Typically, I ride the bus during rush hour, when it’s so crowded that it’s sometimes hard to nab a seat. I’ve done a quick passenger count a number of times, and the bus almost always has at least 35 passengers. Sometimes, especially on the big articulated buses on a busy afternoon, it’s more than 50.
But apparently it’s pretty unusual for a transit bus to be so full. According to nationwide statistics (see table 2.10, here), transit buses carry fewer than 9 passengers, on average. That average includes both the busy routes and the sparsely-ridden ones; apparently there are lots of buses that have just a few riders, especially in spacious suburbs or during late-night runs.
The average transit bus gets a little around 3.5 miles per gallon in regular service. When I do the math, I find that buses average about 30 passenger-miles per gallon of fuel, or thereabouts.
Curiously, that makes transit buses—at least, at current levels of occupancy—just about the least efficient form of personal travel around!
Find this article interesting? Support more research like this with a gift!
Take a look at this graph, derived from the US Department of Energy’s Transportation Energy Data Book.
Intercity buses—which are typically pretty full—are just about the most efficient form of transportation in use today, as measured by energy consumption per mile of travel. But transit buses—which are typically pretty empty—are the least efficient.
The only difference is the number of vacant seats.
The lesson here is not—I repeat, not — that transit buses are inherently inefficient. It’s that all too often they’re underutilized. A reasonably full bus can get something like 120 passenger-miles per gallon. But a reasonably empty bus is less efficient than a Hummer. It all depends on how many seats are filled.
The same goes for just about any other form of transportation: the easiest way to ramp up efficiency is to fill the empty seats. Vanpools are particularly efficient because they average 6 riders per van; most of the seats are full. Just so, we could substantially increase the net efficiency of car travel by ramping up carpooling back to what it was in the early 1970s. On net, a full, 4-passenger car that gets 30 mpg does about as well, in terms of fuel consumed per passenger mile, as a full 40-passenger bus that gets 3 mpg.
And note another curiosity: rail transit, and transit buses, cars, and even SUVs are all pretty darn close in terms of the energy that’s currently used to transport a passenger one mile, at current levels of ridership. (SUVs & other light trucks do a bit better than would have thought, because their average ridership is higher than for cars. Go figure.)
But what’s true for an individual vehicle might not be true for a transportation system. For example, places with high rates of transit ridership tend to consume less energy overall for transportation. But that’s not really because of an inherent efficiency of any particular transportation mode (buses vs. trains vs. cars). Rather, transit-intensive cities also tend to be dense places—since it’s in denser urban areas where transit can attract enough riders to be convenient and cost efficient. And dense cities tend to be places where a) the seats on the buses and trains are likely to be filled, and b) people don’t have to go very far to get to work, stores, and services.
In other words, it’s the land use patterns that make a transportation system efficient or inefficient; to a large degree, transportation mode is an afterthought.
Matt the Engineer
Combining this data with the Free Ride Zone story, it seems to me that a demand-based system is needed. Run a lot of busses during commute hours, few in off hours, and none in hours when few ride the bus. At those times make people call a taxi.Sure there’s an issue with finding drivers that will work only a few hours a day, but that’s not an unsolvable problem. Perhaps a city-run taxi service, running hybrid cars?
There is also a huge equity issue created by not running buses during low demand periods. Low income people (who often work jobs outside of ordinary 8-5 business hours) make up a higher proportion of riders during off-peak hours and are least able to afford a taxi.If the political will could be found to subsidize the taxi service enough to make taking an off-peak taxi cost similarly to riding the bus, then this objection is removed, but I have a hard time seeing that happen.This discussion is somewhat academic, though, as I ride the bus semi-regularly during off-peak hours and sometimes fairly late at night, and even with those rides included, I literally can’t remember the last time I was on a metro bus that was less than half-full for most of the run. Maybe the routes that serve areas outside Seattle are less full, but bus service inside the city is well utilized at all hours of the day and night.
