Diagram of proposed I-5 HOV lanes

One of the quirks of I-5 is the lack of high-occupancy vehicle (HOV) lanes between I-90 and Northgate. Instead, there are reversible express lanes. This may have made sense in the 1950s, but it’s much less useful in today’s world of frequent two-way buses and dispersed urban centers.

Even the Washington State Department of Transportation (WSDOT) recognizes the problem. From their website:

HOV lanes do not continue on the mainline where there are Reversible Express Lanes. At the time of design it was thought that Reversible Express Lanes would serve the HOV needs of the area between Northgate and downtown Seattle. HOV needs for this area are now being re-examined.

As part of East Link, WSDOT is building two-way HOV lanes along the center of I-90. But the I-5 express lanes are not disappearing, and so unlike with I-90, there isn’t an imminent need to replace lost capacity. In addition, I-5’s frequent left exits make it harder to design a free-flowing HOV lane.

But hard doesn’t mean impossible. It can be done, and all it would take is a bit of paint, signage, and Jersey barriers.

Southbound, the existing left-side HOV lane would be extended along the missing span, between Northgate and the downtown express lane merge. Single-occupant vehicles (SOVs) would be allowed to enter the HOV lane to enter or exit the freeway at the Northgate express lane entrance, the left exit onto SR-520, and the left entrance from Mercer Street. (The 65th Street left entrance is already HOV-only.)

Northbound, between I-90 and Northgate, an HOV lane would be created on the right side of the freeway. The lane would start at the I-90/Dearborn northbound entrance ramp. Jersey barriers, starting along the ramp itself, would prevent vehicles from entering the HOV lane directly from the I-5 mainline, or vice versa. (A similar situation exists at the southbound I-5 entrance ramp at NE 65th Street.) This protected lane would continue northward, ending just a little bit north of the Mercer Street exit. HOV traffic would use the rightmost collector-distributor lane where it exists, and the rightmost mainline lane otherwise, protected by barriers in either case. As a consequence of this separation, the ramps at James Street, Madison Street, and Olive Way would be HOV-only. Conversely, vehicles in the HOV lane would not be able to access the left exits at Seneca Street or Mercer Street. After Mercer Street, the physical separation would end, to allow SOVs to use the right-side exits. After the N 85th Street exit, the HOV lane would again be physically separated, forcing HOV traffic to either exit at Northgate or to reenter the mainline. North of Northgate, the regular left-side HOV lane resumes.

All these physical barriers serve two purposes. First, they make it harder for SOVs to opportunistically use the HOV lanes when they think they won’t get caught. Second, they discourage drivers from rapidly crossing 4 lanes of traffic to get between the left-side HOV lanes south of I-90 (or north of Northgate) and the right-side HOV lanes in downtown Seattle.

Right-side lane restrictions aren’t always appropriate. For example, on SR-520, the long queues to enter I-405 made the right-side westbound HOV lanes nearly useless. What’s different about I-5 is the frequency and importance of its left-hand exits. The distance between I-90 (and the northbound express lane entrance) and Seneca Street is about 1 mile, and it’s just another mile to Mercer Street. Modern freeway design manuals suggest that exits should not be spaced more closely than a mile, and I-5 northbound covers that spacing with left-hand exits alone. Separating transit and emergency vehicles from the mounds of downtown traffic exiting the mainline highway on the left would be much more productive than separating them from the much smaller number of vehicles exiting on the right. In addition, once these buses leave downtown, they will eventually exit the freeway on the right, whether at SR-520 (e.g. the 545), or NE 45th Street (e.g. the 73), or 65th (e.g. the 522), or Northgate (e.g. the 41). A right-side northbound transit-only lane means that these buses can stay in one lane, rather than having to cross 4 lanes of high-speed traffic.

The benefits of this plan would extend beyond the freeway. Today, James Street is a parking lot during peak, and Madison Street and Olive Way aren’t much better. Without general-purpose I-5 ramps, these streets would have a lot less traffic, making them more useful as urban arterials (and for the HOVs that could still use these ramps). Instead, SOVs heading into downtown would exit northbound I-5 at Seneca Street, and SOVs leaving downtown would enter northbound I-5 at University Street. Like Mercer and N 85th (among others), University is cut off by I-5, making it useless for any purpose except entering the freeway, but perfectly suited for that role.

Finally, it’s worth noting that the James Street and Madison Street exits play a vital role for emergency response vehicles, which use them as a quick way to access First Hill’s hospitals. The proposed plan would guarantee that ambulances would never get stuck in traffic while trying to exit the freeway at these two ramps.

Undoubtedly, this plan isn’t perfect. The northbound lane would be be discontinuous with the existing left-side HOV lanes; buses that travel through downtown without stopping would have a difficult time making full use of both facilities. The loss of SOV access to James/Madison/Olive would make some car trips longer, and might create more traffic on local roads. But on the whole, I see these as minor problems. The bigger problem is the congestion that slows down buses between downtown and Northgate every single day. Even an imperfect workaround is far better than the status quo.

79 Replies to “I-5 HOV Lanes”

    1. A little bit of paint, moving of jersey barriers and changing of signs.

      It sounds so simplistic.

      Much of the barrier through the corridor isn’t pre-cast Type 2 barrier (WSDOT terminology)…aka Jersey barrier. It’s actually cast-in-place barrier. Barrier would need to be knocked out, new substrate would need to be built and then the pavement can be placed. Any barrier modification on top of structures may require further review from structural engineers to determine whether added loads from the new heavier barrier, meeting current NCHRP 350 TL-3 standards, would require further retrofitting or widening of the bridge or structure. In addition, impact attenuators would need to be installed at the leading ends in attempts of reducing crash severity (ie REACT 350, Universal Tau II).

