The phrase,'Unsound Transit', was coined by the Wall Street Journal to describe Seattle where,"Light Rail Madness eats billions that could otherwise be devoted to truly efficient transportation technologies." The Puget Sound's traffic congestion is a growing cancer on the region's prosperity. This website, captures news and expert opinion about ways to address the crisis. This is not a blog, but a knowledge base, which collects the best articles and presents them in a searchable format. My goal is to arm residents with knowledge so they can champion fact-based, rather than emotional, solutions.


Tuesday, January 25, 2005

The Case for Running BRT in HOV lanes in the Puget Sound

Earlier comment letter on the Sound Transit Long Range Plan:

January 25, 2005

Steve Kennedy
Senior Environmental Planner

Sound Transit

401 South Jackson

Seattle, Washington 98104


Sound Transit Regional Transit Long-Range Plan

Draft Supplemental Environmental Impact Statement

December 2004

Dear Mr. Kennedy:

This Draft SEIS is supplemental to the Regional Transit System Plan, Final Environmental Impact Statement, March 1993. My comments on that earlier EIS [1] primarily concerned congestion pricing and the capacity of bus vs rail systems. In the ten plus years since that EIS several significant developments have occurred. The first is that pricing has come into its own, with several projects, both domestic and foreign having been built, and many more in planning [2]. The second is that what is now called Bus Rapid Transit (BRT), systems of which have been in place in South America for decades, has belatedly been recognized for having capacity similar to heavy-rail. [3] The third is that "inside out" commuting has developed in this region to such a level that it is equal to traditional, "outside in" commuting.

My remarks here are primarily addressed to the confluence of these developments, a confluence not recognized in the DSEIS. Specifically, to meet the reality of these developments, a plan for running BRT on High Occupancy / Toll (HOT) lanes should be developed as a formal third alternative within the Final SEIS. If it is, I believe a fair appraisal [4] will show that such a BRT/HOT Alternative has far more people-moving potential than light rail (LRT), can be built at a fraction of its cost, and better serves the evolving transportation needs of the region. As it stands the DSEIS is essentially a advocacy for extension of the Link light rail system, with the alternatives, such as they are, being little more than "straw men."

Transportation systems are inherently tied into land use patterns. And any document such as the DSEIS, while naturally replete with facts and figures, almost of necessity expresses a philosophy or world view connecting the two. In this case, the embedded philosophy is one with which I am in substantial agreement, namely Transit Oriented Development (walkable, mixed-use neighborhoods and activity centers connected by quality, high capacity transit). [5]

Advantages of the BRT/HOT Alternative

The question is, is light rail the best technology for making connections between these neighborhoods and activity centers? While that is currently the "conventional wisdom" among transportation aficionados in this region, the newer BRT/HOT Alternative has some major advantages [6]: #9;

1) It is cheaper per mile, and thus the whole 125 mile HCT network can easily be built within the identified 25 year time frame. With the LRT system, due to budget constraints, only a small portion of the network could be built within this period. [7]

2) In its turn this means that far more neighborhoods, communities, and activity centers can be connected by HCT, better meeting the goal of creating Transit Oriented Development (TOD).

3) It has far more capacity than the LRT system, particularly with the limitations imposed by the confluence of a) at-grade travel in the Rainier Valley, b) the merge of an at-grade (south) and grade-separated (east) line, and c) the relatively short stations in downtown Seattle.

4) It better serves the region’s evolving "inside-out" commuting pattern.

5) It is faster than the light rail system, particularly where LRT is at-grade.

1. Cheaper per mile. The DSEIS envisions LRT connecting Everett, Seattle, Tacoma, Bellevue and beyond. But because it doesn’t attach a price tag to this vision, this is more a wish list than an objective. If cost no object, then why not heavy rail as envisioned in the 1993 EIS, indeed why not maglev? The price tag for the downtown Seattle to Northgate section, which is assumed under all alternatives in the DSEIS hasn’t even been arrived at yet. None the less we can make what is admittedly a guess, albeit an informed one: $30 - $40 Billion for the full system. [8] The corresponding figure for a HOT lane network is approximately $2 billion. [9]

This latter figure doesn’t include things such as converting the reversible lanes on I-5 to HOT lanes (a WSDOT rather than a Sound Transit responsibility) or building a new bus tunnel in downtown Seattle (on which more later). But it would be sufficient to provide the entire metropolitan Puget Sound region with a BRT system, most of it running at highway speeds (50 - 60 mph) even during the height of rush hours. Of course these are "back of the envelope" figures, but note that the full BRT system is grade-separated while important links in the LRT counterpoint (eg along Martin Luther King Jr. Way) are at-grade. Despite this qualitative difference in favor of BRT, for a similar-length system LRT is fifteen times (15 x) more expensive than BRT. That order-of magnitude differential leaves a lot of room for back-of-the-envelope error, and strongly suggests that a third, BRT/HOT Alternative should be fully vetted in the Final SEIS.

2. Better meets the goals of Transit Oriented Development. The DSEIS says that the Plan Alternative would "result in less dependence on automobile travel and greater transit- and pedestrian-friendly development activity." (4.10-10) This is tantamount to saying that it would cause Transit Oriented Development. The No Action Alternative, on the other hand, allegedly could result in "increased pressure for suburban sprawl" and "pressure to build highway improvements." (4.10-14) However, for tens of billions of dollars the Plan would have only the modest effects on automobile traffic. [10] And the more likely outcome is that because of its gargantuan expense and extraordinarily modest results, the pressure for auto-mobility, with associated "sprawl" would continue unabated. [11]
Thus the Plan Alternative could achieve the exact opposite of the result projected in the DSEIS. And while it’s not clear that any transit system can reduce automobile travel, if the goal is to connect walkable communities [12] with swift, quality transit, then the BRT/HOT Alternative is superior, simply due to the fact that it can easily be built within the 25 year DEIS time-frame [13], while the Plan Alternative is more likely a hundred-year project.

3. More capacity. That buses with off-line loading have far more capacity than rail was one of the two main points I made in my comments on the 1992 DEIS [14], and the past decade has just made this point stronger. The maximum seated [15] throughput for the Link LRT system north of downtown Seattle, according the DSEIS is 7,400 seats per hour per direction. [16] The line capacity of a bus lane with off-line loading [17] (or of a HOT lane that has become a bus lane due to extraordinary transit demand) is approximately 44,000 seats per hour per direction, about six times the LRT figure. [18]

This is not to say that downtown Seattle, to take the most salient example, can handle this level of bus through-put absent major upgrades to transit facilities. It cannot. But then, according the DSEIS it doesn’t need to. What it needs to accommodate is, at most, 7,400 seats per hour. And this buses can do rather easily. [19] I will develop this conclusion under my consideration of the I-90 HCT alternatives.

4. Inside-out commuting. Commuting in the region is currently as much "inside-out" (Seattle to the Eastside) as it is "outside-in." A claimed benefit of the Plan Alternative is that it will handle, traditional, outside-in commutes better because rail passengers to downtown Seattle business district wouldn’t have to transfer on the destination side of their trip (although, due to inherent limitations of rail, they would normally have to transfer on the origination side). As for inside-out commuting, however, since the Eastside has less concentrated employment centers, rubber-tired transit (buses, vanpools, carpools) are inherently more convenient than rail, since the latter will almost always require a destination-side transfer -- and in most cases require a origination-side transfer as well.

5. Faster. Because they always give absolute priority to transit, and because they are priced to keep traffic moving at highway speeds, HOT lanes move BRT buses at 50 to 60 mph 24/7. LRT moves at various speeds, estimated in the DSEIS to be an average of 25 to 30 mph. [20]

Obviously the above represents only the most cursory examination of a comparison of the BRT/HOT vs the LRT Alternatives. The Final SEIS should examine these honestly [21] and in depth.

