MBTA Buses & Infrastructure

I know this is just a blog post (from yesterday, 4/14) and tough to verify some of what's in here, but it supposedly summarizes an SL extension working group meeting from earlier this week and I've not seen this written up elsewhere (if there's a better summary, please share):

First reaction, similar to some sentiments in the post itself, is that those headways are too long, even for an initial pilot.
Second reaction is that any additional service to Kendall from the north is welcome, but please let's do this right.
 
Hopefully the recording of the session will be posted publicly soon. Let's review the six options on the map:
  • SL3
    • Chelsea to Malden Center
      • significant street-running on Ferry St
    • Chelsea to Wellington
      • significant street-running on Route 16
    • Chelsea to Sullivan Square
  • SL6
    • Glendale to Kendall via E Somerville
    • Glendale to Kendall via Community College
    • Glendale to Haymarket via Rutherford Ave
Let's also review the specific pieces of infrastructure discussed here:
  • Bus lanes on Alford Street Bridge across Mystic River
  • Bus lanes on Broadway to Lynn St
  • Bus lanes on Rutherford Ave
In terms of where bus lanes have so far been proposed, there is a clear winning route: Glendale to Haymarket via Rutherford Ave; this route would run almost entirely in dedicated bus lanes.

That list of proposed bus lanes may not be exhaustive. The SL3 alts might see bus lanes between Chelsea and Everett Sq, and perhaps the SL6 alts would see bus lanes southwest of Sullivan; I guess we'll need to wait for the full recording to know for sure.

The article also mentions the possibility of local bus routes using the Silver Line bus lanes as well. Let's review those:
  • Bus lanes on Alford Street Bridge across Mystic River
    • 104, 105, 109
  • Bus lanes on Broadway to Lynn St
    • 97, 110, 112 (partial)
    • 104, 109 (majority)
  • Bus lanes on Rutherford Ave
    • none
    • the following routes originate north of Sullivan and could potentially through-run to downtown if high reliability could be assured: 104, 105, 109, 95, 101, 89
We can see that the SL6 alignments overlap heavily with existing routes. In particular, the 104 and 109 layer on top of each other to form (pre-covid) ~8-min headways, indicating high demand along this corridor. No existing route runs between Sullivan and Lechmere, but EZRide does link Lechmere and Kendall roughly along the proposed alignment.

The SL3 alignments have much lower overlaps with existing bus routes. No routes mirror the Chelsea-Malden or Chelsea-Sullivan alts. The 112 broadly mirrors the Chelsea-Wellington alt; the 112 is low-frequency (45 min headways even at peak) and operates highly circuitously through Chelsea.

Let's momentarily set aside the question of routing, and instead approach this from the perspective of the capital investments under review here. The SL6 alignments reflect existing rider patterns, and build on an existing network of routes which would all see improved reliability and travel times with the construction of dedicated bus lanes. The SL3 alignments do not closely adhere to any existing routes, and have much less evidence of demand.

It's also unclear whether the SL3 alignments would provide speedier service to Everett. SL3 from Chelsea Station to South Station is timetabled in the morning rush at 35 minutes. Let's assume (perhaps generously) that an extension to Everett Sq adds 8 minutes, bringing total travel time from Everett to Downtown to 43 minutes.

By contrast, going from Everett Sq to State via the 109 + Orange today takes 25 minutes; according to Google, at least one 104 bus through runs from Sullivan directly to State as a 92 bus, and is able to complete that journey in 38 minutes. Meaning: the Everett Sq-Sullivan-Haymarket journey today -- without bus lanes south of Sullivan and with 20 min worth of local stops in Charlestown -- is already competitive with an SL3 extension, which already has bus lanes, and which needs to contend with the unreliability of the drawbridge and the Ted Williams Tunnel. The improvements described here would undoubtedly shorten that runtime further.

Finally, it's worth pointing out that an SL3 extension will be limited to rolling stock that can operate in the Transitway underground; a separate SL6 extension would not, and therefore would have greater operational flexibility. (An SL6 service to Haymarket would also be easy to extend to South Station via Congress St, and integrate to the SL4.)