      So…was anyone on the 301 this morning from Shoreline? Did anyone notice where it stalled out in the tunnel portion (underneath the Yale/Howell on-ramp to South I-5) of the exit to Pike Street near the DSTT entrance? I sure did. It blocked traffic for a short while. Barrier separating HOV lanes have shown benefit in reducing crossover collisions where drivers cruising down the HOV lane collide with vehicles entering the HOV lane from a near dead stop. It’s been an issue for motorcyclists for some time. Any incident in the closed off section of the HOV lane would render it useless and force traffic into the GP lane. The unfortunate barricaded HOV traffic would queue until the scene is cleared. While it would be great to enforce, emergency crews would have to double back to access a crash scene or access a scene over the barrier increasing the number of lanes taken, thus increasing traffic delay. Not to mention, barrier may eat up existing limited shoulder width as barrier is not only 2 ft wide at the base, but it requires 2-3 feet of shy distance (this is described in the WSDOT Design Manual Chapter 1610). Not to mention, shoulder deviations for this type of reconstruction would need to be submitted to FHWA for their buyoff and that’s no guarantee.

      Many of the sign bridge truss structures through the urban core are built in the 1960s of aluminum. They’ve exceeded their life expectancy and some have been removed even though replacement funding has not been secured. Sign bridges and sign cantilevers must be designed for a specific windloading (sign square footage) at 90 mph. As such, the cost of a single sign bridge range around $300k. Sign bridges mounted on structures can bump that figure higher as foundations must be doweled/cast into an existing structure or fabricated with new construction. More than likely many new sign structures would need to be built as drivers will need ample time to enter and vacate the right (or left) lane to make an exit.

      What is being proposed southbound would invariably produce a double lane drop at or just south of Northgate and a wicked lane shift south of the on-ramp from SR 520. Additionally, there would be some gnarly congestion where the southbound general purpose lanes would be necked down to two lanes by NE 85th Street in Green Lake. Southbound backups when the Express Lanes are northbound could extend well into Shoreline and Mountlake Terrace. There is limited shoulder on both sides and what shoulder exists most likely would not be able to sustain traffic bearing loads. Full reconstruction and barrier protection of the retaining wall supporting the terraced Express Lanes would be needed.

      The issues related to James, Mercer, Olive, Madison and Seneca are a classic issue of demand versus capacity. You’re talking about large volumes of traffic entering a city grid from a limited access facility. The city grid is already congested with ped movements limiting turning movements at many of the signalized intersections in the urban core. I see all of these routinely ramps back up everyday. No matter how much you try to widen the ramps, you’re going to have an issue. Signal timing changes could improve things (all ped/barn dance phase), but signal changes in the CBD could have a detrimental halo effect.

      Now I’m not suggesting that HOV facilities between Northgate and Downtown isn’t needed, but I think this could be evaluated a little differently. Barrier separation can have bad repercussions and might not meet the demands of FHWA (USDOT).

  1. I don’t see this happening, as WSDOT is all about keeping exit backups off the mainline, and I’ll bet their modeling would show that Seneca would suffer unacceptable backups if James and Madison closed to general traffic.

    Converting the express lanes to two way high occupancy toll lanes is already in the long term unfunded improvement plan and would accomplish most of the goals of this scheme with the addition of ramps to 520 and full time off and on ramps at Lake City and 42nd.

    1. I like the idea of converting the express lanes to the feeder collector for 520, with no through north or south traffic past 520. It would take re-designing the interchange at Roanoke (no small task). But 520 would end up with it’s own ramps downtown. Since the buses will be evicted from the bus tunnel in 2016 they don’t need to use the express lanes any more.

      Northbound 520 traffic gets off I-5 at the current express lane ramp on the left. Downtown traffic heading to 520 uses the express lane ramps. Mercer Traffic to 520 uses the express lane ramps.

      Southbound 520 traffic shifts onto the express lanes at Northgate or in the Ravenna area where all three lanes are at the same grade. Reconfigure 50th or 45th to have a right turn on ramp mid-bridge onto the express lanes.

      Traffic from 520 northbound would merge back onto the main line in Ravenna or Northgate. Traffic from 520 southbound would use the downtown exits or would merge back into the mainline where the express lanes currently merge.

      You’d get rid of the Mercer-520 leap that causes so much disruption to traffic flow.

  2. Unless they’ve changed it in the past couple years, the 65th St. entrance is not HOV only. Access to the express lanes from 65th st. is HOV only, but the entrance is normal occupancy, is metered, and often the metered traffic blocks access to the express lanes for buses and HOV’s.

    1. Only the 65th St. entrance to the express lanes is HOV. The 65th St. entrance to the regular lanes are open to anyone.

      1. When the Express Lanes are configured southbound, carpools and HOV traffic can shift left and enter the express lanes. I think it was a snafu in the post.

  3. Interesting idea. I’d love to see what happened if you put this through a computer model that simulated traffic. Would downtown become a mess because people had to drive farther to get to their freeway entrance/from their freeway exit? Would it stay the same, but just shift traffic to other streets?

    I really like the idea of having all day, two way HOV lanes on I5, but i’m also worried about legibility. It would be frustrating to get on the freeway and always think, so wait, I can’t use these exists, but I can use those – which gets me there fastest? For downtown commuters, though, it that would probably not be a problem.

    Also, hopefully we’ll get reroutes so that we won’t need busses on this stretch once link is open to Northgate. 45th, 65th, and Northgate will all be link stops, and we can run high frequency, east-west only lines along those streets to feed into link. With luck, it will help us transition to a more gridded system.