I now very briefly examine the application of the BRT/HOT Alternative as it relates to a couple of specific corridors referenced in the DSEIS.

BRT/HOT on I-405

The "I-405 Congestion Relief and BRT Project" report calls for BRT on HOV or future HOT lanes (Appendix O of the DSEIS, p 19). Given the above discussion, as well as the fact that there is plenty of room within this corridor for off-line bus stops, BRT/HOT is the obvious HCT choice for this corridor. [22] The only significant issue is HOV to HOV (which would become HOT to HOT) interchanges on I-405. But regardless of the HCT option chosen, these interchanges need to be built in any event. And since they are the responsibility of WSDOT (and thus are exterior to Sound Transit’s budget), this is really a non-issue. Sound Transit does have the responsibility to build the BRT stations and the direct-access ramps, but these are relatively minor items compared to constructing a whole new right of way for trains (or, for that matter, for buses)

BRT/HOT on I-90

The I-90 HCT corridor is a strong, if largely implicit focus of the DSEIS. The DSEIS contains a BRT option for this corridor, which is sketched out in Appendix N, having the following elements:

1. The I-90 center span becomes a busway.
2. At least three freeway BRT stops are on-line (in the parlance of the document, "in-line")
3. The central portion of the north-south BRT line is over a Burlington Northern railroad right of way (ROW).
4. Buses run in mixed-traffic city streets in downtown Seattle.

There are major problems with all these elements:

1. The I-90 center span becomes a busway. Converting the I-90 center span to a busway certainly creates huge transit capacity. [23] In fact the capacity is such that to equal the seated capacity of rail on the bridge there need to be only one bus traveling each direction on the bridge at a time. [24] Needless to say, if this were the only traffic on the center span of the bridge, that would be a huge waste of the capacity

Also, because the center span under this configuration would appear nearly empty (one bus traveling each direction at a time), this is probably not "politically sustainable," since commuters would be irate to see a nearly empty center span. However, with the center-span converted to HOT instead of bus-only lanes, center-span capacity would not to be wasted. To the extent that buses and vanpools didn’t use the available capacity, it would be made available to non-transit users, taking some pressure off the general-purpose lanes.

2. At least three on-line freeway BRT stops in the I-90/I-405 corridors are envisioned in the DSEIS. On-line stops hugely reduce the potential through-put of a BRT system. With such stops express buses have to operate in the same manner as trains: when a train is stopped in a station another train can’t pass it, and therefore sufficient "headways" must be built in to the system to avoid collisions. The result is that the capacity of rail systems are determined by the headways necessary to avoid collisions at stations. Such headways are typically no less than every few minutes. However, with off-line stops buses on HOT lanes can operate every few seconds. [25] I made this point in my comments on the 1992 DEIS [26]

3. The central portion of the north-south BRT line is over a Burlington Northern ROW. First, as mentioned above, if the HOV lanes on I-405 are converted to HOT lanes, they will have enormous transit capacity, and therefore the use of the BN ROW is unnecessary. Second, a railway ROW is not a good fit for a busway: it is too narrow, and of course the railroad trestles would have to be completely rebuilt. Ironically this ROW is one place in the region that may be a relatively good fit for light rail or DMU (diesel multiple units). After all, as in San Diego or Saint Louis, the tracks are already in place. [27]

4. Buses run in mixed-traffic city streets in downtown Seattle. The DSEIS contemplates the completion of the Sound Move program, which includes the annexation of the DSTT by trains, with buses being "turned out on the street." While such annexation is, in my opinion, misguided, [28] I will not discuss it further here. But even with this annexation, are there alternatives to buses fighting their way through downtown Seattle traffic? The answer, of course, is yes.

First, third Avenue could remain a bus-mall, either full-time or rush hours only, as it will be during conversion of the DSTT to rail use.

Second, the region has made the decision to tear down the Alaska Way Viaduct and replace it with a tunnel. If this is carried forward, Alaska Way will gain substantial, above-ground capacity, which could be used by buses.

Third, the cost of LRT in the I-90/Eastside corridor is enormous [29], and in order to make a valid comparison of alternatives, reasonably equal funds must be assumed for each. Otherwise one side is, metaphorically speaking, fighting with a spear, a dagger, a net, and a trident, while the other is unarmed: not exactly a fair fight. For a fraction of the funds dedicated to the Eastside/I-90 LRT alternative, Sound Transit could build a second transit tunnel in downtown Seattle, this one optimized for buses. [30]

Eventual Further Development of BRT/HOT Alternative

The BRT/HOT Alternative sketched out there is hardly the be all and end all. Further development of the principles embodied in this alternative could lead to congestion pricing of all freeway lanes as I advocated in my 1992 comments to the former EIS. [31] General Congestion Pricing, coupled with demand-responsive transit and instant carpooling, is the ultimate – albeit optional -- development of the principles underlying this third, BRT/HOT Alternative. [32]


However, I am not advocating general Congestion Pricing here. Rather I am advocating that Bus Rapid Transit running on a system of HOT lanes be fully and fairly developed as a formal, third Alternative in the Final Supplemental EIS. Anything less constitutes a breach of duty to fairly consider all reasonable alternatives, as well as a willful ignore-ance of the situation facing our region.

Thank you for consideration of these remarks.


Donald F Padelford



comment on 1992 DEIS
See Appendix A.


BRT projects
In-place projects include:
- SR 91 "Value" lanes in Orange County, California
- I-15 HOT lanes in San Diego, California

- Katy Freeway HOT Lanes in Houston, Texas

- London, England, cordon pricing
- Singapore, cordon pricing
Many pricing projects are in planning, including the Washington State SR167 HOT lane Pilot project


BRT through-put similar to heavy rail
According to FTA documentation the San Paulo and Bogotá BRT systems carry 30,000 + people per hour per direction (see Appendix D). This is comparable to most heavy rail systems (Appendix E), and is well above the through-put for any light rail line. The 1993 Regional Transit Project EIS estimated the capacity of its "rapid" (ie "heavy") rail Preferred Alternative at 22,400 people per hour per direction.

Note that the preferred alternative in the 1993 EIS was a totally grade-separated system, similar to BART or the DC Metro. That document examined and rejected at least two iterations of non-grade-separated "light" rail such as Sound Transit is now constructing along Martin Luther King Way and rejected those alternatives. See also the first paragraph of footnote 2, Appendix B.


fair appraisal

There can be no doubt that Sound Transit and its predecessor agencies have frequently been guilty of "stacking the deck" in favor of whatever rail alternative was currently in favor, indeed the deck is clearly stacked in the current DSEIS. To take a couple representative examples:

In order to justify the Sounder commuter rail line to Everett ST needed to show a time savings compared to competing bus service. It managed to do this by comparing the then-prospective rail line, not to then-existing express buses running between Everett and downtown Seattle but rather to a "milk run" that stopped every few blocks and required a transfer to boot.

When politics required a diversion of the preferred grade-separated rail system running down the Duwamish to a system running through the Rainier Valley (currently under construction, and contrary to the findings of the 1993 EIS, being built at-grade) this was accomplished by covertly undermining the Duwamish route (see footnote 2 in Appendix B).

In these and numerous other instances Sound Transit has acted in abject and flagrant bad faith.


TOD vs "sprawl"

The dominant form of development in post WWII America is characterized by

- Separation of work, shopping and housing

- Near-exclusive dependence on the automobile

- Abundant free parking.

- Spread-out development to accommodate the automobile.

- Dominance of the automobile over the pedestrian

Together these are generally characterized by the term "sprawl."

