yet if we had a light rail system of Portland’s size in place. After a couple of recent trips to/from Seatac on both 194 bus and Link, I can say I’d never ride the 194 again (even if weren’t being eliminated.) The space and ride quality on Link are substantially better.

      I don’t understand how Link on I-90 reduces capacity. Today there are 8 lanes of vehicle traffic across the bridge (and the “reverse” commute is a nightmare.) After construction is complete there will be 8 lanes of vehicle traffic plus light rail. That’s an increase in total capacity, and with permanent HOV lanes both ways, the reverse commute will function a lot better.

      1. The outer spans can each be 4 lanes with either light rail or brt in the middle. brt gives you much higher capacity in the center spans than ST’s plan for light rail.

        In other words, the choice is for 8 lanes of vehicle traffic plus light rail in the middle; or 8 lanes of vehicle traffic plus brt in the middle; or even 8 lanes of vehicle traffic plus HOV lanes with SWIFT-type bus service in the middle.

        No matter what you do with the middle span, the outer spans can each carry 4 lanes of vehicle traffic — a total of 8 lanes of vehicle traffic.

        The space on Link is “better” because it is never filled to anything close to capacity, as the 194 often is. If they had added eqivalent capacity to the 194, then space on the 194 would be better, too.

      2. Norman, the bus service over I-90 doesn’t all go away because of East Link. In fact, it’ll have 2-way HOV in a way it doesn’t have today.

        It may replace the 550, but even you wouldn’t argue that Link has lower capacity than that.

      3. You could keep all existing buses and other traffic which now crosses the bridge, and move all of that traffic, including the current buses, in four outer lanes each way, instead of the current 3 outer lanes each way.

        Then, adding just 50 buses per hour in each direction in the middle lanes gives you more capacity than the ST plan for light rail in 2030.

        Or, move all existing buses to the middle lanes, while adding another 50 buses per hour per direction in the middle lanes, for 100 buses per hour per direction in the middle lanes, opening up more room in the outer spans for carpools and van pools. One hundred buses per hour per direction in the middle lanes is more than double the capacity of the planned ST light rail.

        So, basically, you could double the number of buses which cross the I-90 bridge during peak hours now (which is currently around 50 per hour per direction), and that would be more additional capacity than the planned light rail. All the additional buses would not have to be on one route — they could be spread out among several existing routes, and maybe create some new routes, as well.

      4. Norman, you’re comparing the number of buses you could put on the roadway to the actual number of trains running based on projected demand in 2030. the trackway could obviously accommodate many more trains. At any rate, the dominant operating cost is operators, and under your assumptions you’ve got 13 times as many operators serving the same corridor.

        Are your buses ever going to stop? Because the spacing constraint is not the physical roadway or trackway but the time spent loitering at stops. Anyone who’s tried to get off a crowded bus and a crowded link train knows which is faster, and that’s before we deal with wheelchair lifts and bike racks. Combined with different routes interlining together, you’re going to have significant bus bunching that prevents full utilization of the capacity.

      5. Those are nice ideas, Norman… but they bear no relationship to what the I-90 bridge was planned for when built, regional transportation planning since, or what voters have approved.

        That also has nothing to do with your original comment — claiming that ST’s light rail plan would decrease capacity on I-90 — which Carl and Martin have pointed out is blatantly incorrect.

      6. The buses are never going to stop on the I-90 bridge, no, that is for sure. They will exit I-90 to get to stops, as they do now. They will not stop right on the highway, just as buses do not stop right in a lane of I-5. They use exits to get to stops.

        Train headways over I-90 are constrained by the fact that they have to share the downtown tunnel with central Link, which will always limit the number of trains which could cross the I-90 floating bridge.

        How do you get “the dominant operating cost is operators”? On one thread here, a bus driver wrote that overtime pay for Metro buses who have the highest seniority is $40 per hour. If overtime is 1.5 times regular pay, that would put bus drivers at about $27 per hour. That is not such a large par of the $130 per hour or so it costs to operate an ST express bus. In any case, that $130 (or so) per hour in operating cost for ST express buses includes all the costs of the drivers.

        And operators are a tiny fraction of the cost of Link light rail operations. Even is you use $40 per hour as the cost of each operator, divided by 2 cars per train, that is $20 per car, or divided by 4 ccars per train, that is $10 per car. Even if you double that, it would be only about $20 per car for 4-car light rail trains. Since ST says that is costs about $370 per hour per car to operate 2-car trains, if $40 per hour of that is operators salary and benefits (totaling $80 per hour, which is way more than it probably is), changing to 4-car trains cuts that in half, to $20 per hour per car. So the cost of operating 4-car trains would still be about $350 per hour per car, compared to around $130 per hour for a bus.

        Unless you have some figures for how much operators cost that is far different from my scenario above, I think that operators cost is a very small factor in comparing operating cost of light rail to buses.

        What figures would you use?

      7. Well for starters $27 an hour completely ignores the cost of benefits, which are something like the wages themselves. The $130 figure is what Metro charges ST when it adds in a share of overhead; Metro itself uses about $100/hr for its costs.

        I don’t really have the time right now to look up Link operating costs, but I imagine that’s the total operating cost, plus Metro overhead, divided by the number of service hours in the system. It certainly doesn’t seem to make sense that salary+electricity+insurance+the marginal cost of maintenance would come out to $370 an hour per car. And again, you have Metro tagging overhead on it. Obviously that will scale much better as the system grows, and is nearly irrelevant to a discussion of adding incremental service.

      8. I take back the $100 figure. Thinking further I’m pretty sure the figure is $125.

    3. The thread above is OK, but let’s keep rail-BRT comments in the context of the article.

      We will not have every thread hijacked into bus vs. rail. See our comment policy at right if you have any questions.

    4. Jarrett writes on:

      Obviously, none of this is an argument against rail rapid transit, in Seattle or elsewhere. Rail can provide higher capacity per unit of operating cost, and where the alternative is non-electric buses it has an obvious emissions advantage. But it does mean that in Seattle, rail is about increasing the efficiency with which we carry people who already have a motive to use transit, and also building the capacity that will be needed if and when congestion pricing becomes realistic.

      The main point of his article is not “bus vs rail”, it’s about natural barriers allowing transit to be competitive with cars.