    1. Computer models are just that–computer models. They rarely reflect actual behviours as people would ultimately respond to a new pattern. It might be interesting, but I wouldn’t get hung up on it, and neither should WSDOT.

  4. The northbound HOV lane plan through Downtown has a very bad side-effect. Shifting SOV traffic from the First Hill-side ramps to the Downtown-side ramps would significantly increase congestion there, since cars would now have to use Downtown avenues (5th & 6th).

    This would add to the current congestion that already spreads down toward 2nd, and would negatively affect the buses running on Downtown avenues.

  5. If current behavior is any indicator, that northbound section that’s “HOV only except for exiting traffic” will collapse by week two due to no enforcement.

  6. One smaller scale improvement that could happen now: physically prohibit SR 520 access from the NE 45th/50th ramps. People are shifting across 5 lanes of traffic in one mile, and it causes horrible backups every day. But I’ve noticed many times anecdotally — the most recent being my smooth trip on the southbound 512 yesterday — that I-5 mainline traffic is way better on weekends when 520 is closed.

    1. With respect, while I believe that the 520 weave from the 45th/59th on ramp is
      a problem–I am not sure that elimination of the possibility (enforcement?) of the
      phenomena is practicable. I fear it would not lead to any traffic semi0-nirvana.
      A giant problem, observable onalmost any day is the slow down arising from the
      combination of the Ship Canal Bridge design in which the roadway which, to
      southbound drivers, gives an appearance of being about to drop off into space
      or at least generates some uncertainty. A supporting role is often played by the
      electronic sign over the highway.

    2. That goes for northbound Mercer to 520 as well with the left (fast) lane onramp of Mercer to I-5 and the right lane connector to SR 520, most of the traffic north out of downtown before the ship canal bridge is a result of this epic fail.

    3. If the 45th St. ramp didn’t have access to 520, then how would drivers from Fremont/Wallingford be expected to get on 520? Without using the 45 St. ramp, they would have to slog it out through the U-district and get on at Montlake, an area that is already quite congested.

      Interestingly enough, the usage of the 45th St. ramp for 520 access isn’t just for cars. Several Microsoft Connector buses enter I-5 at the 45th St. ramp every day to get on 520. I’ve also seen route 555 buses use the 45th St. exit on occasion to bypass bad traffic. (Aside: it would be really awesome if the 555 could actually stop at the 45th St. freeway station. It would be great for transferees from the 44 trying to get to the eastside, and with normal rush-hour traffic conditions, the additional overhead of exiting the freeway to serve the bus stop would be negligible).

    4. One way you could stop the weave of traffic from 45th/50th southbound to the 520 on ramp would be to switch the 520 and Roanoke exits.

      The Roanoke exit would be excavated and dug under i5 south to connect to the existing tunnel. The 520 exit would ramp up and T to the Roanoke Street overpass.

  7. I get that 1960s thinking was that the main priority was to get SOVs to Downtown, and that the express lanes would suffice. But SR 520 was built around the same time and it seems designed for exactly the opposite. Whose idea was it to have a left-hand entrance onto I-5 near Roanoke, then have no left-hand exits? ALL SR 520–>Downtown traffic/transit has to merge across 5 lanes to exit at Mercer/Stewart/Union/James/Dearborn. I don’t want left-hand exits, but I also don’t want the insane merging patterns that SR 520 access creates. Am I right to think that the new I-5/SR520 connection (if ever funded built) will fix with right-hand entrance ramps and direct express lane access?

    1. Which bus routes (including ones not yet in service) do you expect to use SR 520 after Overlake TC Station opens?

    2. My recollection is that the idea for 520 was that it would be tied into the expressway that Mercer St was to become, which probably anticipated reconfiguring that interchange.

      I agree that the “enter from 45th, exit onto 520” traffic is a major cause of slowdowns on southbound I-5. I’m not sure how viable prohibiting this is, as the alternate route would push more cars into the already congested (and bus-filled) area around UW. Will the (eventually) reconstruction of 520 help this?

      Here’s an alternate proposal for speeding up buses along this route:
      1. Allow HOV (or buses only) to exit from the express lanes from the second lane from the right ontto Denny/Stewart (widening that ramp). I see way too many buses stopped in the right lane of the southbound express lanes during AM rush.
      2. Make one lane of southbound Roosevelt/University Bridge/Eastlake bus-only during the PM peak. In many sections, this would replace parking.
      3. During PM peak, Roanoke becomes bus-only between Bolyston and Eastlake. This would allow buses to exit the 520 at Roanoke, and proceed downtown by a bus-only route.

    3. 1960s thinking was that most traffic would come from Northgate/Lynnwood in the morning and return in the afternoon. 520 was not expected to have much traffic; that’s why it was four lanes without HOV lanes or a sidewalk. They underestimated the Eastside population explosion by a factor of three or more. In the 1960s the Eastside was like… Marysville. The 520 tolls were removed seven years early because growth exceeded expectations that much.

      1. I came here to write exactly this. You can pretty much date houses in much of the Redmond/North Bellevue area to the construction of 520.

      2. I suspect some observers might note a class bias in the planning of the
        reversible express lanes (access to and from downtown from the north
        end). The freeway route south of downtown did not pass through many
        middle class neighborhoods. You may recall that the “Blue Streak”
        express services introduced by Seattle Transit also served the north end.