Transit Oriented Development (TOD) is simply a return to an earlier built-space form, one characterized by

- Mixing of work, shopping and housing in a single district or neighborhood

- Ability to move about conveniently via transit

- Parking that is subordinated to the "walkability" of a neighborhood
(although it may – or may not still be "abundant")

- More compact development along with characteristics, such as small blocks,
that increase "walkability."

- Dominance of the pedestrian over the automobile


major advantages

The advantage of the LRT system is that it has, at least in the eyes of its proponents more "cachet," often being described as being "world class," and thus might better address whatever civic insecurities the region labors under. A remark that has stayed with me was made at a Camber of Commerce conference in Vancouver BC by a Portland attorney, who said that that latter city’s light rail system had "make it ok" to use transit again. This is non quantifiable, but may be a real phenomenon. On the other hand, studies have conclusively shown that people are not so much concerned about what kind of wheels their "transit box" runs on as they are about convenience, safety, headways, and door to door travel time. Some Paris subway use rubber tires. There is no indication that if they left their tracks for the last mile or so of the journey to deliver people to their precise destination that people would see this as anything but a plus.


budget constraints

As I understand it the increase of costs of the Link light rail system vis a vis the Sound Move document is currently at 170%, ie it is 2.7 times the cost indicated in that document. At the time, figures in Sound Move were represented to the public as being "consciously conservative."

the full network

see Appendix F

and approximately $6 Billion (2.4 + 3.6) for SeaTac to Northgate

cost of a HOT network

The cost to complete the "core" HOV lane network is, according to WSDOT, approximately $1,650 million. Other costs are $290 million for associated HOV to HOV ramps, and $60 million for a bus-only Ravenna to Stewart south-bound lane on the I-5 express lanes. There is no public cost to convert HOV to HOT lanes, as that is born by paying, non-transit users. (See also Appendix B)


modest effects on traffic congestion.

According to Jim MacIsaac (12/9/04) (Appendix F) $28 Billion in rail capital costs and a probably additional $10 Billion in O&M (Operations and Maintenance) costs by Sound Transit through 2030 would serve to offset ONE years growth in traffic.


pressure for auto-mobility would continue

According to John Tierney of the New York Times, "highway congestion increased in Portland more than any other American city in the 15 years after the first light-rail line opened." ( Therefore the argument that rail decreases the rate of growth of automobile travel is at best unsubstantiated, and more likely is simply false.

Another important point that Mr. Tierney makes in this article is that pricing keeps the "traffic pipes" at their maximum dimension at all times. Normally during rush hours these "pipes" reduce flow to a trickle. Therefore, pricing is a powerful form of capacity enhancement – sort of freeway-onramp metering on steroids.


reduce congestion

See previous footnote in reference to Portland’s failure to reduce congestion.

Connect walkable communities

Note that there are various cities that have Transit Oriented Development with no trains, and others with trains but no TOD. Therefore rail is neither necessary nor sufficient for TOD.


easily built in the next 25 years

The cost of a 125 mile basic HOT and bus-only lane system is approximately $2 Billion (see footnote 9), ie less than the cost of Initial Segment of Link from Westlake to north of SeaTac Airport. (Note that Initial Segment is the starter system of Phase I of the Link system, which itself was characterized as a starter system within Sound Move. So Initial Segment is a starter system of a starter system.)


my comments on the 1992-3 EIS

See Appendix A


why seated

It is essential that comparisons be made on an "apples to apples basis." Seated capacity in each case is the easiest comparison to make, which is why I use it. It is also possible to compare seated plus standing capacity in each case. Note that in comparing light rail to buses in the DSTT, however, Sound Transit has been guilty of comparing a percentage of bus seats against a multiple of rail seats. (see footnote 6 to Appendix B)


7400 seats per hour

60 minutes per hour / one train every 2.4 minutes (page 4.9-26) = 25 trains per hour

25 trains per hour x 4 cars per train x 74 seats per car = 7400 seats per hour


off-line loading

As is the case with rail, buses with on-line unloading/loading must maintain sufficient headways to prevent a bus stopped at a station from being run into by another bus coming up behind. On-line loading for what we now call Bus Rapid Transit seems to have been a premise of the 1993 EIS (see Appendix C, top illustration of a "Transitway"). This premise also appears to have carried over to the current DSEIS, as Appendix N to that document envisions "in-line" loading for BRT several places in the I-90 / I-405 corridors.

According to Sound Transit, light rail could run on as 2.4 minute headways (see footnote 16) north of downtown Seattle. With on-line loading it is likely that a similar (or somewhat better) time between buses could be maintained. However with off-line loading the time between buses is reduced from minutes to seconds, and capacity is commensurately increased:

"Capacity of a route or of a group of routes on the same facility is almost always determined by conditions at stops areas rather than line conditions. ... When stops are made off the main line or artery, capacity is determined by the safe separation between transit units. Thus, on exclusive busways or bus lanes on freeways, with off-line bus stops, headways of 5 s[econds] can be achieved. Theoretically rail systems could operate at headways of perhaps 60 s[econds] under similar conditions, but such situations are not found in practice"

Observed bus capacities:
I-495, New York: 490 / hour;
Lincoln Tunnel, New York: 735 / hour
(1994 ITE Highway Capacity Manual, p. 193)

On-line loading: 60 minutes per hour / one bus per 2.4 minutes = 25 buses per hour
Off-line loading: 3600 seconds per hour / one bus per 5 seconds = 720 buses per hour


44,000 seats per hour

720 buses per hour x 61 seats per hybrid bus = 43,920 seats per hour per direction.

Cf light rail at 25 trains per hour x 4 cars per train x 74 seats per car = 7,400 seats per hour per direction.


buses could easily meet transit demand in the DSTT

125 buses per hour x 61 seats per bus = 7625 seats per hour per direction. 125 buses can be accommodated by the DSTT (see Appendix B, page 2 -- there I used the figure of 63 seats per bus, but my understanding is that the new, hybrid buses have 61 seats)


25-30 mph
page 4.9-25

Also: table 4.9-17 on page 4.9-48 of the DSEIS states that travel time from downtown to SeaTac airport with LRT will be 42 minutes, which is stated to be a 14% savings "from existing".

X - 14% = 42min

X = 49 min ("existing")

However the 194 bus makes it from downtown to SeaTac in 29 min at 5 pm, according to its schedule

examine honestly

See footnote 4.


obvious choice for this corridor

Much information about HOT lanes will be gleaned when the SR167 HOT pilot project is put in place. The BRT/HOT Alternative would build on and extend the 167 project methodology, having rail-quality off-line bus stops as well as dedicated off and on ramps, neither of which will be part of the SR167 HOT project.


huge transit capacity

720 buses per hour (see footnote 17) x 61 seats per bus = 43,920 seats per hour per direction


require one bus traveling each direction on the bridge at a time

According to Figure 4.9-10 (page 4.9-26) rail headways on the I-90 span could eventually be as little as 5 minutes.

60 min per hour / 5min per train = 12 trains per hour

12 trains per hour x 4 cars per train x 72 seats per car = 3456 seats per hour

To equal this, the number of buses required is:

3456 seats per hour / 61 seats per hybrid bus = 57 buses per hour per direction

3600 seconds per hour / 57 buses per hour = 63 seconds per bus

Assuming a HOT lane in free-flow, bus spacing at 60 mph becomes

60 miles per 3600 seconds / 1 bus per 63 seconds = 1.05 miles per bus

1 bus per mile

The length of the floating portion of the I-90 span is about a mile. So for buses to equal the seated capacity of Link on the bridge, as one bus is driving onto the bridge from Mercer Island another will be driving off of it towards Seattle. Alternately if one bus headed to the eastside is in the middle of the bridge, the eastbound portion of the bridge will otherwise be unoccupied by buses.


every few seconds

See footnote 17


made this point

Appendix A


maybe a good fit for LRT or DMU’s

On the other hand the best use of this right of way may by for a bike and pedestrian trail.


annexation of DSTT misguided

See discussion in Appendix B


cost of LRT in the I-90/Eastside corridor enormous

According to one knowledgeable individual "at least" $5Billion including reserves and finance cost through 2016


this one optimized for buses

The existing DSTT cost approximately $600 Million. If we double that cost to $1.2 Billion this is still a small fraction of the Eastside Link cost (see previous footnote).