    5. No, it doesn’t. Jarrett’s point about chokepoints is exactly why fixed transit has an advantage, because rail can bypass topographical chokepoints (and geographical, albeit at a higher capital investment). You can’t drive a bus over a wooded mountain forest, but you can bore a rail tunnel under it.

      1. Here is what he actually wrote:

        “This is part of why various forms of Bus Rapid Transit have particular potential in Seattle: if you give transit an advantage through the chokepoint, you can achieve a lot of mode shift. The bus services across Lake Washington (between Seattle and its eastern suburbs) on I-90 do well because they have preferential access through a major chokepoint. East of the lake, they spread out to serve many suburbs directly, something buses do easily and rail does only with a required transfer.”

        He said that buses can “easily” “spread out to serve many suburbs directly”, which rail “does only with a required transfer.”

        This is a distinct advantage of buses over fixed transit across Lake Washington.

      2. Except no one is advocating for trains to “spread out”, the whole point is serve (or create) the very highest demand corridors with rail.

        To connect low-density points, everyone agrees we should use buses separated from traffic. For the high demand corridors, as we’ve pointed out to you, you can only run about 12 buses an hour due to spacing limitations, and that’s simply not enough.

      3. 12 buses per hour? Where did that come from? Right now, in peak afternoon hours, they are running about 50 buses per hour in each direction over the I-90 bridge.

        I’m really curious if you actually think they could operate only 12 buses per hour over the I-90 bridge, when they are actually running four times that many right now, and that is nowhere near capacity. Or is that some sort of typo?

        I think they are running somewhere around 40 – 50 buses per hour in the downtown tunnel right now in peak hours, even sharing it with light rail.

      4. Most I-90 routes run only in the peak direction with a few bi-directional peak-only routes. Only the 550 and 554 provide service in both directions all day (total 8 buses per hour midday beginning Feb 2010). I used OneBusAway to get the combined schedules to find the following figures. The peak hours are assumed to be between 7 and 8 am and between 5 and 6 pm. All services are assumed to run in the Express Lanes when possible, which means WB in the AM and EB in the PM.

        Westbound at the Rainier/I-90 stop:

        AM Peak 42 buses/h
        PM Peak 12 buses/h

        Eastbound at the Rainier/I-90 stop:

        AM Peak 10 buses/h
        PM Peak 26 buses/h

        The only cross-lake I-90 buses that don’t serve the Rainier stop are:

        Rt 202 S Mercer Island-Downtown, 2 buses (AM WB), 2 buses (PM EB)
        Rt 205 UW-First Hill-Mercer Island, 1 bus per hour only in the peak direction

        For the DSTT at University St Station:

        SB AM 52 buses/h and 8 trains/h, PM 44 buses/h and 8 trains/h
        NB AM 53 buses/h and 8 trains/h, PM 51 buses/h and 8 trains/h

        Under joint operations the plan is 60 buses per hour per direction and 10 trains/h/direction maximum.

      5. Oran, thanks for that information. I’m surprised, and not exactly convinced it’s accurate, about the apparent low number of buses in the “off-direction” (can’t think of the correct term) in the peak hours over the I-90 bridge, but I think you have about 45 buses per hour over the I-90 bridge as the heaviest travel.

        A year or two ago, I stood on the view point above the west end of the bridge (right above the east end of the tunnel on the west side of Lake Washington) and counted buses crossing the I-90 bridge for about an hour. It was either from 4:30 to 5:30 or 5:30 to 6:30, I don’t remember exactly. But that day there were around 50 buses in that hour heading east over the I-90 bridge in late afternoon. So, I don’t think your pm east-bound numbers are correct if you only count about 30 per hour.

        But, that bridge could easily accomodate hundreds of buses per hour in each direction. I took some 3-minute videos of the traffic that day, and in one of them there are 5 east-bound buses entering the bridge at the west end in those 3 minutes. Along with a lot of other vehicles. so, 5 buses every 3 minutes would give you 100 buses per hour. That is just a fraction of what that bridge could handle.

        I’m going to see if there is any video on the internet of traffic over the I-90 bridge, because when you see a video of actual traffic on the I-90 bridge, it is absolutely obvious to anyone watching it that that bridge could handle hundreds of buses per hour each direction with no trouble at all.

      6. Make that 28+3 buses for PM EB. I counted again. You can count it yourself if you’re still not convinced. OneBusAway makes it easy by condensing all the schedules for a stop in a single page.

        It is not that I-90 cannot handle a hundred or two hundred buses per hour, it very well can. It’s that downtown surface streets and the tunnel cannot handle that volume of bus traffic, especially when they all originate from a single roadway (I-90 or I-5). Downtown becomes the chokepoint. We all see how chaotic rush hour can be downtown with buses weaving in and out of stops on 2nd, 3rd, and 4th. I do not disagree with the principle of having service branching out but it has its limitations.

        This is a similar problem to what’s happening in Ottawa, well known for its extensive busway network, which has its busways feed in to bus lanes on downtown surface streets. They made the choice to replace that with a light rail line and downtown tunnel with the busway feeding into the LRT.

      7. Norman all the writers at STB take a lot of pride in being accurate and factual. When you say you aren’t “convinced” you are wasting our time because we actually do spend the time to dig up the facts. From now on please do the same and your comments will be meet with less skepticism.

        You have completely highjacked this thread. It won’t happen again.

      8. all the writers at STB take a lot of pride in being accurate and factual

        Here’s a great example of a poster being browbeaten – and threatened with moderation – for posting factual information (negating your claim of “12 buses per hour”; analysis of the Human Transit article that HT ITSELF backs up in today’s response that simply disagrees with a premise posted by one of the writers.

        Norman has not “hijacked” this thread – and the very idea is ludicrous. He’s aptly challenged some laid out assumptions, and done it well, respectfully, and thoughtfully.

        This type of admonishment is what demonstrates a clear bias in what at times feels like a mis-named blog (should be Seattle RAIL ONLY AND DON’T YOU DARE QUESTION THAT Blog), and the at times thin-skin and heavy hand exercised against folks who dare raise legitimate questions about underlying assumptions.

        Shame on you.

