As Orphan Road brought up in their own great post on the same subject, light rail skeptics are commonly cite “transit share” as a way of dismissing serious investment in our transit infrastructure. But what is “transit share”? Well, first we have to see what it’s a “share” of: the total amount of trips in the region. Walking down the block to the grocery store is a trip. Taking the bus to work is a trip. Bicycling to Pike Place Market is a trip. The transit share is the percentage of all trips made by transit.
I dismiss this measure as a distraction in the headline. And here’s why: evaluating every trip with equal weight is not meaningful. There are work trips and non-work trips. Unsurprisingly, non-work trips have a very low transit share (see the graph to the right) but they also have a much higher walk/bike and carpool shares. These non-work trips drag the over-all transit share down, allowed a significant investment in light rail and bus service to be spun as a waste.
But non-work trips are the trips made during off peak times where congestion is minimal. Congestion is the largest motivator of highway spending, and when transit provides a sensible alternative to driving to work alone the region can not only save money but help the planet at the same time. Note how 75% of work trips are in single-occupancy vehicles. I call that “low-hanging fruit.”
Besides congestion, non-work trips differ substantially from work trips in length. According to the PSRC, work trips are expected to be more than two and half times in length than non-work ones by 2040 (around 13 miles compared to around 5 miles). Those single-occupancy vehicles driving for longer distances in congestion for work trips represent the largest collective emitter of CO2 in our region.
So, work trips are where the congestion is, they’re where the single-occupancy drivers are, and they’re where the length is. Just like when the highway lobby talks about congestion they mean during peak hours, so should an honest conversation about transit reflect its peak use. As Orphan Road points out, one alternative for the region’s 2040 plans, which focuses heavily on light rail investment, nearly doubles the transit share of work trips to 19%. That’s the number that matters.
(Once again, this post was inspired by a blog entry by our friends at Orphan Road. Be sure to read their original report.)
Amen!
Note that congestion during commute hours is also the excuse for many local road expansion projects which are paid for by our sales and other taxes (like Bridging the Gap).
Nice post. Good to see you and Orphan Road are not letting these shysters get away with their funny number business unchallenged.
Another often overlooked fact (but very well understood by Niles and his ilk) is the fact that “trip” counting is completely skewered towards drivers. Under the current way of counting “trips,” if you take your child to day care, then drive to the store and then to work and then to pick up your child and then home, you have just made seven “trips.”
That’s what allows Niles/Kemper/etc. to make the ridiculous claim that most weekday trips in the region are not work/commute trips. In the above example, only one of those seven “trips” is a commute/work trip.
But in the end, who cares? We won, they lost. We build.
What is wrong with this trip counting? The fact that most trips in any region are not work trips has been well documented as stated in NCHRP 365 Page 29 (Page 41 in pdf) Table 9. (You can take a look at it in this 12 MB PDF file). As long as any type of trip is made during rush hour, why should it not be counted? Also, in your example, two of those trips are work trips.
Great post. What I’d also like to see is these percentages in numbers – how many people are taking transit to work every day and how many more cars on the road this would be if everyone drove.
Here in Vancouver BC, journey-to-work mode share by vehicle (driver) decreased from 55.1% to 51.5%, with transit increasing from 23.7% to 25.1%, walking increasing from 10.7% to 12.2%, and cycling from 3.3% to 3.8%. I expect transit percentages to increase with the recent completion of the Canada Line from downtown Vancouver to the suburb of Richmond as well as the international airport.
For downtown residents, walking made up 38.8%, transit 20.4%, and cycling 2.6%. Vehicle (driver) made up 32.7%, and I’m willing to bet that most of these people work outside of the downtown core.
In the rest of Metro Vancouver, however, Vehicle (driver) mode share was still a whopping 73.4%, although this was down from 2001. Full report here:
http://vancouver.ca/ctyclerk/cclerk/20080513/documents/tt1.pdf
One factor though is that you can’t put a car in your pocket for part of the day and take it out later. You either have to take it out for a whole set of trips, or none of them. One of those trips may be work, but then you have to stop by the store on the way home, or go to an evening meeting. If adequate off-peak transit exists, you might leave the car home for the day, or even not have a car. But if it doesn’t, you’ll be driving all day. Some transit fans think only peak-our service and work-trip transit are necessary. But you can’t fully separate peak from off-peak. (Not to mention the people that work odd hours.)