A tunnel optimized for buses would have two passing lanes in the stations (instead of one under the current DSTT), allowing "skip-stops," with the consequent at-least doubling of through-put. (See also footnote 13 in Appendix B)


as I advocated in my 1992 comments

See Appendix A


ultimate development

Others have also called for the pricing of all roads at all times. Because the overwhelming number of roads are uncongested, this proposal, which I do not support, is primarily a revenue-raising mechanism meant to replace the gas tax, rather than a congestion-regulation and capacity-enhancement mechanism.

Appendix A

Donald F. Padelford

320 Marketplace One

2001 Western Avenue

Seattle, Washington 98121-2114

fax: 441-9546

November 23, 1992

Gregory M. Bush

Municipality of Metropolitan Seattle

Environmental Compliance Division

821 2nd Avenue


Seattle, WA 98104-1598

re: Draft Environmental Impact Statement
Regional Transit System Plan, October 1992

Dear Mr. Bush:

I have been, and remain, an advocate of Congestion Pricing. My remarks are directed first at the Draft Environmental Impact Statement's (DEIS) analysis of this concept and then to other aspects of the DEIS.

It is stated in section 2.6.9 that Congestion Pricing "could not by itself meet the objectives of the System Plan". I take the word objectives to mean goals as delineated beginning on page x of the DEIS, the most important of which is the first, ensure the ability to move around the region. If this assumption is correct, then the statement of non-attainment is either false, meaningless, or grossly misleading.

To see that the statement is false or meaningless, ask yourself the following question, "Will there be enough parking in downtown Seattle in the year 2020?" There certainly won't be enough free parking to meet the demand, as there isn't today, but there certainly will be enough parking for those who are willing and able to meet the price. Likewise in a Demand Management alternative there would, by definition, be the ability to move around the region for those who are willing and able to meet the price (and for others if the money thus collected were dedicated to making transit frequent and cheap). The questions are, what is that price, and how does it compare to the alternatives delineated in the DEIS, considering all relevant factors such as the cost of congestion, and cross-subsidation of one form of transportation to another. The DEIS doesn't even attempt to answer these questions.

To see that the statement is grossly misleading it is only necessary to repeat it with emphasis changed: Congestion Pricing could not by itself meet the objectives of the System Plan. But then neither could the other alternatives. None of the alternatives by themselves will preserve communities and open space, improve the region's economic vitality, or preserve environmental quality. "Implementing the System Plan by itself will not solve congestion on the region's freeways" (p. xiv). "No transit alternative will have much of a land use impact unless other complementary actions are associated with it." (p. 3-119)

Of all the alternatives by themselves Congestion Pricing (CP) comes the closest to meeting the four objectives. In theory CP could meet all the mobility needs of the region. More realistically, it would likely be teamed with Transportation System Management (TSM) as delineated in the DEIS, since even in a the freely flowing system created by CP, the region's freeways would most likely continue to have "choke points" where traffic slows, and it would be good public policy to give high occupancy vehicles (HOVs) a way around these. If TSM were implemented before CP, then the technology used to enforce TSM would be in place to implement Congestion Pricing at a later date.

In section 2.6.9 the DEIS also mentions equity, privacy, and land use as issues it indicates would need looking into.

The equity issue can be addressed in a couple of ways. The first way is to note that if funds collected in a TSM/Congestion Pricing (TSM/CP) alternative were dedicated to transit, making it frequent and inexpensive, possibly even free, then the relatively well off would in fact be paying the transportation tab for the less well off. Contriwise local sources for the other alternatives, sales tax, motor vehicle excise tax, gasoline tax, parking tax, are to one degree or another "regressive". The second way is less obvious: "As paradoxical as it may seem, by raising the cash cost of certain types of trips, virtually all trips can be made more economically." (Market-Based Solutions to the Transportation Crisis, Bay-Area Economic Forum, May 1990, p. 12)

I have to assume staff didn't look into privacy issues since 45 minutes investigating it would have shown that this issue has been met.

Land use issues would indeed need to be investigated since there would be spillover from priced to non-priced roads. However, the DEIS made no attempt to investigate such spillover in the alternatives considered. "By 2005, 'normal' congestion will cause about 75 million annual person-hours of delay" (p. 1-4), much of which will obviously spill over onto secondary streets as it does in Los Angeles today. Since, as I will later discuss, even the minimal effect the other alternatives are projected to have have on congestion is highly suspect, spillover is something that should be studied under all alternatives.

Additionally, under TSM/CP demand elasticities would have to analyzed.

Turning to other aspects of the DEIS. The bedrock of the evidently preferred alternative, Rail/TSM, is that rail is stated to have much greater capacity than the bus alternatives. Since this greater capacity has been presented here and elsewhere as a fact, I was surprised to find that not all experts agree with this "fact". Their argument is as follows: as long as busses have somewhere to go at the other end, a "busway" can carry more passengers than a railway.

Think of a stretch of road, like the floating bridges, on which there are no stops. Then consider that busses going, for instance, to Mercer Island can exit there without impeding the busses following. Likewise busses going to Factoria or Isaquah. This assumes that there is "elbow room" at the exits for busses to stop, discharge and take on passengers, and then get back on the freeway. On the other hand in a rail or transitway situation, as sketched on page xiii, when one vehicle stops, it impedes all following vehicles, thereby creating the need for adequate "headways", so that they don't run into each other.

On page xvi of the DEIS it is stated that one HOV lane can carry between 4,800 and 5,700 people/hour, a busway up to 9,400 people/hour. A "back of the envelope" calculation of a bus-only HOV lane (i.e. an HOV lane fully occupied by busses) is as follows:

Bus @ 40 mph, 44 seated people/bus

800 busses/hour x 44 people = 35,200 people / hour,

or nearly four times the number indicated in the DEIS.

Obviously this hangs on the ability of busses to freely exit the freeway, and freely get back on. In certain cases, most notably downtown Seattle, this may present a question. I will deal with that question later.

One issue that I didn't see analyzed in the DEIS is that of the flexibility of busses verses trains. This shows up in the relative ability to change concentrations of service, as well as in the relatively greater need for transfers in a train system.

One reason the downtown Seattle bus tunnel was built is that it was perceived that people don't like to transfer. A train system maximizes the number of people who have to transfer. Since with a train system "feeder busses" can't get on the "main line", people either have to transfer from bus to train, or a much more extensive parking system has to be built around the stations, obviating a significant portion of the business concentration assumed for them, as well as defeating whatever environmental gains were expected in getting people out of their cars altogether.

Turning to rail. The DEIS repeatedly states that rail has capacity for 22,000 people/hour for a single rail line. I have been told that BART carries 72 seated people in 10 cars, 3 minute headways, 20 trains an hour, or 14,400 people/hour, about 65% of the projections in the DEIS. Does the DEIS assume standing people? Will the downtown Seattle transit tunnel accommodate 10 car trains? Are shorter headways than have been able to be achieved in the Bay Area assumed? Or is this two lanes of rail?

If the above calculations are even "ballpark" correct, then a bus-dedicated HOV lane (a "double diamond" lane?) which is unconstrained as to exiting and entering, could carry something on the order of 2.4 times as many people per hour as a rail system. The "extra" capacity could obviously be used by vans and carpools.

Now some miscellaneous comments on the DEIS.