      9. Adam,

        Note as well that ORAN HIMSELF acknowledged that the actual (his) number of buses in the off-peak direction was “28+3” – more than twice the original number claimed. It appears Norman was correct in not being convinced by the original claim – and right to say (again, respectfully) that he was skeptical of the claim.

        Your comment to him was completely off-base, and you owe the man an apology.

      10. Jeff, Norman said in response to Oran…

        “I’m surprised, and not exactly convinced it’s accurate, about the apparent low number of buses in the “off-direction”

        I might have misunderstood what Norman wrote but from my understanding he was dismissing the information that Oran spent time researching.

        The things that we all care about, improving transit in Seattle, are not served by indulging in bloodbaths like this.

      11. Jeff,

        Spare us your “shame”. The idea that this is a pro-rail echo chamber, sanitized of dissent, is disproved by your continued (and welcome) posting here.

        It should be possible to have a discussion about something other than bus vs. rail; that’s why we have a comment policy. In spite of my earlier threat, the thread has stayed pretty close to the chokepoint issue, so I’ve let things slide.

        He didn’t “negate” my 12 buses/hour claim; as I indicated, I chose the wrong word in haste.

      12. The 12 buses per hour number is about the maximum you can have serving a single route without bunching. You can combine multiple bus routes over a segment so there are more than 12 per hour passing a single point but due to the need to stop you end up with the buses all bunched up anyway rather than shorter headways. Furthermore any particular stop is going to have an upper limit to the number of buses it can effectively serve. For example the Rainier freeway station can only handle so many buses before it becomes a source of delay for all of the routes serving it.

        To really make the I-90 center lanes similar to LRT service it would need to be turned into a two-way transit way with and the non-bus users moved to the outer roadway HOV lanes. In addition some form of direct connection between the I-90 transit way and/or HOV lanes and the I-405 HOV lanes would be needed. This is a real capital expense and it isn’t going to be cheap.

        To look at it another way Link has a capacity of about 9600 people per hour per direction across I-90 (4 car trains limited to 5 minute headways due to DSTT capacity issues and loading issues on the bridge). This is roughly 120 60′ hybrid coaches at crush load (which BTW you won’t get on every bus in the real world).

        As Oran points out above you have the problem of what to do with all of those buses when they get to downtown and you have a similar problem on the other side of the lake since a large portion of those buses are going to need to stop at places like the Mercer Island P&R, S. Bellevue P&R, Bellevue TC, Eastgate P&R, Issaquah P&R, etc.

        BTW even before there were trains in the DSTT there was an upper limit to the number of buses per hour per direction the DSTT could handle. I don’t remember what the actual number was but I’m willing to bet it was around 100 buses per direction per hour which means the current Link capacity through the DSTT with shared operations is the same as the DSTT had when it was bus-only.

      13. For east-bound buses over the I-90 bridge between 5 and 6 pm, I get 43, which is somewhat lower than the 50 I counted a year or two ago.

        The routes I counted are: Metro 202, 211, 212, 215, 216, 217, 225, 229, 210, 111, 114, 218, ST 550, ST 554.

        I don’t know for sure if this is all the routes which use the I-90 bridge, or not.

        When the downtown tunnel was closed for a year or so, a couple of times I counted buses on 3rd Ave downtown around Pike or Pine, and in the pm peak hour I counted 20 buses in each direction every 10 minutes. That comes to 120 buses per direction per hour. Of course, that is on 2 lanes in each direction, not a single lane. And, even with that many buses, there were times when there were no buses at all in that block in one direction, or the other. So, that many buses was not causing terrible delays.

        One 4-car Link train every 5 minutes gives 48 Link cars per hour. At the 140 riders per car that ST says it will limit ridership over the I-90 bridge to, that gives 6,720 passengers per hour per direction over the I-90 bridge. Of course, ST has no plans to actually do this. Their plan is for 4-car trains every NINE minutes by 2030 — twenty years from now! That is 3,733 riders per hour, or about 42 bus-loads per hour.

        But, if ST ever did run trains every 5 minutes over the I-90 bridge, the 6,720 riders per hour that gives is eqivalent to 75 articulated buses per hour at 90 riders per bus. That is only about 30 buses per hour more than cross the I-90 bridge now in the peak direction in the afternoon.

        Not all buses crossing the I-90 bridge have to go through downtown Seattle. Some could go directly to neighborhoods, such as First Hill-Capitol Hill.

        The idea that there would not be enough room to put 100 buses an hour or so at either end of the I-90 bridge is amusing to me, since there are around 6,000 vehicles per hour which cross that bridge in peak hours, and they all find someplace to go. A bus can go almost anywhere a car or truck can go. If 6,000 vehicles per hour find someplace to go at the west and east ends of the I-90 floating bridge, don’t you think there is room for 100 buses per hour? Especially since one bus can eliminate up to 90 cars.

        The DSTT can handle 125 buses per hour per direction, according to Metro. That is a lot more capacity with buses only, than the tunnel now has with 8 2-car trains and 50-60 buses per hour.

      14. Jeff,

        read the post again. It’s not “twice the original number claimed”. My original count for PM EB was 26+3 and the new count was 28+3, a difference of +2 buses. The other numbers were unchanged.

        AM peak travel tends to have more intense service because everyone’s going to work while PM peak travel tends to be more spread out (people doing other errands instead of going straight home), hence less service per hour in the PM peak but a wider spread.

      15. Oran,

        Still not seeing the words “I’m not convinced” translating into a thread hijacking.

        My error though as it was Martin’s number (12) that was being questioned.

      16. Martin,

        It’s not MY shame at issue here, but what should be a correction/retraction for accusing Norman of “hijacking the thread” and told “it will not happen again” simply for using the words “I’m not convinced”.

        If you guys can’t handle even that polite, three-word ‘challenge’ to your facts and figures (or anecdotal reporting, or whatever), I believe I’m on-target as questioning just how welcome any type of contructive debate or sharing of information/perspectives you’re really willing to allow.

        Are you seriously agreeing that Norman’s post including “I’m not convinced” deserved to be accused of “hijacking the thread” and threatened with moderation? I’d like some clarity on this, as I found this instance unfounded, and validating of my ongoing observation that you host-types are so ardently pro-rail that you tend to overreact and even drop the – what was it – “ban hammer” as Chris called it – on people just for either disagreeing with a pro-rail argument, challenging pro-rail assumptions or heaven-forbid (as in this case) challenging the accuracy of an anti-bus comment.