      3. From what I’ve read Seattle was planning a freeway from Northgate to downtown, and 1st Ave NE was a ribbon of swamp so it didn’t displace many houses. When the Interstate system came and most importantly federal funding, they hastily added a south end extension along the west side of Beacon Hill. I don’t know if that side of the hill had houses or was buildable: it looks now like it was all trees. Perhaps people didn’t build houses because they didn’t want a view of the Industrial District. South Seattle was middle class (“Boeing Hill” and all that), but more multiethnic than the north end. White flight was just starting when the freeway was planned. The CD was considered a blight area, and in the 1972 Forward Thrust map it was the only neighborhood without its real name; instead it said “Model Neighborhood”, as if destined for some future urban renewal. In my opinion the CD and Rainier Valley are lucky the freeway didn’t run on that side. The west side of Beacon Hill is the best place it could have gone. The south end and southern suburbs had a lot fewer people than then north end, and they also have more highways (509, 99, 181, 167), so that’s probably why express lanes were considered unnecessary.

        When I moved to Bellevue in 1972, downtown Bellevue, Kirkland, and Redmond were like Marysville is now, and Crossroads/Overlake where I lived was like Smokey Point but with a few more department stores: Sears, White Front (now Fred Meyer). The suburban ring was Lynnwood – Bothell – Redmond – Somerset – Renton – Kent – Federal Way. Tacoma and Auburn were separate metropolitan areas: not many people commuted to Seattle from there. Everett was separate except for the cross traffic to Everett Boeing. Issaquah was a small town.

      4. That sounds about right to me as well, Mike. I’m a little over fifty and grew up in Seattle. I agree with just about all of your assessments on neighborhoods, with a couple minor quibbles. While some folks in power surely considered the central area a blighted area (after all — lots of black people = blight was a pretty common assumption) I don’t think that attitude was shared by many in the city. I have no idea why the Forward Thrust map called the area “Model Neighborhood”, but neighborhood names are fluid, and change quite a bit. Growing up, people used the terms “Central Area” and “Central District”, but they used “area” a lot more (e. g. “CAYA” was short for Central Area Youth Association). I always considered Capitol Hill a subset of the Central District and few would say that now. Partly it is because of gentrification, of course. Schools like Minor and Meany were 90% black (as was the surrounding neighborhood) but you sure wouldn’t guess that now. Despite the obvious results of such redlining, the area wasn’t as run down as similar areas in the country. Part of that was due to the geography. To the east and west you have million dollar views. To the north you have Montlake, which is right across the water from the UW. The public schools were good, which meant that a school like Garfield, for example, had kids whose parents taught at the UW, other parents who had million dollar homes as well as the folks that weren’t as well to do. It helped that the city invested in the school as well — hiring some extremely bright, young teachers around that time. All of that meant that the most “inner city school” in the city was not only good in basketball, but good in scholastic competition as well (chess, debate, music, etc.). Maybe some of the investment that might have gone towards tearing down buildings (AKA urban renewal) went instead into the schools or community programs (e. g. that CAYA I mentioned earlier). Another sign that while Seattle sometimes flirts with some really stupid ideas, they often come to their senses eventually (I wish I could say that about the 99 viaduct of course).

        As far as the freeway route is concerned, I think there weren’t that many houses there just because of the slope. That being said, I seem to remember some “Now and Then” pictures showing that hillside with some houses. The freeway took down a lot of good neighborhood land on both sides of the freeway, so I don’t really see any cultural or class bias with the design.

        I agree completely that the Central Area was lucky that they didn’t run the freeway through there. Of course, if we had been even luckier, they would have skipped the freeways through the city altogether (as they did with Vancouver B. C.).

    4. Am I right to think that the new I-5/SR520 connection (if ever funded built) will fix with right-hand entrance ramps and direct express lane access?

      My understanding is that those have been considered but the answer is basically “no” on the second and “maybe” on the first. There’s no way to get to the express lanes that far south without a massive, I-5 shuttering, risky tunnel project.

      1. The FEIS shows that the preferred alternative is the same ramp configuration as today, but with the addition of a direct access HOV ramp connecting the the express lanes to the south.

        In my opinion, this is where the real Mercer mess is and nothing will be done to fix it.

      2. I thought they had ruled that out in the DEIS, but I know people were really adamant they look at it. Good for them, sometimes it feels like WSDOT is almost capable of listening.

    5. My understanding is that the left-exit from I-5 to 520 was necessitated by space constraints, and building a flyover ramp from the right would have significantly increased the number of homes that would need to be leveled to build the interchange. The current design, by freeway interchange standards, is actually remarkably compact.

    6. 1960’s thinking was that most 520 traffic would use the Thomsen Expressway and that traffic from 520 to downtown would use the Bay Freeway.

  8. Seneca would get crushed – the buses that exit at Seneca would have to switch to the SODO busway route to access downtown. SOV access from I-5 to Capitol Hill would be extremely difficult. Dearborn and Lakeview would have to assume most of the displaced SOV traffic heading to Capitol Hill. Right now that traffic can choose Dearborn/James/Madison/Seneca/Olive/Lakeview which helps to spread out the load on any single street.

    Ultimately, it means a lot more cars on other surface streets getting to/from I-5 which will shift much of the traffic from James, Madison, and Olive to those streets.

  9. Finally, a way to keep Capitol Hill all to ourselves!

    Seriously though, one of those three first/cap hill exits would have to be open to SOV. I guess Olive, since it’s last and isn’t on the collector.

  10. If the southbound HOV lanes were on the left the 512’s access to the 45th St freeway station would be as difficult as its access to the 145th St freeway station, except in an area with a lot more entrances and exits. Forward-peak buses (510, 511) would now be physically able to make it to the exit, but it would probably be a bad idea given typical traffic conditions, so service patterns would probably remain the same as today.