That is not entirely true, just as John Niles states in the comment thread. The baseline scenario is the same year as Alt 5. The baseline already has a 16% transit share for work trips. So overall, Alt 5 increases transit share from 16% to 19% which is a 15% increase in transit share for work trips – not double.
Good point.
No, the baseline includes all this light rail. Your statement about light rail investment is accurate, as that’s where the 6% difference between 2006 and baseline is coming from.
Exactly, so the whole point of the argument is that this 6% increase is going to happen regardless of which alternative is selected. Therefore any variation on top of that 6% is what matters.
That’s a good point in that we need to be more aggressive than the PSRC suggests. :)
The baseline assumes some things that aren’t happening – like continuing increases in Metro service hours.
Or perhaps Wamu failing or perhaps Russell leaving Tacoma for that matter. The forecast is always wrong. Furthermore since it was made during a time of significant economical change before all the data is available or a new household travel survey or Census 2010 is done these are all likely going to cause unforeseen effects. But unless some blatant error is identified, we have a baseline and five alternatives to compare that all share these inherent problems.
Remember the baseline is assuming a full build-out of ST2. On the other hand the Base Year (2006) is before the opening of Link, so the 19% is a doubling vs. the 2006 base year. At best this argues that additional investment on top of ST2 may be marginal in a regional context. However with additional ST investment in Link there may be some corridors (say Tacoma/Federal Way) where transit share does increase but since it is over such a low baseline it doesn’t move the regional numbers much. Conversely in Seattle you have the opposite problem as the commute share for transit is already high (17%) many of the users of rail are already transit users. They are just switching from congested, slow, and unreliable buses to rail. Indeed this follows the historical pattern where as transit use increased in a corridor it was upgraded to handle more traffic until we get to the subways and streetcar tunnels of the worlds major metro areas.
I found this quote by Serial Catowner the most interesting:
“””The number of trips made by suburban drivers is almost unbelievably huge- these are people who might make 4-5 trips a day without even noticing it. This is much different from transit riders, who tend to be more efficient and avoid trivial trips, such as you might make to save 10 cents on a gallon of laundry detergent.
Transit will never compare on a “number of trips” metric because the suburban driver wants to make many trips, and the transit rider doesn’t.
“””
Remember,
The private automobile allows one the “freedom” to do ones shopping, etc. in multiple locations without any resistance save for time.
All you have to do is push on a pedal. There is no incentive to be efficient.
If I drive to QFC for this, Trader Joes for that, and Wal-Mart for something else and then Costco for a hot-dog in my car, the car earns 5 “trips”. If I ride my bike to the Pike Place Market and buy everything there, because I obviously don’t want to bike to QFC, TJ’s and Wal-Mart, the bike only gets two “trips”.
Heck, if I stop at Bartells while transferring from the Monorail to Link on my way home from work, that’s “One Trip”.
Modeling: justifying the asphalt and concrete orgy for over 60 years!
No, it would be two trips. The phenomenon you’re describing is called trip chaining.
That’s debatable. If the stop at Bartell’s was during the transfer wait time it could just be one linked transit trip.
Actually, now that I think about it, the Bartells would have to be in the exact path of the transfer, so unless it was at the Mezzanine it would otherwise probably be an extra trip. These small trips are known to be seriously under-reported.
And that underreporting will get even worse when we have little coffee stands and such in stations!
What if I save enough time taking transit from the P&R instead of sitting in traffic as a SOV that I can swing by the gym on my way home from the P&R. Has transit just enabled doubling the number of automobile trips?
I guess that’s the real reason I hate anything to do with metrics involving trip counting. I just don’t understand it.
It is tricky. What if both transit services were very frequent? How is it different from driving? Would a stop at the drive-thru espresso stand on the way to work make it two trips, like a stop at the stand in the station?
In the travel diary PSRC uses to collect trip data, it tells you to list every trip/transfer and every activity you do.
Whoa whoa whoa. Can someone please tell me how many trips it is if I walk from work to Dick’s by the bus stop in Lake City, ride the 522 downtown, use the restroom in Nordstrom, ride Link to Columbia City, walk to the farmers market, then walk home? I always thought that was one really amazingly fun trip, but now I’m having trouble counting. Is it work-related? What if I drop off some work-related papers along the way?
The farmer’s market alone probably counts as about 20 trips.