Yeah, I agree: there is a big equity issue with transit; cutting service during off-peak hours would improve the fuel efficiency of the system, but at the cost of denying basic transportation services to people who aren’t able to drive. Of course, the buses do only a so-so job of providing transportation services to people who can’t drive, or can’t afford to; but I’m not sure that there’s a more cost effective solution. Maybe there is, but I haven’t seen an analysis of other options.
Do you know whether those figures account for energy used in manufacturing or are operational energy alone?
Matt the Engineer
Perhaps the equity issue can be solved by downsizing busses instead of removing them. During off-peak hours, assuming an area where ridership is low, have drivers switch to small fuel efficient busses. Maybe they can’t get up to 30 mpg, but they can certainly beat 3.
it’s the land use patterns that make a transportation system efficient or inefficient; to a large degree, transportation mode is an afterthought. Absolutely. IIRC, KingCo Metro uses 7 DU/ac as a minimum to deliver service that is efficient. For example: the 914 line to Enumclaw must average 1 person past Auburn (except at peak) & thus is subsidized transit. Buckley’s 406 line is the same, with a limited schedule due to low ridership. DS
It’s worth remembering that these figures are the US average as a whole. If Seattle has higher average ridership than the country, then its transport system is more efficient than the chart would suggest.
To Steve – just transportation energy, not manufacturing.Matt – nice idea!Dan & Eldan—nice points, as always.
I’ve thought that Metro could (a) simplify their fare structure and (b) encourage more off-peak riders by charging a peak fare all the time, but allowing ‘two-for-the-price-of-one’ during off-peak times. This would also provide more opportunities to introduce new bus riders to how the system works.Regarding taxis…part of the problem is that we don’t encourage ridesharing in our taxi system. That is, one person per destination. Why not get multiple people in a taxi going to similar destinations and have them split the fare? Of course, we’d have to ensure that the taxi companies have an incentive to play along.I’ve also read claims that construction of mass transit—such as the Beacon Hill tunnel—consumes huge amounts of power…so huge, in fact, that the energy consumption breakeven for this tunnel comes 70 years after opening…eek! (Of course, it’s mostly a sunk cost now…but in retrospect, couldn’t they have run the mainline around the north end of Beacon Hill, and then build a surface spur streetcar line down Beacon Ave S?)
Perhaps the equity issue can be solved by downsizing busses instead of removing them. During off-peak hours, assuming an area where ridership is low, have drivers switch to small fuel efficient busses.Metro already does this on some routes. The 222 in Bellevue typically has one of 3 different sized buses. For peak runs they use a full sized bus. Mid-day runs use a shortened version which has fewer seats and no rear door. In the evenings and on weekends they switch to a “mini-bus” which seats probably no more than 12 people.
To some extent, nationwide statistics are irrelevant when considering this issue for specific transit systems. For example, I live in Washington, DC, in a neighborhood served by two of the city’s busiest bus lines. Combined, these two lines have a ridership that exceeds the Baltimore Light Rail line by at least 25%. (And probably the DC region’s 6-7 top bus lines by ridership exceed the light rail ridership of Portland.)Most of the in-city buses I ride have decent ridership at a variety of times during the day. Some are articulated buses, some regular size, and some small. Depending on the route and demand.Similarly, the buses I rode in Portland when there last fall for the National Trust conference (6, 14?–Nob Hill?, and others) had decent ridership—given my experience on bus routes with close to 20,000 daily riders, and my empirical bent, I have an ability to “read” ridership.The point is you lop together high ridershp systems with low ridership systems when you look at national statistics.The point should be to assess high ridership systems and figure out what conditions are favorable to high usage and how to duplicate these conditions in your own situation.I think one of the major problems is that we need to move transit planning towards what I call the “mobility shed” or “transit shed” and address planning and moving away from SOV trips towards multiple modes.Unfortunately, the computer I am on at the moment doesn’t load my blog too well, so I can’t pull off cites for entries where I’ve discussed the transit shed concept.
“Fill those seats!” is a Very good point. Another way to achieve this is by improving dial-a-ride service. Currently, Metro’s DART program is not very popular beyond senior citicens due to limitations such 2 hours advance reservation. Such systems work very well in German suburbs, especially where they can be ordered directly from the connecting bus.