It is stated on the same page that "In areas where there is efficient mass transit, business and residents cluster near transit facilities." Referring to the previously quoted Market-Based Solutions to the Transportation Crisis (MBS), page 5: "Less than four percent ... of the workers at offices around Walnut Creek's BART stations use BART to get there."

The DEIS makes no attempt to cost the system on a per person basis. However using the data on page xxxvi, one can make an attempt. The capital cost of the Rail/TSM alternative is stated to be $11.5 Billion. The cost of TSM, is $4.7 Billion. Assuming 7% bonds, the interest on the Rail/TSM alternative is $805 million/year; interest on the TSM alternative is $329 million/year; the difference being $476 million/year. Likewise subtracting the operating cost of TSM from rail gives a difference of $93 million/year. Together this equals $569 million/year, the cost attributable to rail. For this one is (allegedly) gaining 23.6 million passengers/year. Dividing $569 million by 23.6 million, the incremental cost/passenger of rail is $24.11. Similar calculations for the TSM alternative give a figure of $5.41/passenger, about 22% of the cost for rail.millions

Riders Capital Costs 7% Operating Costs

Rail/TSM 157.3 $ 11,500 $ 805 $ 492

TSM 133.7 4,700 329 399

--- ----- ---- ---

23.6 $ 6,800 $ 476 $ 93



$ 569

$569/23.6 = $24.11/passenger

Assume a commuter using the system workdays:

$24.11/trip x 2 trips/day x 5 days/week x 50 weeks/year =$12,055 commuter/year.

The equivalent cost per bus commuter year comes to $2,715/year.

This assumes that a Rail/TSM system will capture 18% more passengers than a TSM system (p. xxxvi). A lot of people don't think this is realistic. Although the Bay Area has much better demographics and geography as well as a more extensive system than is proposed in the DEIS, there "transit use has shown an average rate of decline of almost 2 percent per year" (MBS p. 4). To the extent that rail captures fewer riders, the incremental cost per passenger of Rail/TSM over TSM will be that much greater. If Rail/TSM captures 9% more riders than TSM, the cost of rail will be $47.03/trip ($23,515/commuter/year); if 5%, $85.05/trip ($42,525/commuter/year). At that point, of course, it would be cheaper to pay people to stay home.

Added to the costs shown here should be the costs of delay. The data to analyze this doesn't appear to be in the DEIS. However it is mentioned on page 1-4 that by the year 2005 "'normal' congestion will cause about 75 million annual person-hours of delay." Presumably this figure will be considerably higher in 2020, say 100 million. If Rail/TSM combined with the Commute Trip Reduction Law increases transit ridership by 13% (p. 3-89), assume (crudely) a decrease in delay by that amount:

100 million less 13% = 87 million person hours of delay.

Assume a delay cost of $15/hour (1992 dollars),

87,000,000 hours x $15/hr = $1.3 billion per year.

So to the $569 million annual cost of the system should be added $1.3 billion in delay cost, for a total of $1.87 billion per year. The question for TSM/Congestion Pricing is, can it be implemented for less than this? While I am sure that the answer is yes, the data to calculate this cost, most notably the "demand elasticity" (at what congestion price level will people switch to transit) is not in my possession. Presumably Metro staff could derive this information.

As outlined in the DEIS there seem to be two flaws to the bus alternatives (TSM, and Transitway/TSM). One is the capacity of downtown Seattle streets to take the bus traffic. The other is the bottleneck that would be created by dedicating the express lanes on I-5 to a transitway. (The second is considered a "fatal flaw" to the Transitway/TSM alternative (p. xxxiii); however, on page 2-20 it is stated "like the TSM alternative, [the Transitway/TSM] would replace the reversible lane on I-90 and I-5 with a two-way transitway" (emphasis added). So apparently the TSM alternative envisions a transitway on I-5 north, even though it is nowhere identified as a "fatal flaw".)

As to the capacity of downtown Seattle streets, first of all one wonders if this is a major problem why wasn't the transit tunnel built with more capacity to begin with? Be that as it may, there are various possible ways around this problem. They include dedicating one or more surface streets to "transit only" either during rush-hours or full-time; making some busses stop only at the north or south tunnel stations and then getting back on I-5 north or south; making some busses stop somewhere else along the HOV (transitway) running through downtown; using double-articulated busses; building another tunnel (expensive, but a hell of a lot cheaper than a rail system).

As to the I-5 north bottleneck "fatal flaw", one wonders why 75 million of hours of delay by 2005 isn't in itself a "fatal flaw". Presumably the bottleneck is "fatal" while the delay isn't because people will see something being taken away from them with the former (the "don't-take-candy-away-from-a-baby syndrome"). Some possible ways around this "baby" include: move the entrances to 520 onto the express lanes north and south (from NE 50th and downtown), increase single occupancy vehicle (SOV) lanes north by increasing the Battery Street tunnel lanes to 6 and improving SR 99 capacity to move traffic, connecting I-5 with SR 99 at the West Seattle Bridge, widening I-5 by one lane each way from north of the Ship Canal bridge to Northgate, or building a SOV/HOV tunnel connecting 23rd Avenue East to the University Village area (all expensive, all a lot cheaper than rail).

The DEIS assumes that improved transit will result in fewer people on the highways (3.8% less, Table 3.9.3, p. 3-90). However experience calls this assumption into question. When the new I-90 bridge opened, the "extra capacity" that was supposed to last 15 years (or whatever it was) lasted about 15 minutes. Everybody wants the other guy to use transit so he can use his car. London, Paris, and New York are all in gridlock.

This being the case, the argument that any of the alternatives, with the probable exception of TSM/CP, would do much for air pollution is a pipe dream. Quite simply, given the existing and projected vehicle fleet mix if you don't get rid of congestion, you don't get rid of air pollution.

Even if you were to believe the projections of diminished car trips and VMT, the Regional Transit Project would be a ridiculously expensive way to tackle the problem. It would be several orders of magnitude cheaper to simply buy up clunkers (which do most of the polluting) and/or convert a part of the vehicle fleet to compressed natural gas or electricity.

A word about technology relative to the TSM or the TSM/CP alternative. The means to electronically collect tolls is being used in various parts of the country. One problem with HOV lanes is that there are an awful lot of "cheaters" using them (50% ?). The technology used to collect tolls could also insure that authorized vehicles use the HOV lanes. If it were desired, these vehicles could be made to "check in" at freeway onramps to make sure there were carrying the right number of people, possibly using something akin to the remote tellers at drive-in banks: the remote "teller" would authorized that vehicle to be on the HOV lanes for the following hour or until it exited the freeway. This same technology could be used to charge tolls under a TSM/CP alternative. As stated earlier the privacy issues on this subject have been met, and staff simply needs to consult the relevant literature.

My conclusions (other than that the belief in "salvation from steel wheels" should possibly be considered a cult and referred to the proper authorities) are as follows:

It appears that under the right circumstances TSM can handle as much traffic as rail if not more.

Busses are obviously much more flexible than trains and can minimize transfers.

The two bottlenecks to a bus (TSM) system, downtown Seattle and the I-5 expresslanes, don't seem to be insurmountable problems. Certainly they are solvable for much less than a rail system.

The technology that would be desirable to enforce a bus (TSM) system could be used at a later date if the region wanted to move to a TSM/CP system.

None of the alternatives examined in the DEIS would have much affect on air pollution. Congestion Pricing could have a significant affect, but that would need to be studied.

Therefore an alternative that should be further examined is TSM including SOV improvements, with a possible later addition of Congestion Pricing (TSM/CP).

However, if the Federal Government is willing to pick up 90% of the tab for trains, kindly ignore the above.