Each activity, even Matt’s restroom stop in Nordstrom, is considered a trip end. Thus the trip to the restroom and from the restroom should be counted as one trip each. As far as building a travel demand model, it really boils down to what is reported or observed in household travel surveys that is inputted into the model. People probably don’t realize that all the little trips enabled by dense development are actually trips.
But the household surveys are like the Radio Diaries that the radio ratings services just dumped. They depend on the user to accurately fill out the log and as we see above, it is far easier to miss a transit trip than it is to miss an auto trip because while the auto trip is begun and ended by (usually) a turn of the ignition key, a non-auto trip can be made sub-consciously.
The move to an electronic measure of radio listening has led to a huge revamp in the radio ratings biz. May I suggest a similar move in the world of trip data collection?
But again it will ALWAYS favor the car, because the car only requires the gentle push of an accelerator.
Even a paraplegic can travel solo by car. They’re not likely to ever travel as much or as far in an unpowered wheelchair!
It sounds like transfers skew the numbers. If I walk to a bus stop, that’s me making a choice. In NYC if you take two subways to reach a destination, that’s technically two trips but effectively one trip, because the system was designed to be one big virtual line, as is shown by the one-trip farecard charge. When Metro split the 43 into the 43 and 44, that suddenly made one trip into two trips without the rider doing anything to cause it.
So the effect of this split is ambiguous. Overall it helps people by making the buses more on time. But it inconveniences those whose trip was suddenly split. But if Metro funding depends on more trips, it’s good. But if Metro funding depends on more one-seat rides, it’s bad. Counting by passenger-miles benefits those who take long trips but disadvantages those who take short trips (by shifting money away from routes that are predominantly short trips; e.g., the 2 vs the 358).
I agree that the method of counting trips is seriously flawed. If someone drives to work in Seattle from the Plateau and back again they’ve made two trips. Both long and taxing peak capacity. If they drive to a P&R and take the bus they’ve added two trips by transit but the number of SOV trips driving remains the same. Yet clearly there is a huge benefit to using the P&R. Peak hour vehicle mile reduction is what’s important and an even sharper focus should look at where that reduction occurs. The choke points in the system are well known, I-5, I-405, I-90, SR-520, SR-167, SR-522…
I just count passenger miles.
Passenger miles has its own flaws, though. A bus with one passenger coming 40 miles from North Bend to Seattle is equivalent in passenger-miles to the SLUT running one mile with 40 passengers.
Which makes it harder to increase transit market share (by trips), numerically, since auto trips don’t decrease. It is a flawed way to measure transit’s effectiveness. Where in fact more people are taking transit than what the numbers say.
I love how the choke points you mentioned are simply the major highways. That about sums it up.
Oh there’s plenty of other choke points but the major highways by definition are going to be where you have the greatest numbers of people and therefore the largest potential market. Mercer would sure be a choke point; Avondale Road & Novelty Hill another, Willows Road, 1st Ave in Seattle… But by picking the low hanging fruit (major highways) you’ll have a significant effect on reducing the secondary choke points (surface streets). If things weren’t routed along high capacity corridors already you’d have a much harder time of it.
I think we should also suggest that as 1/2 of trips are hov, that converting up to 1/2 of chokes point lanes would be a cost effective option.
While rail better than HOV, HOV is better than SOV.
It seems obvious to me that clogged HOV lanes (like those on I-405 through Bellevue) should be converted from HOV 2+ to HOV 3+. If the HOV lanes are the same (sometimes slower) than the general purpose lanes then something is seriously out of whack. Highways need a certain amount of merge room. I don’t think you could convert 2 out of 3 lanes say to HOV and have it work for anybody. But, in areas where there are 4+ lanes in each direction once you’ve gone to HOV 3+ it might make sense to restrict the “fast lane” to HOV 2+. One problem with this though is that HOV lanes in really dense areas tend to work only when they have dedicated entrance and exit ramps. These are a lot more expensive than simply painting diamonds. If you create a pair of HOV 3+ and HOV 2+ do you force the HOV 2+ to weave across two general purpose lanes to exit or do you allow them to merge and utilize the (usually inside) HOV 3+ lanes?
They generally do convert HOV2 to HOV3 the worse traffic gets. Maybe 405 hasn’t kept up.
I don’t know of anywhere that has more than one HOV lane on the same road. That would be a big social change to limit SOV traffic to one or two lanes. Maybe someday. (Unless Tim Eyeman passes an initiative to prohibit it.)