    As around 145th (a right-side freeway station in an area with left-side HOV lanes), drivers on the 512 might choose to keep right if traffic is moving quickly, but use the HOV lane as much as possible in heavy traffic. When there’s heavy non-forward-peak traffic into Seattle (most Friday evenings, many games and events) it would be really tempting to skip the 45th St stop, though it is fairly important to general network connectivity for now.

    1. I don´t see I-5 being rearranged faster than Lynnwood Link gets built. Once Lynnwood Station is open (planned for 2023), I don´t expect any buses to be running on I-5 between Lynnwood Station and downtown Seattle.

    2. Certainly, but tons of 200s, 300s, 400s, 500s, 800s, and other local expresses enter and exit I-5 at various points. They only have to begin a merge at most one-half-mile before, so there’s no point in trying to consider all scenarios–it won’t be perfect. But I can tell you that bit from 145th to 45th on the 512 would cut off 5 to 10 minutes each way during the peak–even with advanced merges to the ramps. It’s that bad. Besides, it ain’t just buses we’re trying to move. Vanpools, long distance buses like Greyhound and Bellair, and carpools.

  11. :) You know you have my backing, Aleks! I didn’t expect the northbound solution though, that’s an interesting modification on the concept of mainline HOV.

  12. Two things:

    This may have made sense in the 1950s, but it’s much less useful in today’s world of frequent two-way buses and dispersed urban centers.

    I-5 traffic is still vastly weighted toward downtown during the mornings and away at night. In fact, the same is true for the combined 520-I-90 when you count people instead of vehicles.

    Second, is this to add new lanes or re-purpose existing ones? I-5 South right about 85th is a disaster because of the four lanes-to-three merge. This would be made worse if one of those was to be made an HOV lane – unless you actually made the four-to-three squeeze happen farther north.

    1. It may very well be weighed that way, but the directional express lanes mean, at least from my limited experience, that traffic speed is usually far worse going the opposite of peak in the area being discussed.

      1. It’s not just because of the express lanes. It’s because 520 doesn’t connect to the express lanes. If it did you’d see the same slowdown in the express lanes as you do in the regular lanes. Lucky for Northgate!

    2. The four-to-three squeeze would have little bearing on HOV. HOV already carries over 40% of the people. You’re just inducing them to not drive solo by this change.

      Also, the more accurate statement is this: “traffic is weighted toward Downtown”. This is regardless of the “peak direction”. Traffic into Downtown in the “reverse peak” during the PM is usually as bad–if not worse–than the “peak direction”.

      1. Right–I’m suggesting there would be a three-to-two squeeze–which is irrelevant. Most of the capacity is in the form of HOV anyway. My point is why does it matter that there is a squeeze at all? An HOV would be unhampered by such a squeeze–and it would induce more to become HOVs. This is actually why I think Aleks should go a step further and advocate for HOV+3 like SR-520.

      2. In practice, a 2+ HOV lane restriction is insufficient to provide any kind of meaningful congestion relief. For example, witness the carpool lanes on I-90 or I-405. The problem is that if 1 in 5 cars is already carrying 2 or more people, designating 1 in 5 lanes as HOV accomplishes nothing in terms of congestion relief – the 1 in 5 cars eligible for the HOV lane simply move there, while the other drivers take the other lanes, and the HOV lane experiences exactly the same amount of congestion as the rest of the roadway.

        This is especially true when you consider that a 2+ HOV restriction includes a lot cars that are not doing carpooling by any stretch of the imagination. For instance, every car trip you will ever take as a child will, by necessity, have 2+ people in the car.

  13. From a transit point of view, the plan above involves a very complicated highway chiefly for automobiles- which may or may not work.

    What transit needs now- and has needed for decades, has been a two-way bus-only right of way from Northgate to Convention Place- or failing that, the 9th and Olive ramp to the CPS staging yard.

    By eyeball, I can’t tell if the “cut” leaving CPS staging for I-5 is wide enough for 2-way bus only traffic. Would be good if tape-measure cooperated. If not, bus-only lanes might work on Fairview between Mercer and CPS.

    Legit point about length of time any rebuild at all might take versus time ’till NorthLINK really cuts back transit traffic on I-5. But two things to remember:

    Several miles to dig to get to Northgate, and a lot of elevated to build to Lynnwood. Schedule adherence is bad enough on the surface.Underground? The waterfront DBT machine will be really late into its next station.

    And if EastLINK figures into this calculation- there are also one or two unknowns about what’s needed where a floating track meets one bolted to the shore.

    Also, when NorthLINK takes over the North, Convention Place Station and the Pine Street section of the DSTT will be done with underground transit.

    Would be good if whatever development is planned there refurbish the cut for a pedestrian passage to Westlake Station. If not still needed for LINK turnback or other rail needs. Whatever I-5 lane setup in decades to come, might not be bad to keep access to I-5, if only for utilities or emergency.

    2023 is nine years away, with some serious engineering ahead. Recalling the nineteen years it took between DSTT opening day and LINK opening, I wouldn’t wouldn’t write the train schedule right now- I mean write the logarithms and app the smart phones. Which by then will be same as cans with string are now.

    Mainly fed up with transit’s habit of using warp speed future service as an excuse to leave current service lame for a glacial age. It wasn’t either Metro or Seattle Transit that built a major freeway to be obsolete the day it was built. But over all those years since, transit’s effort to get a few miles of southbound lane hasn’t seemed to leave any lingering smell of sweat.