Donald F. Padelford

Appendix B

Donald F. Padelford

1221 First Avenue

Suite 2111 98101

POB 2846
Seattle, Washington 98111


fax 262-1930

March 19, 2004

Douglas B. MacDonald

Secretary of Transportation

PO Box 47300

Olympia, Washington 98504


Thank you for your letter of February 6, 2004. [1] Having mulled this over through the kids’ winter break, my annual all-guy ski trip and some elective surgery, I continue to believe that the central issue here is the statement, made in your letter to Secretary Mineta, that the National Review Panel (NRP), which was convened to study this region’s transit choices, concluded that "a new rail corridor was the only practicable choice for meeting the need to move more people in the I-5 corridor." There are two aspects to this statement. The first and less important is, did the NRP say this? The second and more important is, is it correct?

Concerning the first aspect, given that your letter was written specifically in support of Sound Transit’s light rail project, and given that the relevant EIS, presumably informed by the NRP, specifically rejected light rail, I will stand by my previous statement that your claim was "mis-informed." [2] However the real issue is not what the NRP did or didn’t say. The real issue is whether or not rail is the "only practicable choice for meeting the need to move more people in the I-5 corridor." It seems clear on its face that it is not. In fact it seems clear that BRT gives at least an order of magnitude more "mobility for the buck" than light rail. [3] And considering the bucks we are talking about ("a billion here, a billion there"), this is no small matter.

The central thrust of your 2/6 letter to me is that the bus capacity I derived for freeway HOT lanes is heavily constrained by the capacity of the Downtown Seattle Transit Tunnel (DSTT). I agree with this. But "the dog that didn’t bark" in your letter is any disagreement with the capacity figures I derived for HOT lanes. I will apply the "silence equals consent" rule (which phrase I may have made up, given that none of my attorney friends seem to know the Latin equivalent), to indicate agreement that buses on, for example I-90 or I-405 HOT lanes, would have far more capacity than rail in those corridors. I do agree that bus capacity on prospective I-5 HOT lanes is constrained by the "pinch point" of downtown Seattle [4] (although, again, assuming HOT lane build-out, it is not similarly constrained in the region’s other freeway corridors).

Taking your figure of a 2-way bus capacity in the DSTT of 125 buses per hour per direction (which number is likely to be low) [5] and 63 seats per articulated bus, gives 7,875 seats per direction per hour. In order to produce an apples to apples result, I will not take some fraction of bus seats vs some multiple of rail seats; that would be more nearly apples to cantaloupes. [6] On the rail side, while it is unlikely to the point of impossibility that Link could ever achieve 2 minute headways, [7] for sake of this discussion, I will take your number of 8,880 seats per hour per direction (74 seats per car x 4 cars per train x 30 trains per hour), an additional 925 seats per hour or about 12% more than the bus number above. [8] Now, given that in order to achieve the rail number, Link will have to be completed through to Northgate at a cost of something like $7 Billion [9] the question becomes, it is possible for bus to achieve this capacity for something less than this figure? [10]

Well, duh. Ignoring for the moment the ways that exist to achieve greater operational efficiencies, [11] and ignoring all the relatively simple and/or inexpensive remedies that exist to increase through-put [12] – ignoring all that and just assuming that we built another bus tunnel once the current one is maxed out some decades hence, that in itself would double and could triple [13] the capacity you indicated for the current DSTT, which by your figures is currently 56% utilized. [14]

Adding this up produces something on the order of an eventual bus through-put in excess of 23,000 seats per hour per direction, [15] well above of any plausible transit demand, and of course many times the number theoretically achievable by Link. [16] The fact that this figure is only half the full bus capacity of a freeway HOT lane, [17] just means that these lanes are never likely to be fully utilized by buses, and will always have excess capacity for non-bus use. [18] Another way of putting this is that a HOT lane system will always have a robust transit capacity buffer, and thus will never need to become a bus-only lanes (busway) as envisioned by the National Review Panel. [19]

The cost of a new DSTT ("DSTT2")? Well, the existing one, if memory serves me, cost about $600MM. So double (or triple) that. As for a bus lane on the I-5 reversible lanes, a WSDOT number for that (Ravenna to Stewart) is $59MM. [20] By any measure, it’s still a fraction of the Initial Segment cost, let alone Link from SeaTac to Northgate, not to mention a 150-mile build-out of the system.

The conclusion I take form the above is that while, as you state in your letter, the DSTT limits bus capacity in the I-5 corridor, given some simple investments, it doesn’t limit it to the point where bus through-put is less than rail. And given major investment (but far less than required for a comparable length of light rail) bus can give substantially more capacity. In sum: for less money, equal capacity / for equal money, more capacity. Either way: "more mobility for the buck." If this was not clear at the time of the National Review Panel, it certainly is clear now. [21]

My only other point concerns the consent of the governed. You said,

Talking with Aubrey Davis about your letter, he points to the strong voter support for the transportation program containing light rail in the 1996 Sound Move vote, especially in the North King Subarea where the tax burden falls to pay for light rail. Even as the plan has been shaped (and unceasingly attacked) [22] in intervening years, voters in the area continue to favor it as recently shown by their tally against Initiative 776. Transportation policy must be rooted in popular support as well as in technical analysis; Link Light Rail, despite the controversy it attracts, seems to stand with a strong measure of both.

There are a number of things to note here. The first is that we the voters were lied to

before the 1996 vote. Sound Transit said that building the light rail system would be the equivalent of adding a 12 (sometimes 16) lane freeway. In fact not a few ST board members, perhaps in a kind of "blow back," appear to believe that to this day. [23] Obviously if you added a 12 lane freeway, you would be taking a big whack out of congestion. [24] Unfortunately the Link system would have no noticeable congestion effect. [25] As for the 776 vote, before the vote we were repeatedly told by Sound Transit supporters that it had nothing to do with light rail. After the vote, it was a referendum on it. [26]

As a summary I could do worse than quoting Emory Bundy, in the title of an essay he wrote before the ’96 vote: "If the [Sound Transit light rail] project [were] defensible, it could be defended by telling the truth." Unfortunately it isn’t, so it isn’t. Yes, if we had tens of billions of dollars burning a hole in our collective pockets, I suppose it would be nice to have a rail system (although in that case it should be the kind of grade-separated system envisioned by the National Review Panel). But I haven’t noticed that we are suffering from such an embarrassment of riches. Since I am the author of the phrase "WPPSS on Wheels," I will end by saying that, in my view, this is pretty much what the Link light rail project amounts to. I realize that you have a different view.

"We wish we had never started the whole thing. Fixed rail is not the answer to the transportation needs of our city. We should stop all this insanity that has gone on these past years."
Richard Riordan, Mayor of Los Angeles and Chair of the Los Angeles County Metropolitan Transportation Authority, on public radio program, Which Way LA?,
June, 1998.


Donald F Padelford

As a postscript let me say something you already know, namely that I am fully on-board with the WSDOT in getting a HOT network established. We may have failed in this legislative session, but I am reassured by the words of Winton Churchill: "The Americans can always be counted on to do the right thing – after exhausting every other alternative." I would say the other alternatives are just about exhausted.

Also, if you would like to get a few people together for lunch, I would suggest (if they will consent) Emory Bundy and Chuck Collins for my, in your word, "seconds." I trust, however, this doesn’t entail, as the term might suggest, a duel.

Entries in this arial font are notes to myself that were added after letter was sent.



This was a response to my email to you of 10/4/03 with clarifying emails sent on 10/25/03 and 10/27/03; which email communication, in its turn was a response to your 7/3/03 letter to Secretary Mineta, written in support of the now-approved FFGA for Initial Segment of Sound Transit’s Link light rail project.