      17. Adam,

        I might have misunderstood what Norman wrote but from my understanding he was dismissing the information that Oran spent time researching.

        He used three words: “I’m not convinced”. Not that he did not use the words “You’re full of crap” or “That’s a lie” – but said that he was not convinced.

        The things that we all care about, improving transit in Seattle, are not served by indulging in bloodbaths like this.

        “I’m not convinced” = “bloodbath”? For crying out loud – are you serious? Now THERE’S three words you can take as a “dismissal”.

        Get a grip. Norman is owed an apology for being accused of “hijacking the thread” and for being threatened with moderation – merely because he DARED to say “I”m not convinced”.

      18. Norman,

        According to Alaskan Way Viaduct planning documents, downtown already handles 600 buses in the N/S direction during peak hour. The limitation is not street capacity, it is station/stop capacity. All those cars are going to some off-street parking spot somewhere, buses stop on the curbside. There’s limited curb space and bus dwell times usually exceed train dwell times. The practical capacity of downtown streets with varying amounts of transit priority range between 150 (bus lane)-200 (Portland Transit Mall) v/h/d. More traffic and the reliability and capacity is reduced.

        The tunnel was never used to its full capacity, 70 vhd max pre-closure. And bus-only capacity is not more than future rail-only capacity with 4-car trains and 30 tph (120 vhd). How much service they actually plan to provide is a matter of meeting demand.

        I believe you are citing the ITR/DMJM+Harris Report. The Metro figure of 125 vhd is correct but that report makes unrealistic assumptions and many errors, especially regarding bus loading and operations, summarized in this technical analysis by People for Modern Transit.

        Buses could go directly to First Hill, as they are now for a total of only 3 trips/hour during peak. Metro is not planning to expand that service. So don’t expect that to ever happen.

        Sound Transit plans to run that amount of service because, that’s the expected demand. Comparing expected demand to theoretical capacity to prove buses are better is meaningless.

      19. Jeff,

        How many of Norman’s comments have been moderated in this post? Zero. Three people can moderate — Ben, John, and me, and we’ve seen fit not to. What anyone else says is their personal opinion. But as usual you’re painting the STB staff with a broad brush (“you guys”) to feed your conspiracy theories that we’re stifling descent. There’s no need to enter moral panic about censorship everytime someone in a comment thread calls for moderation.

        I have more of a mind to delete this tangent about our comment policy. In fact, if you have any more complaints about our comments policy either email us privately or wait for an open thread.

      20. Hmm, interesting quote from the link Oran cites above:

        Capacity calculations violated Metro’s operational standards by having standing passengers on express buses.

        If this is still true it rather lowers the capacity for buses across I-90 (OTOH the 41, 71, 72, 73, and 74 are all regularly crush loaded at AM and PM peak and the 71, 72, and 73 are at crush loads entering/leaving downtown well after the end of both peak periods.). BTW a DE60LF in Metro/ST configuration has seats for 56 passengers.

      21. Norman, your arguments make some sense until you use numbers that seem to be obviously misrepresented. For example, you persist at using a high bus crush load (90) vs. a low Link standing load of 140. You also persist at comparing general traffic flow to buses without any consideration of longer vehicle length, required following distances, and increased loading/unloading times for fewer and narrower doors.

        (And by the way, you forgot the 214 in your list of routes)

      22. 90 is nowhere near “crush load” for an articulated bus. I have been on buses with 120 or more people on them leaving Husky Stadium after a football game. “Crush load” for a bus is probably over 120.

        90 passengers on an articluated bus is accepted as “capacity.” Buses are expected to have standing passengers in the aisles — that is why they have straps hanging overhead, for standing passengers to hold on to. 90 passengers on an articulated bus is 1.5 standers for every seat.
        140 passengers on a Link train is 1.9 standing passengers for every seat.

        Why would people think that normal capacity on a 90-foot Link train is 200, while on a 60-foot bus it is only 60? Does this make sense? You think that a Link train is about 1.5 times as long as a bus, but it has 3.3 times the capacity? Really? If so, please explain that. As I already wrote, buses are expected to, and often have, passengers standing in the aisles.

        Buses in this area carry 90 people at some points on some routes on a regular basis. This is not some theory, it actually happens regularly.

        There is no city in the U.S. where light rail carries over 135 riders per car at the peak point in the route at the peak run in the peak hour. It doesn’t happen, from everything I have read on the subject.

        If you can find any study where any light rail system in the U.S. carries anything close to 200 riders per car past any given point during the peak of the peak, please send me the link to that article. Otherwise, I will assume this is just a theory based on what the manufacturer of the vehicle claims, and not based on any actual experience.

        There is plenty of unused capacity on Seattle streets for more buses. This was proven when the bus tunnel was completely closed. If necessary, they could even eliminate parking on some streets to create even more capacity.

        So, by 2030, ST is planning to have about the same capacity on the I-90 bridge, as there is currently in the peak direction, just on the current buses using it! But, with buses instead of light rail on the center span, you could have hundreds of carpools and van pools using those lanes at the same time as the buses. And you could easily have two or three times the number of buses per hour that you have now.

        Using the center span for light rail is really a waste of capacity on the I-90 floating bridge.

      23. oops. On buses 90 is 1.5 passengers for every seat — not 1.5 “standers.” Likewise, 140 on a Link car is 1.9 passengers for every seat — not 1.9 “standing passengers”.

      24. Buses, carpools and vanpools will get their HOV lane in both directions, unlike the current one-way express lanes.

        And you could easily have two or three times the number of buses per hour that you have now.

        And is there any plan from any agency to run that amount of buses in the future? No

        Any realistic, workable plan from proponents? No

        We can debate all day about which mode theoretically can carry more people but the fact remains, the people chose light rail as the regional high capacity transit mode for our busiest corridors.

        If necessary, they could even eliminate parking on some streets to create even more capacity.

        Parking on 3rd, 2nd, 4th, 5th Avenues & Stewart St are already prohibited during peak hours. 1st Ave has peak hour restrictions, too.