So, I say, build the bus lanes first, and then figure out changes you want to make to the routes. Bus lanes on Broadway and the bridge serve routes that sorely need them now, and which could integrate with a new Glendale-Downtown service to provide high-frequency headways within Everett. These would also leave open the door for subsequent SL3 extensions to Glendale, Sullivan, or even further south.

Bus lanes on Rutherford would enable one-seat rides from Everett to downtown; Everett is the only community that directly borders Boston which does not have a (regularly scheduled) one-seat ride to downtown. Lanes on Rutherford can also be leveraged for a service to Kendall, either now or in the future.

By contrast, the SL3 alignments rely more heavily on mixed-traffic running, use a more circuitous route to reach downtown, integrate poorly into the existing network, and will have kneecapped reliability due to the drawbridge, Ted Williams Tunnel, and more limited rolling stock.

1) Build the bus lanes, reflecting present and validated rider demand
2) Institute SL6 from Glendale to downtown or Sullivan, and combine and coordinate SL6, 104 and 109 schedules.
3) See where things go from there
 
Do we know how much demand there is from Everett to Seaport and maybe South Station? The SL3 alignments may have an advantage for those commuters.

Regardless, I agree with the guess that SL6 alignments will indeed benefit Everett more. As for the southern terminus, Haymarket vs Kendall will be a very tough choice - both have tremendous value.
 
Ridership question about the SL3 today (if anyone knows).

Do people really sit on the SL3 all the way to South Station downtown, or do they hop off at Airport and take the much faster Blue Line route into downtown?

It probably depends on final destination, but the Ted Williams - transit way combo is really slow.
 
Ridership question about the SL3 today (if anyone knows).

Do people really sit on the SL3 all the way to South Station downtown, or do they hop off at Airport and take the much faster Blue Line route into downtown?

It probably depends on final destination, but the Ted Williams - transit way combo is really slow.
It looks like about 2/3rds of riders who board north of Airport continue through the Ted. If you look at Peak only it's closer to 1/2.
However, the people who exit at Airport are replaced by a roughly equal load of people who board at Airport and use SL3 to get from Airport to SLW/South Station/Red Line. If you want to get from Orient Heights to Seaport or Quincy Center, SL3 is a reasonable choice.
743 Ridership.jpg
 
Last edited:
Ridership question about the SL3 today (if anyone knows).

Do people really sit on the SL3 all the way to South Station downtown, or do they hop off at Airport and take the much faster Blue Line route into downtown?

It probably depends on final destination, but the Ted Williams - transit way combo is really slow.

I can only speak anecdotally, but I live in Quincy and when my now wife was living near Airport station I would frequently take the SL3 between SS and Airport. Although had R/B connector been built I would have likely opted for Blue
 
It looks like about 2/3rds of riders who board north of Airport continue through the Ted. If you look at Peak only it get's closer to 1/2.
However, the people who exit at Airport are replaced by a roughly equal load of people who board at Airport and use SL3 to get from Airport to SLW/South Station/Red Line. If you want to get from Orient Heights to Seaport or Quincy Center, SL3 is a reasonable choice. View attachment 23492

Thanks for pulling those numbers, you beat me to it.

I made a couple of Sankey diagrams (using sankeymatic.com, which was the first thing that showed up when I googled; code in spoiler box below).

(Note that the "Load" numbers were automatically calculated by the diagram tool and basically represent a hypothetical number -- specifically, the load on the bus after everyone has boarded at a stop and before anyone has disembarked at a stop -- i.e. a sort of max load that never occurs in the real world. So ignore those numbers, and instead pay attention to the width of the bars on each side of the markers.)

The first goes stop by stop:

SL3 ridership sankey.png


From here, we can see that the lion's share of SL3 riders -- whether originating at the Airport or in Chelsea -- disembark at South Station. We can also see that no particular station in the Seaport is higher ridership than the others, though if we combine the on-street World Trade Center stop with the tunnel stop, then we do see a stand-out destination, with just under 500 alightings -- on par with Airport as a destination.

To simplify things, I made a second diagram that groups stations together by segment:

SL3 simple ridership sankey.png


About 2,000 riders board at stops in Chelsea. About a third disembark at Airport, who are then replaced nearly one-to-one, as @737900er points out, with new passengers.