I think that this is adequately covered in previous communications, but to briefly recap: The 3/93 EIS rejected "light" rail, as well as busways, and supported "heavy" rail. I think it’s a fair inference that what the EIS was rejecting in "light" rail is lack of grade-separation. Thus the EIS (informed by the NRP) rejected non-grade-separated rail transit. It seems likely, and is in conformance to my memory of the then-discussion, that over-head-feed (ie "light") rail that is grade-separated in it core would have passed muster. However the EIS examined at least two iterations of "real" (non-grade-separated) light rail, such as Sound Transit is now beginning to construct, and explicitly rejected that modality. Bus rapid transit running on HOT lanes (BRT on HOT) was not on the radar screen at that time and was thus not considered. It is also a fact that neither the ’93 EIS nor any other EIS or similar analysis compared specifically Light rail to Bus Rapid Transit. Repeat: LRT has never been directly compared to BRT. Never happened.

As communicated earlier I believe that the reason we have "real" (non grade-separated) light rail running down MLK is that the Seattle City Council, during the mayoralty of Norm Rice, "played the race card" to divert the system, which was to run elevated (like Vancouver’s Skytrain) down the Duwamish, to the Rainier Valley, where, due to some suspect budget constraints, it became an at-grade system. The irony of this, of course, is that, thanks to the race card, the predominantly minority residents of the valley get a dangerous at-grade system while the white, middle-class residents of the north end get tunnels. Besides being dangerous, such an at-grade system is also slow, low-capacity, and puts a monkey wrench in way of east-west travel. In a pattern that has been repeated numerous times, the analysis that was produced to support the MLK route was in essence fraudulent.

If Link had gone where and how objective analysis at the time said it should go (elevated down the Duwamish) it is likely that the opposition that has formed around this project would never have coalesced, and that the monorail movement, which has as its mantra that monorails "don’t get stuck in traffic," would never have been born. Thus the reason that Link has, in your words, been "unceasingly attacked" is a direct result of disingenuousness in the process that gave it birth.


Of course there are lot of unknowns, and any cost comparisons are somewhat "back of the envelope." For one thing Sound Transit doesn’t know (or says it doesn’t) how much the extension to SeaTac and Northgate will cost (which doesn’t stop Ron Sims from being "confident" that it can be built within existing taxing recourses -- now wait a minute, if your don’t know how much it will cost, how can you …. oh, never mind). None the less our best guess (via the good offices of Jim MacIsaac) for the SeaTac to Northgate line is in the $7B range. If a full system is 150 miles, and if the rest costs about the same per mile, this would produce a system cost of about $50B. As for BRT on HOT, the main thing that needs to be done is to finish the HOV lane system, which according to the WSDOT would cost on the order of $1.65B for the "core" system (not including the cost of ramps and connectors to make it seamless), and then convert the HOV lanes to HOT, which can be financed from HOT revenues. While special ramps and connectors are not needed right away, a WSDOT number for that is approximately $290MM. Thus finishing the core HOV system and building the HOV connections should cost on the order of $1.9B, about 1/26th of the "back of the envelope" cost for Link above. Given HOT conversion (to keep the lanes in free-flow in perpetuity), for this you would be getting a BRT system running 150 miles and covering the entire metropolitan Puget Sound region at 60 mph 24/7. Even if the relative cost figures are off by a factor of 100%, this still is an order of magnitude difference between the two.

There two, off-setting, errors above: Link SeaTac to Northgate is about 24 miles, not the 21 miles used in the letter. However, this is off-set by updated cost data: the SeaTac to Northgate cost, according to new (to me) information from MacIsaac (email from JM 4/24/04) is $7.9B vs the reported $7B

$7.9B / 24 miles = $329MM / mile; $329 x 150 miles = $49.5B, essentially equal to the $50B figure mentioned in the letter.

Also the above figures do not include the cost of a southbound bus lane on I-5 at about $59 MM. Thus the total cost for a BRT roadbed would be approximately $1,650 + 290 + 59 = $1,999MM

1,999 / 49,500 = 1/25, not 1/26 whatever


While not all I-5 buses stop in downtown Seattle (eg airporters from Skagit do not), almost all do; it seems like a reasonable simplifying assumption to assume they all do.


Previous Metro analyses prior to 8/01 (see footnote #8) showed a capacity of 150 buses per hour per direction. Our understanding of the 125 figure is that Metro was instructed to assume no operational changes or new efficiencies (I believe there is documentation to this point). Having said this, my knowledge of the nuts and bolts of bus operations is very limited. If you want to get the real scoop on achievable bus efficiencies you need to talk to someone who is both knowledgeable and is an honest advocate for that technology. In this region you couldn’t do better than Chuck Collins, the former head of Metro.


Either you have to compare theoretical capacity to theoretical capacity, or achievable capacity to achievable capacity. It just isn’t kosher to compare the theoretical capacity of rail to the achievable capacity of bus. However, I’m afraid that this is exactly what your letter does. In my case I am using a form or theoretical capacity, seated capacity in each case, which is apples to apples. I leave achievable capacities in each case (oranges to oranges) to those with more expertise than myself.

However, the concept that if you put a bus on rails, passenger capacity goes from 80% of the number of seats to 185% is sort of ridiculous.

If, as you aver, this is partially a function of the "shorter trips" for light rail, that just shows that what Sound Transit is building is not a transit "spine" but a gold-plated local bus line substitute (instead of a bus line down MLK, we in effect get an extension of the Waterfront Trolley down that arterial).


BART is of course a completely grade-separated system. It is my understanding that the best it has been able to achieve, after decades and billions and billions of dollars, is slightly better than 3 minute headways. In order to get 2 minute headways for Link, the concept that I have read is that a train would come from the south, then one from the east on I-90, then one from the south, etc. The south and east trains would be on 4 minute headways, and combining them would produce 2 minute headways going north from downtown. But can trains running at-grade through the Rainier Valley really reliably maintain 4 minute headways, and merge, with the precision of a Swiss watch, with trains from the east? Of course not. And what would a train every 4 minutes do to cross traffic in the valley? Devastate it. (From the 1993 EIS, page 2-50: Non-grade-separated "[t]rain frequency is also limited, since there must be time to clear intersections between trains. Conventional transit practice and highway standards suggest that when train frequencies are under 6 minutes, cross traffic on arterials will be affected to the extent that grade separation is necessary.") So the concept that Link will run on 2 minute headways north of downtown under such a scenario is not credible (if the grade-separated BART can’t do it …). If Link were to eventually achieve 6 minute headways with both feeder lines (south and east), that could produce 3 minute headways north of downtown, which, as I understand it, is the figure that Sound Transit originally posited before it figured out that 6 car trains won’t fit into the DSTT stations (see below).

The 1996 Sound Move document envisioned 6 car trains. When it was pointed out that 60 foot rail cars haven’t been manufactured since WWII, the six 60’ railcars morphed into four 90’ cars (the stations are 380’ long). But interestingly the 60’ and 90’ foot cars each were to carry about the same number of people (125 vs 137), so while the length of the cars increased 50%, the number of passengers increased a mere 10%. Didn’t someone bother to measure the stations before asking the taxpayers for billions? Or were they perhaps contemplating double-decker rail cars in Sound Move? Anyway, this little fiasco required Sound Transit to decrease the headways in order to maintain (theoretical) capacity. This was accomplished as follows: 6 cars trains running every 3 minutes produces 120 cars per hour (20trains/hr x 6cars/train = 120cars/hr) This was changed to 30trains/hr x 4cars/train = 120cars/hr. The only problem is that there never was any realistic prospect of running trains every 2 minutes.


My understanding is that the 8/21/01 "Evaluation of Joint Operations in the DSTT," referenced in your letter, is, to put it charitably, a flawed document. Metro at this point has become essentially a captive agency of Sound Transit. Therefore work product that comes from Metro should be examined with a jaundiced eye. People working there know the score. Transportation professionals in other agencies who showed insufficiently-attuned political instincts were shown the door.

