The Secret Worldwide Transit Cabal
Friday, May 16, 2003
Wendell Cox Is Not Mediocre, But Sometimes, He's Dense -- 9 and 10
Home of More Transit Links Than You can Possibly Check(tm), Unless you have no life other than websurfing
"Truth passes through three phases: 1) It is ridiculed. 2) It is violently opposed. 3) It is accepted as self-evident." Albert Schopenhouer. In the United States, rail is currently passing through Phase Two.
From the Cabalmaster:
Continuing with our sophisticated, erudite analysis of urban transit traffic density, comparing and contrasting with the weak and lame “analysis” attempted by, well, you-know-who.
(More shameless self-promotion in the mold of the good Mr. Cox.)
FOR THOSE OF YOU TIRED OF THIS SUBJECT, THIS IS OUR LAST POST ON "COX IS DENSE."
The fact that “boarding density” and “traffic density” are not equivalent may be illustrated as follows:
---Tokyo’s Teito Rapid Transit Authority (“Eidan” -- Urban HRT) reports twice as many boardings per route-km as the Odakyu Electric Railway (Suburban). However, traffic densities are almost identical, because the average Odakyu passenger travels twice as far as the average Eidan passenger.
---The Paris Metro (Urban HRT) reports more than five times as many boardings per route-km as Japan’s Nishi-Nippon Railroad Co (Suburban), but traffic densities are virtually identical. The average Nishi-Nippon passenger travels much farther than the average Paris Metro passenger.
---The Lyon Metro reports nearly three times as many boardings per route-km as Washington, DC, Metrorail (and the Hamburg S-Bahn). However, traffic densities are identical, for the average passenger travels much farther in Washington (and Hamburg) than in Lyon.
(We’re eagerly awaiting Wendell’s flavorful (i.e. fudge-laden) exposition of why rail transit is not justified in Hamburg and Lyon . . .)
In terms of traffic density, the world’s three busiest urban heavy rail systems -- Hong Kong (150 million pass-mi per route-mi per year), Moscow (137) and Sao Paulo (120) stand out -- as outliers. Anyone who normalizes an index on an “outlier” (www.publicpurpose.com/ut-wrail.htm) risks revealing themselves as “statistically challenged” -- or, in this case, just plain dense (. . . given the topic at hand . . .).
The highest traffic density carried by ANY urban rail system in the U.S., Canada, Australia or Western Europe are found in Lisbon (40) and Vienna (39). The urban heavy-rail average for these countries is 14.4; BART scores an even 14, Washington, DC scores 12; Philadelphia and Boston score 10.
The light-rail average for the U.S., Canada, Australia and Western Europe is 4 million pass-km per route-km per year; Los Angeles scores 5.4, San Diego scores 5.2, and St. Louis scores 4.7.
It should not surprise that traffic density levels in the developed world -- with the single exception of Japan -- are significantly lower than in other countries. High traffic densities are associated with peak-hour crowding levels that most passengers in the developed world will not tolerate. The traffic densities carried by Russian streetcar systems are (or were; traffic has plummeted since the dissolution of the USSR) associated with very low levels of auto ownership and very high levels of peak-hour crowding -- sometimes to the point where doors could not be closed.
Another issue, which tends to inflate annual traffic-density statistics for Japan (and probably for other countries): weekday and Saturday traffic levels do not differ much, and Sunday traffic levels throughout much of the day resemble those carried on weekdays.
A “Public Purpose” page has a table, “Metro and Light Rail: Selected International Systems Ranked by Passenger Kilometers per Route Kilometer” (see: www.publicpurpose.com/ut-intl-pmrt.htm). This betrays the compiler’s total lack of understanding of the concept, and of its significance. This table is also a typical example of the poor presentation which we Opinionated Ones find typical of “The Public Purpose” annd “Demographia.”
The table headings read as follows: “Urban Area,” “Type of System,” “Passenger Miles (KMs) per Route Mile (KM),” “Year.”
Now, is that “Passenger Miles (KMs) per Route Mile (KM)” per second, minute, hour, day, or ?
The “Public Purpose” figures APPEAR to be “per day” or “per weekday,” but this cannot be determined from the table.
The “Public Purpose” table also contains this note: “Estimated daily one directional ridership (two way ridership would be double).”
Traffic density is not “ridership,” although it is of course related to same. It is possble for two systems with very different ridership levels to have equivalent traffic density -- and for equivalent traffic-density statistics to mask substantial differences in boarding counts. In addition, we do not see the point in calculating “one-way” traffic densities . . . unless, of course, one wishes to “lowball” this statistic.
The Secret Worldwide Transit Cabal has this advice for our friend, the good Mr. Fudge:
Don’t be dense, Wendell! Learn the difference between “traffic density” and “boarding density!”
Wendell Cox Is Not Mediocre, But Sometimes, He's Dense -- 10
We Opinionated Ones hope that we’ve managed a reasonable job so far of enlightening you about a serious transit planning issue, in reasonably-accessible fashion, all while engaging in our favorite sport, which is ridiculing Wendell. (He makes such an inviting target . . .)
Transit planners in other countries pay careful attention to traffic density (passengers traveling over each mile of route). But, this country, for reasons of simplicity, politics, convenience, politics, established practice, politics, and so forth, boarding density (passengers entering or boarding along each mile of route) is the standard yardstick.