So about 2,000 riders head through the Ted Williams Tunnel. In the Seaport, about 1,000 riders disembark and about 1,200 riders board. We can probably assume that everyone boarding in the Seaport is traveling on to South Station -- at most, perhaps the 215 riders boarding at Silver Line Way are traveling to Courthouse, but that seems unlikely.

South Station sees 1,700-2,000 passengers disembarking (the T's numbers are weird here, but oh well). If roughly 1,000 of those come from the Seaport, that leaves 700-1,000 passengers who must come from Airport or Chelsea. That matches up with 2,000 passengers leaving from Airport, about 1,000 of whom disembark in the Seaport.

Of those 2,000 riders coming to the Seaport and South Station from the Airport, 1,400 must be coming from Chelsea. If 1,000 people are disembarking at the Seaport and only 600 board at Airport, then there must be at least 400 Chelsea-Seaport riders (probably more, since at least some of Airport's 600 riders will instead go to South Station).

I think the breakdown looks like this:
  • Airport-Seaport: no more than 600
  • Airport-South Station: no more than 600
  • Chelsea-Airport: about 600 (we know this number for sure)
  • Chelsea-Seaport: at least 400, maybe as high as 700-800 depending on Airport's South Station/Seaport split above
  • Chelsea-South Station: at least 400, maybe as high as 700-800 depending on Airport's South Station/Seaport split above
  • Chelsea-South Station or Seaport: 1,400 total
So, yes, there definitely are folks who are riding all the way from Chelsea through the Ted Williams Tunnel -- about 2,000 inbound each day. To me, the question remains: for Chelsea riders, what's the breakdown in Seaport vs South Station journeys? And likewise for Airport riders? Seaport-destined journeys probably end in the Seaport, but South Station-destined journeys may carry on further.

In the context of the Everett proposals, this matters: if most Chelsea SL3 riders are in fact going to South Station, then faster service via Sullivan would be an improvement, all the moreso because of the better transfers and access to downtown. If many Chelsea SL3 riders are instead going to the Seaport, then the story is a bit more complicated.

(This also reminds me of an analysis I did last year, looking at how many Seaport riders originate at the Logan Airport terminals. That number is vanishingly small.)

Detailed diagram:

Chelsea ON [736] Load 0
Load 0 [736] Load 1
Bellingham ON [532] Load 1
Load 1 [37] Bellingham OFF
Load 1 [1261] Load 2
Box District ON [453] Load 2
Load 2 [59] Box District OFF
Load 2 [1657] Load 3
Eastern Ave ON [282] Load 3
Load 3 [105] Eastern Ave OFF
Load 3 [1838] Load 4
Airport ON [595] Load 4
Load 4 [612] Airport OFF
Load 4 [1822] Load 5
Congress St ON [88] Load 5
Load 5 [372] Congress St OFF
Load 5 [1538] Load 6
Silver Line Way ON [215] Load 6
Load 6 [315] Silver Line Way OFF
Load 6 [1435] Load 7
World Trade Ctr ON [399] Load 7
Load 7 [127] World Trade Center OFF
Load 7 [1734] Load 8
Courthouse ON [521] Load 8
Load 8 [253] Courthouse OFF
Load 8 [2009] Load 9
Load 9 [1768] South Station OFF

Simple diagrams:
Chelsea stops ON [2003] Load 1
Load 1 [201] Chelsea stops OFF
Load 1 [1838] Load 2
Airport ON [595] Load 2
Load 2 [612] Airport OFF
Load 2 [1822] Load 3
Seaport stops ON [1223] Load 3
Load 3 [1067] Seaport stops OFF
Load 3 [2009] South Station OFF
 
Thanks for pulling those numbers, you beat me to it.

I made a couple of Sankey diagrams (using sankeymatic.com, which was the first thing that showed up when I googled; code in spoiler box below).

(Note that the "Load" numbers were automatically calculated by the diagram tool and basically represent a hypothetical number -- specifically, the load on the bus after everyone has boarded at a stop and before anyone has disembarked at a stop -- i.e. a sort of max load that never occurs in the real world. So ignore those numbers, and instead pay attention to the width of the bars on each side of the markers.)