        Here’s an interesting read on the modal capacity published at a Transportation Research Board LRT conference:

        From the abstract:

        Rail modes attract greater utilized capacity per unit of offered capacity during peak period than bus modes, and this aspect of consumer choice may be quantified by regression analysis. Data and observations fail to support alternative hypotheses to a consistent and observable consumer preference for rail.

      25. I’m getting 200 from ST’s web site for Link standee load. Quote and use a source for a rated standee load on an articulated coach and I’ll be happy.

      26. Community Transit’s Swift BRT page gives “Seating for 43 passengers and overall capacity up to 80 riders.” And that’s a spacious bus with three doors, less seats and better circulation. Metro and Sound Transit’s buses have 13 more seats but a lot less standing room, narrower aisles, and only 2 doors.

      27. Here is one link you can take a look at:

        I believe these specs are right from the various manufacturers. Starting on page 17:

        Nabi 60 LFW — seats 62; standees 31; total 93

        New Flyer DE60LF — seats 62; standees 53; total 115

        NABI 60 BRT — seats 62; standees 30; total 92

        New Flyer DE60LF-BFT — seats 47-53; standees 53; total 100-106

        Van Hool AG300 — seats 45; standees 57; total 102

        APTS Phileas 60 — seats 29; standees 111; total 140

        Irisbus CIVIS — seats 27; standees 90; total 117

      28. So, can you explain why a 90-foot Link car would have so much more capacity than a 60-foot bus in your opinion? The Link cars are about 50% bigger than articulated buses, and I believe they have a little more than 50% more capacity than articulated buses: buses around 90; Link cars around 140. That makes sense. 200 for Link cars vs 60 to 80 for buses does not make sense.

        Can anyone cite a study which shows any light rail system in the U.S. which carries more than 135 riders per car past the peak point in the peak run of the peak hour? Or, is that just a theory that this could happen somewhere sometime in the future?

      29. Norman,

        On capacity, three words: no wheel wells. The width of the rail car vs. bus also allows more standees and internal mobility while in operation.

        On the # of riders per car during off-peak operation, I’m guessing that responses will tag that concept as largely irrelevant in people-moving comparisons, as no bus, rail car or any transit means remains at peak capacity during any and all hours of operation, and there are other arguments for transit options overriding simple capacity issues – such as service area. The short and sweet there is that if you only want buses/trains to run at maximum capacity most of the time – the simple answer is to eliminate 80% of all service anywhere and only focus on central locales during peak hours.

      30. If you prefer New Flyer’s own website, here is the link:

        Capacity for 60-foot buses: seats up to 62; standees up to 53.

        There are indeed wheel wells in Link cars. In the middle section of Link cars, the seats face each other with only a narrow aisle between them because the backs of those seats are about a foot from the sides of the train walls because of the wheel wells in the center section. That means if those center seats are filled, there is no room for any standees in the middle section of a Link car, because there is only enough room in the aisle there for the legs and feet of the people sitting in those seats.

        A Link car is only about 2.5 inches wider than an articulated bus.

      31. Norman,

        There’s the passenger/vehicle (car) and the normalized passenger/meter vehicle length. * = estimate

        The “passengers per meter of vehicle length” standard for peak-period vehicle occupancy, which places all systems and modes on an equal footing, was suggested by Parkinson and Fisher (1996).

        Boston Green Line 111 pass/veh 5.1 p/m
        Calgary 1994 150 pass/veh 5.5 p/m
        Denver 1994 150 pass/veh 6.2 p/m*
        LA Green Line 2001 139 p/v 5.1 p/m
        Portland 1990 134 p/v 5 p/m
        St Louis 1996 140 p/v 5 p/m*

        Seattle Bus * = assumes 50% artics
        Convention Place 2000 NB PM 54 p/v 3.0 p/m
        I-90 1990 37 p/v 2.4 p/m*
        SR 520 1990 AM 56 p/v 3.7 p/m*

        Median values
        HOV and Busway 2.9 pass/m
        LRT 4.1 pass/m

        Now, lets say you could run an equivalent or better capacity with buses. Would they attract the same amount of ridership and therefore utilization of available capacity? The people who compiled this list and done analysis (see TRB published paper I linked to above) suggested they would not. This phenomenon is called “rail bias”. With the latest BRT implementations copying rail to take advantage of the “rail bias”, this may change.

        The median for both heavy rail and light rail — outside the five most crowded and congested urban centers — is 40 percent higher than the median for HOV, busway, and freeway-express bus services. The rail median is higher than the rapid-bus maximum (3.7, excluding Ottawa). It should be noted that the average railcar is not 40 percent wider than the average bus. (A series of case studies by Tennyson (1989) found that rail would attract 34-43 percent greater ridership given equivalent service conditions.)

        Even if it did work, I find it aesthetically unpleasing to run hundreds of diesel-powered buses as opposed to a few electrically powered light rail cars or trolleybuses. For the I-90 case, trolleybuses can’t be used due to speed limitations.

        Look at these two videos of Ottawa’s busway system: (bus intersection edge of downtown) (Downtown Ottawa bus loading with conga line of diesel buses)

      32. Norman, I believe your figures for the maximum seated and maximum standees for various 60′ buses aren’t for the same seating configuration but rather represent seating configurations optimized for maximum seats or maximum standees.

        For example take the DE60LF, in the seating configuration used by Metro there are 56 seats. Now based on daily experience riding the bus these things tend to get really full at about 80 passengers or so. I have a hard time believing you can get 53 standees on one in any sort of real world conditions.

        There really is only one way to normalize capacity between a Link car and a DE60LF. First calculate the amount of standee area for the DE60LF using the same methodology as Sound Transit uses for Link vehicles. Second calculate the number of standees using the same area per passenger Sound Transit uses for various loading conditions on Link vehicles.

        I’m willing to bet using this methodology that each Link vehicle has somewhere between 2 and 2.5 times the capacity of a DE60LF.

        In any case it takes anywhere from 8 to 10 buses to equal the capacity of 1 4 car link train. For service with 5 minute headways this translates to anywhere from 96 to 120 buses to equal the capacity of one train.

        Of course this again will be a theoretical capacity and not one based on what will be seen in real-world conditions (such as the study Oran cites above). The short answer is people are much more willing to stand on a 30 minute light rail ride than they would be willing to on buses.