There’s an argument that HOV lanes don’t exactly do what they intended to. A carpool is when multiple people who would otherwise drive separately, drive together. But a family outing is not a carpool because the kids can’t drive and ma and pa want to travel together (that’s the purpose of the trip). And UW students pick up people next to campus so that they can qualify for the cheap HOV parking spaces. In most cases people don’t have a choice whether to take one person or multiple people on the trip; that depends on outside factors. You can find somebody to carpool to Microsoft with, but you can’t find somebody who wants to go to the store at the same time you do. HOV lanes are great for providing a speedy lane for transit, and they discourage SOV driving in the abstract (although that hasn’t cut the number of SOVs on the road!), but their ability to increase carpooling has been less effective.
BTW Seattle is #3 nationally for the share of commuters biking to work, #6 for walking, #9 for transit use, and #8 for using a mode other than SOV. Since the data is from 2008 Seattle did by far the best of any city without rail transit (no the SLUT doesn’t count).
Doesn’t Sounder count as rail transit?
I suppose it is, but I was thinking in terms of not having a metro system or a meaningful amount of light rail (which is why I didn’t count the SLUT). I’ve seen Seattle refered to as having an “all-bus” transit system while the Waterfront Streetcar was still running and even after Sounder service started.
A good source of information on trips and travel in the Puget Sound region is the Household Travel Survey series of studies from PSRC — literally world famous, verified by checking Google.
Underway for years, the latest results (2006) are described at http://www.psrc.org/assets/1979/hhactivitysurvey.pdf (1979 in the URL is a serial number, not a year.)
A comparison report covering the entire series going back decades is at http://www.psrc.org/assets/2128/_09-44_Household_Travel_Surveys.pdf .
This one points out that home-based work TRIPS (A to B) are at about 20% share, while TOURS (trip chains, A to B to C to …) centered on work locations are lately at 35% share of travel purposes.
Thanks for the Link (I think). That’s a whole lot of information and from what I’ve plowed through so far is mostly subjective rather and objective data. We have easily tracked and verifiable date like numbers of registered vehicles and thanks to emissions controls we have mileage data for all vehicles in the greater Puget Sound region. We can even track their gas mileage and emissions. Shoot, we even have in the computer what address they’re registered at. I’m hard pressed to understand the value of trip diaries, topology of tours and chaining when we can just read odometers.
One thing the report did expressly address; a trip from home to a P&R to catch a bus is “linked” as one trip. I haven’t dug in far enough to understand how that relates to number of transit trips vs number of automobile trips but the other question it answers is that all stops between the P&R and home are now counted as “extra” trips which is patently absurd. Someone is much more likely to stop and do the grocery shopping if they’ve been able sit and relax (perhaps even do billable work) riding the bus or train than having spent an hour or more battling traffic.
The whole point of a household survey is typically to obtain parameters for use in developing a travel demand model. While it is possible to measure such discrete things such as the cars people own and how far they drive, it is also important to eventually model where all this activity takes place so the number of trips can be estimated in the travel demand model. Using the output is supposed to help regional planning organizations and cities plan for future growth.
Digging deeper (wear your boots) I’m looking at the map on page 9, “travel survey household locations, 1999 vs 2006”. I would expect the density map to follow growth trends. Maybe they do but I’m seeing a 1999 West Seattle density similar to east King County. Major increases in density in downtown Tacoma (maybe I missed it) and a decrease in density for Mercer Island, Bellevue and Kirkland (didn’t see that one coming). Maybe this reflects a shift in relative density but it sure raises a question mark for me as to how much you can relate the results between the two surveys. Not only are there major physical differences between a “trip” for someone in Silverdale vs. Mercer Island vs. Capital Hill but the very nature of the way they will tend to fill out the diary I believe will be different.
This isn’t a shift in density, just a different sampling size. What you see are households that completed a household survey. Not every household will complete a survey, but based on the total population obtained from a census for example, the results from the households that were sampled can be expanded to a value for that particular area.
I know this won’t go over well here — but the notion that commute trips are more important than non-work trips, or that battling congestion is more important than other transit objectives doesn’t sit well with me. That’s a car-oriented approach. You can’t just put that out as a truism.