The first goes stop by stop:

View attachment 23493

From here, we can see that the lion's share of SL3 riders -- whether originating at the Airport or in Chelsea -- disembark at South Station. We can also see that no particular station in the Seaport is higher ridership than the others, though if we combine the on-street World Trade Center stop with the tunnel stop, then we do see a stand-out destination, with just under 500 alightings -- on par with Airport as a destination.

To simplify things, I made a second diagram that groups stations together by segment:

View attachment 23496

About 2,000 riders board at stops in Chelsea. About a third disembark at Airport, who are then replaced nearly one-to-one, as @737900er points out, with new passengers.

So about 2,000 riders head through the Ted Williams Tunnel. In the Seaport, about 1,000 riders disembark and about 1,200 riders board. We can probably assume that everyone boarding in the Seaport is traveling on to South Station -- at most, perhaps the 215 riders boarding at Silver Line Way are traveling to Courthouse, but that seems unlikely.

South Station sees 1,700-2,000 passengers disembarking (the T's numbers are weird here, but oh well). If roughly 1,000 of those come from the Seaport, that leaves 700-1,000 passengers who must come from Airport or Chelsea. That matches up with 2,000 passengers leaving from Airport, about 1,000 of whom disembark in the Seaport.

Of those 2,000 riders coming to the Seaport and South Station from the Airport, 1,400 must be coming from Chelsea. If 1,000 people are disembarking at the Seaport and only 600 board at Airport, then there must be at least 400 Chelsea-Seaport riders (probably more, since at least some of Airport's 600 riders will instead go to South Station).

I think the breakdown looks like this:
  • Airport-Seaport: no more than 600
  • Airport-South Station: no more than 600
  • Chelsea-Airport: about 600 (we know this number for sure)
  • Chelsea-Seaport: at least 400, maybe as high as 700-800 depending on Airport's South Station/Seaport split above
  • Chelsea-South Station: at least 400, maybe as high as 700-800 depending on Airport's South Station/Seaport split above
  • Chelsea-South Station or Seaport: 1,400 total
So, yes, there definitely are folks who are riding all the way from Chelsea through the Ted Williams Tunnel -- about 2,000 inbound each day. To me, the question remains: for Chelsea riders, what's the breakdown in Seaport vs South Station journeys? And likewise for Airport riders? Seaport-destined journeys probably end in the Seaport, but South Station-destined journeys may carry on further.

In the context of the Everett proposals, this matters: if most Chelsea SL3 riders are in fact going to South Station, then faster service via Sullivan would be an improvement, all the moreso because of the better transfers and access to downtown. If many Chelsea SL3 riders are instead going to the Seaport, then the story is a bit more complicated.

(This also reminds me of an analysis I did last year, looking at how many Seaport riders originate at the Logan Airport terminals. That number is vanishingly small.)

Detailed diagram:

Chelsea ON [736] Load 0
Load 0 [736] Load 1
Bellingham ON [532] Load 1
Load 1 [37] Bellingham OFF
Load 1 [1261] Load 2
Box District ON [453] Load 2
Load 2 [59] Box District OFF
Load 2 [1657] Load 3
Eastern Ave ON [282] Load 3
Load 3 [105] Eastern Ave OFF
Load 3 [1838] Load 4
Airport ON [595] Load 4
Load 4 [612] Airport OFF
Load 4 [1822] Load 5
Congress St ON [88] Load 5
Load 5 [372] Congress St OFF
Load 5 [1538] Load 6
Silver Line Way ON [215] Load 6
Load 6 [315] Silver Line Way OFF
Load 6 [1435] Load 7
World Trade Ctr ON [399] Load 7
Load 7 [127] World Trade Center OFF
Load 7 [1734] Load 8
Courthouse ON [521] Load 8
Load 8 [253] Courthouse OFF
Load 8 [2009] Load 9
Load 9 [1768] South Station OFF

Simple diagrams:
Chelsea stops ON [2003] Load 1
Load 1 [201] Chelsea stops OFF
Load 1 [1838] Load 2
Airport ON [595] Load 2
Load 2 [612] Airport OFF
Load 2 [1822] Load 3
Seaport stops ON [1223] Load 3
Load 3 [1067] Seaport stops OFF
Load 3 [2009] South Station OFF
Excellent work to you both!