        Since the buses will be serving different destinations with different demand profiles you’ll probably need more than 96-120 buses per hour to serve the same transit markets with the same frequencies as light rail. Likely on the order of at least 144-180 buses per hour.

        Again you have the issue of limits on the number of buses that can reliably serve each major destination or freeway station and the issue of how many buses can reliably go between the I-90 and I-405 or I-405 and SR 520 HOV lanes per hour, especially in peak traffic conditions.

      33. The documents I linked to are manufacturers’ claims about capacity. I think 90 passengers for an articulated bus is a valid real-world number, which I have seen used in several studies. And I have seen around 90 passengers on articulated buses in the Seattle area on a regular basis on some routes in peak hours. Articulated buses certainly can carry more than 90 passengers, but they don’t during regular commutes.

        At the same time, even Sound Transit uses between 137 and 148 as the capacity of Link light rail cars during normal commutes. ST uses 200 as the capacity of Link cars for special events, such as after football or baseball games, where people don’t want to wait for the next train, and will accept being crammed into trains like cattle for one trip home. They will not accept this on a daily commute.

        So, I am convinced that the real-world, accepted capacities during normal commutes is around 90 for articluated buses, and about 140 for Link light rail cars.

        I have seen 90 passengers on buses in the Seattle area during normal commutes. I have not seen more than about 75 passengers on a Link light rail car during normal commutes, so far. The highest capacity Portland Max has been able to attain is about 135 passengers per light rail car, and that is on only one route.

        Therefore, it takes about 6 buses to equal the capacity of 4 light rail cars.

      34. I finally found a document which I have been searching for since this thread started, and, with my luck, probably nobody I have been discussing this will visit this thread after I post this. Here is a link to a document by People for Modern Transit, in which they trash the bus tunnel study by John Niles. This document claims that, if you use 2 standees per square meter for standing room in both Link light rail cars, and Metro articulated buses, you come up with a capacity of 90 for buses, and 137 for light rail cars. So, at least now you can seem one of the documents from which I got those numbers — I did not just make them up.$FILE/DMJMReport-PMTAnalysis.doc

        QUOTE: Report’s bus capacity calculations contain unrealistic assumptions and significant errors. The stated seating capacity of 65 is close to that of a Breda articulated, high-floor dual-mode tunnel coach (currently used in the DSTT) that also has the same stated overall dimensions. We will base our analysis using a low-floor version of that coach (proposed for the new tunnel fleet) although the low-floor design in itself causes the loss of 6 to 8 seats. Total available standing area instead of the 145 sq. ft. stated is actually 125 sq. ft. Using a figure of 5.1 sq. ft. per person comparable to Link light rail design load factor, the bus capacity would be 65 + 125/5 = 90 passengers. The figure of 25 standees is comparable to what is actually observed in packed buses and compares well to the Breda specification of 20 standees for the existing tunnel buses with step-wells. For passengers with large packs and carry-bags plus packages, 25 standees is already an uncomfortable density.

        Light rail vehicle capacity is underestimated
        The Report bases its light rail vehicle capacity on a generic high-floor light rail vehicle. Sound Transit has been basing their capacity calculations on a generic low-floor vehicle similar to the Siemens/Duewag low-floor car now in service in Portland. Their actual capacity, using a figure of 5.1 sq. ft. (0.5 sq. m.) per person, is 137 passengers (74 seated, 63 standing).


      35. I already linked to that report above at 2010-01-07 13:06:04 in this very same thread.

        You can see where the 200 passengers figure comes from. 137 (74 seated+63 standing) passenger capacity assumes 2 standees/sq.meter (0.5m^2/prs) and 200 passenger (74 seated+126 standing) capacity assumes 4 standees/sq.meter (0.25m^2/prs).

        My transit textbook (Vuchic) says that 4 prs/m^2 (0.25 m^2/prs) is used as the standard “in most cities in industrialized countries” (read: Europe) and the conditions for standees are “extensive body contacts, difficult movements”. The Rail Capacity Manual says that 0.3 m^2/prs is “a reasonable service load with occasional body contact; moving to and from doorways requires some effort” and that 0.2 m^2/prs is “an uncomfortable near crush load for North Americans”.

        Sound Transit describes 150 as “comfortable capacity” and 200 as “crowded capacity” which matches the capacity manual’s description of the various densities.

        It is therefore reasonable to state that the capacity of a light rail vehicle is between 74-200 passengers, depending on what load factor you want to use.

      36. “I have not seen more than about 75 passengers on a Link light rail car during normal commutes, so far.”

        Really? I’ve been on Link light rail cars during normal commutes where all 74 seats are filled with at least 20 people standing.

      37. TriMet also lists the capacity of the low-floor Siemens Type 2/3 LRVs mentioned in the report as 166 at a loading condition of 4 prs/m^2. Siemens also lists their capacity as between 150-180 passengers.

        According to Kinki-Sharyo the Link cars have around 30 m^2 of usable floor space. That is room for between 60 and 120 standees depending on loading conditions. Based on this and the capacity ratings of the Siemens LRVs I think it’s safe to assume that the capacity of the larger Kinki-Sharyo LRVs used on Link is somewhere between 150 and 200 passengers.

        Anecdotally, I read on the APTA website that during the Houston Super Bowl MetroRail was experiencing passenger loads of 250 passengers per car on their Siemens Avanto S70 which is a comparable LRV to the Kinki-Sharyos used on Link. The Siemens Avanto has a capacity of 220-236 according the Siemens website.

      38. If you use the same 2 standees/sq meter for buses and for light rail cars, you get a capacity of 90 for buses, and 137 for light rail cars. These are the numbers that are used in one document after another. I am not interested in what people are willing to accept in Europe. What has that got to do with Seattle? Studies show that in the U.S., outside of New York City, people will not tolerate that sort of crowding. In the U.S., the accepted capacity used is 2 standees per square meter. I have read studies that in S. America, they get up to 270 people on one 80-foot bus. Does that have any relevance to Seattle? Is it ok if I use 3/4 of 270, or 202, as the capacity of a 60-foot bus in Seattle, just because that might be possible in S. America?

        So, now we all know that when you use the same “density” for buses as for light rail, a light rail car will carry just about exactly 1.5 times as many people as a 60-foot bus. Why do you think that People for Modern Transit used 90 for buses and 137 for light rail? These are people who obviously really favor light rail over buses.