When I think of great rail cities I’ve lived in or visited, the most important advantage that rail has provided is to be able to make *all* my trips on transit rather than only my work trip. It’s a no-brainer to use transit (whatever mode) to get to work. It’s another thing entirely to tour a city by transit (needs a network and stops wherever pedestrians want to be concentrated), or to use transit to visit friends across town or take your kids to school. For those trips, the visibility, frequency and simplicity of rail give it an advantage that isn’t as important in the decision to use transit to work – the same trip at the same time every day.
I don’t particularly care, living mostly in cities or metropolitan areas, whether rail extends 40 miles out – it doesn’t add a lot of value over buses. The place where rail adds value, in my experience, is for transforming the way I can make all the other trips in life – the non-work trips that make up the vast majority of trips, tours, whatever. And the cities that have done that particularly well have adopted a dense urban network rather than a model of getting to cities 40 miles away.
What you put forth is altruism. It’s the commute based trips that make transit economically feasible. The reason a tourist can go to London and take the tube anywhere in the city is because that system gets millions of people to work everyday in a density where cars simply can not work. Of course high capacity transit also benefits from consistent demand so enhancing off peak trips where ever possible through routing or TOD is synergistic.
No, the big takeaway from all of this is that work or school trips are the important measure. People tend to make those trips 5x/weekly, and if they’re using transit, they’ll probably do some of their errands on the way to and from work. This is where you get 80% of your results in moving people from 20% of your effort.
Transit relieves congestion, for the person using the transit. Regional society, which is paying the bill, does not experience this relief- road usage rises to congestion levels that act as an informal tax, and stays there. Congestion is just a characteristic of road vehicles, and should more properly be treated as a turbulent flow condition.
If you were to read the history, you would find that the cities with dense networks did not build dense networks- they simply built one line at a time, over and over again. Naturally, there will be a few exceptions to this rule, but that is generally the case.
It’s important to remember that your ability to make all your trips by transit is largely the result of destinations being built where the transit goes, rather than transit going to the destinations. The instances of a tram stop being built for the Opera House are vastly outweighed by the housing and commerce built near the tram line.
Finally, the ability to ride transit to work makes or breaks the deal for most of us. If you need a car to get to work, you’re probably going to own a car. If you don’t, it becomes a dispensable luxury.
I don’t agree with your view of history. Yes, rail cities build one line at a time, of course. But few feel that they need each line to get 40 miles away to the next city before they start on the next line in their network. That’s a political deal that is unique to the Puget Sound, in order to distribute the money around a 70-mile long service area encompassing three cities. There may be one or two cities in the world with 70-mile-long light rail lines, but in most cases that is well into the domain of commuter rail and express buses — since the station spacing required to make rail competitive over that distance is a lot longer than what is considered desirable for light rail.
It should be pointed out that Link will get neither to Tacoma or Everett before we get started on the next line.
I totally agree with Greenwood Rider, and that’s what I meant above about how you can’t put a car in your pocket and take it out later. When you have a chain of trips to make, you have to do them all by car or none by car. You can drive to work, to the store, to home. Or you can ride transit to work, to the store, to home. But you can’t drive to work, ride transit to the store, and drive home, without going back to work to get the car.
Peak-hour transit to work sites is essential, and it increasingly attracts suburbanites who are willing to ride to work but not to use transit otherwise. Though after time they’ll start using transit for a few other things, especially going downtown, or when their car breaks.
Off-peak transit is important, not just for shift workers and carless households, but so that people can use transit for multi-trip chains. If the bus goes everywhere I want to go today and tonight, I might ride it. If it stops at 9pm or doesn’t go where my evening activity is, I’ll drive all day, even though I could take the bus to work and back and then get out the car. Or maybe I can get a bus to the activity at 7pm but not back home at 10pm, so I’ll have to drive both ways.
London’s tube was built primarily to get working-class workers to jobs, but it was consciously extended to 18 hours a day to make carless living feasable. (Of course, motor cars didn’t exist when it was originally built.) Likewise, the PATH trains from NYC to New Jersey are called “commuter trains” but they run 24 hours every 5-15 minutes, making a carless life feasable. The Caltrain in SF runs one an hour till 10:30pm, and every two hours on Sundays. Not so good for carless living. And Sounder runs only peak hours: forget it unless you’re going to work in those hours.
I’m with John Niles on this one – PSRC needs to come up with a model that doubles work trips AND non-work trips. Based on the experience of just about everywhere outside the US, much higher transit mode shares are quite feasible. And since 80% of the total trips are non-work trips, we need to provide people reasonable alternatives for these trips to make a region-wide dent in VMT and greenhouse gases.