We Opinionated Ones admit that “boarding density” is a useful measure for ridership forecasting purposes – call it “passengers attracted per mile of route.” However, estimating “how many” while failing to take into account “how far” – that is, the average distance traveled by each passsenger – may lead to serious design errors that will cost a small fortune to fix. Case in point: the Blue Line in Los Angeles. “Big Blue” was built for a maximum train length of two cars – and escaped a crippling car shortage during the initial years of operation only because a second batch, ordered for the Green Line, was on hand. Peak-hour trains quickly filled up. Then, after platforms were extended, three-car trains began operation early last year – and these are now full during peak hours, and some off-peak hours, too. We’re opinionated, but we think this line should have been built for four-car trains.
(Oops, how foolish of us! Train length doesn’t matter because “no amount of ridership” would make the Blue Line successful!! The esteemed Dr. Peter Gordon of the University of Southern California said so!!! Right here on our blog!!!!)
Scroll down to August 2, 2002 posts. If you don't want to do that, here is the item, reproduced:
From the Cabalmaster:
"Whoever you are, you do not have the courage to identify yourself. I do not know what to make of this but it does not look good. In any case, the world would be better off without the Blue Line. I said it again."
rest snipped for brevity...
NumberMan, he CONFIRMS he said what he was reputed to have said about the Blue Line. Let's run with it, man!
By now, we anticipate that at least one websurfer is now mumbling: If those smarta . . . er, “TransitCabalists” . . . are so smart, then how come they can’t explain the forecasting error in ten words or less?
Well, we can: Planners used a ridiculously low number for average travel distance.
(There! Exactly ten words!!)
So ridiculously low, in fact, that an FOC laughed out loud upon finding printed documentation in an early-planning document.
As a result, traffic density – which exists even if U.S. planners ignore the issue – was vastly underestimated.
The original forecast was about 55,000 passengers per weekday. Planners assumed that the average passenger would travel no farther than bus passengers do – about four miles – even though the rail line would provide much faster service.
So, 55,000 passengers per weekday, times four miles per passenger, divided by the 21-mile line length, gives about 10,500 passenger-miles per mile of route per weekday, or about 3 million per year.
Three million per year . . . say, isn’t that the same as light-rail traffic densities in Gothenburg and The Hague?
(Yes, Wendell, it is, and we’d just LOVE to hear you explain why these two cities should, well, “stop building rail systems, dammit!”)
The “problem” (at the good Dr. Gordon might see it) is that the Blue Line attracts a much longer average travel distance than four miles. An FOC reports that onboard survey data showed an average travel distance of about 8 miles during the early 1990s, exceeding 10 miles during peak periods.
The current “annnual average” is about 60,000 per weekday. Multiply by eight miles, divide by the 21-mile line length, and get almost 23,000 passengers traveling over each mile of route per weekday, or nearly 7 million per year. In other words, more than double the forecast “workload.”
(Hey, Wendell! Did you know that the L.A. Blue Line carries more traffic per mile of route than heavy-rail systems in Oslo, Norway, and Newcastle, England?
(The same traffic per mile of route as light-rail systems in Frankfurt and Stuttgart, Germany?
(The same traffic per of route as the Berlin S-Bahn?
(Almost as much per mile of route as Chicago’s heavy-rail system??
(We’re not surprised. We think that you and your cohorts don’t have the slightest idea of what this is all about . . . but we suspect you’ll learn, the hard way, as others catch on . . .)
A significant issue for planners is what level of traffic density justifies investment in rail transit. Pinpointing a threshold figure is difficult. This question has never been studied systematically in this country (as it has elsewhere), and there are site-specific issues that influence the answer. However, the U.S. “threshold” for low-cost light rail, without tunneling or extensive grade separation, is about 5,000 pass-mi per mile of route per weekday, or about 1.5 million per year. This represents the minimum traffic-density level at which low-cost light rail becomes economic, paying back the cost of construction through increased transit operating efficiency and social benefits (reduced travel time, cost savings over auto travel, and so forth).
If significant tunneling or grade separation is required, the threshold is about 10,000 pass-mi per mile of route per weekday, or about 3 million per year.
On the other hand, the practical limit or “ceiling” for U.S. light rail is about 20,000 pass-mi per mile of route per weekday, or about 6 million per year. This density level implies rush-hour volumes that are beyond the effective capacity of most light rail systems (owing to limitations on train length, service frequency, and the fact that U.S. consumers will not tolerate the sort of “crush” loadings that Tokyo residents stoically endure). So, in the U.S., traffic densities significantly greater than these figurees suggest the need for heavy rail.
The L.A. Blue Line is the proverbial exception that proves the rule. Although this line carries 60,000 – 70,000 per weekday, busiest-hour passenger volumes are not much greater than 3,000 per hour. This is a remarkably low figure in light of the weekday ridership. It may reflect characteristics of the corridor, or may arise simply because of limitations on peak-period service (three-car trains) and lack of tolerance among passengers for Tokyo-style crush loads. We don’t believe that anyone really knows what would happen if this line could operate, say, six-car trains.
Wendell Cox may be incurably dense with reference to transit traffic density (after all, he’s Wendell).
But, seriously, we Opionated Ones hope that others will give this subject the careful study it deserves.
Light rail in Los Angeles carrying the same traffic density as light rail in Stuttgart, Germany, and almost as much as heavy rail in Chicago!??!
Sounds too good to pass up, doesn’t it?
Comments: Post a Comment