        As far as I am concerned, these are the only reasonable numbers to use: 60-foot bus capacity 90; LRT capacity 137. You have not made any kind of argument to use any other numbers. Especially when you consider that ST itself uses 137 as the capacity of their light rail cars in their EIS’s.

        Sorry, but you don’t get to use 4 standees per square meter for light rail and 2 standees per square meter for buses. That is apples to oranges.

        Please tell me where and when you saw Link cars with 94 people on them during normal commutes. Thanks. I have seen LInk cars with people standing, but there have always been at least as many unoccupied seats are there were standees. Are you sure all the seats were filled with people (not luggage in some)? Or did you just assume that they were?

      39. That was before Airport Link opened and it was SB at International District station around 4 pm and yes, I scanned and looked at every seat.

        It is reasonable to use 137 for a system wide average but such capacities are not absolute numbers, they depend on the circumstances.

        2 standees per square meter is the average peak-hour according to TCRP Report 13, but “typical maximum standing loads in North America range between four and five passengers per square meter” and “a generally accepted average for short-distance sustainable peak loading is 4 passengers per square meter (2.6 sq ft per passenger), this may be reduced for longer distance trips, or where service policy or local conditions dictate otherwise.”

        It is reasonable to assume such conditions (200 passengers/car) can happen between Westlake and the UW station where it’s only a 6 minute subway ride. Similar conditions already occur on buses that take at least twice as long. Do you agree?

        Since you cite and accept the Link car capacity from the EIS, do you accept the statement that “The East Link Project has the capacity to comfortably carry 600 persons per 4-car train and 800 persons with crowded conditions” according to the East Link EIS?

      40. Norman, I think your numbers are off for the DE60LF buses used by Metro and Sound Transit on I-90. The seated capacity is 56 rather than the larger seated capacity for the Bredas and D60 high floor buses. While in theory this should be offset by a higher standee capacity that doesn’t seem to be the case in actual practice. With 25 standees a DE60LF is rather crowded, it certainly seems more crowded than the description of 2 passengers per square meter. Certainly 25 standees is about the point where the 41, 71, 72, 73, and 74 will leave passengers at the stop. Either because the passengers decide to wait for the next bus or the driver refuses to load them.

        We’re also not discussing what various load factors do to stop dwell times. Even in the RFA when buses do all-door unloading and loading, dwell times degrade quickly with even a light standee load. Due to differences in interior layout I’d say stop dwell times don’t degrade nearly as quickly at a given standee density as they do for buses.

      41. I have ridden south bound through the downtown tunnel on Link many times between noon and 6pm, and have never seen more than about 75 people in any car I was in.

        Standees per square meter is the same for buses as for light rail. If you can have more than 2 standees/square meter for short distances on light rail, you can have more than 2 standees/square mile for short distances on buses. Just use the same for both. As I have given you several examples of, the official documents I have read, when discussing “capacity” of light rail and buses. use 137 for light rail, and 90 for buses. You can try all the rationalization you want, but when I read the ST SEIS’s, an SDOT document, a People for Modern Transit document, a PSRC document, a Light Rail Now web article, and other organizations all using 90 for the capacity of articulated buses (when they mention bus capacity), and 137 for the capacity of light rail cars, that is what I am going to use. Why do all these organizations use those figures if they are not the accepted figures? The answer is that they are the accepted capacity figures.

        When you discuss the capacity of a light rail line, or a bus route, past a given point, using passengers per direction per hour, you use 137 for a light rail car, and 90 for a bus. Period. End of story.

        If you think that 150 is “comfortable” for light rail cars, and 200 is “crowded” with light rail cars, then the comparable numbers for buses is 100 for “comfortable” on a bus, and 133 for “crowded” on a bus. You can use those numbers if you want, although they are not the accepted figures. The key is that an articulated bus has 2/3 the capacity of a light rail car. As long as you use that ratio, then you can do what you want. But the ratio is a light rail car has 1.52 times the capacity of an articulated bus. I will use the accepted numbers of 90 for a bus and 137 for a light rail car.

        In general, I take everything ST says with a grain of salt, unless it is confirmed by other sources.

        When comparing buses to light rail, I think we should all start talking about the new brt-style buses which are being used on SWIFT and will be used on rapidride, which have 3 large doors per bus. Supposedly that allows for shorter dwell times. Whether that is actually the case, or not, we should get to see pretty soon. Anyone have any numbers for the SWIFT route so far? How are the dwell times working out? Are the dwell times shorter on SWIFT than on routes using “regular” articluated buses? I have meant to go ride SWIFT and check it out, but have not done so yet.

      42. Oran,

        [b]the people chose light rail as the regional high capacity transit mode for our busiest corridors.[/b]

        I disagree. Or if you like – I’m not convinced.

        The people did not “choose light rail as THE regional high-capacity transit mode”, they chose it as *A* regional high-capacity transit mode. It’s not as if they had a ballot measure that said “Check “X” if you want rail; check “Y” if you want BRT”.

        In fact, “the people” have chosen light-rail AND BRT as the regional high-capacity transit mode(s), as both have had their own ballot funding measures placed before the voters.

        I’m sure you’ve done your research and all, but to state unequivocably that “the people have chosen rail” (over other options) is inaccurate hyperbole. That “rail uber alles” thing I keep mentioning.

      43. Fine, Jeff, nitpick on semantics if you like.

        It’s not hyperbole, the people did choose rail for the I-5 and I-90 corridor long term. If they didn’t want it, then they would’ve rejected it. I left buses out because it was so obvious that they are already serving a regional function.

      44. Oran,

        Wasn’t nitpicking – and you’re backtracking. The Human Transit analysis favord BRT as the more efficient solution to address the “chokepoint” issue – and even posted a follow up in direct response to some of the discussion here to clarify its own point in that regard. You appear to have dismissed their analysis, saying that “We can debate all day about which mode theoretically can carry more people but the fact remains, the people chose light rail as the regional high capacity transit mode for our busiest corridors”.