With apologies to Shakespeare, the dialogue between Sound Transit and Metro may have gone something like the following:
Sound Transit: Do you see yonder cloud that's almost in shape of a camel?
Metro: By the mass, and 'tis like a camel, indeed.
ST: Methinks it is like a weasel.
M: It is backed like a weasel.
ST: Or like a whale?
M: Very like a whale.



See footnote # 3

Alternately, if you threw $7 Billion at a bus system, could you do more?

As stated earlier, I am certainly no expert in this realm, but I am given to understand that if one were to use the tunnel only for outbound and "inter-lined" runs (ie runs that start north of the downtown and, sometimes with a change of number, continue on south of downtown, or visa versa, as opposed to runs that terminate in downtown), that efficiency would be improved. I am further given to understand that currently even dead-headed buses use the DSTT. I believe that I am only scratching the surface here.


These would include, from mundane to perhaps fanciful:

Using 3rd Avenue as a bus mall, either during rush hours of full-time

Platooning buses

Using double-articulated buses

Creating bus-trains in controlled-access situations like the DSTT by electronically connecting buses as has recently been demonstrated in Japan.

Intercepting some buses north or south of downtown and transferring passengers to a shuttle (monorail, personal rapid transit, moving sidewalks, gondolas, bike/Segway lanes, etc)

Narrowing lanes (with the implementation of self-guidance technology in the tunnel) and narrowing sidewalks to produce two passing lanes in the stations, thus allowing "skip stops," with the consequent doubling of through-put from the maximum now envisioned for the DSTT.

Punching out station walls into adjoining properties to get an additional passing lane as above.



ie if we had two passing lanes in the stations as opposed to the existing one passing lane, it would be possible to implement skip stops, thus doubling the capacity of the new tunnel when and if needed. Thus, using your figures, the through-put could be 125 (the current DSTT) + 2 x 125 (a new DSTT) = 375 buses per hour per direction.

(Using the normally accepted figure of 150 buses per direction per hour for the DSTT, the eventual through-put could be 450 pdph.)


Thus from today’s through-put, capacity could be increased by a factor of
375 / (125 x 56%) = 5.4 or 540%. That should do it!


375 buses per hour per direction x 63 seats per bus = 23,625 seated passengers per hour per direction.

According an email from Jerry Schneider 2/25/04 Bogota and San Paulo BRT systems carry 30,000 pphpd.


Calculated above at 8,800 seats per hour per direction, assuming (unrealistic) 2 minute headways. Therefore, I don’t agree with your statement that "We do know at least one virtual certainty. Within a short horizon – say twenty years of so into the future – such a [bus] system would not the capacity to serve the needs in our most congested, highest demand corridors."
Bus/rail seats = 23,625/8,800 = 2.7 times.

Assuming more realistic through-put numbers for bus and rail of 450 (two tunnels) and

20 (3 minute headways) respectively, the numbers become:

Rail: 20 trains per hour x 4 cars per train x 74 seats per car = 5,920 seats per hour

Bus: 450 buses per hour x 63 seats per bus = 28,350

Bus / Rail seats = 28,350 / 5,920 = 4.8 times


HOT lane:

3600 seconds per hour x 1 bus per 5 seconds x 63 seats per bus = 45,360

Two tunnels:

375 buses per hours x 63 seats per bus = 23,625


Vanpools, qualifying carpools, emergency vehicles, paying non-carpools.


As communicated in my 10/4/03 email, it appears to me that the NRP envisioned a busway with "on-line" loading/unloading, thus severely constraining its through-put.


As I envision it, a HOT lane build-out would eventually include the now-reversible lanes on I-5 (and I-90). One concept would have a 2x2 HOT format (ie 2 HOT lanes in each direction) south of SR520 and a 1x1 HOT format north of 520. In this concept there would be a HOT to HOT interchange at the intersection of I-5 and SR520. Obviously something of this magnitude would cost in the hundreds of millions of dollars. The above-referenced Ravenna to Stewart contra-flow bus lane on I-5could be thought of as an interim solution. I would add that this is one of the benefits of a BRT approach, it allows interim solutions. You don’t have to build out the whole system at once since buses can, for example use HOV lanes before they become HOT lanes (or use GP lanes). Thus we could have a pretty good 150 mile system quickly and a superb system over time. In contrast light rail in this region obviously cannot run anywhere new right of way isn’t built. The existence of an existing right of way is why the San Diego light rail system was built at such a bargain price. Likewise with St. Louis and commuter rail in Chicago. If we had kept the Interurban ROW, this might have been true of this region as well, but of course, we didn’t.

One, somewhat esoteric point about HOT lanes, is prompted by footnote #1 in your letter to me. You say that HOT lane "would certainly limit the number of cars that the lane would simultaneously accommodate." This subject has been discussed extensively within the HOT lane community. I think it’s a fair characterization to say that we have concluded the opposite – that by avoiding the "backward bending" portion of the flow vs density curve, pricing can maximize through-put. It works much the same way as ramp metering, which increases flow on the freeways, only more so. -- Also I don’t see why HOT lanes would need to annex any GP lanes as you state, other than the reversible lanes. And by simultaneously changing, for example, the Mercer Street entrance to the right-most lane, it may be that no GP flow need be taken away. In any event it is always possible to add GP capacity unless a "pincer movement" against adding or subtracting I-5 capacity is undertaken (see footnote #2 in my 10/4/03 email to you).

Also, while it is certainly desirable to have some kind of separation of HOT (or for that matter HOV) and general purpose lanes a la the SR167 prototype, it isn’t absolutely necessary. After all our HOV lanes operate without them today.


Note that it’s not even clear that 3 minute headways (20 trains per hour) are achievable. But if it is, then the rail seated capacity number is 74 seats per car x 4 cars per train x 20 trains per hour = 5,920 seats per hour, considerably less than bus. So if rail headways of 3 minutes are the best achievable, then bus has more seated capacity today without any further investment, than rail would have after the expenditure of funds necessary to get from SeaTac to Northgate (approximately $7B, see footnote #3).


See footnote #2


I got into somewhat lengthy discussion with Greg Nickels on this very subject before he was elected mayor. He was convinced at the time that it was the gospel truth. (As chair of the finance committee, he also assured me that he bird-dogged the cost figures very closely and that they were absolutely solid – this was before the initial billion dollar cost-overrun – haven’t heard from him on this recently.)

the ’96 election:

- What was promised by ST was what polls said people would go for, not what analysis said was possible.

- Gerrymandering (per Dan Norton)


In the recent SEIS for the north section of Link this egregious piece of mis-information has again made an appearance. Obviously the people who put together that document know better. At long last, have they no shame?


From the 1999 Central Link FEIS, page 3-5: "Therefore, based upon traffic forecasts, the light rail system will not result in a significant difference in regional traffic volumes … Peak hour volumes for the No-build and light rail alternatives are also expected to be similar [almost identical] across all screenlines." In other words, it won’t do diddley for traffic congestion.

The cute circumlocution that Link would provide congestion relief "for those who use it" isn’t even true. Trundling down MLK on light rail would be no faster "for those who use it" than doing so in a bus or car.

26. ]

Also Rob McKenna, the Seattle Times editorial department, and many others opposed 776 and currently oppose Link. Obviously for them 776 was not a referendum on Link.

The articles are posted solely for educational purposes to raise awareness of transportation issues. I claim no authorship, nor do I profit from this website. Where known, all original authors and/or source publisher have been noted in the post. As this is a knowledge base, rather than a blog, I have reproduced the articles in full to allow for complete reader understanding and allow for comprehensive text searching...see custom google search engine at the top of the page. If you have concerns about the inclusion of a specific article, please email for a speedy resolution.