    One more try might be worth it for the exercise. Because if the present Depression doesn’t ever improve into the “mild recession” King county fears, resulting return to living conditions of late 1929 could motivate similar political will for remedy. Including some Civilian Conservation Corps projects for transit. ‘Tis an ill wind…

    Mark Dublin

    1. I’m not too worried about ST running into a show stopper on North Link. The soil is far less challenging and the size of the TBM used a fairly common one compared to Bertha and the waterfront tunnel.

      Lynwood Link is even less of a challenge.

      You are right that East Link has a fair amount of risk with the bridge, but if they can make rail work on suspension bridges there is no reason to think a floating bridge is any more of a challenge.

      I believe the plan for the CPS site is to expand the existing convention center.

      I-5 HOV is worth it simply for the vs pools and other HOV users, though the two-way express lane HOV idea might be a better investment.

  14. I-90 SOVs heading downtown would basically have a choice of 4th Ave S or Mercer at extreme opposite ends of downtown, and your diagram makes the northbound HOV lane look worthless between Mercer and 85th.

  15. The Southbound HOV would require a complete re-laying out of the lanes. The existing entrances from 520 and Mercer don’t merge with the mainline, but rather add a new lane on the left-hand side of the roadway. So without re-striping the entire length from 520 south, the new HOV lane would be required to shift a lane left twice, and both times just as traffic was trying to merge with them.

    1. That isn’t necessarily a problem if you actually end and start the carpool lane and force a merge yield for vehicle entering the highway from the left. It’s not necessarily conventional, but then neither is right merging from lefthand ramps. I’d say force the carpool left with the “new” lane gently and “end” the carpool lane in its mainline–or what is currently a generous purpose lane. You can put markings and barriers to “hold” the lane until the total merge is complete for both the ramp
      and HOV lane, and then allow weaving for non-HOVs out into the” new” general purpose “held” lane. Alternatively, you could just do the merge and create a new outer lane on the righthand side, but that has a more potential weaving issues with it.

  16. I highly enjoyed this thought experiment, and so I’m loathe to point out any more flaws, but…

    The very first thing that caught my eye was that the plan pushes all Mercer-bound northbound traffic from I-90 back into the general-purpose lanes for the entire multi-mile interim stretch — and likely all the way back to Rainier. These general lanes are likely to become even more troublesome as a result of the HOV taking and the physical barrier.

    I know this is irrelevant for I-90 bus traffic, since they all take the direct ramps to 5th, but it strikes me as counterproductive to incentivize HOV usage for other 90-to-5 users, but then penalize carpooling to Mercer by merging those particular users into every worsened downtown trouble spot.

    I hope there is a way to refine the plan to address some of its above-noted flaws, because bi-directional HOV and transit pathways that actually work are already decades overdue.

  17. Instead of trying to control and limit access along the outside edge of northbound I-5, put the effort into reworking the southern end of the express lanes to accommodate two way flow with a lane in each direction? Once you’re north of the 5th AV/Columbia onramp to the express lanes, then you could run two way traffic in existing lanes, with NB exits just north of the ship canal bridge, convert the 65th street access to a NB offramp, and then again at northgate. If there was a need, there could be an offramp/ flyover added at Lake city to complement the existing one. NB onramps would be 5th/Columbia and Mercer.
    Southbound that means there would be onramps at Northgate, Lake City Way, and one could probably be figured out from 45th street connecting to the north end of the ship canal bridge. SB offramps downtown would be Mercer, Stewart, Pike.

  18. Not quite old enough to personally remember bankers jumping off buildings- meaning days when profession allowed enough conscience to do this. But old enough my memory can’t handle antique agency names.

    Really like the idea of guys in undershirts swinging pick-axes to get those lanes into CPS. But WPA probably would have done it. CCC did parks and trails. So maybe they could extend Freeway Park through whole CBD- like Jim Ellis intended.

    But following the example of the whole Interstate System, transit projects should all be handled by the Department of Defense. That way, the farther to the right Congress goes, the more lanes and track we’ll get- even if we won’t like putting Biloxi in ST3. Subway trains can haul flatcars with missiles too- didn’t railroads do that for awhile?

    And best of all NSA should start harassing anti-transit people but not really caring if any of them join ISIS. Because great PR coup for transit to have these people on Facebook dressed in murdering-desert-guerilla uniforms and screaming that transit is Satan’s own commuting mode. Probably both groups’ sole example of true belief.

    And these US-born fanatics will also be correct that transit threatens ISIS’ main source of funding: oil sales.


    1. Mark- If transit projects were defense projects we’d have subways to ever suburb! :) I am with you on this one 100%

  19. I’m very glad that WSDOT is considering this–some form of transit priority on the chronically-congested I-5 corridor has been needed for a long time now.

    Question: would the southbound 71/72/73 be rerouted to I-5 when the express lanes are not open? It would be a bit tricky since buses would then have to negotiate surface traffic to get from Campus Parkway to I-5, and then they would need to cross all lanes of I-5 twice to utilize the HOV lanes.

    1. There’s no reasonable way to accommodate that unfortunately. They would continue as Express limited-stop on surface streets. Plus, the Express 70 series will be greatly changed by the time U District Station opens anyway–changes are likely planned for the series due to UW Station in 2016 as it is. Aleks’s solution for other routes is not a solution for the 70 series, and that’s okay. Getting back to 45th is murder and trying to enter and exit from the Boylston interchange isn’t a time-saver.

    2. The northbound 70s use the I-5 regular lanes from at least 6am-7:30am and 9am-11am. The freeway is sometimes clear, sometimes a 520 slowdown, and sometimes a full-width slowdown. But even when it’s clear the 45th segment can be the worst part, from the exit ramp to Roosevelt. Sometimes that alone takes ten minutes. The HOV lane would help on the freeway but 45th also needs to be addressed.