        As I read your comment – you are saying that the HT analysis doesn’t matter – even if accurate, because “the people chose light rale” as THE solution – i.e. over and above BRT. I’m finding your claim that you left out buses because they’re already serving a regional function to be a bit cofusing – because you *didn’t* leave out buses in your original comment – you outright said it didn’t matter whether or how they compared to LR with regard to capacity: “We can debate all day about which mode theoretically can carry more people but. ..”

        If I’m mis-reading your comment, my apologies, but it sure sounds as if you were making a case for rail OVER BRT because in your view “the people chose” it as “THE regional high capacity transit” solution.

      45. Jeff,

        I haven’t dismissed HT’s analysis. In fact I completely agree with it. HT does favor BRT but it does in no way dismiss LRT or claim that BRT is a substitute for LRT like what Norman is trying to do.

        Lifting a few quotes from the follow up HT piece:

        “As often happens, discussion quickly turned to my references to rail and Bus Rapid Transit, as readers argued over whether my real agenda was to advance one of those modes.”

        “As regular readers will know, it’s rarely that simple. But chokepoints do point to an advantage for Bus Rapid Transit if you’re trying to do things cheaply.”

        “you’d use rail only where you need rail’s higher capacity. You’d also be likely to use rail in corridors where the demand doesn’t spread out over a large area, the way it does in this example north of the Ship Canal.”

        Norman disagrees that rail has a higher capacity over buses. And Jarrett doesn’t suggest using buses as the one-for-all solution.

        “Best-practice transit agencies have short-term improvement programs running concurrently with their thinking about long-term major projects.”

        So you start running buses now and as demand grows you can justify building rail later.

      46. I said that Link light rail has lower capacity over the I-90 floating bridge than that bridge could have without light rail, and using buses on the center span, instead. Places where they can operate 8-car trains every two minutes would be hard to match with buses, but ST can not do that over the I-90 bridge, and everyone here knows that.

        The capacity that ST can put over the I-90 floating bridge on light rail is quite small, and can easily be exceeded with buses, van pools and car pools over that bridge instead of light rail.

      47. Norman,

        In 2005, Sound Transit hired consultants to evaluate potential options for HCT in the Eastside/I-90 corridor, including HOV/BRT, Busway/BRT, LRT, monorail and rail-convertable BRT. It found that LRT would have 60% more ridership across I-90 than the cheaper bus alternatives and 33% more across I-90 than the similarly priced rail-convertable BRT. The LRT/HOV hybrid will have 40% more cross-lake ridership than the other bus only alternatives.

      48. Oray, that’s pages of PDF links and each one is pages of “stuff”. 60% more ridership? At what cost? Because there’s 20% less lanes available to buses? Unless ST makes the assumptions and modeling metrics available to the public their “hired consultants” numbers mean nothing more than Joe Sixpacks guesstimates. In fact I’d say less since they were hired to produce a result. Keep in mind that one of their pet hired guns is Parsons Brinckerhoff who make millions off of the project (even if it doesn’t get built).

      49. Bernie,

        this is the full report

        and the updated summary

        On ridership models, it doesn’t say much:

        Based on the scenario definition (service plan, station location, etc.), ridership forecasts were prepared. All scenarios made use of the same population and employment forecasts contained in the Sound Transit EMME/2 ridership forecasting model. This is the same model that was used for the Long-Range Plan, North Link and Central Link ridership forecasts and has been reviewed by the FTA for reasonableness for project level planning. … Operational issues such as forced transfers or bus operations through congested streets are embedded in the ridership estimation model.

        I wouldn’t be so quick to discredit their numbers and claim bias. Yeah, I’m being defensive here. I’d say any hired consultant could have bias in favor of their employer. Even so-called “independent” studies may be influenced by the person or group who commissioned them depending on their agenda. As someone who’s going into this industry, this casts doubt on whether my work is going to be meaningful or even ethical. If not, then I should change my career path.

      50. Not going to make a “quick claim” and that’s for the additional info/links. It will take time to look them over.

        A hired consultant does have a bias… Well at least any published result will reflect that bias. Yeah, “so called” independent studies. Going into this field your work is most certainly meaningful. Ethics are up to you and it’s not a yes/no black/white decision. NOBODY hires an consultant with the expectation of an unbias result. It’s like our legal system. You’re hired to represent the interests of your employer. There are many legitimate interests and opinions on both sides of any issue. You may land a job with an employer that wants you to portray the best of the position you adhere to. Then again, you may find yourself working for an organization that fundamentally you disagree with. Ethically you just need to step above a forgone conclusion and just present facts. If the person that hires you is only looking for facts to support their conclusion it doesn’t mean that supplying those “facts” are unethical any more than a one sided presentation of “facts” supporting the alternative view are unethical.

      51. You can’t drive a bus over a wooded mountain forest, but you can bore a rail tunnel under it.

        You can bore a bus tunnel under it, too, but in either case, it would be a major budget buster.

  5. “This article makes a good argument for bus raped transit in Seattle, instead of light rail.”

    I’m not sure the prospect of bus rape will have a positive impact on ridership. ;-)

  6. Over the ship canal to the north: there are 34 lanes of traffic! At optimum capacity, thats 74,800 vehicles per hour, or roughly 90,000 people per hour (excluding buses). And 2 commuter rail lines, with an existing capacity of 2,160pph. And some day, two lanes of Link at 38,400pph! If we had 4, imagine the capacity with EXPRESS tracks! (This all, of course, doesn’t take into consideration direction or traffic.)

    Dugg for this:
    “Nowhere in Seattle can you travel in a straight line for more than a few miles without going into the water or over a cliff.”

      1. I really didn’t want to dig though 3 different transit agencies numbers. It’s back-of-napkin anyways.

    1. I’m not sure where two-track commuter rail has a capacity of 2,160 ppdph. The usual figure for mainline or grade-separated rail is closer to 40,000, with 30 trains per hour. A few lines even surpass 50,000.

      You don’t need four tracks to run express trains. Unless you think you’re actually going to need 30 tph soon, a two-track line with bypass sections in the middle should be enough. The most constrained space would have two tracks either way for construction cost reasons, unless for some reason you needed enough capacity for 60 tph.

  7. 1. Interesting article. Makes sense, and while never directly tied to transit, I’ve always thought that Seattle has benefited greatly from it’s geography.

    2. Another thread derailed by a Norman antirail tangent.

    What. A. Surprise.

Comments are closed.