This also shows that even in a world with a GL branch to Everett, Chelsea and maybe Airport (via SL3's ROW), it would still be important to maintain a link between South Station/Seaport and Chelsea, or at least Airport BL for a transfer to Chelsea. Though that can also be solved with through-running GL trains like Chelsea-Everett-Sullivan-Park-Bay Villege-South Station-Seaport.
 
Excellent work to you both!

This also shows that even in a world with a GL branch to Everett, Chelsea and maybe Airport (via SL3's ROW), it would still be important to maintain a link between South Station/Seaport and Chelsea, or at least Airport BL for a transfer to Chelsea. Though that can also be solved with through-running GL trains like Chelsea-Everett-Sullivan-Park-Bay Villege-South Station-Seaport.
This is probably more suited to Crazy Transit Pitches, but it also depends on how many people are choosing that option because there is no Red-Blue Connector.

SL3 is Chelsea's only connection to Red (which is not connected to Blue, obviously), so it might not surprise me to find out that a lot of SL3 passengers through the Ted are in fact going to Cambridge. 116 and 117 go the the Blue. 111 goes to the Orange and Green. If you want a 2-seat ride from Chelsea (or Airport) to Cambridge, SL3 is your only choice. It looks asinine on a map, but it's not crazy.

If it turns out that a lot of Chelsea Passengers are going to Kendall, Harvard, etc., it might be worth exploring a combination of 2 of the proposals they have there: an urban ring style Kendall-Lechmere-Sullivan-Everett-Chelsea-Airport-Transitway-South Station route.
 
Last edited:
Kendall is a decidedly tertiary job destination behind Downtown and Back Bay --- approx the same size as the Seaport in job numbers. A SL6 route from Everett to Kendall via Sullivan is really a circumferential with a major transfer point at Sullivan, which seems a dubious proposition.

If we want circumferential transit to Kendall, we should design it carefully and not compromise transit for Everett.
 
Kendall is a decidedly tertiary job destination behind Downtown and Back Bay --- approx the same size as the Seaport in job numbers. A SL6 route from Everett to Kendall via Sullivan is really a circumferential with a major transfer point at Sullivan, which seems a dubious proposition.

If we want circumferential transit to Kendall, we should design it carefully and not compromise transit for Everett.
One challenge in evaluating employment centers, is we are largely basing our analysis on pre-pandemic commuting patterns.

We don't actually know yet whether Downtown and Back Bay are going to remain huge centers of in-person employment. Hybrid working structures seem to be settling in for a lot of the traditionally office-based industries.

We do know, however, that the lab-based industries (that dominate Kendall) remained in-person throughout the pandemic, and will continue to do so. We might want to exercise a bit of caution in our analysis of future transit needs based on "employment centers".
 
One challenge in evaluating employment centers, is we are largely basing our analysis on pre-pandemic commuting patterns.

We don't actually know yet whether Downtown and Back Bay are going to remain huge centers of in-person employment. Hybrid working structures seem to be settling in for a lot of the traditionally office-based industries.

We do know, however, that the lab-based industries (that dominate Kendall) remained in-person throughout the pandemic, and will continue to do so. We might want to exercise a bit of caution in our analysis of future transit needs based on "employment centers".

While an imperfect measure, one way to check in-person employment trends might be to look at existing transit station passenger boardings/departures versus their pre-pandemic baseline compared to elsewhere in the city/metro. IIRC, Kendall was/is at a relatively high percentage of its pre-pandemic baseline compared to other 'employment center' transit stations (and that was with MIT being largely remote for the 1st year of the pandemic). Also, Kendall is still growing considerably as an employment center, which will mean growth in employees spanning various job functions/education levels, from support functions (local food service, security, lab techs, etc) thru scientists. My point, in conjunction with Jeff's, is that I don't know how sure we are that Kendall will be a "tertiary" employment center going forward...

(as a side note, I still don't think hybrid work spells as much of a doomsday for central business districts as people are saying)
 
My picks that would most benefit Everett IMHO:

SL6 from Glendale Square to Haymarket with local service along the Broadway stretch from Glendale to Alford Street bridge and express service from Alford to Haymarket along Rutherford/Washington Sts (no stopping at Sullivan, Community College, North Station, or any other intermediate points). Fastest service to downtown Boston and GL/OL connections to Back Bay CBD at Haymarket. If the proposed North Station to Seaport BRT happens, then connection to Seaport CBD as well.