The PSRC alternatives assume minimal transit improvements beyond baseline (ST2), and may not accurate model changes in land use, and that is why they don’t deliver the mode share.
Some ideas:
– Buses running at 10 minute intervals on all arterials in the UGA
– A frequent, two-way Sounder network for longer distance travel (incl. Eastside rail)
– An extensive light rail network (i.e. more than a single 60 mile spine) that goes to most sections of the city and most suburbs, for medium-length trips.
– Supportive land use and street connectivity
Expensive, yes. But expensive compared building highway infrastructure for all of these trips throughout an ever sprawling metro area?
With an integrated system like this it becomes convenient for many throughout the metro area to use transit (or to walk, bike or unicycle) for everyday errands. And no, a clean, well-lit low-floor bus that runs every 5 minutes and goes exactly where you need to go is not intrinsically less attractive than a car.
Land use modeling is not my forte, but I was surprised to find that the variations in land use observed in each scenario were very minimal. In the PSRC Transportation 2040 DEIS Appendix D, the graph on page 77 (79 in the pdf) showing the variations was really surprising. There is virtually no variation in land use regardless of the alternatives. It’s hard to say whether the land use growth model is insensitive to transportation infrastructure or if the land use development strategy is so rigid that growth will occur in these ways regardless of what infrastructure is built.
There is no shortage of opportunity to implement bus lines that will receive high ridership. What there is a shortage of is funding to create these routes. Public transit is subsidized to the point that whatever you build (if it’s not just plain stupid) will attract near capacity use during peak commute hours. Right now buses are the best value on attracting new transit users, reducing VMT and offering an alternative to congestion. Sounder is great (well, at least south of Seattle). Link should work out fine; replacing bus miles at a savings, improving air quality and hopefully attracting new riders and aiding in urban renewal along much of it’s route. But better bus service is the bread and butter of increased transit mode share.
Bernie, you statement obviously depends on what “high ridership” is, but my view is that you’re pretty much maxed out unless you’re talking about qualitatively improved service like RapidRide.
The next quantum leap in ridership is going to come from quality (RR, rail, whatever) and better land use.
My viewpoint is obviously eastside oriented but mainly what I see is chock full P&R lots. I’m not convinced expanding any more of these with expensive multilevel parking garages makes sense but I think there’s lots of opportunity to buy more outlying land on say the SR-522 corridor, SR-202, and such which would fill if we had the funds to support them with more bus routes. I can’t really think of anywhere except Aurora/99 from Shoreline north that’s really ripe for true BRT. The Transit/HOV lanes on 405 and SR-520 need immediate attention. I don’t think it’s much more than making it HOV 3+ on 405 as soon (or sooner) as they finish the current lane expansion projects. On SR-520 it’s the switch to an inside HOV lane that’s required. I was told by WSDOT reps at the last open house I attended on the SR-520 corridor that this was partially funded and would proceed independant of the bridge construction. They were hopeful of a 2012 time frame.
I wonder if there are any opportunities for neighborhood feeders to take the load off some of the P&R lots? Obviously with full P&R lots the choice is to either build more storage for cars at great expense or make it so at least some suburban transit riders don’t need to drive to the nearest P&R.
As for Aurora/99, I suspect there is enough demand to justify going straight to rail. Besides rebuilding it along the same lines as MLK was rebuilt (though with more through lanes and possibly with an elevated guideway and stations) would do a lot to make the street much more pedestrian and bike friendly. On top of that due to all of the auto oriented businesses Aurora is relatively low hanging fruit for TOD.
Maybe for some of the major close in lots like Eastgate, South Kirkland, Mercer Island. But the ongoing cost of running the bus service (~$6 per boarding would be my guess) gets you pretty close to the cost of funding a bigger garage. At some point I’d hope that there would be a fee for parking. Charging a reduced fee (or free) for carpools would help free up spaces.
It’s the “far out” locations for new P&R lots I was thinking of. Mini lots of a couple hundered cars would support Eastside routes into Bellevue for transfers or perhaps direct to Seattle. The advantage is you intercept traffic before it gets to the most congested points. Several small lots along say SR-522, SR-203 or SR-202 would be sort of like local service rather than requiring a drive into Woodinville, Mill Creek or Redmond. I think it’s also easier to locate small scale developments. And land always seems to be a good capital investment. Maybe some of lots along the BNSF Woodinville Subdivision could someday become train stations.