      Southbound, I don’t think the 70s ever use the regular lanes. I don’t know why, because northound is significantly faster than Eastlake. Maybe it’s because of two left turns on 45th. (Northbound has two right turns.) So with the HOV lane it would still have to make two left turns, and then cross all the freeway lanes. I imagine Metro never wants buses crossing from right to left or vice-versa in the inner city or just after entering; it’s just asking for an accident eventually.

  20. Closing all right-side exits in downtown (to anyone who isn’t an I-90 HOV) will cause epic traffic problems downtown, as anyone who want to go east will have to go through downtown streets. Also, there isn’t enough room to fit both a HOV and general purpose lane side-by-side at the north end of the collector, is there? It merges down to one lane currently…

  21. Factor nobody’s mentioned: look at the number of times just these last few weeks when weather, single-or-few-vehicle accidents, or just too many cars made everything coming down I-5 from Everett into a linear bathroom-free prison for thousands of people.

    On the other end of the region, before 6AM I-5 is a carpet of tail lights. And one lane blocked anywhere near the Nisqually River…I know how to get around it, enjoyably, and also faster than everybody that isn’t moving. But anybody else will have to pay a lot of gold to the old guy with the hat, and huge jugs of mead to all those dwarves.

    Region-long, this ain’t your grandfather’s I-5. Pretty soon it won’t even be your older brother’s. That’s why I’m not wasting any worry about cars getting stuck. Are there, doing that. But for sake of both people whose bosses value arrival time more than method, and for those who still need cars- those bus lanes might get the whole first group out of the way of the second.


  22. HOV lanes can be just as slow as mixed-flow lanes. Drive around the LA freeway system and you can see many examples of that!

    I am also mystified how this post relates to transit service, because the I-5 north corridor is going to have light rail in a few years and surely many of the buses on I-5 today are going to disappear. I’d even suggest that we should instead be advocating for major street bus-only lanes around new Link stations to prevent feeder buses from getting caught in local traffic around some of the interchanges.

  23. One thing WSDOT could do right now at no cost would be to modify the schedule so that the I-5 express lanes point southbound all day on weekends. Because of the 520 weave, I-5 south has less capacity than I-5 north with a fixed number of lanes, so routing the express lanes southbound when traffic is roughly equal in the two directions makes sense. It wouldn’t do anything for weekdays, but it’s at least something.

    As to what to do for weekdays, I wouldn’t spend too much money on this since in another 7 years, buses between downtown and Northgate will be replaced by Link, so it won’t matter nearly as much.

    As another aside, the proposed connection from 520 to the I-5 express lanes is bad, and will accomplish little except to slow down the express lanes.

  24. This is a very interesting post, but I agree with this comment: https://seattletransitblog.wpcomstaging.com/2014/11/17/i-5-bus-lanes/#comment-557961

    Doing this would take a lot of time and money (not because the physical work is that difficult or time consuming, but because the planning is) and might not be done before North Link. I see no reason for a bus to go beyond Lynnwood on the freeway once Link is done. So, from a transit perspective, this would be nice, but hardly worth it.

    That being said, I think car pooling is a legitimate transportation option, even if it isn’t as popular as it used to be. But I think redoing the HOV lanes should occur when we redo the express lanes. Once North Link is built, HOV lanes on the express lanes won’t be nearly as important. Plus, as mentioned, traffic is more bidirectional, so express lanes don’t gain you that much anymore. The city would rather have them be regular purpose (two way) lanes. That is a big change, and will be quite disruptive. But it opens up a lot of possibilities, such as eliminating the HOV lanes on the express lanes, but adding them on the general purpose lanes. Changes such as those mentioned on this post should be considered as part of that.

    Then you have 520. 520 already carries lots of buses. But when I-5 and I-90 take a lot less (because of Link) the importance of 520 based buses increases. Since Sound Transit didn’t consider a 520 bus connection to North Link (oops), Metro (and Sound Transit) are left for several challenging options in connecting the two (force people to walk from the freeway to Husky Stadium, run the buses downtown, run the buses to a more convenient station to the north). No one knows exactly how 520 will connect to I-5, but this connection (especially from an HOV standpoint) is way more important than HOV service coming from Lynnwood.

    From a planning and construction standpoint, it would make sense to build things in this order:

    1) Build North Link
    2) Rearrange the express lanes and I-5 mainline at roughly the same time 520 is completed.

    Unfortunately, the 520 work will probably occur first. This leaves this approach:

    1) Plan to change the express lanes, mainline I-5 and 520.
    2) Make the 520 changes.
    3) Build North Link.
    4) Make the express lanes and mainline I-5 changes.

    Again, a lot depends on that “Link to 520” connection. How exactly should that work? Should buses exit at Montlake and turn around? How easy is that to do? If not, should the buses continue to downtown? How would that work? A bus could go north as far as 65th or Northgate (mainly to turn around). In both cases there are good exits and good ways to turn around.

  25. This seems fine, but it doesn’t go far enough.

    Better, to make the current reversible HOV lanes bus only, signalize where there is crossing traffic, and make them full time two-way.
    Add a northbound convention center exit,

    No need to do anything for 520, since UW link and East link make 520 buses on I-5 redundant.

  26. Won’t express bus service to downtown Seattle still be needed once light rail has reached Lynnwood and Redmond?

    Seems like a one seat express ride from the suburbs will still be faster than light rail with its stops and lower top speed.

    Buses will use I-5, I-405, I-90 and SR520 forever.