SL3 to from existing Chelsea terminus to Sullivan via Everett Square / Broadway. SL6 / SL3 / local bus transfer hub created at Everett Square. Plenty of local bus connections to points west at Sullivan and Airport / Seaport CBD east along existing SL3 service. Could consider extending SL3 for one stop ride to Kendall with cross connection at GLX East Somerville station; however street routing is a little wonky and indirect (hence the need for Sullivan to GJ light rail connect, but that’s for a different thread). Probably just rework the local bus routing to better serve Kendall from Sullivan.

The point should be really to keep it simple. Often times these initiatives get bogged down or don’t meet expectations because they try to do too much or forget the project objective, which in this case is to improve access for Everett riders.
 
These numbers are....not close:

Office:
35M RSF in the Financial District (not counting nearby parts of downtown)
14M in Back Bay
10M in the Seaport
6.7M in Kendall

Lab:
5.8M in Boston
11M in Kendall

Remember, too, that lab space has lower employee density than office space.

We should think very carefully before supplying Everett with lower quality transit because of the need for real circumferential transit to Kendall. That, too, is important but should be properly addressed and planned.
 
These numbers are....not close:

Office:
35M RSF in the Financial District (not counting nearby parts of downtown)
14M in Back Bay
10M in the Seaport
6.7M in Kendall

Lab:
5.8M in Boston
11M in Kendall

Remember, too, that lab space has lower employee density than office space.

We should think very carefully before supplying Everett with lower quality transit because of the need for real circumferential transit to Kendall. That, too, is important but should be properly addressed and planned.
Lab space had historically lower employee density than office space.

Right now offices are seeing 1/3, 1/4, even 1/5 their historical employee density (which is a huge looming issue for commercial office space that cannot hold.).
 
I can’t imagine many people staying on an SL6 bus when they could transfer to OL at Sullivan. Paralleling the OL to Haymarket has value, but it is far better served with improved OL headways.

One of the biggest weaknesses of the MBTA network is that it is exclusively a hub and spoke network. That is the strongest form for a small network, but at some point it is beneficial to expand as a grid instead of adding more pure radial spokes. The SL6 concept that hits the growth centers of Kendall, Lechmere, Rutherford corridor, and Sullivan with transfers to 3 RT lines seems to me to promote network affects and new growth. These areas are just outside the CBD and arguably can be assimilated if they get serious transit capacity. I don’t see a big compromise for Everett riders who get their pick of transfer point to reach downtown or just about any other destination.

I’ll grant you that the first cardinal sin of transit planning is to build a line that misses the CBD. The second cardinal sin is to miss a transfer connection. I see this as a case where the value of network effects can outweigh the CBD gravity well.
 
I see your point, but I defer to Alon Levy that mixing radials and circumferentials is a bad idea...
 
I wrote this previously about Rutherford: It directly parallels the OL and the 92, and there's not any useful local stops to add that the OL doesn't serve other than City Square. The only way I can really see it making sense is with a dedicated busway on Rutherford, a lot of infill development on Rutherford to justify a stop at Essex Street, and use of a Congress Street BRT to get to South Station.

However, I do like the idea of that Everett-Sullivan-Lechmere-Kendall corridor. Given how many different nodes that connects, I don't think mixing the radial and circumferential elements would be a big issue
 
I wrote this previously about Rutherford: It directly parallels the OL and the 92, and there's not any useful local stops to add that the OL doesn't serve other than City Square. The only way I can really see it making sense is with a dedicated busway on Rutherford, a lot of infill development on Rutherford to justify a stop at Essex Street, and use of a Congress Street BRT to get to South Station.

However, I do like the idea of that Everett-Sullivan-Lechmere-Kendall corridor. Given how many different nodes that connects, I don't think mixing the radial and circumferential elements would be a big issue
The way I see it is that you can treat a Rutherford-Haymarket SL6 as an express route from Everett to downtown, just like 111 and the 500 series. The value of intermediate stops does not matter as much.
 

Back
Top