    1. Some of the 520 downtown expresses might stay. The I-90 express busses will likely be truncated at Mercer Island P&R.

      From the north there is no reason to continue to run express bus service. Link nearly matches free flow speeds between Lynnwood and Westlake. During peak it isn’t even a contest as Link won’t be getting stuck in traffic.

      Not to mention that eliminating downtown expresses frees up a ton of service hours for Link feeder service.

    2. Lynnwood is 28 minutes, Everett is an hour; these are the same as ST Express in spite of the stations. That’s the advantage of grade-separated rail. East Link should be marginally faster than the 550; that should be a wash with the 554 if it’s truncated. ST, Metro, and CT are already leaning toward truncating all north I-5 and I-90 routes, so it doesn’t matter if people want them unless they step up to fund them. But when you start talking about the 520 routes, south I-5 routes, and the 522, Link’s travel time becomes arguably inadequate. That’s where some or all of the buses may remain. I suggest a peak/non-peak tradeoff. Keep the one-seat rides peak hours when time is money and too many people will complain, but truncate them off-peak because Link’s frequency makes up for slower travel time and is a more efficient way to spend our transit dollars.

    3. Yes, as the two commenters said above, I-90 and I-5 bus service in the central core will pretty much go away. There is a trade-off, but those who miss the old express buses will just have to deal with it, just as those who want an express to the other end of downtown have to deal with it every day.

      But 520 buses are a completely different story. That is why I wrote a very long comment about that very subject above. There is no simple solution. If done right, it will require proper planning by various agencies on different roadways (I-5, 520 and possibly Seattle streets). I don’t think there is a consensus on the issue, and I don’t think I’ve heard of a solution that sounds great.

      1. I think the service pattern we’ll see is most all-day 520 routes becoming U-District routes and the 520 routes to downtown becoming peak-only.

        While there are still issues with this service pattern it does avoid Montlake at the times with the worst congestion and avoids potential issues with overloading Link during peak.

      2. I think that is a good prediction.

        But where exactly in the U-District will they go, and how will they get there? For example, let’s assume that 520 HOV lanes merge into the mainline. Also assume that a bus wants to spend as little time as possible on the surface streets. So now a bus has smooth sailing from the east side all the way until it merges with I-5. It then encounters one of the worst stretches of traffic in the city (the ship canal). Then it takes the 45th street exit (also a congested area). Then it turns left a couple times and gets back on the freeway. Those two turns might take a while, but it gets even worse once the bus gets back on the freeway. From the southbound on-ramp, the bus would have to work its way over to 520 (changing lanes three times) on general purpose traffic that is usually congested (all day). In other words, a bus from Kirkland to the UW could easily spend well over half of its time getting from the Montlake station back to the Montlake station (heading the other direction). All this so that the bus serves a stop right next to the freeway (which is a long ways from the businesses in the U-District, let alone the campus). Of course, if the bus actually served the U-District, it would be even slower.

        I don’t mean to suggest that this is what you had in mind. All I’m saying is that the devil is in the details. I’m genuinely interested in hearing ideas, because I have yet to hear or figure out a great solution. Here is another example:

        1) Make the express lanes bidirectional,
        2) Make 42nd street exit a northbound exit only.
        3) Make the exit HOV only.
        4) Create a ramp connecting HOV westbound lanes of 520 to the northbound express lanes.
        5) Make the right lane between the 520 merge and the 42nd street exit HOV only.
        6) Ravenna remains a southbound HOV only ramp.
        7) Create a ramp connecting southbound express lanes to HOV eastbound lanes of 520.
        8) Make the section between Ravenna and the 520 exit HOV only.

        I think you can see where I’m going with this. Buses would get off at 42nd, then get back on at Ravenna. This is no picnic, but the time spent slogging at slow speeds would at least be serving the neighborhood. It would connect two stations in an area where there would be plenty of demand. So the 520 bus could easily serve double duty as a local connector (even one with limited stops). Perhaps the biggest drawback to this approach is the number of ramps. You would connect 520 with both the express lanes and the mainline. But again, that is what should be discussed. Is it worth it to connect those two freeways so that buses can go faster? If not, what are the alternatives? I would hate to spend billions on a light rail line and billions on a new 520 bridge (with its own HOV lane) only to see 520 buses still spend the bulk of their time stuck in traffic.

      3. What all-day 520 downtown routes are there besides the 255 and 545? We don’t have to get overly preoccupied about two little routes. The 545 Microsifties will probably gladly switch to East Link, which is practically the same. The 255 is part of the intractable “Kirkland problem”: Link is just not close enough to Kirkland to be a viable alternative. It’s not the end of the world if the 255 keeps going downtown, and terminating at UW Station has the problems of Montlake congestion and inadequate layover space. Although a Metro planner told me that through-routing is the solution to layover-space problems. He suggested a 271/32 route. What would you want to connect the 255 to?

        There’s no way they would route 520 buses to 45th: that’s the long slow way around. Try driving from Kirkland to Latona and you’ll find it isn’t fast and easy.

      4. Actually my thought would be to have the all-day 520 routes exit at Montlake, travel through the U-District, layover somewhere and reverse the process for the trip East.

        The peak downtown routes would more or less follow the current routing of the 545 in Seattle.

  27. Left exits, yeech. Early freeways were full of these (there’s one infamous freeway in Connecticut where the exits alternate between left and right sides). But it was pretty quickly figured out that they’re a terrible idea; nervewracking for drivers and disruptive of traffic, with no benefits.

    Is there any way to simply get rid of the left exits? The Seneca and University exits on the northbound could probably just be closed. Something could be done with the Olive Way exit to replace the function of the Mercer exit on the northbound. The SR520 exit on the southbound, however, is a nightmare.

Comments are closed.